British Journal of Dermatology
July 2007 - Vol. 157 Issue 1 Page1-214
Snippets
RESEARCH SNIPPETS pages xix–xix
Editorial
The alcohol hand rub: a good soap substitute? pages 1–3
Review articles
The evolution of the psoriatic lesion pages 4–15
Skin manifestations of intravascular lymphoma mimic inflammatory diseases of the skin pages 16–25
Original articles
Cutaneous Biology
DNA repair capacities of cutaneous fibroblasts: effect of sun exposure, age and smoking on response to an acute oxidative stress pages 26–32
A cytotoxic analysis of antiseptic medication on skin substitutes and autograft pages 33–40
Original articles
Clinical and laboratory investigations
Differences in survivin location and Bcl-2 expression in CD30+ lymphoproliferative disorders of the skin compared with systemic anaplastic large cell lymphomas: an immunohistochemical study pages 41–48
Association of functional gene variants in the regulatory regions of COX-2 gene (PTGS2) with nonmelanoma skin cancer after organ transplantation pages 49–57
Sentinel lymph node biopsy in melanoma: a micromorphometric study relating to prognosis and completion lymph node dissection pages 58–67
Prevalence of metabolic syndrome in patients with psoriasis: a hospital-based case–control study pages 68–73
Original articles
Contact dermatitis and allergy
How irritant is alcohol? pages 74–81
Artificial reduction in transepidermal water loss improves skin barrier function pages 82–86
Original articles
Dermatological surgery and lasers
Randomized, double-blind, prospective study to compare topical 5-aminolaevulinic acid methylester with topical 5-aminolaevulinic acid photodynamic therapy for extensive scalp actinic keratosis pages 87–91
Lipomas after blunt soft tissue trauma: are they real? Analysis of 31 cases pages 92–99
Original articles
Epidemiology and health services research
Impetigo in epidemic and nonepidemic phases: an incidence study over 4½ years in a general population pages 100–105
Hairdressing is associated with scalp disease in African schoolchildren pages 106–110
Original articles
Photobiology
Variable pulsed light is less painful than light-emitting diodes for topical photodynamic therapy of actinic keratosis: a prospective randomized controlled trial pages 111–117
Original articles
Therapeutics
Use of oral glycopyrronium bromide in hyperhidrosis pages 118–121
Oral R115866 in the treatment of moderate to severe facial acne vulgaris: an exploratory study pages 122–126
Mycophenolate mofetil for severe childhood atopic dermatitis: experience in 14 patients pages 127–132
Vehicle-controlled, randomized, double-blind study to assess safety and efficacy of imiquimod 5% cream applied once daily 3 days per week in one or two courses of treatment of actinic keratoses on the head pages 133–141
Corticosteroid-induced clinical adverse events: frequency, risk factors and patient’s opinion pages 142–148
A multicentre, randomized, controlled study of the efficacy, safety and cost-effectiveness of a combination therapy with amorolfine nail lacquer and oral terbinafine compared with oral terbinafine alone for the treatment of onychomycosis with matrix involvement pages 149–157
Original articles
Concise communications
Endothelial cells in infantile haemangiomas originate from the child and not from the mother (a fluorescence in situ hybridization-based study) pages 158–160
Transforming growth factor-β receptor II is preferentially expressed in the companion layer of the human anagen hair follicle pages 161–164
Associations of promoter region polymorphisms in the tumour necrosis factor-α gene and early-onset psoriasis vulgaris in a northern Polish population pages 165–167
Case reports
Acquired palmoplantar keratoderma and immunobullous disease associated with antibodies to desmocollin 3 pages 168–173
Bartonella-related pseudomembranous angiomatous papillomatosis of the oral cavity associated with allogeneic bone marrow transplantation and oral graft-versus-host disease pages 174–178
Gene corner
Novel COL7A1 mutations in a Japanese family with transient bullous dermolysis of the newborn associated with pseudosyndactyly pages 179–182
Correspondence
Cutaneous lesions in neurofibromatosis 1: confused terminology pages 183–184
Lentigo maligna involving the tumour nests and stroma of a nodular basal cell carcinoma pages 184–188
Eruptive vellus hair cysts presenting as bluish-grey facial discoloration masquering as naevus of Ota pages 188–189
Angiomyofibroblastoma of the vulva with a penile appearance pages 189–191
Successful treatment of severe psoriatic arthritis with infliximab in an 11-year-old child suffering from linear psoriasis along lines of Blaschko pages 191–192
Two siblings with neonatal pemphigus vulgaris associated with mild maternal disease pages 192–194
An infantile case of pityriasis lichenoides et varioliformis acuta
pages 194–196
Gomez–Lopez–Hernandez syndrome: another consideration in focal congenital alopecia pages 196–198
A case of phosphaturic mesenchymal tumour (mixed connective tissue variant) that developed in the subcutaneous tissue of a patient with oncogenic osteomalacia and produced fibroblast growth factor 23 pages 198–200
Unilateral periorbital oedema due to sarcoid infiltration of the eyelid: an unusual presentation of sarcoidosis with facial nerve palsy and parotid gland enlargement pages 200–202
Pityriasis rubra pilaris in a mother and two daughters pages 202–204
Cutaneous Mycobacterium neoaurum infection causing scarring alopecia in an immunocompetent host
pages 204–206
The use of intravenous immunoglobulin in cutaneous and recurrent perforating intestinal Degos disease (malignant atrophic papulosis) pages 206–207 A novel deletion mutation in the EDAR gene in a Pakistani family with autosomal recessive hypohidrotic ectodermal dysplasia pages 207–209
Mondor’s phlebitis after using tadalafil
pages 209–210
Cutaneous angiokeratoma and venous malformations in a Hispanic-American patient with cerebral cavernous malformations pages 210–212
Folate with methotrexate: big benefit, questionable cost
pages 213–213
News and Notices
News and Notices
pages 213–214
RESEARCH SNIPPETS DOI 10.1111/j.1365-2133.2007.08079.x
A cytotoxic analysis of antiseptic medication on skin substitutes and autograft Antibacterial agents are frequently used on human skin substitutes (HSSs) for treating difficult-to-heal ulcers, as infection or bacterial colonization can cause the graft to fail. Twelve different antiseptic agents were tested on two different HSS models (autologous reconstructed epidermis on fibroblast-populated dermis and allogeneic reconstructed epidermis on a fibroblast- populated collagen gel). The degree of cytotoxicity was analysed by detrimental changes in histology, metabolic activity and RNA staining of tissue sections. ActicoatÒ, Aquacel AgÒ, DermacynÒ, FucidinÒ,0Æ5% silver nitrate solution and chlorhexidine digluconate were not cytotoxic for either HSS or autograft, and can therefore be used as required. FlamazineÒ and zinc oxide cream resulted in moderate cytotoxicity. However, application of BetadineÒ, cerium-silver sulfadiazine cream, silver sulfadiazine cream with 1% acetic acid and FuracineÒ resulted in a substantial decrease in cell viability and a detrimental effect on tissue histology when applied to autograft and especially to HSS. Le Duc Q, Breetveld M, Middlekoop E et al. A cytotoxic analysis of antiseptic medication on skin substitutes and autograft. Br J Dermatol 2007; 157:33–40.
Sentinel lymph node biopsy in melanoma: a micromorphometric study relating to prognosis and completion lymph node dissection Sentinel lymph node (SLN) biopsy still remains controversial. However, this large series of patients studied (n ¼ 455) has confirmed the prognostic value of SLN biopsy. It also confirmed the relationship between the tumour burden in the SLN (evaluated by the maximum diameter of the largest metastasis) and the clinical outcome. Debarbieux S, Duru G, Dalle S et al. Sentinel lymph node biopsy in melanoma: a micromorphometric study relating to prognosis and completion lymph node dissection. Br J Dermatol 2007; 157:58–67.
Prevalence of metabolic syndrome in patients with psoriasis: a hospital-based case–control study The aim of this study was to investigate the prevalence of metabolic syndrome in patients with psoriasis. Adult patients with chronic plaque psoriasis (n ¼ 338) and patients with skin diseases other than psoriasis (n ¼ 334) were studied. Metabolic syndrome was significantly more common in psoriatic patients than in controls (30Æ1% vs. 20Æ6%; odds ratio 1Æ65; 95% confidence interval 1Æ16–2Æ35; P ¼ 0Æ005) after the age of 40 years. Psoriatic patients had also a higher prevalence of hypertriglyceridaemia and abdominal obesity, whereas hyperglycaemia, arterial hypertension and plasma levels of high-density lipoprotein cholesterol were similar. Although patients with psoriasis were more frequently smokers, the association of psoriasis with metabolic syndrome was independent from smoking. There was no correlation between severity of psoriasis and prevalence of metabolic syndrome. Psoriatic patients with metabolic syndrome were older and had a longer disease duration compared with psoriatic patients without metabolic syndrome. Gisondi P, Tessari G, Conti A et al. Prevalence of metabolic syndrome in patients with psoriasis: a hospital-based case–control study. Br J Dermatol 2007; 157:68–73.
Post-traumatic lipomas: fact or fiction? Is there a link between blunt soft tissue trauma and the formation of lipomas? Thirty-four cases of post-traumatic lipomas in 31 patients were studied. The mean time elapsed between soft tissue trauma and lipoma formation was 2Æ0 years (range 0Æ5–5). Twenty-five of the 31 patients reported an extensive and slowly resolving haematoma after the traumatic event. Eleven of 34 lipomas were found on the upper extremities, five on the lower extremities, 13 on the trunk, and two on the face. All tumours were located subcutaneously. Fourteen patients presented with an elevated partial thromboplastin time. There are two potential explanations to correlate blunt soft tissue trauma and the formation of post-traumatic lipomas: (i) the formation of so-called post-traumatic ‘pseudolipomas’ by prolapsing adipose tissue through fascia resulting from direct impact, and (ii) lipoma formation as a result of preadipocyte differentiation and proliferation mediated by cytokine release following soft tissue damage and haematoma formation. Aust MC, Spies M, Kall S et al. Lipomas after blunt soft tissue trauma: are they real? Analysis of 31 cases. Br J Dermatol 2007; 157:92–99.
Mycophenolate mofetil for severe childhood atopic dermatitis: experience in 14 patients Mycophenolate mofetil (MMF) has been used in adults with severe atopic dermatitis with some success. Heller et al. retrospectively studied 14 children with recalcitrant atopic dermatitis. Four patients (29%) achieved complete clearance, four (29%) had > 90% improvement (almost complete), five (35%) had 60–90% improvement and one (7%) failed to respond. Initial responses occurred within 8 weeks (mean 4 weeks), and maximal effects were attained after 8–12 weeks (mean 9 weeks) ) ) at MMF doses of 40–50 mg kg 1 daily in younger children and 30–40 mg kg 1 daily in adolescents. The medication was well tolerated in all patients, with no infectious complications or development of leucopenia, anaemia, thrombocytopenia or elevated aminotransferases. MMF can be a safe and effective treatment for severe, refractory atopic dermatitis in children. Heller M, Shin HT, Orlow SJ, Schaffer JV. Mycophenolate mofetil for severe childhood atopic dermatitis: experience in 14 patients. Br J Dermatol 2007; 157:1271–32. EDITORIAL DOI 10.1111/j.1365-2133.2007.07786.x The alcohol hand rub: a good soap substitute?
Good hand hygiene prevents cross-infection in healthcare detergent washing as one might see in practice, the authors environments but adherence is poor among healthcare work- used standardized single and repetitive patch testing with sev- ers.1 Short-chain aliphatic alcohols, ‘alcohol-based hand rubs’ eral alcohols including the commonly used isopropyl alcohol (AHRs) are now in general use by healthcare professionals for (isopropanol, propan-2-ol). They assessed erythema, skin bar- hand disinfection in order to prevent transmission of patho- rier function (by measuring transepidermal water loss, TEWL) gens. They are indicated in clinically clean hands and are and also determined skin hydration. Erythema or skin barrier designed to be used after washing hands with detergent/soap disruption was not induced by the alcohols in the patch tests and water.2 AHRs are reported to be superior to traditional although skin hydration did decrease significantly. Interest- hand washing as they require less time to apply, are faster act- ingly, isopropyl alcohol produced the least change. ing, are less irritating to the skin and have been shown to Application of alcohols to skin pre-irritated by the detergent contribute significantly to decreased infection rates.1 There is a sodium lauryl sulphate (SLS) also did not produce an increase continued impression, however, that compliance rates are sub- in skin barrier disruption over that of SLS alone. optimal3 possibly because of the subjective stinging and skin Additionally, and to give a more clinically applicable result, discomfort induced by AHRs.1 standardized wash tests were performed with ethanol and the Occupational hand dermatitis may occur in up to 30% of irritant detergent SLS on the forearms with either each agent healthcare workers.4 This figure is higher for nursing staff5,6 alone or both agents in a tandem design. These tests demon- but the problem is also reported to be common in healthcare strated that alcohol application caused significantly less skin workers without direct patient contact.7 In a U.S. hospital irritation than washing with a detergent. Even on pre-irritated study 26% of nurses had damaged skin of the hands while skin, alcohol did not enhance irritation. The authors surmised 86% reported ever having skin problems.5 These changes that alcohol used after washing the skin with a detergent correlate with the use of soap at work, the number of brought about a protective effect by washing away irritating hand washes per shift and the frequency of glove changes. detergent molecules. They stressed that in this study alcohol The figure is higher again for atopic individuals.5 The clinical used in hand rubs did not induce further skin irritation and, picture is predominantly that of an irritant contact dermatitis importantly for compliance, emphasized that, despite alcohol and is mainly due to soaps and detergents used in routine application causing subjective sensory symptoms, the patho- hand cleaning. After using an AHR, healthcare workers may physiology of the skin was not altered. report burning, which may be explained by disruption of the Do these findings concur with recent literature that alcohol skin barrier by frequent hand washing. This can trigger a is indeed less irritating to normal skin and what, if anything, vicious cycle whereby the healthcare worker increases the is known about the situation in pre-irritated and atopic skin? frequency of hand washing and reduces the frequency of hand In clinical practice, cutaneous exposure to a wide variety of disinfection.1 Staphylococcus aureus colonization is common in chemical irritants such as surfactants and detergents is fre- dermatitis and can further exacerbate the disease. Patients with quent. Detergents are regarded as having relatively high skin atopic dermatitis also carry higher amounts of S. aureus on their toxicity. They can induce clinically relevant barrier disruption hands.8 The importance of using an AHR is therefore arguably and inflammation even after a single patch test and both the more important in those with damaged skin than it is in those duration of application and the concentration of the irritant with clinically normal skin. are important. More recently developed detergents such as sodium laureth sulphate and alkyl polyglucoside are less irrita- How irritant is alcohol? ting and have a lesser effect on skin permeability and barrier function than traditionally used agents such as the anionic The study reported by Lo¨ffler et al.9 in this issue of the Journal detergent SLS.10 Some irritants have persistent or late effects of suggests that AHRs cause less skin irritation than hand wash- up to 1 week.11 However, nonspecific skin irritation may be ing with detergents/soaps and should therefore be preferred due to different mechanisms of action at tissue level. Irritants from the dermatological viewpoint. They go on to state that such as retinoic acid cause (delayed) impairment of the stra- AHRs may even decrease rather than increase skin irritation tum corneum but produce different and distinctive biological after a hand wash due to a partial mechanical elimination of responses compared with SLS.12 The effects of simultaneous the detergent. application of different irritants has been shown to induce sig- In comprehensively investigating both the skin irritation nificantly stronger reactions than those caused by application caused by alcohols alone and also in combined usage with of each irritant on its own. Mixed application of, for example,
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp1–3 1 2 Editorial an anionic detergent and an organic solvent has an additive the skin than both disinfectant applied alone and disinfectant/ effect on skin irritation.13 This interaction between potential detergent alternately. irritants in the workplace is important because workers are Many health workers with damaged skin are atopic.5 Given not usually exposed to a single irritant, but to multiple poten- that atopics carry higher concentrations of S. aureus, any tially harmful substances. Physical irritant contact dermatitis, adverse effect of AHRs should increase skin irritation and caused by friction or mechanical abrasion, is a further occupa- therefore bacterial colonization.8 tional problem.14 In combined tandem repeated irritation tests Basketter’s group has previously argued that atopics (defined using both physical irritants and detergents, the type of irri- broadly by high IgE reactivity) were only marginally more sus- tant, occlusion, length of application and order of tandem ceptible to experimentally induced skin irritation and that cuta- application all affect the degree of barrier disruption.15 neous reactions to irritants in atopics were similar to those seen Irritation with alcohols is said to be common and many in nonatopics.22 They predicted that atopics and nonatopics healthcare workers complain about nonacceptable skin irrita- would give similar results in human tests of acute skin irritation. tion caused by AHRs. The assumed irritation due to alcohol- This was borne out in a study earlier this year where the dermal based hand antiseptics may impair their widespread use. tolerance to five AHRs was assessed among atopic and nonatopic However, experimental findings contradict the subjective subjects in a controlled, repetitive, occlusive patch test.23 No sig- experience and repeatedly demonstrate low skin toxicity due nificant difference was found and tolerance to the five AHRs was to alcohol especially when compared with detergent washing. good among both atopic and nonatopic subjects. Alcohol decreases hydration but does not induce barrier dis- Some groups argue that the addition of emollients can ruption or erythema.16 When tandem application of SLS and decrease the irritant effect of AHRs as patients with atopic skin n-propanol (propan-1-ol) is performed in healthy humans, n- have a defective barrier function both in affected, xerotic skin propanol does not enhance cumulative skin irritation when and in clinically normal skin. This was addressed in a study of used with SLS17 – although this does occur with stronger 15 patients with atopic dermatitis who were treated for irritants such as toluene. As n-propanol is the active ingredi- 20 days with a moisturizing cream.24 On day 21 the skin was ent in many AHRs, this is of particular interest in occupa- exposed to SLS. Skin hydration was significantly increased by tional irritant contact dermatitis in healthcare workers. A the treatment. Barrier function, measured as TEWL, improved small, prospective, randomized trial with crossover design and skin susceptibility to SLS was significantly reduced. A comparing the frequency of skin irritation and dryness asso- prospective, randomized double-blind study with intraindivid- ciated with using an AHR regimen for hand antisepsis ual comparison assessed the influence of an emollient on the against using soap and water for hand washing showed that acceptability of a mixed n-propanol (50% v/v) and isopropa- skin irritation and dryness decreased only slightly when nur- nol (30% v/v) AHR.16 Repeated application of these antiseptic ses used an AHR but increased substantially when nurses preparations caused significantly less irritation when the anti- used soap and water.18 Another study in nurses assessed both septic contained emollient. More recently, a prospective, ran- skin tolerance and antimicrobial effects of 8-day periods of domized, controlled double-blind trial of 35 subjects, half of either an AHR or a hand wash with a nonantiseptic soap. whom were atopic, compared an AHR with a propan-2-ol– Skin condition and grade of skin damage worsened signifi- emollient mix applied in a repeated open application test.25 cantly more in the group using soap than in the group using Erythema and dryness were significantly lower for the AHR the AHR. The AHR was also significantly more effective than with emollients in comparison with that for the AHR without liquid soap in removing transient contaminant microorgan- emollients. There was again no difference between the atopic isms. Of further relevance, the total bacterial count increased and nonatopic subjects, leading the authors to conclude that with the increasing number of hand washes in the soap the addition of emollients to a propanol-based hand rub group and with the degree of skin damage in the AHR can significantly decrease irritant contact dermatitis under group.19 frequent-use conditions.
What happens in already damaged skin? Should we be using alcohol-based hand rubs instead of detergents? Many healthcare workers may already have clinically or sub- clinically damaged skin.5 A study assessing the irritancy of It can be concluded that AHRs have a less deleterious effect on n-propanol 60% induced experimental low-grade irritant con- the skin than other physical irritants, which enhance skin tact dermatitis by overnight patch exposure to SLS. On skin reactivity. Hand hygiene using AHRs has become central to sites pre-irritated by SLS, n-propanol increased TEWL. The irri- hospital infection control programmes and is often considered tant potential of n-propanol 60%, the concentration used in synonymous with hand washing. However, healthcare workers daily practice, was significant only on detergent-irritated still do not wash their hands frequently enough.2 AHRs are skin.20 In a study of the short-term effects of intensive, repea- simple and quick to use, are economical,26 and there are now ted exposure to an alcohol-based disinfectant, to a detergent increasingly frequent calls for their use not only in replacing alone and to an alcohol-based disinfectant/detergent alter- detergent washes in basic hand hygiene,27 but also as part of nately,21 the detergent caused more redness and irritation of surgical hand disinfection in the operating theatre.28,29 Maybe
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp1–3 Editorial 3 the answer is that we can replace detergent washing with 13 Wigger-Alberti W, Krebs A, Elsner P. Experimental irritant contact AHRs if it leads to increased compliance with hand hygiene dermatitis due to cumulative epicutaneous exposure to sodium lau- measures30 but the evidence is that this compliance needs to ryl sulphate and toluene: single and concurrent application. Br J Dermatol 2000; 143:551–6. be encouraged with continued education and reinforcement.3 14 McMullen E, Gawkrodger DJ. Physical friction is under-recognized The possibility that with increasing use of AHRs the less irrita- as an irritant that can cause or contribute to contact dermatitis. Br J ting effect may be outweighed by greater exposure needs to Dermatol 2006; 154:154–6. be studied. The current major disadvantage of an AHR-based 15 Fluhr JW, Akengin A, Bornkessel A et al. Additive impairment of hand disinfection regimen is in the inefficacy of AHRs against the barrier function by mechanical irritation, occlusion and sodium bacterial spores.26,29 The current rise in hospital-acquired lauryl sulphate in vivo. Br J Dermatol 2005; 153:125–31. cases of antibiotic-associated colitis due to Clostridium difficile 16 Rotter ML, Koller W, Neumann R. The influence of cosmetic addi- tives on the acceptability of alcohol-based hand disinfectants. J Hosp infection means that old-fashioned hand washing with soap Infect 1991; 18 (Suppl. B):57–63. will be around on the wards for some time to come. 17 Kappes UP, Goritz N, Wigger-Alberti W et al. Tandem application of sodium lauryl sulfate and n-propanol does not lead to enhance- Department of Dermatology, G.A. JOHNSTON ment of cumulative skin irritation. Acta Derm Venereol (Stockh) 2001; Leicester Royal Infirmary, J.S.C. ENGLISH* 81:403–5. Leicester LE1 5WW, U.K. 18 Boyce JM, Kelliher S, Vallande N. Skin irritation and dryness associ- *Department of Dermatology, Queen’s ated with two hand hygiene regimens: soap-and-water hand wash- ing versus hand antisepsis with an alcoholic hand gel. Infect Control Medical Centre, Nottingham Hosp Epidemiol 2000; 21:442–8. University Hospitals NHS Trust, 19 Winnefeld M, Richard MA, Drancourt M et al. Skin tolerance and Nottingham NG7 2UH, U.K. effectiveness of two hand decontamination procedures in everyday Correspondence: John S.C. English. hospital use. Br J Dermatol 2000; 143:546–50. E-mail: [email protected] 20 Lu¨bbe J, Ruffieux C, van Melle G et al. Irritancy of the skin disin- fectant n-propanol. Contact Dermatitis 2001; 45:226–31. 21 Pedersen LK, Held E, Johansen JD et al. Short-term effects of alco- References hol-based disinfectant and detergent on skin irritation. Contact Dermatitis 2005; 52:82–7. 1 Pittet D. Compliance with hand disinfection and its impact on hos- 22 Basketter DA, Miettinen J, Lahti A. Acute irritant reactivity to pital-acquired infections. J Hosp Infect 2001; 48 (Suppl. A):S40–6. sodium lauryl sulfate in atopics and non-atopics. Contact Dermatitis 2 Widmer AF. Replace hand washing with use of a waterless alcohol 1998; 38:253–7. hand rub? Clin Infect Dis 2000; 31:136–43. 23 Kampf G, Wigger-Alberti W, Wilhelm KP. Do atopics tolerate 3 Creedon SA. Healthcare workers’ hand decontamination practices: alcohol-based hand rubs? A prospective, controlled, random- compliance with recommended guidelines. J Adv Nurs 2005; ized double-blind clinical trial. Acta Derm Venereol (Stockh) 2006; 51:208–16. 86:140–3. 4 Kampf G, Lo¨ffler H. Dermatological aspects of a successful intro- 24 Loden M, Andersson AC, Lindberg M. Improvement in skin duction and continuation of alcohol-based hand rubs for hygienic barrier function in patients with atopic dermatitis after treatment hand disinfection. J Hosp Infect 2003; 55:1–7. with a moisturizing cream (Canoderm). Br J Dermatol 1999; 5 Larson E, Friedman C, Cohran J et al. Prevalence and correlates of 140:264–7. skin damage on the hands of nurses. Heart Lung 1997; 26:404–12. 25 Kampf G, Wigger-Alberti W, Schoder V, Wilhelm KP. Emollients 6 Smit HA, Burdorf A, Coenraads PJ. Prevalence of hand dermatitis in a propanol-based hand rub can significantly decrease irritant in different occupations. Int J Epidemiol 1993; 22:288–93. contact dermatitis. Contact Dermatitis 2005; 53:344–9. 7 Kavli G, Angell E, Moseng D. Hospital employees and skin prob- 26 Rotter ML. Arguments for alcoholic hand disinfection. J Hosp Infect lems. Contact Dermatitis 1987; 17:156–8. 2001; 48 (Suppl. A):S4–8. 8 Williams JV, Vowels B, Honig P, Leyden JJ. Staphylococcus aureus isola- 27 Jungbauer FH, van der Harst JJ, Groothoff JW et al. Skin protection tion from the lesions of the hands and anterior nares of patients in nursing work: promoting the use of gloves and hand alcohol. with atopic dermatitis. Emerg Med 1999; 17:207–11. Contact Dermatitis 2004; 51:135–40. 9 Lo¨ffler H, Kampf G, Schermund D, Maibach HI. How irritant is 28 Hu¨bner NO, Kampf G, Kamp P et al. Does a preceding hand alcohol? Br J Dermatol 2007; 157:74–81. wash and drying time after surgical hand disinfection influence the 10 Lo¨ffler H, Happle R. Profile of irritant patch testing with deter- efficacy of a propanol-based hand rub? BMC Microbiol 2006; gents: sodium lauryl sulfate, sodium laureth sulfate and alkyl 6:57. polyglucoside. Contact Dermatitis 2003; 48:26–32. 29 Hu¨bner NO, Kampf G, Lo¨ffler H et al. Effect of a 1 min hand wash 11 Ale SI, Laugier JP, Maibach HI. Differential irritant skin responses on the bactericidal efficacy of consecutive surgical hand disinfec- to tandem application of topical retinoic acid and sodium lauryl tion with standard alcohols and on skin hydration. Int J Hyg Environ sulphate: II. Effect of time between first and second exposure. Br J Health 2006; 209:285–91. Dermatol 1997; 137:226–33. 30 Maury E, Alzieu M, Baudel JL et al. Availability of an alcoholic solu- 12 Effendy I, Weltfriend S, Patil S et al. Differential irritant skin tion can improve hand disinfection compliance in an intensive care responses to topical retinoic acid and sodium lauryl sulphate: alone unit. Am J Respir Crit Care Med 2000; 162:324–7. and in crossover design. Br J Dermatol 1996; 134:424–30.
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp1–3 REVIEW ARTICLE DOI 10.1111/j.1365-2133.2007.07907.x The evolution of the psoriatic lesion P.C.M. van de Kerkhof Department of Dermatology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands
Summary
Correspondence Psoriasis comprises a host of abnormalities, and various aspects of the pathogene- Peter C.M. van de Kerkhof. sis of psoriasis have been suggested to be of primary relevance. The aim of this E-mail: [email protected] review is to identity driving factors in the pathogenesis of psoriasis and to explore the dynamics of processes eventually resulting in a psoriatic lesion. In Accepted for publication 17 January 2007 this review observations on the evolution from the symptomless skin to lesional skin in patients with psoriasis will be integrated with observations in various ani- Key words mal models of psoriasis. epidermal proliferation, immunology, keratinization, psoriasis
Conflicts of interest None declared.
For decades the psoriatic lesion and the symptomless skin of Observations in lesional and symptomless skin patients with psoriasis have been compared and contrasted and peripheral blood of patients with psoriasis with the skin of normal healthy subjects. The peripheral blood of patients with psoriasis has also been studied extensively. Inflammatory changes and abnormalities in epidermal differ- Many abnormalities have been recorded; however, the ques- entiation characterize the psoriatic lesion. Symptomless skin tion of the order of events during the evolution of the lesion and peripheral blood of patients with psoriasis have been could not be resolved using these ‘static comparisons’. Also reported to be abnormal in various respects. the question of whether an abnormality can drive the evolu- tion of the psoriatic lesion and whether such factors are The psoriatic lesion obligatory remains to be answered. During the last decade our insights into the development of the psoriatic lesion have been The psoriatic lesion is characterized by numerous abnormal- expanded by animal models, which permit studies of the phe- ities. Abnormalities in innate immunity and adaptive immun- notype resulting from the overexpression of genes that have ity but also the characteristics of regenerative skin, as observed been suggested to be relevant in the pathogenesis of psoriasis during wound healing, constitute the complex reality of the and by xenograft models which have provided the opportun- psoriatic lesion. ity to study the transition of symptomless psoriatic skin to It has been shown that CD4+ T cells (helper T cells) out- lesional skin. In order to understand the evolution of the pso- number CD8+ cells (cytotoxic T cells) in the dermis, whereas riatic lesion, the transition between symptomless and clinically CD8+ T cells are more abundantly present in the epidermis involved skin provides an important opportunity to improve of the psoriatic plaque.1 The vast majority of T cells are our understanding of the pathogenesis of psoriasis. This transi- of the memory effector subset (CD45RO+), rather than the tion can be studied in the margin of psoriatic plaques with CD45RA+ (‘naive’) population.2 Upon activation, these cells centrifugal expansion, in sequential biopsies following stan- strongly upregulate CD25 [the a subunit of the interleukin dardized injury of the symptomless skin, or during follow-up (IL)-2 receptor], CD69, HLA-DR and the costimulatory recep- of relapses after a successful antipsoriatic treatment. tor CD2.1 Using immunofluorescent double staining, it was In this review, observations in lesional and symptomless shown that all pathogenic T-cell subsets (CD4+ CD25+, psoriatic skin and peripheral blood of patients with psoriasis CD4+ CD45RO+, CD8+ CD25+ and CD8+ CD45RO+ cells) will firstly be summarized. Observations in transgenic mice, were significantly increased in the dermis and in the epider- knockout mice and xenograft mice will be reviewed subse- mis of lesional psoriatic skin, as compared with normal skin.3 quently. Further, the transition between symptomless and Although the mentioned T-cell subsets are of critical import- lesional skin will be described based on experiments in the ance in the pathogenesis of psoriasis, none of these cells is margin of spreading psoriatic plaques and during relapse after psoriasis specific. Bovenschen et al.4 found that in plaque psor- treatment. Finally, a model for the evolution of the psoriatic iasis, substantially fewer activated lesional CD4+, CD8+ and lesion will be presented. CD45RO+ T cells are present in the dermis as compared with
4 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 5 lichen planus. In contrast, psoriatic epidermis exhibits a pro- Full activation of the T cell following interaction with the nounced CD8+ epidermotropism, which was associated with APC requires three signals: (i) TCR stimulation, (ii) stimula- epidermal hyperproliferation and abnormal keratinization.5 tion by costimulatory receptors (relevant receptor interactions These changes were less prominent in atopic dermatitis and and potential targets for antipsoriatic treatment are LFA3–CD2, lichen planus.4 Nickoloff et al. described interesting T-cell sub- B7–CD80 and CD28–CTLA41) and (iii) a third signal delivered sets (CD3+) bearing NK receptors on their surface (CD94, by various cytokines: for psoriasis IL-2 and IL-12 are of par- CD161), the so-called natural killer T cells, or NK-T cells.6 ticular importance.1 These cells are part of the innate immune system, as they do Neutrophils are present in ‘early’ and ‘active’ psoriatic not require previous sensitization for activation and have an lesions. In the mature psoriatic plaque neutrophils are more extremely restricted T-cell receptor (TCR) repertoire. CD94 sporadic. It is intriguing that accumulations of neutrophils may receptors are expressed on most NK cells and some T cells, adopt a pathognomonic morphology for psoriasis. Spongiform whereas CD161 is an obligate NK-T cell receptor. It has been pustules (micropustules of Kogoj) can be seen in the epidermis shown that a significantly higher number of NK-T cells is pre- and microabscesses of neutrophils in the stratum corneum, sur- sent in the psoriatic lesion in comparison with uninvolved or rounded by a parakeratotic stratum corneum (microabscesses normal skin. It is not clear whether these cells mainly initiate of Munro), are indicative for the diagnosis of psoriasis. Human psoriasis or have regulatory properties that might in fact sta- leucocyte-derived elastase is released from neutrophils and has bilize the psoriatic process.7 It has been suggested that NK-T been shown to induce keratinocyte hyperproliferation.14 cells interact with CD1d, expressed on keratinocytes in psoria- Monocytes and macrophages have been observed in the der- sis, leading to production of interferon (IFN)-c by NK-T mis. These cells express CD11a, CD14 and CD36. Activated cells.8 Various T-cell subsets with immunoregulatory functions macrophages are situated just underneath the basement mem- have recently been described. Intriguingly, these cells are brane.15 Protrusions of these cells may be in direct contact CD4+ T cells that constitutively express CD25 and the tran- with keratinocytes and stimulate keratinocytes via IL-6 and scription factor Foxp3, the so-called regulatory T cells (Treg). IL-8. A large body of evidence, derived both from in vitro and in vivo The microvasculature in the psoriatic lesion shows dilated studies, suggests that Treg have suppressive capacity via a tortuous capillaries. Adhesion molecules, including E-selectin, mechanism that requires cell–cell contact and the secretion of intercellular adhesion molecule-1 (ICAM-1) and vascular cell anti-inflammatory cytokines [e.g. IL-10, transforming growth adhesion molecule-1 (VCAM-1) are upregulated on endothel- factor (TGF)-b]. Because CD4+ CD25+ Treg are able to abro- ial cells of the psoriatic lesion and interact with LFA1 gate immune responses, they have recently been under scru- (CD11a/CD18) present on all leucocytes and Mac-1 (CD11b/ tiny in several disease entities of (auto-)immunological origin, CD18) present on monocytes, eosinophils and neutrophils.16 such as systemic lupus erythematosus, diabetes mellitus and In the stroma several changes have been reported. In lesion- myasthenia gravis. In these diseases Treg were found to be al and clinically uninvolved skin an increase of glycosamino- clearly dysfunctional. Recently, we were able to reveal CD4+ glycans was observed.17 Tenascin is an extracellular matrix CD25+ Foxp3+ Treg in the psoriatic dermis, but not in the molecule which has been shown to have increased expression dermis of normal skin, using multicolour immunofluorescent if the epithelium is activated, such as in psoriasis, skin cancers staining.3 Moreover, Sugiyama et al. reported the finding of an and wound healing.18 aberrant suppressive function of Treg, both in peripheral Epidermal proliferation in the psoriatic lesion is character- 9 blood and in psoriatic lesional skin. ized by an increase of cycling cells from the resting G0 popu- In psoriasis antigen-presenting cells (APCs) are of import- lation. Cell cycle times in the lesional epidermis are essentially ance for activation of T cells.10 CD1a+ Langerhans cells normal. By double labelling using anti-5-bromo-2¢-deoxyuri- (HLA-DR+), CD1a– dendritic cells (HLA-DR+), HLA-DR+ dine (BrdUrd) and Ki-67 it was shown that the number of keratinocytes and dermal dendrocytes are all APCs that may be DNA-synthesizing cells and the number of cycling cells recrui- relevant to psoriasis. Dendritic cells are activated in the psori- ted from the G0 compartment are increased but the ratio of atic lesion. Furthermore, inflammatory dendritic epidermal BrdUrd+ cells to Ki-67+ cells proved to be the same as in cells and plasmacytoid dendritic cells have been shown in pso- normal skin.19 This was confirmed by using a histone probe riatic lesional skin.10,11 T-cell activation may follow classical to calculate the number of cells in S phase.20 In an earlier antigen presentation by APCs to the TCR. Indeed, clonal study, Bata-Csorgo et al. observed increased responsiveness of expansion of T cells and repetitive TCR rearrangements, which b1 integrin-positive K1/K10-negative psoriatic keratinocytes are conserved within one patient during many years, suggest vs. b1 integrin-positive K1/K10-expressing normal keratino- that specific antigen stimulation is relevant to the pathogenesis cytes following stimulation with supernatants of psoriatic of psoriasis.12 T cells in the psoriatic lesion preferentially use T-cell clones.21 It remains to be resolved whether stem cells Vb genes, which indicates clonality of T cells and suggests that (bright cells) or transit amplifying cells were hyper-responsive (super)antigens are involved in the pathogenesis of psoriasis, in this study. Supernatants of the T-cell clones had high levels such as stratum corneum autoantigen, human papillomavirus of granulocyte/macrophage colony-stimulating factor and 5, streptococcal superantigen and keratin 17 (for review see IFN-c, low levels of IL-3 and IFN-a, and variable levels of Bos et al.13). However, the antigens have not been identified. IL-4. Anti-IFN-c counteracted the growth-promoting effect of
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 6 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof
Table 1 Molecules that are expressed in the psoriatic plaque but are dense aggregates and can be regarded as a late marker for ter- absent or have a restricted expression in normal skin minal differentiation. Several proteins are expressed during the differentiation of the suprabasal compartment with protein Molecule Reference precursors of the cornified envelope, for example involucrin, CD1d 8 envoplakin and periplakin. These proteins are early markers of Complement proteins C3a, C4a, C5a, C4d, 17,22,23 terminal differentiation. Transglutaminases are an important Bb, C5b-9 group of enzymes in the cornified envelope. These enzymes Keratin 6, 16 and 17 24 cross-link the above-mentioned precursors together with other Skin-associated antileucoprotease 25 protein such as loricrin, members of the small proline-rich Psoriasis-associated fatty acid-binding protein 26 Psoriasin 27 family, cystatin a, elafin, desmoplakin late envelope proteins Transforming growth factor-a 28 and cytokeratins to form the cornified envelope. The process Amphiregulin 29 of keratinization results in a stratum corneum with corneo- Epidermal growth factor receptor 30 cytes without a cell nucleus filled with proteins (keratins and Interleukin-1ra 31 filaggrin) surrounded by a cornified envelope and a lipid Interleukin-1b 32 bilayer, which is formed by lipids extruded from the cell Interleukin-6 32 envelope.40 Interleukin-8 33 Growth-related oncogene a/b/c 34 Cytokines are a large family of extracellular protein media- Fibronectin 35 tors. These include ILs, colony-stimulating factors, IFNs, b-defensin-2 36 tumour necrosis factors (TNFs), chemokines and growth fac- b-defensin-3 36 tors. A large series of cytokines has been demonstrated in the Cathelicidin 37 psoriatic lesion. Increased secretions of cytokines are relevant TLR1 38 for the perpetuation of the lesion. Cytokines are secreted by TLR2 38 various cell types. In the psoriatic lesion, increased production TLR5 38 HSP27 39 of type 1 T-cell cytokines has been reported, including IFN-c, 36 HSP60 39 instead of type 2 T-cell cytokines such as IL-4. Supernatants 37 HSP70 39 of psoriatic lesional keratinocytes can activate CD4+ T cells. T-cell supernatants induce epidermal proliferation, indicating TLR, toll-like receptor; HSP, heat shock protein. the complexity of cytokine signalling in psoriasis.21 Of par- ticular interest are the molecules IL-1a, IL-6, IFN-c, TGF-a, the supernatants, which is at variance with the well-estab- keratinocyte growth factor (KGF) and vascular endothelial lished growth-inhibiting effect of IFN-c. The suprabasal com- growth factor (VEGF), which show increased expression in partment of the epidermis comprises keratinocytes in various the psoriatic lesion; transgenic mice that have been developed phases of differentiation. The entire population of keratino- with overexpression of these molecules show a psoriasiform cytes of the psoriatic lesion is different from keratinocytes of phenotype. TNF-a, IL-12 and IL-20 show increased expression symptomless psoriatic skin or of normal skin. In the psoriatic in psoriatic lesional skin and these molecules are of particular lesion several molecules have increased expression but are importance as therapeutic targets. absent or have a restricted expression in symptomless skin (Table 1).8,17,22–39 The spectrum of proteins illustrates that Uninvolved skin psoriatic epidermis is abnormal with respect to many func- tions, in particular inflammation control. These molecules The micromorphological picture of the distant uninvolved skin, comprise cytokines and chemokines such as IL-8, melanoma as visualized by haematoxylin and eosin staining, is essentially growth-stimulatory activity/growth-related oncogene and normal. However, slightly increased numbers of CD4+ and complement. The host defence proteins, toll-like receptors CD8+ cells have been observed.5,41 Also, small numbers of (TLRs) and heat shock proteins (HSPs), are part of innate HLA-DR+/CD1a) APCs have been reported.42 Furthermore, immunity and are overexpressed in the psoriatic lesion. Host dendritic cells in the uninvolved skin of patients with psoriasis defence proteins that are highly expressed in lesional psoriatic already seem to be activated, as shown in oligonucleotide array skin include b-defensin (HBN)-2, HBN-3 and cathelicidin.36,37 analysis.43 In the lesional psoriatic skin, genes related to TLR1, TLR2 and TLR538 are overexpressed in lesional skin immune responses and epidermal proliferation were upregu- of patients with psoriasis. HSP27, HSP60 and HSP70 and one lated.43 Furthermore, foci of CD11b+ cells have been repor- of their ligands CD91 have been reported to be increased ted.44 The endothelium of the distant clinically uninvolved skin in lesional psoriatic skin.39 The process of keratinization is has been shown to have some abnormalities: increased presence abnormal in many respects. In the epidermis filaggrin is of high endothelial venules, decreased capillary resistance, and expressed, which expression is marked, in particular in the decreased endoglin expression.45–47 stratum granulosum, just underneath the stratum corneum. In the dermis increased proliferation and glycosaminoglycan Filaggrin, which shows a decreased expression in lesional pso- secretion have been reported.48 Compared with the skin of riatic skin, is involved in packaging of keratin filaments into healthy persons, the distant uninvolved skin showed a relative
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 7
Table 2 Changes in symptomless skin. For review see van de in psoriasis patients9,54 with increased numbers of NK-T cells 50 Kerkhof in lesional skin7 and decreased numbers of these cells in the peripheral blood.52 In psoriasis, the number of peripheral Epidermis blood cells expressing NK-T cell markers CD16, CD56, CD94 52 Thymidine incorporation I or CD158a is reduced as compared with healthy controls.
% S/G2/M phase I This finding corresponds with the results of similar studies in Ki-67+ nuclei N autoimmune disease, possibly suggesting a defect in the mech- Involucrin+ cells N anisms of ‘self-tolerance’ in patients with psoriasis. Filaggrin+ cells N Increased amounts of several type 1 cytokines have been Protein kinase C activity D 13 Ornithine decarboxylase activity I found in peripheral blood of patients with psoriasis. Expres- Arachidonic acid release I sion of IL-1, IL-6, IL-8, IL-12, TNF-a and IFN-c was found to
Phospholipase A2 activity I be augmented. But also IL-10, an anti-inflammatory cytokine, Calmodulin N/I was shown in increased amounts in peripheral blood of patients with psoriasis. This overproduction of proinflamm- N, normal; I, increased; D, decreased. atory cytokines, induced by activated T cells, is thought to stimulate the proliferation of keratinocytes and, via this pathway, seems to be of importance in the development of decrease of the surface of cell layers expressing b1 integrin as psoriatic plaques. a percentage of the total surface of the epidermis. There was, The number of circulating neutrophils in psoriasis was however, an increase in the absolute surface of b1 integrin- found to be increased in white blood cell counts.55 Further- expressing cells, mostly consisting of cells expressing b1 inte- more, elevations of elastase and transferrin levels have been grin with a weak intensity, so-called b1 integrin-dim cells, reported. As these are markers of neutrophil activation, this most likely comprising transit amplifying cells. Also keratin indicates a role for neutrophils in the inflammation in psoria- 15, normally expressed in basal keratinocytes but downregu- sis. However, it has to be evaluated if this role is a cause or lated in activated keratinocytes, was absent in the distant unin- an effect of the disease. volved skin as compared with the skin of healthy controls.49 The correlation between peripheral blood (systemic) and In many respects the epidermis of distant uninvolved skin skin (local) parameters in psoriasis has only rarely been inves- shows abnormalities (Table 2).50 In the distant uninvolved tigated. It is hypothesized that changes in both compartments skin slight, but statistically significant, abnormalities can be are the consequence of a recompartmentalization of cells in seen at many levels. This strongly suggests that a multisystem psoriasis. abnormality exists in the earliest phase of psoriasis, most likely 13 a dysregulation of innate immunity. Animal models for psoriasis
Various animal models have been developed and have provi- Peripheral blood ded important information on the pathogenesis of psoriasis. Several abnormalities have been observed in peripheral blood of patients with psoriasis. Although data on T cells are not Spontaneous mouse mutations unanimous, there is accumulating evidence pointing towards this cell type as one of the key players in the pathogenesis of In spontaneous mouse mutations some psoriasis-like pheno- psoriasis, both systemic and local. It has been suggested that cir- types have been described. Mice homozygous for the asebian culating T cells are activated and subsequently recruited from mutation share some characteristics with psoriasis such as epi- the circulation during the development of psoriatic plaques.5 dermal acanthosis, increased vascularity and an infiltrate of In psoriasis, there is an apparent deviation of T-cell differ- macrophages and mast cells.56 However, the characteristic entiation towards type 1 helper T cells and cytotoxic T cells. inflammatory events, such as intraepidermal accumulations of This is reflected by the increased percentages of T cells expres- neutrophils, do not occur in this model. Furthermore, the sing activation markers in peripheral blood.51–53 Elevated model proved to be of limited use in screening of some top- levels of CD45RO+, CD4+ CD45RO, CD4+ CD25+, CD8+ ical antipsoriatic treatments. Mice homozygous for chronic CD25+, CD4+ CD54+, CD4+ HLA-DR+ and CD8+ HLA- proliferative dermatitis (cpd) or flaky skin (fsn) mutations DR+, as compared with healthy controls, have been reported. have a more psoriasiform appearance with epidermal acantho- The level of cutaneous lymphocyte-associated antigen (CLA)- sis, focal parakeratosis, increased vascularity and accumulations expressing T cells in psoriasis was also found to be increased. of neutrophils in the epidermis.57,58 Although corticosteroids CLA is expressed on skin-homing cells, indicating that in psor- improve the skin abnormalities in the flaky skin, ciclosporin iasis more cells are intended to migrate to the skin, in com- has no effect in this model, which is in line with absence of a parison with healthy subjects. significant role of T cells in this model. These models may be Additionally, a loss of function of CD4+ CD25(high) useful to study some aspects of the pathogenesis of psoriasis CTLA4+ Foxp3(high) cells, the so-called Treg, has been found (epidermal proliferation, parakeratosis, increased vascularity,
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 8 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof neutrophil accumulation). The value of these models to study during the first 5 months. Lesions can be elicited by incisions the pathogenesis of psoriasis is limited as T cells do not play a (Koebner phenomenon). Older animals have dramatic lesions. significant role in these models. Furthermore, the aspects of Epidermal proliferation, parakeratosis, T-cell accumulation the pathogenesis of psoriasis which are expressed in these (predominantly CD4+ cells), neutrophil accumulation, mast spontaneous mouse mutations do not provide a reliable cells and monocytes/macrophages and increased vascularity screening model for antipsoriatic treatments. are observed. With respect to endothelial changes, E-selectin, VCAM-1 and ICAM-1 are expressed, compatible with psoriasis. VEGF Trap, a potent VEGF antagonist, proved to reverse the Transgenic mice with targeted cytokine expression psoriatic phenotype. Transgenic mice with targeted cytokine expression have provi- K14/IL-20 transgenic mice show a phenotype characterized ded important new insights in the pathogenesis of psoriasis. by a wrinkled hyperkeratotic skin. The animals live for only a Cytokines with a major relevance to the pathogenesis of psor- few days. Hyperkeratosis and epidermal proliferation without iasis, which have been overexpressed in transgenic mice, immune infiltrates are observed. Therefore, this phenotype is are IL-1a, IFN-c, IL-6, TGF-a, KGF, TGF-b, VEGF, IL-20 and not comparable with psoriasis although IL-20 is dramatically amphiregulin. upregulated in psoriasis and stimulates signal transduction Transgenic mice with overexpression of K14/IL-1a show through signal transducer and activator of transcription infiltrates of macrophages and monocytes in the dermis.59 Epi- (STAT)3.66 dermal hyperproliferation, some parakeratosis, T-cell infiltrates K14/amphiregulin transgenic mice approach a psoriatic and intraepidermal neutrophil accumulations can be seen only phenotype.67 Amphiregulin is part of the epidermal growth in some severely affected animals bearing psoriasiform lesions. factor (EGF) family and is an autocrine growth factor for No information is available on antipsoriatic treatments. Trans- human keratinocytes. The animals have a shortened life span. genic mice with high expression of involucrin/IFN-c show, Prominent scaling, erythema and papillomatosis suggest a in up to 20% of the animals, a psoriasiform phenotype with a psoriasis-like phenotype. Focal parakeratosis, acanthosis, der- marked increase of epidermal proliferation, parakeratosis, in- mal and epidermal infiltration of neutrophils and lymphocytes creased vascularity, T cells (exclusively dermal) and occasion- as well as tortuous dilated capillaries are suggestive for psoria- ally microabscesses of neutrophils.60 No information is sis. Also neutrophil accumulations as micropustules of Kogoj available on the effect of antipsoriatic treatments. and microabscesses of Munro have been observed in this ani- Transgenic mice expressing K14/IL-6 and K14/TGF-a show mal model. However, the typical psoriatic architecture with no or only marginal changes of keratinocyte proliferation and elongated epidermal ridges is lacking. No effects of antipsori- differentiation.61,62 Only some severely affected K14/TGF-a atic treatments have been observed in this model. mice show a psoriasis-like phenotype with increased epider- Transgenic mice with targeted cytokine expression have mal proliferation, accumulation of T cells, neutrophil micro- provided important information on the pathogenesis of abscesses and increased vascularity. No information is available psoriasis. A relatively poor induction of the psoriatic pheno- on the effects of antipsoriatic treatments in these models. type was induced by IL-1a, IFN-c, IL-6, TGF-a, KGF, TGF-b, K14/KGF transgenics show some abnormalities in epider- IL-20 and amphiregulin. This by no means excludes the rele- mal growth and differentiation but do not show significant vance of these cytokines in the pathogenesis of psoriasis as inflammatory changes.63 these cytokines may well contribute to the pathogenesis of Transgenic mice with overexpression of K10/bone morpho- psoriasis, although not as a single driving force. Furthermore, genic protein-6 (BMP-6) provided intriguing information.64 the relatively low penetrance of the psoriasiform phenotype BMP-6 is a member of the TGF-b superfamily. Strong and makes these models less practical as screening models. The homogeneous expression was associated with reduced epider- K14/VEGF transgenic mice, however, show features closely mal proliferation. However, weaker discontinuous expression resembling psoriasis, including the Koebner phenomenon and resulted in marked hyperproliferation, parakeratosis, increased the characteristic epidermal psoriatic architecture with elonga- vascularity, T-cell infiltrates and neutrophil microabscesses. tion of rete ridges. The continuous overexpression of VEGF is Psoriasiform lesions were observed only in some of these ani- likely to be of primary significance in the pathogenesis of mals. No information is available on the effects of antipsoriatic psoriasis. treatments in these animals. K14/VEGF transgenic mice develop a phenotype in 100% of Modifications of integrin expression the animals which shows striking similarities to the psoriatic lesion.65 In particular, the typical architecture of the epidermis Modifications of integrin expression can induce psoriasiform with elongations of rete ridges, the microabscesses of Munro phenotypes in animals. When a hypomorphic mutation in and micropustules of Kogoj and the positive Koebner phe- CD18 was back-crossed on to the PL/J inbred strain, all nomenon are entirely compatible with psoriasis and have not homozygous mice developed a chronic inflammatory skin dis- been described in other transgenic models overexpressing ease with a mean age at onset of 11 weeks after birth.68 The cytokines. The young K14/VEGF transgenic mice overexpress animals develop erythematous scaly patches which are charac- VEGF and will spontaneously develop psoriasiform lesions terized histologically by psoriasiform hyperplasia with elonga-
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 9 tion of rete ridges, parakeratosis, subcorneal accumulations of expression.72 However, 10 weeks after transplantation, there neutrophils, T cells (CD4 and CD8) in the dermis and epider- was some decrease of HLA-DR expression and ICAM-1 mis, macrophages and some mast cells. The lesions resolve expression73 and after 22 weeks there was a decrease of following treatment with a topical corticosteroid. Depletion of parakeratosis, Munro’s microabscesses and infiltration of CD4+ cells by depleting monoclonal antibodies to CD4+ cells lymphocytes. Intradermal or intravenous injections of T cells resulted in a complete clearing of the psoriasiform derma- from psoriatic plaques, but not from peripheral blood, main- titis.69 Such was not observed with antibodies to CD8+ cells. tained all pathological changes for longer.74 Induction of a This study suggests that a reduced expression of CD18 results typical psoriatic phenotype was observed when autologous in lesions compatible with psoriasis, which respond to corti- peripheral blood mononuclear cells (PBMC), activated with costeroid treatment. In involucrin/b1, a2b1ora5b1 trans- staphylococcal enterotoxins and IL-2, were injected into genic mice, psoriasiform lesions appeared in 36–89% of the grafts of nonlesional skin from patients with psoriasis.75 All animals.70 The lesions are characterized by marked epidermal histological characteristics of a psoriatic lesion were induced, proliferation, focal parakeratosis, increased vascularity, T-cell including elongation of rete ridges, neoangiogenesis and infiltrate and neutrophil microabscesses, although only in epidermal expression of HLA-DR, ICAM-1, b integrins and severe cases. The typical psoriasiform architecture with elonga- involucrin.75 Furthermore, it was shown that only autolo- tion of rete ridges was not seen. Although the lesions only to gous activated CD4+ cells from peripheral blood and not the some extent approach psoriasis it is important that b1 integrin CD8+ cells induce the psoriatic phenotype in nonlesional expression may induce epidermal proliferation and inflamma- psoriatic skin grafted on SCID mice.76 However, CD4+ cells tory changes as seen in psoriasis. stimulated resident CD8+ cells (CD69+ and CD25+) and, furthermore, CD94+ and CD161+ cells, which are NK cells, appeared in the grafts.76 It was shown that a CD94+/ Histocompatibility mismatched naive CD4+ T cells CD161+ T-cell line, established in a patient with psoriasis, Reconstitution of scid/scid mice with minor histocompatibi- could induce a psoriatic plaque on nonlesional psoriatic skin lity mismatched naive CD4+ T cells resulted in skin lesions grafted on SCID mice.6 Furthermore, it was shown that both with striking similarity to psoriasis in all animals, including hyperplastic human and murine keratinocytes expressed the marked epidermal proliferation, parakeratosis, a psoriasiform major histocompatibility complex class I-like CD1d protein, epidermal architecture, increased vascularity, mast cells, T cells which has been shown to be a specific ligand of cells and neutrophil infiltrates with microabscess formation.71 The expressing NK cell receptors.77 lesions disappeared by coinjection of memory T cells. Ultravi- Injection of PBMC following stimulation by superantigen olet B and ciclosporin treatment improved the lesions mark- in nonlesional psoriatic skin resulted in the development of edly. It is important that CD4+ cells may induce a psoriatic a psoriatic lesion.78 Activated allogeneic PBMC from patients phenotype in the absence of epidermal abnormalities and with psoriasis have been shown to induce the full psoriatic CD8+ cells. phenotype in transplanted normal skin from healthy individ- uals on SCID mice.76 However, another group did not suc- ceed in inducing the psoriatic phenotype in normal skin Xenotransplantation models transplanted on SCID mice.78 Within 2–4 weeks following As psoriasis is a polygenic disease, it is unlikely that it can be injection of immunocytes in grafted symptomless skin, created by manipulation of one single gene. In order to microvascular endothelial cells are activated and become approach the psoriatic phenotype as closely as possible, xeno- hyperproliferative and express avb3 and endothelial leuco- transplantation models have been developed. cyte adhesion molecule-1, similar to the situation in psoria- tic plaques.6 A series of antipsoriatic treatments has been shown to reduce psoriatic lesions grafted into SCID mice: Nude mouse dexamethasone, clobetasol propionate, ciclosporin, 1a,25-di- 79–81 The first xenotransplantation model was in the nude mouse, hydroxyvitamin D3 and efalizumab. Interestingly, treat- which lacks the T-cell arm of the immune system. However, ment of the peripheral blood-derived immunocytes the transplants were not stable and psoriatic lesions lost their pretreated with ciclosporin or 1,25-dihydroxyvitamin D3 characteristics in a few weeks, whereas normal skin became also rendered the immunocytes incapable of inducing the hyperproliferative following transplantation. psoriatic phenotype in psoriatic uninvolved skin transplanted to nude mice.82 Although this xenograft model creates a lesion which con- Severe combined immunodeficient mice tains the major characteristics of the psoriatic lesion, the Severe combined immunodeficient (SCID) mice lack both hu- model has some limitations. The intradermal injection of stim- moral and cellular immunity, and psoriatic plaques clinically ulated immunocytes remains artificial and does not permit and histologically retain their characteristics following trans- study of the activation of the immunocytes. The Koebner phe- plantation on SCID mice for several months, including the nomenon cannot be studied. Furthermore, no arthritis occurs composition of CD4 and CD8 cells, ICAM-1 and VCAM-1 in this model.
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 10 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof
been shown to drive the process of T-cell activation and Xenograft model using AGR 129 mice proliferation. A further development in the xenotransplantation model is a xenograft model using AGR 129 mice.83 Due to a defici- Transgenic mice with overexpression and deletion ency in type I (A) and type II (G) IFN receptors these mice of signal transducers have immature NK cells and furthermore they lack T and B cells.83 Grafts of clinically uninvolved skin of 12 patients Transgenic mice with overexpression and deletion of signal with plaque psoriasis were transplanted on to AGR 129 transducers have provided important new information on the mice. A full psoriatic phenotype developed in 28 of 31 pathogenesis of psoriasis. STATs are cytoplasmatic proteins grafted mice within 8 weeks, with epidermal proliferation, which transmit external signals to the nucleus. Stat3 is acti- focal parakeratosis, elongation of rete ridges and dermal vated by phosphorylation of Tyr 705 followed by dimerization papillae, increased vascularity and numerous mononuclear and nuclear translocation. Stat3 is involved in signal transduc- cells. The xenograft model using AGR 129 mice clearly tion related to proliferation and has been studied during demonstrates proliferation of T cells, already present in the wound healing. More recently, it was shown that Stat3 is acti- symptomless skin of patients with psoriasis, with CD8 cells vated with nuclear staining for phosphorylated Tyr 705 Stat3 and CD4 cells predominantly in epidermis and dermis, (pY-Stat3) in lesional psoriatic skin and symptomless psoriatic respectively, corresponding with the distribution pattern in skin adjacent to the lesion.87 Furthermore, it was shown that the psoriatic lesion. The proliferation of T cells was pre- other inflammatory conditions (chronic dermatitis, prurigo dominantly in the dermis and peaked at 4 weeks. Injection and lichen planus) did not show the activated staining pattern. of anti-CD3 monoclonal antibodies blocked development of Transgenic mice expressing a constitutively active Stat3 the psoriatic lesion. Furthermore, anti-TNF-a treatment also (K5.Stat3C mice) proved to develop, either spontaneously or prevented the development of psoriatic lesions in this following wounding, skin lesions which closely resembled model. Although T cells appear early in the development of psoriasis.87 This phenotype was expressed in all animals. These psoriasis in this model and although anti-CD3 inhibits the mice had increased levels of Stat3 and pY-Stat3. These trans- development of the lesion, this model does not prove that genic mice were normal at birth, except for a thicker skin T cells are the only driving force in the pathogenesis of with a more prominent subcutaneous vasculature. By the age psoriasis. In a later study the same group provided of 2 weeks already hyperkeratotic, erythematous lesions evidence, using the same animal model, that plasmacytoid appeared, histologically characterized by acanthosis with predendritic cells (PDCs) drive the T-cell proliferation via elongation of rete ridges, focal parakeratosis, a dense dermal IFN-a production.84 PDCs are key to innate immunity but inflammatory infiltrate, increased numbers of capillaries and are capable of driving Th1 responses, affecting also adaptive intraepidermal accumulations of neutrophils. Ki-67 expression immunity. The TLR7 agonist imiquimod has been shown to was increased also in the suprabasal cell layers and keratin 6 produce massive amounts of IFN from TLR7-expressing was expressed at the expense of keratin 10 expression, entirely PDCs, while aggravating psoriasis.85 PDCs have been compatible with the classical psoriatic lesion. Following full- observed in psoriatic lesions and IFN-a may aggravate psor- thickness wounds scaly erythematous lesions developed, again iasis. Human IFN-a mRNA expression levels were increased with the histological characteristics of the psoriatic lesion. Pso- already at 7 days after transplantation. Whereas T cells riatic lesions also resulted from tape stripping and application expanded in parallel to the increases in IFN-a, epidermal of a phorbol ester. Keratinocytes of the K5.Stat3C mice hyperplasia appeared much later. Neutralizing antibodies to showed elevated mRNA levels for Stat3 but also upregulation IFN-a/b decreased and recombinant IFN-a again induced of several psoriasis-related genes: Vegfa, Tgfa, Icam1 and Nfkbia. activation and expansion of pathogenic T cells in the AGR Already in the newborn mice Vegfa was increased twofold, 129 model, indicating that in this model IFN-a is a driving which was compatible with the more prominent subcutaneous force of T-cell proliferation. Further studies revealed that microvasculature in the newborn. It is of interest that trans- early, 4 weeks after engraftment, CD91-expressing cells genic mice with increased expression of VEGF also displayed accumulated in symptomless skin engrafted on AGR 129 psoriatic lesions and also showed the Koebner phenomenon.87 mice.86 CD91 was expressed predominantly by APCs show- However, the spontaneous expression of psoriasis in the VEGF ing activation of nuclear factor-jB (NF-jB) signalling and transgenics was much later, which suggests that at least other expressing TNF-a. In psoriatic lesions CD91-bearing APCs Stat3-dependent genes are involved in the induction of the were observed in close vicinity to HSP70-expressing kera- psoriatic phenotype, which perhaps include Nfkbia, Ccnd1 and tinocytes. It is attractive to speculate that HSPs are released Icam1. Skin grafts from K5.Stat3C mice were transplanted on from keratinocytes by trauma or infection, activating APCs, athymic nude mice and SCID mice. Tape stripping following releasing TNF-a. The AGR 129 mouse has been shown to grafting did not produce the psoriatic phenotype. Simulta- provide an environment for transition of psoriatic unin- neous injection of activated T cells from the spleen of volved to lesional skin. The model has shown that T-cell K5.Stat3C mice was required to induce psoriatic lesions in proliferation is a prerequisite for the transformation in this these animals. T-cell subsets showed a distribution similar to model and so far APCs as well as IFN-a and TNF-a have the psoriatic lesion. Remarkably, these grafts on SCID mice,
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 11 showing psoriatic lesions, had nuclear pY-Stat3 staining observation that treatment of the double-mutant Jun B/c-Jun whereas nonconverted grafts failed to have this. Topical treat- mice with a broad-spectrum antibiotic significantly delayed the ment with a Stat3-specific decoy oligonucleotide inhibited the onset of psoriatic lesions, which is in line with the clinical tape stripping-induced generation of psoriatic lesions in observation that psoriatic lesions can be triggered by bacterial K5.Stat3C mice. Several cytokines and growth factors may acti- infections in humans. vate Stat3 in skin: IL-6, EGF, hepatocyte growth factor, IL-20, NF-jB(IjB) is an important part of signal transaction rela- IFN-c. Therefore, activation of Stat3 may occur during wound ted to inflammatory processes. By epidermis-specific deletion healing and inflammatory processes with activation of T cells. of inhibitor of NF-jB(IjB) kinase 2 (IKK2) a skin phenotype Jun B and c-Jun are components of AP-1 transcription factor is induced which approaches the clinical and histological hall- and are involved in regulation of cell proliferation, differenti- marks of psoriasis to a large extent.89,90 By specific elimin- ation, stress responses and cytokine expression. In the psoriatic ation of inflammatory cells, skin macrophages proved to be lesion, Jun B expression has been reported to be markedly obligatory in the induction of the psoriatic phenotype. Indeed, decreased, whereas c-Jun expression is increased as compared when K14-Cre/IKK2FLFL mice were crossed with TCRa-knock- with normal skin.88 Inducible, conditional, single and double out mice the K14-Cre/IKK2FLFL/TCRa)/) mice developed the knockout mice for Jun B and c-Jun were designed.88 Single same skin lesions with similar severity, which eliminate ab T knockout mice for jun B or c-Jun did not show any abnormal lymphocytes as obligatory for the disease. Elimination of skin phenotype. However, lesions closely resembling psori- macrophages by subcutaneous injection of clodronate lipo- asis were observed 8–10 days after tamoxifen induction of Jun somes resulted in resolution of the lesions. As targeted dele- B/c-Jun double-mutant mice. The expression of this psoriasis- tion of CD18 did not result in an improvement of the lesions, like phenotype was seen in all double mutants. The pheno- granulocytes are not obligatory to maintain the psoriatic typical characteristics of the lesions were acanthosis with lesion. Targeted deletion of the receptor for IFN-c revealed prominent rete ridges, parakeratosis, increased subepidermal that classical IFN-c-mediated macrophage activation is not vascularization, intraepidermal T cells, intraepidermal accumu- involved in the induction of the lesions in this model. This lation of neutrophils and increased numbers of macrophages in model provides evidence for macrophages as a driving force the dermis. Most importantly, arthritic lesions with features of in the pathogenesis of psoriasis. psoriatic arthritis were observed with bone destruction and periostitis. The expression of various cytokines in the lesions in The development of the psoriatic lesion these animals approached the psoriatic lesion: upregulation of in patients with psoriasis IL-1a, IL-1b, IFN-c, TNF-a, macrophage inflammatory protein (MIP)-2 (IL-8 in humans), MIP-1a, MIP-1b, IFN-inducible To study the development of the psoriatic lesion in patients protein-10, monocyte chemoattractant protein (MCP)-1, with psoriasis, several models are available, which all have
IL-12p40 and TGF-b2, and downregulation of IL-18. Further- their opportunities and limitations. more, the following molecules were upregulated in the In the first model, the artificial elicitation of psoriatic double-mutant Jun B/c-Jun animals: chemotactic proteins lesions following injury of the symptomless skin induces a S100A8 and S100A9, epidermal fatty acid binding protein, psoriatic lesion in 25% of patients with psoriasis.91 This phe- secretory leucocyte protease inhibitor and calmodulin, whereas nomenon proved to be an ‘all or nothing’ phenomenon; serial keratin 15 and caveolin were significantly reduced in mutant challenges revealed that a patient will develop psoriatic lesions epidermis. These changes were entirely compatible with the at all challenged sites or at none of the challenged sites. Fur- psoriatic lesion. Three days following the Jun B and c-Jun dele- thermore, it was shown that a dermal injury (intradermal tion the chemotactic proteins S100A8 and S100A9 were injections of chymotrypsin) never resulted in a psoriatic induced, so far in the absence of any other abnormality. These lesion: the epidermis has to be injured as well.90 Vice versa, proteins, mapped to the psoriasis susceptibility region PSORS4, removal of the stratum corneum by tape stripping of the epi- strongly induce chemotaxis by neutrophils. In order to ascer- dermis rarely results in psoriatic lesions. We have observed tain the role of T cells in this animal model, deletions of Jun B psoriatic lesions within stripped areas, confined to the biopsy and c-Jun were induced in Rag2-deficient mice. Rag2-deficient site. Another point of interest is the observation that the use Jun B/c-Jun double-mutant mice still developed psoriasis-like of a local anaesthetic, while causing vasoconstriction, signifi- lesions, although milder. This suggests that T and B cells do cantly reduced the occurrence of a positive Koebner response. not have a prominent role in the pathogenesis of the skin Several reports have indicated that pressure on the skin signifi- lesions. However, in these animals arthritis did not occur. cantly inhibits the development of a positive Koebner phe- Deletion of Jun B and c-Jun in mice deficient for TNFR1 devel- nomenon. Tape stripping has been used to study in particular oped a similar phenotype: skin lesions but no arthritis. These the epidermal growth and differentiation characteristics in experiments suggest that, in contrast to psoriatic arthritis, the normal subjects but also in symptomless skin of patients with development of psoriatic plaques is not fully dependent psoriasis. In normal subjects a regenerative response is seen, upon T cells and TNF-a. In xenograft studies no experiments 2 days later, characterized by epidermal proliferation, keratin have been performed where skin of healthy volunteers was 16 expression, parakeratosis, an increased number of involu- challenged by only T cells. A feature that is of interest is the crin-positive cells and a decreased number of filaggrin-positive
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 12 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof cells.92 Intraepidermal accumulation of neutrophils is seen third phase is the simultaneous appearance of Ki-67+ nuclei within 1 day and after 4 days mononuclear cells dominate.93 in the basal cell layer, accumulation of CD4+ cells and NK-T The symptomless skin of patients with psoriasis largely shows cells and increased expression of involucrin with decreased the same response; however, the proliferative response is expression of filaggrin. So far, the position of monocytes and enhanced.94 The accumulation of neutrophils is decreased in Langerhans cells has not been studied in this context. patients with stable disease and increased in patients with More recently it was shown that plasmacytoid dendritic unstable disease.93 The observations following standardized cells are increased in the ‘nearly’ symptomless skin, suggesting injury of the symptomless skin in patients with psoriasis sug- early involvement of this class of APCs.84 Furthermore, the gest that both epidermis and stroma are driving the pathogen- nearly symptomless skin showed Stat3 activation as identified esis of psoriasis and that both have to be affected in order to by nuclear staining of pY-Stat3, suggesting early involvement induce psoriatic lesions efficiently. The involvement of the of this pathway in psoriasis.87 endothelium is obligatory as vasconstriction inhibits the devel- Studies in the margin zone suggest that stroma, microvascu- opment of the lesion. Hyper-regeneration of the epidermis of lature and mast cells are involved early in the pathogenesis of patients with psoriasis is a constitutive abnormality and psoriasis. Endothelial changes are obligatory for the develop- increased or decreased neutrophil accumulation is highly vari- ment of the psoriatic lesion. Activated CD8+ cells and able depending on the activity of the disease. CD45RO+ cells are early invading T cells, whereas NK-T cells Studies during relapse of psoriasis following discontinuation are relatively late in the development of the lesion, although of treatment represent another model to obtain insight into the latter cells may induce psoriatic lesions in a xenograft the order of events. Schubert and Christophers described the model.6,77 It is of interest that both Stat3 and plasmacytoid histological changes in symptomless skin following discon- dendritic cells are observed in the ‘nearby symptomless skin’, tinuation of corticosteroid treatment.95 In these studies, the suggesting early involvement of these aspects, which have first changes were accumulation of monocytes and mast cells. been shown to be relevant to the pathogenesis of psoriasis in Mast cells may reflect an early role of the functioning of the animal models. microvasculature in the development of the lesion. With respect to the early appearance of monocytes it might be rele- Discussion vant that these cells line up close to the basal cells, in particu- lar under the tips of the ridges where epidermal cells express From the comparison between lesional psoriatic skin, symptom- MCP-1.96 MCP-1 is expressed on stimulation with IFN-c and less psoriatic skin and normal skin it is evident that a large reper- TNF-a. toire of abnormalities characterizes the psoriatic lesion and to a Before appearance of overt lesions following discontinuation lesser extent symptomless psoriatic skin and peripheral blood. of corticosteroid treatment, suprabasal expression of Ki-67 and The focus of the suggested leading abnormality has been an increased fraction of cells in S/G2/M phase in the supraba- strongly inspired by an upcoming new treatment principle: at sal layer can be seen, with complete epidermal proliferation in present an abnormality in innate immunity,13 inspired by the the pinpoint lesions.97,98 Studies following discontinuation of efficacy of anti-TNF-a, but previously an abnormality of treatment suggest that mast cells and monocytes are involved acquired immunity inspired by the success of ciclosporin and an early in the pathogenesis of psoriasis and that the suprabasal abnormality in epidermal proliferation and differentiation compartment anticipates full epidermal proliferation. inspired by methotrexate efficacy. Major questions are what are In a third model, the margin zone of the spreading psoriatic drivers in the pathogenesis of psoriasis, are these drivers plaque has been used to study the transition between symp- involved early or late in the pathogenesis of psoriasis and are tomless and lesional psoriatic skin.99 The changes in front of these abnormalities obligatory in the development of the lesion? the transition may be expected to represent the most primary Animal models of psoriasis have provided important infor- phenomena, whereas the changes closer to the lesion are likely mation on driving forces in the pathogenesis of psoriasis. to be secondary. Earlier studies have suggested a three-stage Animal models with a complete expression of the micromor- transition model. The first phase of transition from symptom- phology of the psoriatic lesions in a considerable percentage less to lesional skin involves the stroma and is characterized of the animals are supposed to harbour a driving force in the by increased expression of tenascin X and alkaline phosphatase pathogenesis of psoriasis. Such has been shown for the K14/ activity, indicative for endothelial cell activation. By means of VEGF transgenic mice, hypomorphic mutation of CD18, Scid/ Doppler flowmetry, it has also been shown that increased Scid mice with minor histocompatibility mismatched naive blood flow is an initial event in the development of the CD4+ T cells, activated CD4+ cells and NK-T cells injected in lesion.100 The second phase is the appearance of the inflam- xenografts from nonlesional psoriatic skin on SCID mice. Xen- matory infiltrate with CD8+ and CD45RO+ cells, CD2+ and ografts from psoriatic uninvolved skin on AGR 129 mice, CD25+ cells as early invaders.6 However, in distant un- missing type I and II IFN receptors and therefore with only involved skin already some increases of CD4+ cells have been immature NK cells and lacking T and B cells, have indicated reported. In the second phase of development suprabasal plasmacytoid dendritic cells and IFN-a to be an early abnor- expression of keratin 16 and suprabasal expression of b1 inte- mality in the development of the psoriatic lesions and in this grin-dim cells and Ki-67+ nuclei have been reported.49 The model these abnormalities were obligatory.
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 13
Transgenic mice expressing constitutively active Stat3 devel- 4 Bovenschen HJ, Seyger MMB, van de Kerkhof PCM. Plaque psoria- oped psoriatic lesions, and, interestingly, psoriatic lesions sis vs. atopic dermatitis and lichen planus: a comparison for le- could be induced by injury of the skin in these animals. Tam- sional T-cell subsets, epidermal proliferation and differentiation. Br J Dermatol 2005; 153:72–8. oxifen induction of Jun B/c-Jun double-mutant mice resulted 5 Baker BS, Swain AD, Fry L, Valdimarsson H. Epidermal T lympho- in psoriatic lesions and arthritis; when the same construction cytes and HLA-DR expression in psoriasis. Br J Dermatol 1984; was realized in Rag2-deficient mice, with impaired T and B 110:555–64. cells, psoriatic skin lesions still occurred, whereas the arthritis 6 Nickoloff BJ, Bonish B, Huang BB, Porcelli SA. Characterization of no longer developed. In animals with an epidermal deletion a T cell line bearing natural killer receptors and capable of cre- of NF-jB kinase 2 psoriasis lesions appeared with macrophages ating psoriasis in a SCID mouse model. J Dermatol Sci 2000; as crucial cells. In conclusion, animal models revealed the 24:212–25. 7 Cameron AL, Kirby B, Frei W, Griffiths CEM. Natural killer and following driving forces for psoriasis: VEGF, CD4+ cells, natural killer T-cells in psoriasis. Arch Dermatol Res 2002; 294:363– NK-T cells, IFN receptors, Stat3, Jun B/c-Jun and IKK2. Some 9. of these models revealed that another factor was also essential 8 Bonish B, Jullien D, Dutronc Y. Overexpression of CD1d by kera- in the development of the lesions, such as T cells in the Stat3 tinocytes in psoriasis and CD1d-dependent IFN-c production by model, whereas other models indicated that this factor was NK-T-cells. J Immunol 2000; 165:4076–85. less relevant: T cells in the Jun B/c-Jun model or NF-jB kin- 9 Sugiyama H, Gyulai R, Toichi E et al. Dysfunctional blood and ase 2 model. The variety of driving factors is compatible with target tissue CD4+ CD25high regulatory T cells in psoriasis: mechanism underlying unrestrained pathogenic effector T cell the polygenic inheritance of psoriasis and the heterogeneous proliferation. J Immunol 2005; 174:164–73. responses to antipsoriatic treatments. 10 Nickoloff BJ, Griffiths CEM. Lymphocyte trafficking in psoriasis: Studies on the development of psoriatic lesions in patients a new perspective emphasizing the dermal dendrocyte with provide some further information as to what extent the active dermal recruitment mediated via endothelial cells above-mentioned driving forces are early or late. Our informa- followed by intra-epidermal T-cell activation. J Invest Dermatol tion on these in vivo studies is limited so far. It has been 1990; 95:S35–7. shown, however, that endothelial involvement is early in the 11 Teunissen MBM. Langerhans cells and other skin dendritic cells. In: Skin Immune System: Cutaneous Immunology and Clinical Immunodermato- pathogenesis and that the appearance of NK-T cells is relatively logy (Bos JD, ed.), 3rd edn. Boca Raton, FL: CRC Press, 2005; late. However, Stat3 activation and plasmacytoid dendritic cells 123–82. have been observed in an early phase of the lesion. With 12 Prinz JC. Disease mimicry – a pathogenetic concept for T cell- respect to the question of which factors are obligatory, induc- mediated autoimmune disorders triggered by molecular mimicry? tion of psoriatic lesions in symptomless skin requires both Autoimmun Rev 2004; 3:10–15. dermal and epidermal injury, and vasoconstriction inhibits the 13 Bos JD, de Rie MA, Teunissen MBM, Piskin G. Psoriasis: dysregu- induction of psoriatic lesions. The response to highly selective lation of innate immunity. Br J Dermatol 2005; 152:1098–107. 14 Rogalski C, Meyer-Hoffert E, Proksch E, Wiedow O. Human treatments provides new information as to what extent some leukocyte elastase induces keratinocyte proliferation in vitro and factors are obligatory in the pathogenesis of psoriasis. Anti- in vivo. J Invest Dermatol 2002; 118:49–54. TNF-a treatments have shown that this key mediator in innate 15 van den Oord JJ, de Wolf-Peeters C. Epithelium-lining macro- immunity is obligatory. However, treatment resistance does phages in psoriasis. Br J Dermatol 1994; 130:589–94. occur in some patients, suggesting that TNF-a is not abso- 16 Smith CH, Barker JN. Pathogenic aspects of psoriasis: cell traffick- lutely obligatory. Even a more variable response has been ing and role of adhesion molecules in psoriasis. In: Pathogenic Aspects shown for anti-CD11a (efalizumab) and alefacept interfering of Psoriasis (de Kerkhof PCM, Bos JD, eds). New York: Elsevier, 1995; 151–60. with LFA3–CD2 and CD2–NK cells. This heterogeneity in 17 Priestly GC, Adams LW. Hyperactivity of fibroblasts cultured from responses reconfirms that psoriasis is a multifactorial polygenic psoriatic skin. I. Faster proliferation and effect of serum with- disease with variable involvement of driving forms. drawal. Br J Dermatol 1983; 109:149–56. For the future, development of biomarkers which identify 18 MacKie EJ, Halfter W, Liverani D. Induction of tenascin in healing responders prospectively will be an important improvement in wounds. J Cell Biol 1988; 107:2757–67. the treatment of psoriasis. 19 van Erp PEJ, de Mare S, Rijzewijk JJ et al. A sequential double im- munoenzymic staining procedure to obtain cell kinetic informa- tion in normal and hyperproliferative epidermis. Histochem J 1989; References 21:343–7. 20 Castelijns FACM, Gerritsen MJP, van Erp PEJ, van de Kerkhof PCM. 1 Krueger JG. The immunologic basis for the treatment of psoriasis Cell kinetic evidence for increased recruitment of cycling epider- with new biologic agents. J Am Acad Dermatol 2002; 46:1–23. mal cells in psoriasis: ratio of histone and Ki-67 antigen expres- 2 Bos JD, Hagenaars C, Das PK et al. Predominance of ‘memory’ T sion is constant. Dermatology 2000; 200:25–30. cells (CD4+, CDw29+) over ‘naı¨ve’ T cells (CD4+, CD45R+) in 21 Bata-Csorgo Z, Hammerberg C, Voorhees JJ, Cooper KD. Kinetics both normal and diseased human skin. Arch Dermatol Res 1989; and regulation of human keratinocyte stem cell growth in 281:24–30. short-term primary ex vivo culture. Cooperative growth factors 3 Bovenschen HJ, van Vlijmen-Willems IMJJ, van de Kerkhof from psoriatic lesional T lymphocytes stimulate proliferation PCM, van Erp PEJ. Identification of lesional CD4+ CD25+ among psoriatic uninvolved, but not normal stem, keratinocytes. Foxp3+ regulatory T cells in psoriasis. Dermatology 2006; J Clin Invest 1995; 95:317–27. 213:111–17.
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 14 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof
22 Pasch MC, Timar KK, van Meurs M et al. In situ demonstration of 42 Prens EP, Benne K, van Joost T, Benner R. The autologous mixed CD40- and CD154-positive cells in psoriatic lesions and keratino- epidermal cell–T lymphocyte reaction is elevated in psoriasis: a cyte production of chemokines by CD40 ligation in vitro. J Pathol crucial role for epidermal HLA-DR+/CD1a) antigen-presenting 2004; 203:839–48. cells. J Invest Dermatol 1991; 96:880–7. 23 Pasch MC, Bos JD, Asghar SS. Activation of complement in 43 Zhou X, Krueger JG, Kao M-CJ et al. Novel mechanisms of T-cell psoriasis. Clin Exp Dermatol 1998; 23:189–90. and dendritic cell activation revealed by profiling of psoriasis on 24 de Jong EMGJ, van Vlijmen IM, van Erp PEJ. Keratin 17: a useful the 63,100-element oligonucleotide array. Physiol Genomics 2003; marker in anti-psoriatic therapies. Arch Dermatol Res 1991; 13:69–78. 283:480–2. 44 Cooper KD, Baadsbaard O, Ellis CN et al. Mechanisms of cyclospo- 25 Schalkwijk J, Chang A, Janssen P et al. Skin-derived antileucopro- rine A inhibition of antigen-presenting activity in uninvolved and teases (SKALPs): characterization of two new elastase inhibitors lesional psoriatic epidermis. J Invest Dermatol 1990; 94:649–56. from psoriatic epidermis. Br J Dermatol 1990; 122:631–41. 45 Heng MC, Allen SG, Chase DG. High endothelial venules in 26 Siegenthaler G, Hotz R, Chatellard-Gruaz D et al. Characterization involved and uninvolved psoriatic skin: recognition by homing and expression of a novel human fatty acid-binding protein: receptors on cytotoxic T lymphocytes. Br J Dermatol 1988; the epidermal type (E-FABP). Biochem Biophys Res Commun 1993; 118:315–26. 190:482–7. 46 Kemeny L, Csato M, Nyiradi J et al. Decreased capillary resistance 27 Madsen P, Rasmussen HH, Leffers H et al. Molecular cloning, of the uninvolved skin in psoriasis. Acta Derm Venereol (Stockh) 1988; occurrence, and expression of a novel partially secreted protein 68:459–60. ‘psoriasin’ that is highly upregulated in psoriatic skin. J Invest 47 Rulo HF, Westphal JR, van de Kerkhof PCM et al. Expression of Dermatol 1991; 97:701–12. endoglin in psoriatic involved and uninvolved skin. J Dermatol Sci 28 Elder JT, Fisher GJ, Lindquist PB et al. Overexpression of trans- 1995; 10:103–9. forming growth factor alpha in psoriatic epidermis. Science 1989; 48 Priestley GC. The fibroblast in psoriasis: master or slave? Eur J 243:811–14. Dermatol 1991; 1:185–91. 29 Cook PW, Pittelkow MR, Keeble WW et al. Amphiregulin messen- 49 Ko¨rver JEM, van Dijnhoven MWFM, Pasch MC et al. Assessment of ger RNA is elevated in psoriatic epidermis and gastrointestinal epidermal subpopulations and proliferation in healthy skin, symp- carcinomas. Cancer Res 1992; 52:3224–7. tomless and lesional skin of spreading psoriasis. Br J Dermatol 2006; 30 Nanney LB, Yates RA, King LE Jr. Modulation of epidermal 155:688–94. growth factor receptors in psoriatic lesions during treatment with 50 van de Kerkhof PCM. Pathogenesis. In: Textbook of Psoriasis (van de topical EGF. J Invest Dermatol 1992; 98:296–301. Kerkhof PCM, ed.), 2nd edn. Oxford: Blackwell Publications, 31 Hammerberg C, Arend EP, Fisher GJ et al. Interleukin-1 receptor 2003; 124–9. antagonist in normal and psoriatic epidermis. J Clin Invest 1992; 51 DePita O, Ruffelli M, Cadoni S et al. Psoriasis: comparison of 90:571–83. immunological markers in patients with acute and remission 32 Schmid P, Cox D, McMaster GK, Itin P. In situ hybridization analy- phase. J Dermatol Sci 1996; 13:118–24. sis of cytokine, proto-oncogene and tumor suppressor gene 52 Koreck A, Suranyi A, Szony BJ et al. CD3(+) CD56(+) NK T cells expression in psoriasis. Arch Dermatol 1993; 285:334–40. are significantly decreased in the peripheral blood of patients with 33 Nickoloff BJ, Karabin GD, Barker JN et al. Cellular localization of psoriasis. Clin Exp Immunol 2002; 127:176–82. interleukin-8 and its inducer, tumor necrosis factor-alpha in psor- 53 Lecewicz-Toru´n B, Pietrzak A, Rolinski J, Brzozowski W. The per- iasis. Am J Pathol 1991; 138:129–40. ipheral blood lymphocyte pattern in psoriasis preceded by an 34 Tettlebach W, Nanney L, Ellis D et al. Localization of MGSA/GRO infection. Med Sci Monit 2001; 7:889–93. protein in cutaneous lesions. J Cutan Pathol 1993; 20:259–66. 54 van der Vliet HJJ, von Blomberg BME, Nishi N et al. Circulating 35 Bernard BA, Asselineau D, Schaftar D, Darmon MY. Abnormal Va24+ Vb11+ NKT-cell numbers are decreased in a wide variety sequence of expression of differentiation markers in psoriatic epi- of diseases that are characterized by autoreactive tissue damage. dermis: inversion of two steps in the differentiation program? Clin Immunol 2001; 100:144–8. J Invest Dermatol 1988; 90:801–5. 55 Rocha-Pereira P, Santos-Silva A, Rebelo I et al. The inflammatory 36 Lew W, Bowcock AM, Krueger JG. Psoriasis vulgaris: cuta- response in mild and in severe psoriasis. Br J Dermatol 2004; neous lymphoid tissue supports T-cell activation and ‘type 1’ 150:917–28. inflammatory gene expression. Trends Immunol 2004; 25:295–305. 56 Brown WR, Hardy MH. Effects of topical triamcinolone acetonide, 37 Chang EY, Hammerberg C, Fisher G et al. T-cell activation is retinoic acid and anthralin on asebia mouse skin. J Invest Dermatol potentiated by cytokines released by lesional psoriatic but not nor- 1984; 82:412–16. mal epidermis. Arch Dermatol 1992; 128:1479–85. 57 Hogenesch H, Gijbels MJJ, Zurcher C. The chronic proliferative 38 Baker BS, Ovigne JM, Powles AV et al. Normal keratinocytes dermatitis (cpd) mutation chromosome? In: Handbook of Mouse express toll-like receptors (TLRs) 1, 2 and 5: modulation of TLR Mutations with Skin, Hair Abnormalities. Animal Models. Bio Medical Tools expression in chronic plaque psoriasis. Br J Dermatol 2003; (Sundberg JP, ed.). Boca Raton, FL: CRC Press, 1994; 217–20. 148:670–9. 58 Sunberg JP, Boggess D, Shultz LD, Beamer WG. The flaky skin 39 Curry JL, Quin JZ, Bonish B et al. Innate immune-related receptors (fsn) mutation chromosome? In: Handbook of Mouse Mutations with in normal and psoriatic skin. Arch Pathol Lab Med 2003; 127:178– Skin, Hair Abnormalities. Animal Models. Bio Medical Tools (Sundberg JP, 86. ed.). Boca Raton, FL: CRC Press, 1994; 253–68. 40 Steinert PM. The complexity and redundancy of epithelial barrier 59 Groves RW, Mizutani H, Kieffer JD, Kupper TS. Inflammatory skin function. J Cell Biol 2000; 151:F5–8. disease in transgenic mice that express high levels of interleukin 41 Placek W, Haftek M, Thivolet J. Sequence of changes in psoriatic 1a in basal epidermis. Proc Natl Acad Sci USA 1995; 92:11874–8. epidermis. Immunocompetent cell redistribution precedes altered 60 Carroll JM, Crompton T, Seery JP, Watt FM. Transgenic mice expression of keratinocyte differentiation markers. Acta Derm Venereol expressing IFN-c in the epidermis have eczema, hair hypo- (Stockh) 1988; 68:369–77. pigmentation, and hair loss. J Invest Dermatol 1997; 108:412–22.
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 The evolution of the psoriatic lesion, P.C.M. van de Kerkhof 15
61 Turksen K, Kupper T, Degenstein L et al. Interleukin 6: insights to parison of antibody to CD11a with cyclosporin A and clobetasol its function by overexpression in transgenic mice. Proc Natl Acad Sci propionate. Lab Invest 2001; 81:1253–61. USA 1992; 89:5068–72. 82 Dam TS, Kang S, Nickoloff BJ. Pharmacological modulation of 62 Vassar R, Fuchs E. Transgenic mice provide new insights into the induction and treatment of psoriasis using human skin grafted role of TGF-a during epidermal development and differentiation. onto SCID mice. J Invest Dermatol 1997; 108:572. Genes Dev 1991; 5:714–27. 83 Boyman O, Hefti HP, Conrad C et al. Spontaneous development of 63 Guo L, Yu QC, Fuchs E. Targeting expression of keratinocyte psoriasis in a new animal model shows an essential role for resi- growth factor to keratinocytes elicits striking changes in epithelial dent T cells and tumor necrosis factor-a. J Exp Med 2004; differentiation in transgenic mice. EMBO J 1993; 12:973–86. 199:731–6. 64 Blessing M, Schirmacher P, Kaiser S. Overexpression of bone 84 Nestle FO, Conrad C, Tun-Kyi A et al. Plasmacytoid predendritic morphogenic protein-6 (BMP-6) in the epidermis of transgenic cells initiate psoriasis through interferon-alpha production. J Exp mice: inhibition or stimulation of proliferation depending on the Med 2005; 202:135–43. pattern of transgene expression and formation of psoriatic lesions. 85 Gilliet M, Conrad C, Geiges M et al. Psoriasis triggered by toll-like J Cell Biol 1996; 135:227–39. receptor 7 agonist imiquimod in the presence of dermal plasma- 65 Xia YP, Li B, Hylton D et al. Transgenic delivery of VEGF to mouse cytoid dendritic cell precursors. Arch Dermatol 2004; 140:1490–5. skin leads to an inflammatory condition resembling human psor- 86 Boyman O, Conrad C, Dudli C et al. Activation of dendritic anti- iasis. Blood 2003; 102:161–8. gen-presenting cells expressing common heat shock protein recep- 66 Blumberg H, Conklin D, Xu WF et al. Interleukin 20: discovery, recep- tor CD91 during induction of psoriasis. Br J Dermatol 2005; tor identification, and role in epidermal function. Cell 2001; 104:9–19. 152:1095–7. 67 Cook PW, Piepkorn M, Clegg CH et al. Transgenic expression of 87 Sano S, Chan KS, Carbajal S et al. Stat3 links activated keratinocytes the human amphiregulin gene induces a psoriasis-like phenotype. and immunocytes required for the development of psoriasis in a J Clin Invest 1997; 100:2286–94. novel transgenic mouse model. Nat Med 2004; 11:43–9. 68 Bullard DC, Scharfetter-Kochanek K, McArthur MJ et al. A polygen- 88 Zenz R, Eferl R, Kenner L et al. Psoriasis-like skin disease and arth- ic mouse model of psoriasiform skin disease in CD18-deficient ritis caused by inducible epidermal deletion of Jun proteins. Nature mice. Proc Natl Acad Sci USA 1996; 93:2116–21. 2005; 437:369–75. 69 Kess D, Peters T, Zamek J et al. CD4+ T cell-associated pathophysi- 89 Pasparakis M, Courtois G, Hafner M et al. TNF-mediated inflamma- ology critically depends on CD18 gene dose effects in a murine tory skin disease in mice with epidermis-specific delation of IKK2. model of psoriasis. J Immunol 2003; 171:5697–706. Nature 2002; 417:861–6. 70 Carroll JM, Romero MR, Watt FM. Suprabasal integrin expression in 90 Stratis A, Pasparakis M, Rupec RA et al. Pathogenetic role for skin the epidermis of transgenic mice results in developmental defects macrophages in a mouse model of keratinocyte-induced psoriasis- and a phenotype resembling psoriasis. Cell 1995; 83:957–68. like skin inflammation. J Clin Invest 2006; 116:2094–103. 71 Scho¨n MP, Detmar M, Parker CM. Murine psoriasis-like disorder 91 Eyre RW, Krueger GP. Response to injury of skin involved and induced by naive CD4+ T cells. Nat Med 1997; 3:183–8. uninvolved with psoriasis, and its relation to disease activity 72 Nickoloff BJ, Kunkel SL, Burdick M, Stricker RM. Severe combined Koebner reactions. Br J Dermatol 1982; 106:153–9. immunodeficient mouse and human psoriatic skin chimeras – val- 92 Gerritsen MJP, van Erp PEJ, van Vlijmen-Willems IMJJ et al. Repea- idation of a new model. Am J Pathol 1995; 146:580–8. ted tape stripping of normal human skin: a histological assessment 73 Gilhar A, David M, Ulmann Y et al. Lymphocyte dependence of and comparison with events seen in psoriasis. Arch Dermatol Res psoriatic pathology in human psoriatic skin engrafted to SCID 1994; 256:455–61. mouse. J Invest Dermatol 1997; 109:283–9. 93 Chang A, de Jong GJ, Mier PD, van de Kerkhof PCM. Enzymatic quan- 74 Sugai S, Iizuka M, Kawakubo Y et al. Histological and immuno- tification of polymorphonuclear leukocytes in normal and psoriatic chemical studies of the human psoriatic lesions transplanted onto skin following standardized injury. Clin Exp Dermatol 1978; 13:62–6. SCID mice. J Dermatol Sci 1998; 17:85–92. 94 van de Kerkhof PCM, van Rennes H, de Grood RM et al. Responses 75 Wrone-Smith T, Nickoloff BJ. Dermal injection of immunocytes of the clinically uninvolved skin of psoriatics to standardized induces psoriasis. J Clin Invest 1996; 98:1878–87. injury. Br J Dermatol 1983; 109:287–94. 76 Nickoloff BJ, Wrone-Smith T. Injection of prepsoriatic skin with 95 Schubert C, Christophers E. Mast cells and macrophages in early CD4+ T cells induces psoriasis. Am J Pathol 1999; 155:145–58. relapsing psoriasis. Arch Dermatol Res 1985; 277:352–8. 77 Nickoloff BJ, Wrone-Smith T, Bonish B, Porcelli SA. Response of 96 Gillitzer R, Wolff K, Tong D et al. MCP-1 mRNA expression in murine and normal human skin to injection of allogenic blood basal keratinocytes of psoriatic lesions. J Invest Dermatol 1993; derived psoriatic immunocytes: detection of T-cells expressing 101:127–31. receptors typically present on natural killer cells, including CD94, 97 Glade CP, van Erp PEJ, Werner-Schlenzka H, van de Kerkhof PCM. CD158 and CD161. Arch Dermatol 1999; 135:546–52. A clinical and flow cytometric model to study remission and 78 Boehncke WH, Zollner TM, Dressel D, Kaufmann R. Induction of relapse in psoriasis. Acta Derm Venererol (Stockh) 1998; 78:180–5. psoriasiform inflammation by a bacterial superantigen in the SCID- 98 Castelijns FACM, Gerritsen MJP, van Vlijmen-Willems IMJJ et al. hu xenogenic transplantation model. J Cutan Pathol 1997; 24:1–7. The epidermal phenotype during initiation of the psoriatic lesion 79 Boehncke WH, Kock M, Hardt-Weinelt K et al. The SCID-hu xeno- in the symptomless margin of relapsing psoriasis. J Am Acad Derma- geneic transplantation model allows screening of anti-psoriatic tol 1999; 40:901–9. drugs. Arch Dermatol 1999; 291:104–6. 99 van de Kerkhof PCM, Gerritsen MJP, de Jong EMGJ. Transition 80 Dam TN, Kang S, Nickoloff BJ, Voorhees JJ. 1a,25-dihydroxychol- from symptomless to lesional psoriatic skin. Clin Exp Dermatol 1996; ecalciferol and cyclosporine suppress induction and promote reso- 21:325–9. lution of psoriasis in human skin grafts transplanted on to SCID 100 Hull SM, Goodfield M, Wood EJ, Cunliffe WJ. Active and inactive mice. J Invest Dermatol 1999; 113:1082–9. edges of psoriatic plaques: identification by tracing and investiga- 81 Zeigler M, Chi Y, Tumas DB et al. Anti-CD11a ameliorates disease tion by laser-Doppler flowmetry and immunohistology. J Invest in the human psoriatic skin SCID mouse transplant model: com- Dermatol 1989; 92:782–5.
2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp4–15 REVIEW ARTICLE DOI 10.1111/j.1365-2133.2007.07954.x Skin manifestations of intravascular lymphoma mimic inflammatory diseases of the skin J. Ro¨glin and A. Bo¨er DERMATOLOGIKUM Hamburg, Stephansplatz 5, 20354 Hamburg, Germany
Summary
Correspondence Background Intravascular lymphoma (IVL) is fatal when it is diagnosed late in the Almut Bo¨er. course. Sometimes skin lesions enable early diagnosis, but criteria for diagnosis E-mail: [email protected] are not well established. Objectives To demonstrate the clinical spectrum of skin lesions of IVL and to corre- Accepted for publication 3 February 2007 late it with clinical outcome; to identify features differentiating between B-cell and T-cell IVL with skin involvement. Key words Methods Review of 97 articles reporting on total of 224 patients with IVL. angioendotheliomatosis maligna, B-cell lymphoma, Results Skin lesions were mentioned in 90 of 224 patients. They were nodules intravascular lymphoma, skin and ⁄or plaques (49%) or macules (22Æ5%) of red (31%) or blue to livid (19%) Conflicts of interest colour on the leg (35%), the thigh (41%) and the trunk (31%). Telangiectases None declared. were present in only 20% of the patients. Oedema (27Æ5%) of the legs and pain (24%) were often accompanying. No criteria enabled distinction between lesions restricted to the skin and skin lesions concurrent with IVL in other organs, but when the disease was restricted to the skin, the prognosis was favourable (10% vs. 85% fatal outcome). Skin lesions of T-cell IVL are indistinguishable from those of B-cell IVL. Conclusions Forty per cent of all patients with IVL have skin lesions, these being red, sometimes painful plaques located typically on the lower extremities, accom- panied by oedema. A clinician risks misinterpreting these changes as thrombo- phlebitis, erythema nodosum or erysipelas. Neither clinical course nor differentiation of the lymphoma can be predicted from the morphology of skin lesions, but involvement of other organs at the time of diagnosis indicates a poor prognosis.
Intravascular lymphoma (IVL) was described first in 1959 by lesions enable a diagnosis to be made early, but diagnostic Pfleger and Tappeiner under the designation ‘Angioendo- criteria for such lesions are not well established, patients being theliomatosis proliferans systemisata’.1 The disease was for reported almost exclusively in the form of case reports. long considered to be a malignant neoplasm of endothelial Several authors of reports on IVL included a differentiation2–18 but immunohistochemical studies finally review of patients published prior to their own revealed the lymphocytic differentiation of the proliferations work,12–14,26,43,57,58,63,73,92,96,98–100 but none of them of atypical cells within vessels.19,20 Histopathological findings focused specifically on the clinical appearance of lesions in the are stereotypical and consist of infiltrates of large blasts within skin, a reason for that being that authors of these articles were vessels of the superficial and deep plexus in the dermis some- mainly oncologists, neurologists and pathologists, but not times also involving vessels in the upper part of the sub- dermatologists. cutaneous fat (Fig. 1). Reactive inflammatory infiltrates and Recently, we diagnosed one patient of our own with IVL. The haemorrhage as well as various amounts of haemosiderophages patient, a man aged 59 years, had presented with recurrent ery- may also be present. thema tender to the touch on the inner aspects of the thighs and Intravascular lymphoma is a rare disease that often has a the legs. For several weeks he had been misdiagnosed clinically fatal course because patients do not present with signs and as having thrombophlebitis migrans, when a biopsy specimen symptoms that allow a correct diagnosis to be made before taken to exclude chronic erysipelas finally revealed the correct the disease is well advanced.1–97 Sometimes, however, skin diagnosis. Figure 1 shows photomicrographs from a biopsy
2007 The Authors 16 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er 17
Fig 1. Findings histopathologic of IVL: Infiltrates of large blasts within vessels of the superficial and deep plexus in the dermis, sometimes also involving vessels in the subcutaneous fat. In this patient, blasts were CD20 positive, proving IVL-BCL. taken from our patient. Large blasts were seen in the lumen of in regard to site of involvement, colour of lesions, shape medium-sized vessels in the subcutis. Blasts were negative for and consistency of lesions, and associated symptoms. Lesions T-cell markers CD3, CD4, CD8 and CD30 but were positive for of patients who presented only with skin lesions were com- the B-cell marker CD20. IgH rearrangement studies by poly- pared with those in patients who also had involvement of merase chain reaction revealed a clonal B-cell population. other organs in order to identify differences between both Our patient prompted us to undertake a review of 224 groups. Lesions of patients with proven IVL-BCL were com- patients published in the literature,1–97 with the aim of pared with those encountered in proven IVL-TCL. Figures improving criteria for clinical diagnosis of skin lesions in IVL. published in articles were collected by us and are repro- Clinical appearance was evaluated on the basis of course and duced here in order to show the morphological spectrum outcome of the disease in order to identify morphological of the disease. characteristics of prognostic value. Additionally, we attempted to identify criteria differentiating between B-cell (IVL-BCL) Results and T-cell (IVL-TCL) variants of the disease, when presenting with lesions in the skin. Articles reporting on a total of 224 patients with IVL were reviewed. In articles before 1986 and many thereafter no dif- Materials and methods ferentiation is made by the authors between B-cell and T-cell lymphoma (105 patients).1–27,29,30,36,49,51,53,77,92 Beginning Reports in the literature were reviewed to identify patients in 1986, authors increasingly included immunohistochemical with IVL who had skin manifestations of the disease. Arti- data of the patients studied by them.19,20,23,25 The majority cles were retrieved via Medline search for the terms ‘intra- of patients had IVL-BCL (86 patients)31,32,34,35,38–40,42,47,50,53, vascular lymphoma’ and ‘angioendotheliomatosis’ and from 55–58,60–62,64,67,68,73,74,77,78,80,82,83,86–89,93,95,97 whereas a minor- references listed in those articles. Descriptions of clinical le- ity had IVL-TCL (33 patients).25,37,40,44,45,47,48,52–54,57,59,63,66, sions were analysed for features in common among them 69,71–73,75,78,79,81,91,95–97
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 18 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er
In 59 patients no signs and symptoms of the skin were lesions was described as red in 31% of the patients, blue to mentioned, and in another 75 patients authors stated specific- livid in 19% and brown to grey in 9%, no colour of lesions ally that the skin had been uninvolved in the lymphomatous being given in the remaining 41%. A frequently accompany- process. In 90 of 224 patients, the disease presented with ing sign was oedema of the legs, being mentioned in 27Æ5% lesions in the skin; nine of these had insufficient data to be of the patients. Pain was a relatively common complaint assessed in detail in this review.78,92,97 The remaining 81 (24%). Lesions were reported to have been waxing and wan- patients comprised 47 women and 34 men, mean age ing in 16% of the patients. Associated findings such as fever, 60 years. Twenty-nine had lesions restricted to the malaise and weakness were described in 47Æ5% of the skin,1,3,16,22,25–27,40,49,50,54,57,60,68,74,75,87,89,92 whereas in the patients. other 52 patients organs other than the skin were Comparison of clinical features in patients with lesions also involved. In 26 of these 52 patients, lesions restricted to the skin with those in whom the disease involved in the skin preceded those in other other organs after or before lesions were manifest in the skin organs5,7–11,13,21,24,26,30,34,37,39,40,44,48,60,61,64,66,69,93 whereas is presented in Table 1. No significant differences between the in the remaining 26, signs and symptoms of other three groups could be determined in regard to the clinical organs preceded the appearance of lesions in the presentation of skin lesions. skin.2,6,17,19,20,24,26,34,36,38,42,51,52,54,56–58,81,82,88 Photographs of clinical lesions were included for only 32 Skin manifestations were said to be localized in decreasing patients,1–3,7,9,10,13,16,19,21,22,24,27,30,34,37,39,45,48–50,52,60,66,68, order of frequency on the thigh (41%), leg (35%), trunk 74,82,87,89 10 of those having IVL-BCL,34,39,50,60,68,74,82,87,89 (31%), arm (15%) and buttock (7Æ5%). When the thighs were six IVL-TCL37,45,48,52,66 and 16 not being characterized as affected, lesions were mentioned to have favoured the antero- either TCL or BCL.1–3,7,9,10,13,16,19,21,22,24,27,30,59 Figure 2 (a–d) medial aspect in nine patients.16,21,27,30,34,39,40,60,88 Lesions shows clinical lesions of patients with IVL from articles were said to have involved both sides of the body in 47 before 1986, when immunophenotyping was not performed. patients.1,3,5,7–9,13,16,21,25,30,34,37,40,42,44,45,48–50,52,54,56,58,60, Figure 3(a–k) shows clinical lesions in patients with proven 61,64,68,69,72,74,82,87–89,93 IVL-BCL (including one patient of our own). In Figure 4(a, b) Most commonly lesions were described as being nodules lesions of patients with proven IVL-TCL are presented. and ⁄or plaques (49%); macules were encountered in 22Æ5% Comparison of sure cases of IVL-BCL and IVL-TCL is provided of the patients. Lesions were said to be indurated or firm in in Table 2. It did not reveal any major differences between both 27Æ5% of the patients. Mention was made of prominent groups, an exception being the involvement of the face and telangiectases in only 20% of the patients. The colour of the anogenital area in four patients with IVL-TCL.37,40,57,66
Table 1 Comparison of clinical lesions in Lesions Skin lesions Lesions in other patients with or without involvement of restricted preceding organs preceding organs other than the skin to the skin lesions in other lesions in the Total (n = 29) organs (n = 26) skin (n = 26) (n = 81) % Site involved Trunk 8 12 5 25 31 Buttock 2 2 2 6 7Æ5 Thigh 14 13 6 33 41 Leg 10 13 5 28 35 Arm 3 4 5 12 15 Others 2 2 2 6 7Æ5 Type of lesions Nodules ⁄ 14 17 8 39 49 plaques Macules 9 2 7 18 22Æ5 Telangiectases 5 7 4 16 20 Induration 12 6 4 22 27Æ5 Colour of lesions Red 7 8 11 25 31 Blue ⁄livid 6 6 3 15 19 Grey ⁄brown 2 4 1 7 9 Oedema 11 8 3 22 27Æ5 Pain 12 7 0 19 24 Lesions waxing 57 1 1316 and waning Fever ⁄malaise 10 14 14 38 47Æ5
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er 19
(a) (b)
(c)
(d)
Fig 2. (a–d) Clinical lesions of patients with IVL from articles before 1986, when immunophenotyping was not performed. (a) Reproduced from Pfleger L, Tappeiner J. 19591, with kind permission of Springer Science and Business Media Copyright (1959). All Rights reserved; (b) reproduced with kind permission from Fievez M, Fievez C, Hustin J,7 Copyright (1971), American Medical Association. All Rights reserved; (c) reproduced with kind permission from Bhawan J, Wolff SM, Ucci AA, Bhan AK,19 Copyright (1985), Wiley & Sons, Ltd.; (d) reproduced with kind permission from Keahey TM, Guerry D, Tuthill R et al.,16 Copyright (1982), American Medical Association. All Rights reserved.
Moreover, patients with IVL-TCL seemed to be younger at the lesions restricted to the skin continued to do well during a fol- time of diagnosis (mean 53 years, range 17–84) compared with low-up period ranging from 3 months to 14 years, the majority patients with IVL-BCL (mean 66 years, range 54–84).45,81 of patients who had lesions in the skin together with lesions of Data on treatment and follow-up were available for 81 IVL in other organs died within 2 years after the diagnosis of patients.1–3,6–11,13,16,20,22,24,25,30,34,36,37,42,44,45,48,50–52,54,56– the disease. Patients with IVL-TCL compared with patients with 58,60,61,64,66,69,72,75,81,88,89,92,93 The vast majority of patients IVL-BCL had a slightly worse outcome based on the few cases received combination chemotherapy such as CHOP (cyclo- available for comparison (69% vs. 52% fatal outcome). phosphamide, doxorubicin, vincristine and prednisone), and some received radiation treatment either as a single treatment or Discussion in combination with chemotherapy. Death from IVL that pre- sented with lesions in the skin was explicitly mentioned in 42 Our review of the literature revealed that 40% of patients with patients and is summarized in Table 3. Whereas patients with IVL present with lesions in the skin, thus demonstrating a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 20 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er
(a) (b) (c) (d)
(e) (f) (g)
(h) (j) (k)
(i)
Fig 3. (a–k) Clinical lesions in patients with proven IVL-BCL. (a) Reproduced with kind permission from Chang AB, Zic JA, Boyd AS,74 Copyright (1998), Elsevier Health Sciences. All Rights reserved; (b) reproduced from Williams REA, Seywright MM, Lever R, Lucie NP. 199039; (c) reproduced with kind permission from Perniciario C, Winkelmann RK, Daoud MS, Su WPD,60 Copyright (1995), Lippincott Williams & Wilkins. All Rights reserved; (d) reproduced with kind permission from Asagoe K, Fujimoto W, Yoshino T et al. 2003,88 Copyright (2003), Elsevier Health Sciences. All Rights reserved; (e) patient of our own; (f, g) reproduced with kind permission from Wilson BB,50 Copyright (1992), American Medical Association. All Rights reserved; (h, i) reproduced with kind permission from Ero¨s N, Karolyi Z, Kovacs A et al. 2002,87 Copyright (2002), Elsevier Health Sciences. All Rights reserved; (j) reproduced from O¨ zgu¨roglu E, Bu¨yu¨lbabani N, O¨ zgu¨roglu M, Baykal C. 199768; (k) reproduced with kind permission from Petroff N, Koger OW, Fleming MG, Fishleder A et al.,34 Copyright (1989), Elsevier Health Sciences. All Rights reserved. need for dermatologists to be aware of the typical clinical by analysing scrupulously the references quoted by other appearance of such lesions. Many of the articles reviewed by authors, thus demonstrating the deficiencies of the search us were case reports: only a few studies included more than engines of the Medline database. Unfortunately, in one of the one patient. Interestingly, many of the articles could not be largest studies, undertaken by Ferreri et al., in which 10 retrieved from a simple Medline search but were found only patients were said to have lesions in the skin, clinical findings
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er 21
(a) (b)
Fig 4. (a, b) Clinical lesions of patients with proven IVL-TCL. (a) Reproduced with kind permission from Lopez-Gil F, Roura M, Umbert I, Umbert P. 1992,48 Copyright (1992), Elsevier Health Sciences. All Rights reserved; (b) reproduced with kind permission from Sepp N, Schuler G, Romani N et al. 1990,37 Copyright (1990), Elsevier Health Sciences. All Rights reserved.
Table 2 Reported patients with intravascular lymphoma (IVL) and lesions in the skin. Differentiation Comparison of patients with proven B-cell IVL-BCL IVL-TCL not known Total lymphoma (BCL) with those with T-cell lymphoma (TCL) Lesions restricted to the skin 8 6 15 29 Skin lesions preceding lesions 6 7 13 26 in other organs Lesions in other organs 11 3 12 26 preceding lesions in the skin Total 25 16 40 81 Pictures of lesions available 10 6 16 32
Table 3 Patients reported to have died from intravascular lymphoma (IVL) that presented with lesions in the skin. Comparison of patients with proven B-cell lymphoma (BCL) with those with T-cell lymphoma (TCL) and comparison of patients with lesions restricted to the skin with those in whom lesions in the skin preceded involvement of other organs and with those in whom lesions in the skin appeared after involvement of organs other than the skin
Differentiation IVL-BCL IVL-TCL not known Total Percentage Lesions restricted to the skin 1 2 0 3 (of 29) 10% Skin lesions preceding lesions in other organs 3 6 8 17 (of 26) 65% Lesions in other organs preceding lesions in the skin 9 3 10 22 (of 26) 85% Total 13 (of 25) 11 (of 16) 18 (of 40) 42 (of 81) 52% Percentage 52% 69% 45% 52%
of skin lesions are given only in the form of a brief summary a full-body examination. Often descriptions of clinical lesions of all of the 10 patients, no patient being presented individu- are not detailed and data about site, signs and symptoms such ally. Therefore, assessment of these patients in regard to the as pain or tenderness are incomplete. clinical appearance is not possible in retrospect.92 Early reports from the 1960s up to the early 1980s did not Many of the reports studied by us did not include photo- include results of sophisticated ancillary laboratory techniques. graphs of clinical lesions but only photomicrographs or Exact classification of patients into lymphoma of B-cell or T-cell electron microscopic images. In numerous patients the authors differentiation was only possible after the advent of immuno- described only a few lesions in the skin but did not document histochemistry and as but one example, the differentiation of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 22 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er
Table 4 Differential diagnoses of skin manifestations of intravascular lymphoma (IVL) and differentiating features
Features in common with skin manifestations of IVL Features differentiating from skin manifestations of IVL Thrombophlebitis Indurated painful strand surrounded by erythema No telangiectases in the area affected No systemic signs and symptoms such as malaise and fever Thrombophlebitis migrans Indurated painful strand waxing and waning Sites involved vary and do not favour the inner aspect of the anterior thigh Erysipelas Erythema, elevated body temperature, oedema Body temperature febrile whereas it is subfebrile in IVL of the limb affected Painful lymphadenopathy is typical of erysipelas but lacking in skin lesions of IVL Erythema nodosum Indurated nodules tender to the touch, oedema Lesions favour the anterior aspects of the shins but not of the legs the ankles or the anterior aspects of the thighs Panarteritis nodosa Indurated nodules tender to the touch, oedema Lesions do not favour the ankles and the anterior aspects of the legs of the thighs as in skin lesions of IVL Leucocytoclastic vasculitis Purpuriform lesions on the legs, oedema of the Lesions are small haemorrhagic and sometimes necrotic legs papules but not patchy indurated erythema or nodules Livedo racemosa Reticulated livid erythema, oedema of the legs Ulceration and scarring are common in livedo racemosa but rare in skin lesions of IVL Septic vasculitis Purpuriform lesions on acral sites, oedema of Lesions favour the digits but not the anterior aspects of the legs the thighs as in skin lesions of IVL
cells in the original report by Pfleger and Tappeiner1 and of by neovascularization. Reorganization of thrombi is usually many patients published thereafter is unclear. accompanied by local increase of the tissue temperature which Our analysis of descriptions of clinical lesions as well as of may simulate an authentic inflammatory or even infectious photographs presented in the articles revealed that the presen- condition of the dermis or subcutis. Telangiectases are clinical tation in the skin of IVL shares many features with inflamma- evidence of recanalization and of recurrences of lesions at the tory diseases of the skin. Lesions may be red, painful, very same site. They are an important clue to the diagnosis of accompanied by oedema, and waxing and waning. Often the IVL, but they are not present invariably and, moreover, they clinical findings are misinterpreted when the patient first may be misinterpreted as varicosities. consults a dermatologist. Common misdiagnoses are listed in Patients who are diagnosed with IVL early in the course Table 4 and include ‘thrombophlebitis’, ‘thrombophlebitis when lesions are still restricted to the skin have a much migrans’, ‘erythema nodosum’, ‘vasculitis’ and ‘livedo rac- more favourable prognosis. In contrast, patients in whom the emosa’.51,56,61 Associated findings such as fever, malaise, and disease is diagnosed only when several organs are already elevation of lactate dehydrogenase and erythrocyte sedimenta- involved have a poor prognosis despite treatment with poly- tion rate may also mislead a dermatologist to suspect an chemotherapy. Not uncommonly, the disease goes entirely inflammatory condition rather than a lymphoma. Whereas undiagnosed and the correct diagnosis is only made on post most patients with IVL present with cytopenias, mainly anae- mortem examination.24,51,53,56,61,69,77 Our comparison of mia and thrombocytopenia, identifiable involvement of the lesions in patients with IVL restricted to the skin with skin bone marrow is rare, especially early in the course of the lesions that appeared together with involvement of other disease. organs by IVL did not reveal any significant differences, indi- The fact that skin lesions of IVL mimic inflammatory dis- cating that the clinical course of an individual patient cannot eases of the skin can be explained by clinicopathological cor- be predicted from the morphology of skin lesions. It is relation. Occlusion of vessels by blasts of T- or B-cell essential, therefore, that patients with lesions of IVL in the differentiation activates the coagulation cascade and thrombi skin are examined fully. Ultrasound of lymph nodes, blood develop within the lumina of vessels. The clinical signs may cell count, bone marrow biopsy, abdominal ultrasound and be indistinguishable from those of thrombophlebitis. If super- chest X-ray are recommended. Neurological examination and ficial vessels are involved in the process the clinical pattern renal function tests should also be performed because the may be that of livedo racemosa; if vessels in the deep dermis brain and the kidney are other sites commonly involved in or the septa of the subcutaneous fat are involved, clinical IVL. Skin manifestations of IVL-TCL seem not to differ from lesions mimic those of erythema nodosum or nodular vasculi- those encountered in IVL-BCL, the only exceptions being that tis. If many vessels are involved in the process, patchy ery- IVL-TCL may also, at times, involve the face and the ano- thema along with haemorrhage may call to mind the clinical genital region and that patients with IVL-TCL seem to be diagnosis of erysipelas. That lesions are waxing and waning younger than patients with IVL-BCL, and have a slightly can be explained by thrombi being recanalized or bypassed worse prognosis.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er 23
It is a matter of controversy whether patients with IVL 12 Petito CK, Gottlieb GJ, Dougherty JH, Petito FA. Neoplastic restricted to the skin should be treated aggressively. Experience angioendotheliomatosis: ultrastructual study and review of the with such patients is limited and no controlled studies exist so literature. Ann Neurol 1978; 3:393–9. 13 Kauh YC, McFarland JP, Carnabuci GG, Luscombe HA. Malignant far. Usually patients receive polychemotherapy for large B-cell proliferating angioendotheliomatosis. Arch Dermatol 1980; 116:803–6. or T-cell lymphoma, and the composition of such treatment 14 Wick MR, Banks PM, McDonald TJ. Angioendotheliomatosis of the regimens changes whenever guidelines for treatment of nose with fatal systemic dissemination. Cancer 1981; 48:2510–17. lymphomas are revised. When blasts of IVL-BCL are strongly 15 Wick MR, Scheithauer BW, Okazaki H, Thomas JE. Cerebral CD20 positive, treatment with antibodies to CD20 is a new angioendotheliomatosis. Arch Pathol Lab Med 1982; 106:342–6. option for treatment. Clinical follow-up is mandatory in order 16 Keahey TM, Guerry D, Tuthill R et al. Malignant angioendothelio- to detect early progression of the disease with involvement of matosis proliferans treated with doxorubicin. Arch Dermatol 1982; 118:512–14. other organs in the disease process. 17 Ansell J, Bhawan J, Cohen S et al. Histiocytic lymphoma and In conclusion, lesions of IVL in the skin may mimic inflam- malignant angioendotheliomatosis one disease or two? Cancer matory conditions such as thrombophlebitis, thrombophlebitis 1982; 50:1506–12. migrans, erythema nodosum or chronic erysipelas because 18 Arnn ET, Yam LT, Li C-Y. Systemic angioendotheliomatosis lesions may present as red, painful patches and plaques, presenting with hemolytic anemia. Am J Clin Pathol 1983; 80:246– accompanied by oedema, situated on the leg or the thigh, and 51. may be waxing and waning. Whenever a patient is diagnosed 19 Bhawan J, Wolff SM, Ucci AA, Bhan AK. Malignant lymphoma and malignant angioendotheliomatosis: one disease. Cancer 1985; with ‘recurrent thrombophlebitis’, ‘recurrent erythema nodo- 55:570–6. sum’ of unknown cause or with an erysipelas unresponsive to 20 Wrotnowski U, Mills SE, Cooper PH. Malignant angioendothelio- antibiotic treatment, the possibility of an IVL should be con- matosis: an angiotropic lymphoma? Am J Clin Pathol 1985; 83:244– sidered. A deep excisional biopsy specimen is requisite to con- 8. firm the diagnosis. 21 Lim HW, Anderson HM. Angioendotheliomatosis associated with Unfortunately, neither the differentiation of the lymphoma histiocytic lymphoma. Response to systemic chemotherapy. JAm nor the clinical course of an individual patient can be predic- Acad Dermatol 1985; 13:903–8. 22 Gupta AK, Lipa M, Haberman HF. Proliferating angioendothelio- ted from the morphology of skin lesions seen clinically. Every matosis: case with long survival and review of literature. Arch patient with IVL in the skin needs a complete staging for Dermatol 1986; 122:314–19. involvement of other organs in the process. However, if sta- 23 Carroll TJ Jr, Schelper RL, Goeken JA, Kemp JD. Neoplastic angio- ging does not reveal any involvement of other organs at the endotheliomatosis: immunopathologic and morphological evi- time of diagnosis of skin lesions, the prognosis is favourable dence for intravascular malignant lymphomatosis. Am J Clin Pathol based on the data provided in articles reviewed by us. 1986; 85:169–75. 24 Elner VM, Hidayat AA, Charles NC et al. Neoplastic angioendo- theliomatosis. A variant of malignant lymphoma. Immunohisto- References chemical and ultrastructural observations of three cases. Ophthalmology 1986; 93:1237–45. 1 Pfleger L, Tappeiner J. Zur Kenntnis der systemisierten endotheli- 25 Sheibani K, Battifora H, Winberg CD et al. Further evidence that omatose der cutanen Blutgefa¨ße (Reticuloendotheliose?). Hautarzt ‘malignant angioendotheliomatosis’ is an angiotropic large-cell 1959; 10:359–63. lymphoma. N Engl J Med 1986; 314:943–8. 2 Braverman IM, Lerner AB. Diffuse malignant proliferation of vas- 26 Wick MR, Mills SE, Scheithauer BW et al. Reassessment of malig- cular endothelium: a possible new clinical and pathological entity. nant ‘angioendotheliomatosis’: evidence in favor of its reclassifi- Arch Dermatol 1961; 84:22–30. cation as ‘intravascular lymphomatosis’. Am J Surg Pathol 1986; 3 Tappeiner J, Pfleger L. Angioendotheliomatosis proliferans sys- 10:112–23. temisata. Ein klinisch und pathologisch neues Krankheitsbild. 27 Willemze R, Kruyswijk MRJ, De Bruin CD et al. Angiotropic Hautarzt 1963; 14:67–70. (intravascular) large cell lymphoma of the skin previously classi- 4 Shtern R, Likhachev I. Endotheliomatosis as a systemic neoplastic fied as malignant angioendotheliomatosis. Br J Dermatol 1987; disease. Arch Pathol 1963; 25:35–40. 116:393–9. 5 Haber H, Harris-Jones JN, Wells AL. Intravascular endothelioma 28 Wick MR, Rocamora A. Reactive and malignant ‘angioendothelio- (endothelioma in situ, systemic endotheliomatosis). J Clin Pathol matosis’: a discriminant clinicopathological study. J Cutan Pathol 1964; 17:608–11. 1988; 15:260–71. 6 Strouth JC, Donahue S, Ross A, Aldred A. Neoplastic angioendo- 29 Ferry JA, Harris NL, Picker LJ et al. Intravascular lymphomatosis theliosis. Neurology 1965; 15:644–8. (malignant angioendotheliomatosis): a B-cell neoplasm expressing 7 Fievez M, Fievez C, Hustin J. Proliferating systematized angio- surface homing receptors. Mod Pathol 1988; 1:444–52. endotheliomatosis. Arch Dermatol 1971; 104:320–4. 30 Berger TG, Dawson NA. Angioendotheliomatosis. J Am Acad Dermatol 8 Okagaki T, Richart R. Systemic proliferating angioendothelioma- 1988; 18:407–12. tosis: a case report. Obstet Gynecol 1971; 37:377–80. 31 Otrakji CL, Voigt W, Amador A et al. Malignant angioendothelio- 9 Scott PWB, Silvers DN, Helwig EB. Proliferating angioendothelio- matosis – a true lymphoma: a case of intravascular malignant matosis. Arch Pathol 1975; 99:323–6. lymphomatosis studied by Southern blot hybridization analysis. 10 Madara J, Shane J, Scarlato M. Systemic endotheliomatosis: a case Hum Pathol 1988; 19:475–8. report. J Clin Pathol 1975; 28:476–82. 32 Domizio P, Hall PA, Cotter F et al. Angiotropic large cell lymphoma 11 Kurrein F. Systemic angioendotheliomatosis with metastases. J Clin (ALCL): morphological, immunohistochemical and genotypic Pathol 1976; 29:347–53.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 24 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er
studies with analysis of previous reports. Hematol Oncol 1989; 53 Glass J, Hochberg FH, Miller DC. Intravascular lymphomatosis. A 7:195–206. systemic disease with neurologic manifestations. Cancer 1993; 33 Lanham GR, Weiss SW, Enzinger FM. Malignant lymphoma: a 71:3156–62. study of 75 cases presenting in soft tissue. Am J Surg Pathol 1989; 54 Mock DJ, Jundt JW, Green JB, Speights VO. Angiotropic lymph- 13:1–10. oma manifested by fever and painful swollen legs. South Med J 34 Petroff N, Koger OW, Fleming MG et al. Malignant angioendo- 1993; 86:1432–5. theliomatosis: an angiotropic lymphoma. J Am Acad Dermatol 1989; 55 Sleater JP, Segal GH, Scott MD, Masih AS. Intravascular (angiotropic) 21:727–33. large cell lymphoma: determination of monoclonality by poly- 35 Jalkanen S, Aho R, Kallajoki M et al. Lymphocyte homing recep- merase chain reaction on paraffin embedded tissues. Mod Pathol tors and adhesion molecules in intravascular malignant lympho- 1994; 7:593–8. matosis. Int J Cancer 1989; 44:777–82. 56 Walker UA, Herbst EW, Ansorge O. Intravascular lymphoma 36 Drobacheff C, Blanc D, Zultak M et al. Malignant angioendothelio- simulating vasculitis. Rheumatol Int 1994; 14:131–3. matosis. Int J Dermatol 1989; 28:454–6. 57 Di Giuseppe JA, Nelson WG, Seifter EJ et al. Intravascular lympho- 37 Sepp N, Schuler G, Romani N et al. ‘Intravascular lymphomatosis’ matosis: a clinicopathologic study of 10 cases and assessment of (angioendotheliomatosis): evidence for a T-cell origin in two response to chemotherapy. J Clin Oncol 1994; 12:2573–9. cases. Hum Pathol 1990; 21:1051–8. 58 Demirer T, Dail DH, Aboulafia DM. Four varied cases of intravas- 38 Molina A, Lombard C, Donlon T et al. Immunohistochemical and cular lymphomatosis and a literature review. Cancer 1994; cytogenetic studies indicate that malignant angioendotheliomato- 73:1738–44. sis is a primary intravascular (angiotropic) lymphoma. Cancer 59 Lakhani SR, Hulman G, Hall JM et al. Intravascular malignant 1990; 66:474–9. lymphomatosis (angiotropic large-cell lymphoma). A case report 39 Williams REA, Seywright MM, Lever R, Lucie NP. Angiotropic with evidence for T-cell lineage with polymerase chain reaction B-cell lymphoma (malignant angioendotheliomatosis): failure of analysis. Histopathology 1994; 25:283–6. systemic chemotherapy. Br J Dermatol 1990; 123:807–10. 60 Perniciario C, Winkelmann RK, Daoud MS, Su WPD. Malignant 40 Stroup RM, Sheibani K, Moncada A et al. Angiotropic (intravascu- angioendotheliomatosis is an angiotropic intravascular lymphoma. lar) large cell lymphoma. Cancer 1990; 66:1781–8. Am J Dermatopathol 1995; 17:242–8. 41 Williams DB, Lyons MK, Yanagihara T et al. Cerebral angiotropic 61 Roux S, Grossin M, De Brandt M et al. Angiotropic large large cell lymphoma (neoplastic angioendotheliosis): therapeutic cell lymphoma with mononeuritis multiplex mimicking systemic considerations. J Neurol Sci 1991; 103:16–21. vasculitis. J Neurol Neurosurg Psychiatry 1995; 58:363–6. 42 Stahl RL, Chan W, Duncan A, Corley CC. Malignant angioendo- 62 Chapin JE, Davis LE, Kornfeld M, Mandler RN. Neurologic mani- theliomatosis presenting as disseminated intravascular coagulo- festation of intravascular lymphomatosis. Acta Neurol Scand 1995; pathy. Cancer 1991; 68:2319–23. 91:494–9. 43 Fredericks RK, Walker FO, Elster A, Challa V. Angiotropic intra- 63 Ghorbani RP, Shokouh-Amiri H, Gaber LW. Intragraft angiotropic vascular large-cell lymphoma (malignant angioendotheliomato- large-cell lymphoma of T-cell-type in a long-term renal allograft sis): report of a case and review of the literature. Surg Neurol recipient. Mod Pathol 1996; 9:671–6. 1991; 35:218–23. 64 Rubin MA, Cossman J, Freter CE, Azumi N. Intravascular large 44 Shimokawa I, Higami Y, Sakai H et al. Intravascular malignant cell lymphoma coexisting within hemangiomas of the skin. Am J lymphomatosis: a case of T-cell lymphoma probably associated Surg Pathol 1997; 21:860–4. with human T-cell lymphotropic virus. Hum Pathol 1991; 22:200–2. 65 Ip M, Chan KW, Chan IKL. Systemic inflammatory response syn- 45 Sangueza O, Hyder DM, Sangueza P. Intravascular lymphomatosis: drome in intravascular lymphomatosis. Intensive Care Med 1997; report of an unusual case with T cell phenotype occurring in an 23:783–6. adolescent male. J Cutan Pathol 1992; 19:226–31. 66 Cho K, Kim C, Yang S et al. Angiocentric T cell lymphoma of the 46 Tateyama H, Eimoto T, Tada T et al. Congenital angiotropic skin presenting as inflammatory nodules of the leg. Clin Exp Derma- lymphoma (intravascular lymphomatosis) of the T-cell type. Cancer tol 1997; 22:104–8. 1991; 67:2131–6. 67 Murase T, Nakamura S, Tashiro K et al. Malignant histiocytosis- 47 Clark WC, Dohan FC Jr, Moss T, Schweitzer JB. Immuno- like B-cell lymphoma, a distinct pathologic variant of intravascular cytochemical evidence of lymphocytic derivation of neoplastic lymphomatosis: a report of five cases and review of the literature. cells in malignant angioendotheliomatosis. J Neurosurg 1991; Br J Haematol 1997; 99:656–64. 74:757–62. 68 O¨ zgu¨roglu E, Bu¨yu¨lbabani N, O¨ zgu¨roglu M, Baykal C. 48 Lopez-Gil F, Roura M, Umbert I, Umbert P. Malignant prolifera- Generalized telangiectasia as the major manifestation of tive angioendotheliomatosis or angiotropic lymphoma associated angiotropic (intravascular) lymphoma. Br J Dermatol 1997; with a soft-tissue lymphoma. J Am Acad Dermatol 1992; 26:101– 137:422–5. 4. 69 Di Giuseppe JA, Hartmann DP, Cossmann J, Mann RB. Molecular 49 Helm TN, Bergfeld WF, Elston MD. Angiotropic lymphoma: detection of bone marrow involvement in intravascular lympho- malignant angioendotheliomatosis. Cutis 1992; 50:204–6. matosis. Mod Pathol 1997; 10:33–7. 50 Wilson BB. Indurated telangiectatic plaques: malignant angioen- 70 Snowden JA, Angel CA, Winfield DA et al. Angiotropic lymph- dotheliomatosis (MAE). Arch Dermatol 1992; 128:255, 258. oma: report of a case with histiocytic features. J Clin Pathol 1997; 51 Kao NL, Broy S, Tillawi I. Malignant angioendotheliomatosis 50:67–70. mimicking systemic necrotizing vasculitis. J Rheumatol 1992; 71 Suh CH, Kim SK, Shin DH et al. Intravascular lymphomatosis of 19:1133–5. the T cell type presenting as interstitial lung disease – a case 52 Setoyama M, Mizoguchi S, Orikawa T, Tashiro M. A case of intra- report. J Korean Med Sci 1997; 12:457–60. vascular malignant lymphomatosis (angiotropic large-cell lymph- 72 Au WY, Shek WH, Nicholls J et al. T-cell intravascular lymphoma- oma) presenting memory T cell phenotype and its expression of tosis (angiotropic large cell lymphoma): association with Epstein– adhesion molecules. J Dermatol 1992; 19:263–9. Barr viral infection. Histopathology 1997; 31:563–7.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 Skin manifestations of intravascular lymphoma, J. Ro¨glin and A. Bo¨er 25
73 Ko YH, Han JH, Go JH et al. Intravascular lymphomatosis: a clin- 87 Ero¨s N, Karolyi Z, Kovacs A et al. Intravascular B-cell lymphoma. icopathological study of two cases presenting as an interstitial J Am Acad Dermatol 2002; 47 (Suppl. 5):S260–2. lung disease. Histopathology 1997; 31:555–62. 88 Asagoe K, Fujimoto W, Yoshino T et al. Intravascular lymphoma- 74 Chang AB, Zic JA, Boyd AS. Intravascular large cell lymphoma: a tosis of the skin as a manifestation of recurrent B-cell lymphoma. patient with asymptomatic purpuric patches and a chronic clinical J Am Acad Dermatol 2003; 48:1–4. course. J Am Acad Dermatol 1998; 39:318–21. 89 Chen TM, Crow MK, Teller C. Angiotropic large-cell lymphoma 75 Chen M, Qiu B, Kong J, Chen J. Angiotropic T cell lymphoma. with striking blue plaques. J Am Acad Dermatol 2003; 48:633–4. Chin Med J 1998; 111:762–4. 90 Rongioletti F, Rebora A. Cutaneous reactive angiomatoses: pat- 76 Rieger E, Soyer HP, LeBoit PE et al. Reactive angioendotheliomato- terns and classification of reactive vascular proliferation. J Am Acad sis or intravascular histiocytosis? An immunohistochemical and Dermatol 2003; 49:887–96. ultrastructural study in two cases of intravascular histiocytic cell 91 Merchant SH, Viswanatha DS, Zumwalt RE, Foucar K. Epstein– proliferation Br J Dermatol 1999; 140:497–504. Barr virus-associated intravascular large T-cell lymphoma present- 77 Calamia KT, Miller A, Shuster EA et al. Intravascular lymphomato- ing as acute renal failure in a patient with acquired immune sis, a report of ten patients with central nervous system involve- deficiency syndrome. Hum Pathol 2003; 34:950–4. ment and a review of the disease process. Adv Exp Med Biol 1999; 92 Ferreri AJM, Campo E, Seymour JF et al. The International Extra- 455:249–65. nodal Lymphoma Study Group (IELSG). Intravascular lymphoma: 78 Kanda M, Suzumiya J, Ohshima K et al. Intravascular large cell clinical presentation, natural history, management and prognostic lymphoma: clinicopathological, immuno-histochemical and factors in a series of 38 cases, with special emphasis on the ‘cuta- molecular genetic studies. Leuk Lymphoma 1999; 34:569–80. neous variant’. Br J Haematol 2004; 127:173–83. 79 Malicki DM, Suh YK, Fuller GN, Shin SS. Angiotropic (intravas- 93 Tomasini C, Novelli M, Ponti R et al. Cutaneous intravascular cular) large cell lymphoma of T-cell phenotype presenting as lymphoma following extravascular lymphoma of the lung. Derma- acute appendicitis in a patient with acquired immunodeficiency tology 2004; 208:158–63. syndrome. Arch Pathol Lab Med 1999; 123:335–7. 94 Mensing CH, Krengel S, Tronnier M, Wolff HH. Reactive angio- 80 Tucker TJ, Bardales RH, Miranda RN. Intravascular lymphomatosis endotheliomatosis: is it ‘intravascular histiocytosis’? J Eur Acad with bone marrow involvement. Arch Pathol Lab Med 1999; 123:952–6. Dermatol Venereol 2005; 19:216–19. 81 Isimbaldi G, Corral L, Songia S et al. An unusual presentation of a 95 Williams G, Foyle A, White D et al. Intravascular T-cell lymphoma case of T cell angiotropic (intravascular) lymphoma. Leukemia with bowel involvement: case report and literature review. Am J 2000; 14:2321–6. Hematol 2005; 78:207–11. 82 Kobayashi T, Munakata S, Sugiura H et al. Angiotropic lymphoma: 96 Takahashi E, Kajimoto K, Fukatsu T et al. Intravascular large T-cell proliferation of B cells in the capillaries of cutaneous angiomas. lymphoma: a case report of CD30-positive and ALK-negative ana- Br J Dermatol 2000; 143:162–4. plastic type with cytotoxic molecule expression. Virchows Arch 83 Yamaguchi M, Kimura M, Watanabe Y et al. Successful auto- 2005; 447:1000–6. logous peripheral blood stem cell transplantation for relapsed intra- 97 Sukpanichnant S, Visuthisakchai S. Intravascular lymphomatosis: vascular lymphomatosis. Bone Marrow Transplant 2001; 27:89–91. a study of 20 cases in Thailand and a review of the literature. 84 Gatter KC, Warnke RA. Intravascular large B-cell lymphoma. In: Clin Lymphoma Myeloma 2006; 6:319–28. World Health Organisation Classification of Tumours, Pathology and Genetics of 98 Ponzoni M, Ferreri AJ. Intravascular lymphoma: a neoplasm of Tumours of Haematopoietic and Lymphoid Tissues (Jaffe ES, Harris NL, ‘homeless’ lymphocytes? Hematol Oncol 2006; 24:105–12. Stein H et al., eds). Lyon: IARC Press, 2001; 177–8. 99 Zuckerman D, Seliem R, Hochberg E. Intravascular lymphoma: 85 Pileri SA, Dirnhofer S, Went P et al. Diffuse large B-cell lymph- the oncologist’s ‘great imitator’. Oncologist 2006; 11:496–502. oma: one or more entities? Present controversies and possible 100 Murase T, Yamaguchi M, Suzuki R et al. Intravascular large B-cell tools for its subclassification. Histopathology 2002; 41:482–509. lymphoma (IVLBCL): a clinicopathologic study of 96 cases with 86 Oei ME, Kraft GH, Sarnat HB. Intravascular lymphomatosis. Muscle special reference to the immunophenotypic heterogeneity of CD5. Nerve 2002; 25:742–6. Blood 2007; 109:478–85.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp16–25 CUTANEOUS BIOLOGY DOI 10.1111/j.1365-2133.2007.07890.x DNA repair capacities of cutaneous fibroblasts: effect of sun exposure, age and smoking on response to an acute oxidative stress S. Sauvaigo, M. Bonnet-Duquennoy,* F. Odin, F. Hazane-Puch, N. Lachmann,* F. Bonte´,* R. Kurfu¨rst* and A. Favier Laboratoire des Le´sions des Acides Nucle´iques, LCIB (UMR-E3 CEA-UJF), CEA Grenoble, DRFMC SCIB LAN, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France *LVMH Recherche, 45804 Saint Jean-de-Braye, France Laboratoire ORSOX – UMR-E3 UJF/CEA, Universite´ Joseph Fourier, UFR de Me´decine et Pharmacie, Domaine de la Merci, 38700 La Tronche, France
Summary
Correspondence Background Sun irradiation causes skin ageing and cancer through the accumulation Sauvaigo Sylvie. of damage to cell components. Intrinsic ageing is also associated with accumula- E-mail: [email protected] tion of oxidized macromolecules. Objectives In this study we investigated the effects of sun exposure on response to Accepted for publication 20 November 2006 an acute in vitro oxidative stress (H2O2) using normal human fibroblasts prepared from biopsies from 10 volunteers taken from sun-protected and sun-exposed Key words sites. ageing, comet assay, fibroblasts, oxidative damage, Methods Time-course experiments measuring repair of DNA strand-breaks and sun exposure formamidopyrimidine DNA N-glycosylase-sensitive sites were conducted using Conflicts of interest the single-cell gel electrophoresis (comet) assay. None declared. Results Our results demonstrated that repair of strand-breaks was slower in sun- exposed compared with sun-protected cells. Interestingly, ageing was also associ- ated with decreased DNA repair capacities for single-strand breaks in both sun-exposed and sun-protected cells whereas for formamidopyrimidine glycosy- lase (Fpg)-sensitive sites, this feature was in evidence only in sun-protected cells. Smoking, associated with age, was shown to have a markedly negative impact on DNA repair. Conclusions Taken together our data suggest that stresses like ageing, sun exposure and smoking might have an additive effect contributing to the overall heterogen- eity and decrease of DNA repair capacities in human cells and so increase the danger of sun exposure for health. They also emphasize the importance of the quality of the biological samples when repair studies on skin cells are to be conducted.
Skin ageing is a complex process that is thought to have gen- Consequently, UV radiation may promote cell photoageing, etic (intrinsic) determinants on which environmental factors death and/or carcinogenesis.2 (extrinsic) are superimposed.1 Extrinsic skin ageing has been UVA and UVB components have distinct biological effects. mainly attributed to the chronic exposure of human skin to At skin level, UVB produces erythema, burns and, eventually, solar irradiation. This photoageing is the consequence of the as a long-term effect, skin cancer. Indeed, UVB radiation is interaction of ultraviolet UVA radiation (320–400 nm), the directly absorbed by DNA at bipyrimidine sites and has been predominant component of solar UV radiation, and UVB radi- identified as the most mutagenic and carcinogenic component ation (290–320 nm) with cells contained in the different of the solar spectrum. UVA, although considered weakly carci- layers of the skin. One of the most important short-term nogenic, penetrates deeply into skin to reach the dermis and effects of UV on human skin cells includes the formation of causes ageing and wrinkling of the skin.3 Hence fibroblasts are DNA lesions that could lead to DNA mutations if not repaired. important cell targets for UVA radiation. UVA radiation affects
2007 The Authors 26 Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 DNA repair in skin fibroblasts, S. Sauvaigo et al. 27 cellular DNA integrity by two distinct mechanisms. The most Materials and methods generally recognized is activation of endogenous or exogenous photosensitizers that consequently generate reactive oxygen Subjects species (ROS). The putative oxygen species formed are pre- 1 dominantly singlet oxygen ( O2) that produces exclusively Panellist recruitment and biopsies were performed at Labora- 8-oxo-7,8-dihydroguanine (8-oxo-Gua) and, as a minor spe- toire DermExpert (Paris, France) in accordance with ethical cies, hydroxyl radical (oOH) that oxidizes both purines and procedures. The group consisted of 10 healthy Japanese pyrimidines.4 Another mechanism recently characterized women with age ranging from 30 to 45 years (mean involves triplet–triplet energy transfer between the excited 36 years). The study was performed according to the most photosensitizer and two adjacent pyrimidines and leads to the recent recommendations of the World Medical Association formation of cyclobutane pyrimidine dimers.5 (Declaration of Helsinki 1964, amended in Hong Kong, On the other hand, ROS show reactivity towards nearly all 1989). Before the beginning of the study, an information cellular components including proteins and lipids.6 They also sheet detailing the study methods and restrictions was given exert their physiological effects as signalling molecules in to each woman. Their freely given, informed and express con- pathways that are normally involved in cellular homeostasis.7 sent was collected on a signed consent form. Clinical evalu- Hence exposure to UV radiation and oxidative stress can have ation and biopsies were performed for each subject by a opposite effects according to wavelength and dose. On the dermatologist. None of the subjects were receiving oestrogen one hand, oxygen radicals induce expression of genes hormone replacement therapy. Data from two smokers (sub- ) involved in antioxidant defence and DNA repair leading to ject 1: age 31, 10 cigarettes day 1 since 10 years; subject 2: ) adaptation to sun exposure.8,9 On the other hand, stress- age 45, 13 cigarettes day 1 since 15 years) and one previous ) induced damage of proteins decreases various enzymatic smoker (age 45, 1 pack day 1 for 15 years, nonsmoker for activities including DNA repair.10 Thus, rise in ROS levels has 9 years) were gathered in the ‘smoker’ group (n ¼ 3). The two important consequences: accumulation of damage to var- nonsmoker group contained seven subjects. ious cell components and activation of a specific signalling pathway. Both of these effects could have an impact on age- Cell culture ing.7,11 Besides its generation upon UV exposure, oxidative stress occurs within cells from multiple different and variable Cultures of human dermal fibroblasts were established by the mechanisms that are highly dependent on the physiological outgrowth of 3-mm punch biopsies taken from SE and SP state, dietary behaviour, lifestyle and physical activity of each parts of the forearm of the 10 volunteers. Cells were preserved individual.12 Overall exposure of cells to stress (oxidative, UV in liquid nitrogen at the second passage and subsequently cul- radiation, exogenous carcinogens) could participate in the loss tured for experiments before passage 5. Cells were grown in of function of cell components thus accelerating the ageing M199 medium (InVitrogen, Cergy-Pontoise, France) supple- mechanism. In any case, ageing is associated with accumula- mented with 10% fetal calf serum (FCS) and containing ) ) tion of oxidatively modified proteins, lipids and nucleic 100 U mL 1 penicillin and 100 lgmL 1 streptomycin. Care acids13 and the oxidative stress hypothesis of ageing is still was taken to standardize the fibroblast culture conditions. currently one of the most popular theories of how the bio- Twenty-four hours before the treatment, the cells were plated chemical changes associated with ageing occur.14,15 Indeed, into 3Æ5-cm diameter dishes at a density of 150 000 cells the balance between oxidative stress-mediated generation of per dish. ROS and the cells’ antioxidant defences, including the DNA repair capacities, would be expected to have a major impact Treatment and comet assay on ageing. Hence overall photoageing is a complex phenom- enon, a consequence of the superposition of environmental Subconfluent cells were submitted to H2O2 treatment ) factors on chronological skin changes with high individual (20 mmol L 1) for 5 min on ice in cold PBS (InVitrogen). variations.16 The viability of the cells was monitored by a colorimetric To document further the effects of ageing and photoage- MTT cell proliferation assay that evaluates the activity of ing and those of life-style on DNA repair capacities, we mitochondrial dehydrogenase, 24 h after H2O2 treatment. used the single-cell gel electrophoresis assay (comet assay) Approximately 85% of cells remained viable after )1 to analyse the induction and repair of H2O2-induced lesions 20 mmol L H2O2 treatment for 5 min at 4 C. in cultured fibroblasts established from sun-exposed (SE) For repair experiments, cells were treated at various times and sun-protected (SP) skins of ten female subjects. The over an 8-h period and collected simultaneously. They were number of single-strand breaks (SSB) and alkali-labile sites then re-fed with culture medium and put in the CO2 incubator (ALS) as well as the number of Fpg-sensitive sites were at 37 C to allow DNA repair. The alkaline single-cell gel elec- evaluated by the measurement of the percentage of DNA in trophoresis assay was used to determine basal and induced SSB the tail of the comet (Tail DNA). Results of the comet assay and ALS. The level of oxidative damage was associated with the were correlated with data obtained from a medical ques- extent of Fpg-sensitive sites using the modified version of the tionnaire established for each volunteer. comet assay.17 For determination of the kinetics of DNA repair,
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 28 DNA repair in skin fibroblasts, S. Sauvaigo et al.
Tail DNA values were measured at 0, 0Æ5, 1, 4 and 8 h after naire on percentage of residual damage. P-values < 0Æ05 were treatment. The assay was essentially conducted as described by considered as significant. Sauvaigo et al.18 Trypsinized cells were embedded in low-melt agarose at 37 C (final concentration 0Æ6% in PBS) and spread Results on microscope slides coated with two dried layers of 1% nor- mal agarose in PBS. After gelling on ice, the slides were ) ) Cell observation immersed in lysis solution (2Æ5 mol L 1 NaCl, 0Æ1 mol L 1 )1 EDTA Na2, 10 mmol L TRIS, 1% sodium sarcosinate, 1% tri- Cells from SE and SP sites had similar morphology and pro- ton X-100, 10% DMSO, pH 10) at 4 C for 1 h in the dark. liferation capacities for all subjects except for subject 9. The slides were neutralized with three washes (5 min) in Indeed cells from subject 9 showed a large and dendritic ) 0Æ4 mol L 1 Tris–HCl, pH 8 and equilibrated in the Fpg diges- morphology from the second passage whereas all the others ) ) tion buffer (3 · 5 min in 0Æ1 mol L 1 KCl, 0Æ5 mmol L 1 remained in spindle shape. The proliferation rate of these )1 EDTA Na2,0Æ04 mol L Tris–HCl, pH 8). Digestion with Fpg cells was also lower than the others. This nonsmoking sub- ) ) (1Æ7 lgmL 1, 100 lL slide 1) for detection of oxidated ject had the highest tanning index of the group and also had purines was performed for 45 min at 37 C. Control slides were the highest level of acute sun exposure (more than 6 weeks mock treated with the Fpg digestion buffer. After digestion, per year). The last participant showed also the highest body the slides were transferred into the electrophoresis tank filled mass index. ) ) with electrophoresis buffer (0Æ3 mol L 1 NaOH, 1 mmol L 1 EDTA Na ) prechilled at 4 C. The slides were left at room 2 Effect of photoexposure on DNA repair temperature for 30 min and electrophoresis was subsequently accomplished for 30 min at 25 V and 300 mA. After migra- Analyses were performed on pooled data within each panel tion, the slides were rinsed in the neutralization buffer (SP and SE). Mean values of Tail DNA were calculated for (3 · 5 min, room temp) and stained with ethidium bromide cells from SP and SE sites. Analysis of the mean Tail DNA of ) ) (100 lL slide 1,20lgmL 1). For each subject, cells from SE the two populations revealed that the steady-state level of and SP zones were treated in parallel. damage (without treatment) was equivalent in cells from whatever site. No significant difference was observed either for SSB + ALS or for Fpg-sensitive sites. Similarly, cells from Single-cell gel electrophoresis assay, scoring SP and SE zones exhibited the same initial level of damage and analysis induced by H2O2 treatment (Tables 1 and 2). Examination We used the Komet 3Æ1 software from Kinetic Imaging Ltd of the mean Tail DNA values over the 8-h repair period (Bromborough, U.K.) to analyse the slides. Fifty randomly (experimental points at 0Æ5, 1, 2, 4 and 8 h) showed that selected nuclei were scored by slide and duplicate slides were for all cell strains, the DNA damage detected decreased with processed for each experimental point. The extent of damage increasing repair interval. However, we observed different was evaluated by the Tail DNA value defined as the percentage rates of return to initial level dependent on the amount of of DNA in the tail of the comet. Four repair curves (Tail DNA damage. During the first repair interval (corresponding to as a function of repair time) were obtained for each subject: 0Æ5 h), approximately 60% of H2O2-induced SSB + ALS were SE vs. SP for SSB + ALS and for Fpg-sensitive sites. repaired. By contrast, repair of Fpg-sensitive sites was slower (30% repair in 0Æ5 h). Repair of SSB + ALS was completed in 2 h in SP cells (P <0Æ05) while it took between 4 and Data analysis 8 h in SE cells (Table 1). Surprisingly, Fpg-sensitive sites For establishment of correlations between repair parameters were not totally repaired over the 8-h period, whatever the and characteristics of the subjects, percentage of residual dam- population of cells considered (P <0Æ05; Table 2). Approxi- age was considered. Untreated controls were called C. For mately 30% of Fpg-sensitive sites remained unrepaired after each time point, residual damage (RD) was calculated as the 8-h recovery. ratio between residual damage at the considered time (It–C) and initial damage of treated cells (I0–C) with respect to the )1 Effect of age on DNA repair untreated control cells (C) [RD–(It–C) (I0–C) ]. The effect of age on the level of residual damage was inves- No significant correlations of age and steady-state level of tigated by linear regression and by one-way ANOVA using damage were established. Similarly the initial amount of individual data, and after samples were grouped into two induced damage (I0–C) was not dependent on subject age categories of age [group 1 (< 35 years; mean 31Æ2; n ¼ 5) (data not shown). Residual damage was considered for the and group 2 (> 35Æ01 years; mean 40Æ6 years; n ¼ 5)]. ANOVA analysis of correlations between damage repair kinetics and is based on the F-ratio, which is the ratio of the population subject characteristics. Correlations between age and residual variance as estimated between groups vs. that within groups. damage were examined using individual data. It was found A one-way ANOVA was also applied to determine the effect of that residual SSB + ALS correlated linearly with age in SP life habit values investigated through the medical question- cells after 4-h repair with an accuracy of 68%. A positive
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 DNA repair in skin fibroblasts, S. Sauvaigo et al. 29
Table 1 Tail DNA values of single-strand breaks and alkali-labile sites (SSB + ALS) obtained at specific time points during repair after exposure of the fibroblasts to H2O2
Time points of analysis 0h 0Æ5 h 1 h 2 h 4 h 8 h Control SP 45Æ3* 23Æ9* 17Æ9* 14Æ613Æ915Æ513Æ3 CI 42Æ8–47Æ821Æ4–26Æ415Æ4–20Æ412Æ1–17Æ111Æ4–16Æ413Æ0–18Æ010Æ8–15Æ8 SE 44Æ5* 24Æ9* 20Æ2* 16Æ8* 15Æ714Æ414Æ2 CI 41Æ9–47Æ122Æ3–27Æ517Æ6–22Æ814Æ2–19Æ413Æ1–18Æ311Æ8–17Æ011Æ6–16Æ8
SP, fibroblasts from sun-protected site; SE, fibroblasts from sun-exposed site; CI, confidence interval. *Data significantly different (P <0Æ05) from the Control (nontreated cells).
Table 2 Tail DNA values of formamidopyrimidine DNA glycosylase-sensitive sites obtained at specific time points during repair after exposure of the fibroblasts to H2O2
Time points of analysis 0h 0Æ5 h 1 h 2 h 4 h 8 h Control SP 75Æ5* 62Æ6* 54Æ7* 58Æ1* 53Æ1* 50Æ1* 41Æ2 CI 71Æ8–79Æ258Æ9–66Æ351Æ0–58Æ454Æ4–61Æ849Æ4–56Æ846Æ4–53Æ837Æ5–44Æ9 SE 74Æ8* 62Æ2* 56Æ8* 54Æ9* 51Æ3* 53Æ2* 43Æ6 CI 70Æ8–78Æ858Æ2–66Æ252Æ8–60Æ850Æ9–58Æ947Æ3–55Æ349Æ2–57Æ239Æ6–47Æ6
SP, fibroblasts from sun protected site; SE, fibroblasts from sun exposed site; CI, confidence interval. *Data significantly different (P <0Æ05) from the Control (nontreated cells). correlation was also found after 1-h repair in the SE group (a) (R2 ¼ 51%). Linear regression curves at 4-h repair for SP and 0·31 SP 4 h at 1-h repair for SE are shown in Figures 1a,b, respectively. 0·21 Residual damage at Fpg-sensitive sites and age were positively correlated in SP cells after 4 h of repair (R2 ¼ 44%). At these 0·11 particular times (4 h for SP and 1 h for SE), correlation ana- 0·01 lyses were also performed by comparing mean values obtained –0·09 from data pooled in two groups as a function of subjects’ Residual damage age (subjects were categorized in two groups: group 1a –0·19 (< 35 years; mean 31Æ2; n ¼ 5) and group 1b (‡ 35 years; 30 33 36 39 42 45 mean 40Æ6 years; n ¼ 5). The results obtained from pooled Age (years) data confirmed the effect of age on residual damage [P <0Æ05 (b) (data not shown)]. 0·48 SE 1 h 0·38
Correlation of DNA repair with other values of the 0·28 medical questionnaire 0·18 Using the medical questionnaire, we evaluated the relationship 0·08 between skin characteristics, smoking, life habits, food intake Residual damage and DNA repair kinetics for each subject. A striking feature –0·02 30 33 36 39 42 45 appeared in SP cells: initial repair (0Æ5 and 1 h) of SSB + ALS Age (years) and Fpg-sensitive sites was significantly faster in nonsmokers (n ¼ 7; mean age 34Æ4 years) than in smokers (n ¼ 3; mean Fig 1. Effect of age on DNA repair [residual damage (RD)]. Level of age 39Æ3 years) (Figs 2a,b). This was also the case for the RD correlates with age in sun-protected (SP) fibroblasts (a) after 4-h repair of SSB + ALS at 8 h. On the contrary, no correlation repair (R2 ¼ 68%) and in sun-exposed (SE) fibroblasts (b) after 1-h was found in the SE cells. More precisely the other parameters repair (R2 ¼ 51%). RD was calculated as the ratio between absolute considered were: phototype, skin tanning, photoageing, fruit level of initial damage (I0 – C) and damage measured respectively )1 and vegetable intake, overall iron intake, meat and fish con- 4 h (a) and 1 h (b) after the stress (RD ¼ [I1h –C][I0 –C] and )1 sumption, beverage consumption (alcohol, tea and coffee). RD ¼ [I4h –C][I0 –C] ), where C denotes untreated control cells.
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 30 DNA repair in skin fibroblasts, S. Sauvaigo et al.
Regarding the repair profile of SSB + ALS and Fpg-sensitive SP-SSB (a) 1·3 sites, substantial differences were observed between indi- viduals. Interindividual variability is a common feature of 1 * * experiments conducted on human primary fibroblasts or 0·7 * lymphocytes using the comet assay19–21 and may be attributed 21 0·4 to intrinsic properties of the cells. Of course, Chazal et al. have emphasized the variability attributable to the methodo-
Residual damage 0·1 Non-smokers logy itself. Despite the differences, some specific features Smokers –0·2 emerged from the overall data. 0 0·5 1 2 4 8 Measurement of steady-state levels of damage showed that Repair time (h) there was no significant accumulation of oxidative damage in (b) SP-Fpg the cultured fibroblasts, caused by previous sun exposure of 1·3 * * the skin. This is not surprising as it may be emphasized that 1 sun exposure could lead to accumulation of mutations over time rather than direct accumulation of oxidative lesions. 0·7 Moreover, accumulation of oxidative damage in fibroblasts 0·4 seems to be a feature of in vitro ageing rather than in vivo ageing.22,23
Residual damage 0·1 Non-smokers Smokers In addition, SP and SE cells displayed similar sensitivity to –0·2 oxidative stress as no statistical difference in induced initial 00·51248 level of damage between SE and SP cells was observed. On Repair time (h) the contrary, from the analysis of mean Tail DNA values and of the absolute amount of damage, it clearly appeared Fig 2. Effect of smoking on DNA repair (residual damage). Time that SP cells repaired SSB + ALS faster than SE cells (2 h vs. course repair over an 8-h period of single-strand breaks (SSB) + alkali-labile sites (ALS) (a) and formamidopyrimidine DNA 4 to 8 h). N-glycosylase (Fpg)-sites (b) in sun-protected (SP) fibroblasts It has been suggested that many of the age-associated func- established from smokers (full line) and nonsmokers (dotted line) tional losses of human cells are accelerated in photoaged using mean values ± SD. Lesions are expressed as residual damage. cells.24 A plausible explanation for the decline of repair in SE The (*) indicates significantly different data between smokers and cells could be that genomic as well as mitochondrial DNA nonsmokers (P <0Æ05) obtained using ANOVA. mutations accumulate upon sun exposure and exposure to ROS25 and therefore impede the normal functions of the cells in response to an oxidative stress. None of these parameters influenced significantly the amount The better ability of SP cells compared with SE cells to of residual damage, but we cannot conclude definitely as the repair DNA was not ascertained for Fpg-sites. While standard number of subjects in the present study is limited. deviations (SD) calculated using data (Tail DNA values) of steady-state and initial levels of damage were similar for Discussion SSB + ALS and Fpg-sites, we observed that SD calculated using data of time-course repair experiments were higher for The aim of this study was to get insights into factors that Fpg-digested samples than for SSB + ALS measurement (data could have a significant impact on DNA repair response using not shown). Hence, whether it is due to experimental bias or skin fibroblasts as a cellular model. The parameters considered to interindividual variability, Fpg digestion introduced disper- were sun exposure, age, life and food habits. Hence, DNA sion of the data. As a consequence significance in the correl- repair kinetics of H2O2-induced damage of fibroblasts from ations was difficult to identify. 10 individuals were analysed using the comet assay. On another hand, contrary to what was observed with Designing the study, we tried to minimize sources of bias. SSB + ALS, for which a return to the steady-state level was As different tissues may have specific characteristics, biopsies observed, we also noticed that repair of Fpg-sites was not were taken from the same body location on each participant complete over the 8-h recovery period. Indeed, approximately (forearm). The selected cohort had quite homogenous charac- 30% of Fpg-sensitive sites remained unrepaired. teristics (small age range, same sex, Japanese origin). In order Very few papers document the repair kinetics of oxidative to understand better the effects of sun exposure on DNA damage in cells in general and in fibroblasts in particular. repair capacities, cells were taken from SP and SE sites in each However, our kinetic data are in agreement with those found subject and the comet assay was conducted in parallel for SP by Osterod et al.26 established from immortalized mouse fibro- and SE cells. An important point to underline is that experi- blasts. ments were performed with early passaged cells (< 6) to According to our results, considering individual data, repair maintain as far as possible the epigenetic modification induced efficiency of SSB + ALS was inversely correlated with age, in by chronic sun exposure. SP and in SE cells. Although the correlation was noted only at
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 DNA repair in skin fibroblasts, S. Sauvaigo et al. 31 particular time points, this was not expected because the range 8 Leccia MT, Yaar M, Allen N et al. Solar simulated irradiation of the cohort age was relatively narrow (30–45 years). How- modulates gene expression and activity of antioxidant enzymes in ever, correlations between DNA repair efficiency and age were cultured human dermal fibroblasts. Exp Dermatol 2001; 10:272–79. 9 Li J, Lee JM, Johnson JA. Microarray analysis reveals an antioxidant also obtained using pooled data that reinforce the very signifi- responsive element-driven gene set involved in conferring protec- cant correlation. Decreased repair capacity with age was also tion from an oxidative stress-induced apoptosis in IMR-32 cells. observed for Fpg-sensitive sites but only in SP cells. It is not J Biol Chem 2002; 277:388–94. clear yet why the correlations could be found only at specific 10 Beal MF. Oxidatively modified proteins in aging and disease. time points. Free Rad Biol Med 2002; 32:797–803. Decline of repair capacities with age has been documented 11 Wlaschek M, Tantcheva-Poor I, Naderi L et al. Solar UV irradiation for both base excision repair and nucleotide excision repair and dermal photoaging. J Photochem Photobiol B: Biol 2001; 63:41– 51. activities by different authors using various techniques and 27,28 12 Stadtman E. Importance of individuality in oxidative stress and different cellular models. We show here that it can be aging. Free Rad Biol Med 2002; 33:597–604. observed even with a small age range provided that strictly 13 Bohr V, Anson, RM, Mazur S, Dianov G. Oxidative DNA damage defined culture and assay conditions are used. The current processing and changes with aging. Toxicol Let 1998; 102–103:47– investigation has also demonstrated that cells established from 52. the same subjects but taken from body sites differing in their 14 Harman D. Aging: a theory based on free radical and radiation history of sun exposure displayed different DNA repair capaci- chemistry. J Gerontol 1956; 11:298–300. 15 Stadtman ER. Protein oxidation and aging. Science 1992; 257:1220– ties. These considerations emphasize the importance of con- 24. trolling fibroblast origin and personal characteristics of the 16 Hadshiew IM, Eller MS, Gilchrest BA. Skin aging and photoaging: subject as well as cell passage number when correlation stud- the role of DNA damage and repair. Am J Cont Der 2000; 11:19– ies have to be addressed. 25. Yin et al.29 have demonstrated that sun exposure, age and 17 Collins AR, Duthie S, Dodson VL. Direct enzymic detection of smoking are independent risk factors that contribute to pre- endogenous oxidative base damage in human lymphocyte DNA. mature skin ageing through alterations of matrix metalloprote- Carcinogenesis 1993; 14:1733–35. 18 Sauvaigo S, Petec-Calin C, Caillat S et al. Comet assay coupled to inase-1 mRNA expression. We now show that these three repair enzymes for the detection of oxidative damage to DNA factors have a significant adverse impact on repair capacities. induced by low doses of gamma-radiation: use of YOYO-1, Therefore, our data support the concept that decline in repair low background slides, and optimized electrophoresis conditions. rates might be ascribed to overall exposure to genotoxic and Anal Biochem 2002; 303:107–9. cytotoxic insults and that photoageing is associated with 19 Brammer I, Zoller M, Dikomey E. Relationship between cellular decreased DNA repair capacities. radiosensitivity and DNA damage measured by comet assay in In conclusion, further work involving a larger cohort would human normal, NBS and AT fibroblasts. Int J Radiat Biol 2001; 77:929–38. permit the more precise assessment of the contribution of 20 Twardella D, Popanda O, Helmbold I et al. Personal characteristics, acquired factors on the modulation of DNA repair capacities. therapy modalities and individual DNA repair capacity as predictive In particular, the effect of smoking should be studied without factors of acute skin toxicity in an unselected cohort of breast can- the bias of age. However, it may be speculated that overexpo- cer patients receiving radiotherapy. Radiotherapy Oncol 2003; 69:145– sure to different types of genotoxic agents has cumulative 53. effects that contribute to photoageing in relation to decreased 21 Chazal M, Roux E, Alapetite C et al. Interexperimental and inter- DNA repair capacities. individual variations of DNA repair capacities after UV-B and UV-C irradiations of human keratinocytes and fibroblasts. Photochem Photobiol 2004; 79:286–90. References 22 Wolf FI, Torsello A, Covacci V et al. Oxidative DNA damage as a marker of aging in WI-38 human fibroblasts. Exp Gerontol 2002; 1 Fisher GJ, Kang S, Varan IJ et al. Mechanisms of photoaging and 37:647–56. chronological skin aging. Arch Dermatol 2002; 138:1462–70. 23 Kaneko T, Tahara S, Taguchi T, Kondo H. Accumulation of oxida- 2 Sarasin A. The molecular pathways of ultraviolet-induced carcino- tive damage, 8-oxo-2¢-deoxyguanosine, and change of repair genesis. Mutat Res 1999; 428:5–10. systems during in vitro cellular aging of cultured human skin 3 Krutmann J. Ultraviolet A radiation-induced biological effects in fibroblasts. Mutat Res 2001; 487:19–30. human skin: relevance for photoaging and photodermatosis. 24 Yaar M, Gilchrest BA. Aging versus photoaging: postulated J Dermatol Sci 2000; 23 (Suppl. 1):S22–S26. mechanisms and effectors. J Invest Dermatol Symp Proc 1998; 3:47– 4 Cadet J, Sage E, Douki T. Ultraviolet radiation-mediated damage to 51. cellular DNA. Mutat Res 2005; 571:3–17. 25 Berneburg M, Gattermann N, Stege H et al. Chronically ultraviolet- 5 Courdavault S, Baudouin C, Charveron M et al. Larger yield of exposed human skin shows a higher mutation frequency of cyclobutane pyrimidine dimers than 8-oxo-7,8-dihydroguanine in mitochondrial DNA as compared to unexposed skin and the the DNA of UVA-irradiated human skin cells. Mutat Res 2004; hematopoietic system. Photochem Photobiol 1997; 66:271–5. 556:135–42. 26 Osterod M, Hollenbach S, Hengstler JG et al. Age-related and 6 Stadtman ER, Berlett BS. Reactive oxygen-mediated protein oxida- tissue-specific accumulation of oxidative DNA base damage in tion in aging and disease. Drug Metab Rev 1998; 30:225–43. 7,8-dihydro-8-oxoguanine-DNA glycosylase (Ogg1) deficient mice. 7 Finkel T, Holbrook NJ. Oxidants, oxidative stress and the biology Carcinogenesis 2001; 22:1459–63. of aging. Nature 2000; 408:239–47.
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 32 DNA repair in skin fibroblasts, S. Sauvaigo et al.
27 Goukassian D, Gad F, Yaar M et al. Mechanisms and implications of 29 Yin L, Morita H, Tsuji T. Skin aging induced by ultraviolet the age-associated decrease in DNA repair capacity. FASEB J 2000; exposure and tobacco smoking: evidence from epidemiological 14:1325–34. and molecular studies. Photodermatol Photoimmunol Photomed 2001; 28 Chen SK, Hsieh WA, Tsai MH et al. Age-associated decrease of 17:178–83. oxidative repair enzymes, human 8-oxoguanine DNA glycosylase (hOGG1), in human aging. J Radiat Res 2003; 44:31–5.
2007 The Authors Journal Compilation 2007 CEA • British Journal of Dermatology 2007 157, pp26–32 CUTANEOUS BIOLOGY DOI 10.1111/j.1365-2133.2007.07990.x A cytotoxic analysis of antiseptic medication on skin substitutes and autograft Q. le Duc,* M. Breetveld,* E. Middelkoop, R.J. Scheper,§ M.M.W. Ulrich* and S. Gibbs* Departments of *Dermatology, Plastic, Reconstructive and Hand Surgery and §Pathology, VU University Medical Centre, De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands Association of Dutch Burns Centres, Beverwijk, the Netherlands
Summary
Correspondence Background There is an increasing demand for the clinical application of human S. Gibbs. skin substitutes (HSSs) for treating ulcers, burns and surgical wounds. Due to this E-mail: [email protected] increasing demand and due to the simultaneous requirement for the administra- tion of topical antiseptic medications, there is a need to determine potential cyto- Accepted for publication 12 March 2007 toxic effects of these medications on HSSs compared with autograft skin. Objectives To perform such an evaluation. Key words Methods Two different HSSs were used (autologous reconstructed epidermis on antimicrobial, autograft, cytotoxicity, skin fibroblast-populated human dermis and allogeneic reconstructed epidermis on a substitute fibroblast-populated rat collagen gel) and were compared with conventional Conflicts of interest full-thickness autograft. Twelve different antiseptics were applied topically to the None declared. stratum corneum in vitro for 24 h. The degree of cytotoxicity was analysed as detrimental changes in histology, metabolic activity (MTT assay) and RNA staining of tissue sections. Results The antiseptic medications tested showed different degrees of cytotoxicity. Acticoat , Aquacel Ag , Dermacyn , Fucidin ,0Æ5% silver nitrate solution and chlorhexidine digluconate were not cytotoxic for either HSS or autograft, and can therefore be used as required. Flamazine and zinc oxide cream resulted in moderate cytotoxicity. However, application of Betadine , cerium-silver sulfadia- zine cream, silver sulfadiazine cream with 1% acetic acid and Furacine resulted in a substantial decrease in cell viability and a detrimental effect on tissue hist- ology when applied to autograft and especially to HSS. Conclusions Due to the potential cytotoxic effect of some antiseptics on HSS, it is advised that clinicians balance the cytotoxicity of the medication, its antiseptic properties and the severity of colonization in choosing which one to apply.
Medical specialists are still challenged in finding the optimal split-skin).2,3 However, wound closure is a lengthy process treatment for difficult-to-heal wounds, e.g. ulcers, trauma- and a therapy-resistant group of ulcers remains open.3 induced wounds and deep burns. Open wounds are sus- In all areas of wound healing, advances are being made in ceptible to invading pathogens such as bacteria and fungi. the field of tissue engineering as it is now thought that opti- Therefore optimal wound healing is dependent on the type of mal wound healing may be achieved by application of tissue- antiseptic medication used in combination with the method of engineered skin products. Developing areas of application wound closure. In the case of small, acute wounds and large include ulcers (venous, diabetic foot, decubitus), burns, burn ⁄trauma wounds an autograft (split-skin or full-thickness) trauma-induced wounds, tumour excision sites and treatment is conventionally applied to close the wound directly.1 Use of of bullous disease.4 Application of tissue-engineered skin con- autograft skin requires relatively large amounts of donor skin structs has certain advantages over conventional methods: allo- which is often limited and the resulting scar formation, par- geneic human skin substitute (HSS) requires no donor skin ticularly when meshed or split-skin is used, is suboptimal. and is an off-the-shelf product; autologous HSS requires less Ulcer patients are conventionally treated with wound donor skin than autograft skin and the use of the patient’s debridement followed by wound dressings and ⁄or compres- own cells ensures a good take. Moreover, tissue engineering sion therapy and in some cases autografts (punch biopsies or is aimed at the modification of skin substitutes towards the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 33 34 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al.
) ) treatment of specific problems in wounds, for instance the stimu- cortisone, 1 lmol L 1 isoproterenol and 0Æ1 lmol L 1 insulin, lation of granulation tissue when applied to nonhealing ulcers and cultures were lifted to the air–liquid interface and cultured for the prevention of scar formation when applied to burns. a further 4 days in standard HSS culture medium [DMEM ⁄ ) As most wound beds are prone to colonization with patho- Ham’s F12 (3 : 1) (ICN Biomedicals), 1 lmol L 1 hydrocorti- ) ) gens such as bacteria and fungi, wounds treated with HSS will sone, 1 lmol L 1 isoproteronol, 0Æ1 lmol L 1 insulin, )5 )1 )2 )1 often also require antiseptics. Ulcers and large burns contain 1Æ0 · 10 mol L L-carnitine, 1Æ0 · 10 mol L L-serine, )1 in the majority of cases aerobic Gram-negative (Pseudomonas 1 lmol L DL-a-tocopherol, penicillin and streptomycin aeruginosa and anaerobic cocci) and Gram-positive bacteria acetate] and enriched with a lipid supplement [containing ) ) (Staphylococcus aureus) (for review see Pruitt et al.5 and Jones 25 lmol L 1 palmitic acid, 15 lmol L 1 linoleic acid, ) ) et al.6). The colonization in a wound can impair the take of 7 lmol L 1 arachidonic acid and 24 lmol L 1 bovine serum the HSS or autograft and in turn impair optimal wound albumin] supplemented with 0Æ2% ultroserG. Hereafter, HSS- healing. If left inadequately treated colonization can lead to allo was cultured in standard keratinocyte culture medium ) morbidity and even mortality. supplemented with 50 lgmL 1 of ascorbic acid for an add- As antiseptics often have to be applied together with HSS itional 14 days. treatment, it is important to evaluate the possible cytotoxicity of these antiseptics on HSS. For this study we used two differ- HSS-auto culture ent full-thickness HSSs, which are constructed in our labora- tory. The first HSS is autologous reconstructed epidermis on For each 4 cm2 HSS-auto to be constructed, three 3-mm fibroblast-populated human dermis (HSS-auto). This HSS is diameter punch biopsies were required. Epidermal sheets routinely used in our hospital for closing chronic wounds.7 and dermal fibroblasts were separated from human adult The second HSS is allogeneic (five pooled foreskin donors) abdominal skin by incubation in dispase II (Roche, Mann- reconstructed epidermis on a fibroblast-populated rat collagen heim, Germany) overnight at 4 C. HSS was constructed gel (HSS-allo). This model is based on that of Bell et al.8 A sim- exactly as previously described.7 In short: epidermal sheets ilar model is also routinely used for closing ulcers9 and has also were placed SC side upwards on dead de-epidermized dermis been described in the treatment of burns.10 Both HSSs have a and cultured air-exposed in standard HSS medium supplemen- fully differentiated epidermis consisting of a basal layer (BL), ted with 0Æ2% ultroserG. After 7 days of culturing the primary stratum spinosum (SS), stratum granulosum (SG) and, import- fibroblast culture (in DMEM supplemented with 1% ultroserG antly, a stratum corneum (SC). The presence of the SC means and penicillin and streptomycin) and epidermal sheet apart, that, similar to autograft, HSSs exhibit barrier competency. the fibroblasts were placed in contact with the reticular surface Therefore antiseptics can be applied topically and must first of the de-epidermized dermis in order to allow fibroblast penetrate the nonviable SC in order to exert a cytotoxic effect migration into the dermis and then the HSS-auto was further on the living cell layers below. These HSSs were compared cultured for 2 weeks. During the 3-week culture period, the with full-thickness autograft skin in this study. Twelve antisep- epidermis expanded to cover the dermis, resulting in approxi- tics that are routinely used in ulcer and burn wound treatment mately 20-fold amplification of the original surface area of the were analysed for their cytotoxic effect upon topical applica- epidermal sheet. tion in vitro to the two HSSs and the full-thickness autograft. Autograft Materials and methods Full-thickness skin obtained from human abdominal reduction within 12 h after surgery was washed in phosphate-buffered Human skin substitute culture and autograft preparation saline, removed of fat and cut into pieces of 4 cm2. All chemicals were derived from Sigma-Aldrich Chemie BV HSS-allo culture (Zwijndrecht, the Netherlands) unless otherwise stated. Dermal fibroblasts were isolated from neonatal foreskins and cultured as described by Ponec et al.11 in Dulbecco’s modified Antiseptic exposure Eagle’s medium (DMEM) (ICN Biomedicals, Irvine, CA, U.S.A.) containing 1% ultroserG and penicillin ⁄streptomycin (Invitro- All experiments were performed with three independent gen, Paisley, U.K.). Fibroblasts from five independent donors donors for autograft and HSS-auto, and three independent cul- were pooled and incorporated into collagen gels (1 · 105 tures (originating from different donor pools) for HSS-allo, ) cells mL 1) essentially as described by Smola et al.12 each being performed in duplicate within each single donor or Epidermal keratinocytes were isolated from neonatal fore- donor pool. Autograft and HSSs were placed upon a 24-mm, skins, essentially as described earlier.11,13 The subconfluent, 3-lm pore size transwell (Corning Inc., Corning, NY, U.S.A.). second-passage cultures of five independent keratinocyte The bottom compartment of the transwell was filled with donors were pooled and seeded on to fibroblast-populated 1Æ5 mL standard keratinocyte culture medium supplemented ) collagen gels. After culturing overnight in medium containing with 50 lgmL 1 ascorbic acid. In the case of autograft and ) DMEM ⁄Ham’s F12 (3 : 1), 1% ultroserG, 1 lmol L 1 hydro- HSS-auto, culture medium was additionally supplemented with
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al. 35
Table 1 Composition of antiseptics used
Name Composition Supplier Acticoat Absorbent dressing with nanocrystalline silver Smith & Nephew BV, Hoofdorp, NL Aquacel Ag Hydrofibre dressing containing silver ConvaTec, Woerden, NL Dermacyn Superoxidized aqueous solution Oculus Innovative Sciences Netherlands BV, Sittard, NL ) Fucidin Ointment; 20 mg g 1 sodium fusidate in lanolin, cetanol, Leo Pharma BV, Breda, NL liquid paraffin and white vaseline ) Furacine Ointment; nitrofurazone 2 mg g 1 in macrogol 300, 1000, Norgine BV, Amsterdam, NL 3000 and water Chlorhexidine digluconate Solution; chlorhexidine digluconate 0Æ5% in alcohol 70% Lommerse Pharma BV, Oss, NL ) Flamazine Cream; silver sulfadiazine 10 mg g 1 cream (polysorbate Solvay Pharma BV, Weesp, NL 60, 80, glyceryl monostearate, cetylalcohol, liquid paraffin; propylene glycol, water) Betadine Ointment; povidone-iodine, macrogol 400, 4000, 6000 Viatris Manufacturing BV, Diemen, and water NL Zinc oxide cream (local Cream; zinc oxide 5% in cetomacrogol In-house pharmacy, Red Cross brand) Hospital, Beverwijk, NL Cerium-silver sulfadiazine Cream; cerium (III) nitrate 2Æ2%, silver sulfadiazine 1% in In-house pharmacy, Red Cross cream (local brand) Cetiol V, cetomacrogol, sorbitol 70% Hospital, Beverwijk, NL Silver sulfadiazine containing Ointment; silver sulfadiazine 1% with 1% acetic acid In-house pharmacy, Red Cross 1% acetic acid (local brand) Hospital, Beverwijk, NL Silver nitrate solution (local Aqueous solution of silver nitrate 0Æ5% in macrogol 400 In-house pharmacy, Red Cross brand) 30% and sorbitol 20% Hospital, Beverwijk, NL
NL, Netherlands.
0Æ2% ultroserG. Antiseptic ointments and creams were spread were transferred to a 96-well microtitre plate containing ) evenly in a thin film over 18-mm filter paper discs (Epitest 200 lL per well isopropanol, acidified with 0Æ04 mol L 1 HCl Ltd, Tuusula, Finland). For antiseptic solutions (see Table 1) 3 : 1, and incubated overnight at 37 C. One hundred micro- and sodium dodecyl sulphate (SDS), 200 lL of the solution litres of the extractant solution was transferred to a 96-well was used to soak the filter paper discs completely. SDS concen- microtitre plate for optical density measurements at 550 nm trations used were 10% for autograft and 1% for HSS. These and a reference wavelength of 650 nm. SDS concentrations resulted in total loss of cell viability in autograft and HSS, respectively. Dressings were cut in appro- Histology and RNA staining priate 3-cm2 pieces. Antiseptics were placed on HSS and auto- graft so that they were in direct contact with the SC in a To assess cytotoxicity with respect to changes in tissue archi- similar manner to applying the antiseptics in vivo. Therefore the tecture, samples were fixed in 4% paraformaldehyde and proc- antiseptic could only penetrate via the SC. Care was taken that essed for conventional paraffin embedding. Sections (5 lm) no leakage occurred around the edges of the HSS or autograft were cut and stained with haematoxylin and eosin for light into the culture medium. Duplicate autograft or HSS cultures microscopic examination. Epidermal cytotoxicity, measured as were exposed within a single experiment for 24 h at 37 Cin a decrease in keratinocyte RNA, was assessed by pyronine Y 15 an atmosphere containing 7Æ5% CO2. staining as described previously. Five-micrometre sections were deparaffinized and then incubated for 30 min at room temperature in a fresh pyronine Y staining solution [0Æ1% Measurement of cytotoxicity pyronine Y (Fluka Chemie GmbH, Buchs, Switzerland) in ) 0Æ2 mol L 1 sodium acetate buffer, pH 4Æ0]. The sections were MTT assay washed three times in water, air dried, and embedded in The MTT assay measures mitochondrial metabolic activity and Depex mounting medium (BDH, Poole, U.K.). Results were was essentially performed as described by Mosmann.14 Biop- analysed by two independent observers. sies were taken from autograft and both HSSs with a biopsy punch 3 mm in diameter (Microtek Medical BV, Zutphen, the Statistical analysis Netherlands) and transferred to a 96-well Microlon 200TM flat-bottom microtitre plate (Greiner Bio-one GmbH, Fricken- The unpaired Mann–Whitney test (nonparametric) was used ) hausen, Germany) filled with 200 lL per well of 2 mg mL 1 for statistical evaluation (GraphPad Prism, San Diego, CA, MTT labelling reagent (Roche). After 2 h at 37 C the biopsies U.S.A.). P <0Æ05 was considered statistically significant.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 36 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al.
Results Table 2 Histological analysis of cytotoxicity
In order to determine the cytotoxic effect of antiseptic agents Antiseptic Autograft HSS-auto HSS-allo on autograft and the two different HSS models, twelve anti- Unexposed ))) septic agents were applied topically and cytotoxicity was Acticoat ))) determined by histology, metabolic activity (MTT) and detec- Aquacel Ag ))) tion of cellular RNA. For each factor, the degree of cytotoxi- Dermacyn ))+ city was compared with that in unexposed cultures and in Fucidin ) ++ cultures exposed to extreme toxic concentrations of SDS. Silver nitrate solution ) ++ Unexposed HSS and autograft consist of fully differentiated Chlorhexidine digluconate + + + epidermis (BL, SS, SG and SC present) on a fibroblast-popu- Flamazine ++ ++ ++ Zinc oxide cream ++ ++ ++ lated dermal matrix (Fig. 1a). Detrimental changes in tissue Cerium-silver sulfadiazine cream +++ +++ +++ histology, for example vacuole formation, condensed nuclei, Furacine ++ +++ +++ and separation of the epidermis from the dermal matrix, are Silver sulfadiazine containing +++ +++ +++ directly related to cytotoxicity (Table 2, Fig. 1b). Exposure of 1% acetic acid autograft and both HSSs to Acticoat , Aquacel Ag and Der- Betadine ++ +++ +++ macyn had no detrimental effects on tissue histology with SDS 10% ⁄1% +++ +++ +++ the exception of very mild vacuole formation in the upper SDS, sodium dodecyl sulphate. Haematoxylin and eosin-stained layers of the epidermis in HSS-allo cultures exposed to Derma- sections were analysed with respect to epidermal cytotoxicity, cyn . Exposure to Fucidin and silver nitrate solution resulted taking into account detrimental effects on tissue architecture in mild vacuole formation in both HSSs, while autograft was compared with control, intactness of basal layer, vacuole forma- still unaffected. Chlorhexidine digluconate exposure resulted tion, and condensed nuclei (see Fig. 1b). Cytotoxicity was in similar mild vacuole formation in all three exposed models. indexed as follows: ) (nontoxic) fi +++ (toxic). Data are Antiseptics that resulted in partial separation of the epidermis derived from six samples from three independent donors. from the dermal matrix in addition to mild vacuole formation in autograft and both HSSs were Flamazine and zinc oxide condensed nuclei, vacuole formation) were found in all mod- cream. Severe detrimental effects on tissue histology (e.g. els after exposure to cerium-silver sulfadiazine cream, silver separation of the epidermis from the dermal matrix, sulfadiazine cream containing 1% acetic acid, Furacine and
Fig 1. Histological analysis of cytotoxicity. (a) Tissue architecture of autograft, HSS-auto and HSS-allo. Haematoxylin and eosin staining of 5-lm paraffin-embedded sections is shown. (b) Tissue architecture of HSS-auto after a 24-h topical exposure to antiseptics showing increasing degrees of cytotoxicity. Cytotoxicity was indexed as follows: ) fi +++ indicating increasing cytotoxicity with ) = unexposed culture, + = chlorhexidine digluconate, ++ = Flamazine , +++ = Betadine (see also Table 2). Data are representative of six samples from three independent donors.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al. 37
Betadine . These effects were slightly less pronounced for autograft exposed to Furacine and Betadine (a) Autograft Cytotoxicity can also be determined by the degree of Unexp decrease in cell viability. Because metabolic activity is repre- Acticoat Aquacel-Ag sentative of cell viability, the effect of the antiseptic agents on Dermacyn ** autograft and both HSS models was analysed in an MTT assay Fucidin * that measures mitochondrial activity (Fig. 2). The antiseptics Silver nitrate Chlorhexidine that resulted in MTT levels comparable with those in un- Flamazine ** exposed models were Acticoat , Aquacel Ag and Dermacyn . Zinc-oxide cer-SS ** MTT levels for this group were > 70% (unexposed cul- Furacine *** tures = 100%), which implies a very mild cytotoxic effect of ** SS-HAc *** these antiseptics on all three models. This result corresponds Betadine *** to our histological observations described above. Antiseptics SDS 0 102030405060708090 100 110 that resulted in a decrease in MTT in the range of 30–70% MTT absorbance (% of control) compared with unexposed models were Fucidin , silver nitrate solution, chlorhexidine digluconate, Flamazine , zinc (b) HSS-autologous oxide cream and cerium-silver sulfadiazine cream. Within this Unexp Acticoat group a certain degree of variation was observed between the Aquacel-Ag * two HSS models, reflecting the intrinsic differences with Dermacyn * respect to penetration and cytotoxicity between the two con- Fucidin Silver nitrate *** structs. The antiseptics that resulted in the lowest MTT levels Chlorhexidine ** *** in HSSs were Furacine , silver sulfadiazine cream containing Flamazine ** 1% acetic acid and Betadine . MTT values decreased to 0– Zinc-oxide ** cer-SS ** 20% compared with unexposed models. These antiseptics Furacine resulted in a similar degree of cytotoxicity to that observed *** SS-HAc *** after exposure to toxic concentrations of SDS. For this most Betadine *** toxic group of antiseptics, autograft was affected to a lesser SDS 0 102030405060708090 100 110 degree than either HSS with values decreasing to 30–55% MTT absorbance (% of control) compared with unexposed models. If a cell is viable, RNA will be present within the cell. (c) HSS-allogeneic Therefore an early indication of the cytotoxicity of a substance Unexp Acticoat penetrating the epidermis before detrimental tissue damage ** Aquacel-Ag * occurs can be visualized by RNA staining of tissue sections Dermacyn * (Table 3, Fig. 3). As a toxic substance penetrates the epider- Fucidin ** mis, a reduction of RNA staining will occur from the SG Silver nitrate Chlorhexidine ** down towards the BL. Exposure to Acticoat , Aquacel Ag , Flamazine ** ** Dermacyn , Fucidin and silver nitrate solution resulted in Zinc-oxide ** only a slight reduction of RNA staining in the SG. Cytotoxicity cer-SS Furacine *** caused by these antiseptics was considered to be extremely *** SS-HAc *** mild and cultures were comparable with unexposed cultures. Betadine *** Antiseptics which resulted in a substantial decrease in RNA SDS 0 102030405060708090 100 110 staining in SG and SS of HSSs after exposure were chlorhexi- MTT absorbance (% of control) dine digluconate and Flamazine . For autograft these effects were less pronounced where only minor RNA reduction in SG Fig 2. Metabolic activity after exposure to antiseptics. MTT levels of occurred. Exposure to the antiseptics zinc oxide cream, cerium- autograft, HSS-auto and HSS-allo after exposure to antiseptics. For silver sulfadiazine cream, Furacine , silver sulfadiazine cream each individual experiment, MTT levels of unexposed cultures are set containing 1% acetic acid and Betadine resulted in a strong at 100% and MTT levels of sodium dodecyl sulphate (SDS)-exposed decrease in RNA staining in all layers of both HSSs and autograft, cultures (total cell death) are set at 0%. Data are presented as indicating a deep penetration and strong cytotoxicity upon top- mean ± SEM of six experiments, from three independent donors. ical exposure. Within this group, Furacine and Betadine were ***P <0Æ001, **P <0Æ01, *P <0Æ05 vs. unexposed (control) cultures less cytotoxic to autograft as observed by less reduction in RNA of equivalent model. cer-SS, cerium-silver sulfadiazine cream; SS-Hac, when compared with either HSS. silver sulfadiazine cream containing 1% acetic acid. In summary, this study shows from all three cytotoxic determinations that Acticoat , Aquacel Ag and Dermacyn are sulfadiazine cream are moderately cytotoxic. Furacine , silver noncytotoxic. Fucidin , silver nitrate solution, chlorhexidine sulfadiazine cream containing 1% acetic acid and Betadine digluconate, Flamazine , zinc oxide cream and cerium-silver are very cytotoxic. Our results demonstrate that both HSSs
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 38 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al.
Table 3 Detection of cellular RNA were affected to a greater degree than the autograft, with this being especially apparent after exposure to the most cytotoxic Antiseptic Autograft HSS-auto HSS-allo antiseptics. Unexposed + + + Acticoat +++ Discussion Aquacel Ag +++ Dermacyn +++ Exposure of autograft and HSSs to 12 antiseptics showed that Fucidin +++ some antiseptics were not cytotoxic whereas others were Silver nitrate solution + + + extremely cytotoxic. The degree to which a substance was Chlorhexidine ) SG )SG, SS ) SG, SS digluconate found to be cytotoxic was verified by all three analytical meth- Flamazine ) SG ) SG, SS ) SG, SS ods: histology, metabolic activity (MTT) and RNA staining. In Zinc oxide cream ) SG, SS, BL ) SG, SS, BL ) SG, SS, BL general, the autograft showed less cytotoxicity than either HSS Cerium-silver ) SG, SS, BL ) SG, SS, BL ) SG, SS, BL after exposure to antiseptics. This is most probably due to the sulfadiazine cream fact that although HSSs have a fully differentiated epidermis ) ) ) Furacine SG, SS SG, SS, BL SG, SS, BL and SC the barrier function is still slightly less competent than Silver sulfadiazine ) SG, SS, BL ) SG, SS, BL ) SG, SS, BL autograft16 and therefore the rate of penetration of the anti- containing 1% acetic acid septics is greater for HSS than for autograft. Betadine ) SG, SS ) SG, SS, BL ) SG, SS, BL Chlorhexidine digluconate solution, Flamazine cream, SDS 10% ⁄1% ) SG, SS, BL ) SG, SS, BL ) SG, SS, BL Aquacel Ag dressing and Betadine ointment are routinely used for the treatment of ulcers (Table 4). Within this group SDS, sodium dodecyl sulphate. Overview showing the presence of antiseptics, chlorhexidine digluconate solution is used for or absence of RNA in different cell compartments of the epider- superficial and immediate cleaning of the wound area. Because mis: stratum granulosum (SG), stratum spinosum (SS) and basal layer (BL). + indicates RNA staining in all compartments. this solution is extremely volatile (ethanol based) and only ) indicates absence ⁄strongly decreased RNA staining in SG, SS mildly toxic for HSS, we can conclude that chlorhexidine di- and ⁄or BL. Paraffin-embedded sections (5 lm) were stained gluconate can be used together with HSS. Flamazine cream, with pyronine Y as described in Materials and methods. Data are Aquacel Ag dressing and Betadine ointment, in contrast to derived from six samples from three independent donors. the volatile chlorhexidine digluconate, remain in contact with the ulcer for a substantial period of time. The major factor determining which of these three antiseptics is applied to an ulcer is the degree of exudation from the wound. For ulcers HSS-auto: unexposed with relatively modest exudation, the use of Flamazine is preferred and occasionally Betadine . Both of these antiseptics are used as a preventative antiseptic to reduce general micro- bial burden. However, Betadine remains active in the wound
Table 4 Summary of results
Wound Antiseptic type Safe to apply HSS-auto: SS-HAc Acticoat Burn Autograft ⁄HSS Aquacel Ag Ulcer Autograft ⁄HSS Dermacyn Burn Autograft ⁄HSS Fucidin Burn Autograft ⁄HSS Silver nitrate solution Burn Autograft ⁄HSS Chlorhexidine digluconate Ulcer Autograft ⁄HSS Flamazine Ulcer Autograft ⁄HSS Zinc oxide cream Burn Autograft ⁄HSS Cerium-silver sulfadiazine cream Burn No Furacine Burn Autograft only Silver sulfadiazine containing Burn No 1% acetic acid Betadine Ulcer ⁄burn Autograft only Fig 3. Detection of cellular RNA. RNA (dark pink) staining of HSS-auto unexposed and exposed to silver sulfadiazine cream Antiseptics are listed in order of increasing cytotoxicity to containing 1% acetic acid (SS-HAc) for 24 h. Five-micrometre human skin substitute (HSS), with Acticoat being nontoxic and paraffin-embedded sections were stained with pyronine Y as described Betadine being severely cytotoxic. in Materials and methods.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al. 39 bed for a relatively short time compared with Flamazine .As range of cocci infection. As these two antiseptics are often we have shown that Betadine is severely cytotoxic on HSS used in an alternating therapy to prevent sensitization it is whereas Flamazine only exhibited moderate cytotoxicity, the advised to use the less cytotoxic antiseptic Fucidin directly use of Flamazine together with HSSs instead of Betadine is prior to transplantation. The antiseptics cerium-silver sulfadia- very strongly advised. The dressing Aquacel Ag is exclusively zine cream and silver sulfadiazine cream containing 1% acetic applied on ulcers with considerable exudation. This is due to acid were extremely cytotoxic to autograft and HSS, and there- the ability of this dressing to gel on contact with wound fluid, fore care should be taken to remove all traces of these antisep- thus creating a large fluid-absorption capacity.17 Furthermore, tics prior to placing an autograft or HSS. it contains the antimicrobial active silver in the dressing that is Our results can be used to advise clinicians in their choice only released in a moist environment. This last property makes of which antiseptic is best to use when applying autograft or Aquacel Ag less suitable for wounds with only minor exuda- HSS during burn and ulcer care. The results are summarized tion. As our data do not show any cytotoxic effect after expos- in Table 4. Due to ethical reasons our study is an in vitro study ure of Aquacel Ag to all three models this dressing can be and therefore carries certain limitations in the experimental used together with application of either HSS or autograft. set-up. Importantly, in vivo, quick inactivation or dilution of an Autografts (full-thickness or meshed split-skin) are often antiseptic may occur if a great deal of wound exudate is pre- required for the treatment of burns.1 The use of HSS in burn sent. Dilution of an antiseptic in vivo will result in the antisep- wound treatment is still infrequent, but there are reports des- tic being less cytotoxic than as determined in our in vitro assay. cribing the use of the full-thickness HSS, Apligraf (Organo- Moreover, the epidermis regenerates from the basal prolifera- genesis Inc., Canton, MA, U.S.A.), alone or in combination tive layer. Therefore even if a great deal of tissue damage is with meshed split-thickness autograft.10,18 It is to be expected observed in histological sections (or a decrease in RNA stain- that the number of reports will increase considerably in the ing) due to cytotoxicity in the upper cell layers, as long as the coming years and therefore prior knowledge concerning the basal cell layer remains intact and undamaged the HSS or administration of antiseptics together with HSS will be benefi- autograft may survive and regenerate an epidermis. In this cial to patient care. The antiseptics used in burn wound care respect, the measurement of metabolic activity (MTT assay) is are listed in Table 4. Three different procedures are common a very reliable measurement of cytotoxicity to the BL as it within the 5–7 day period after application of an autograft: measures metabolic activity only in undifferentiated keratino- (i) a solution of Betadine is applied when burn wounds do cytes in the BL and first suprabasal layer of the epidermis19 not exceed 10% of total body area; (ii) silver nitrate solution and the fibroblasts in the dermis. is applied when burn wounds exceed 10% of total body area In conclusion, antiseptics which are classed as safe to use or when wounds are infected with P. aeruginosa; and (iii) Fura- in the in vitro study are almost certainly safe to use in vivo. cine is applied when wounds are infected with S. aureus. Beta- Antiseptics which are classed as cytotoxic in our in vitro study dine and Furacine are moderately cytotoxic for autograft may be slightly less cytotoxic in vivo, particularly if a great and severely cytotoxic for both HSSs. Therefore, Betadine deal of wound exudate is present. This study gives an indica- and Furacine can be applied to an autograft but it is not tion of the cytotoxic effect of antiseptics and, taken together recommended to apply either Betadine or Furacine to HSS with the antiseptic properties of the particular substances, a when treating burns. The antiseptic silver nitrate solution clinician may use these data to help decide on an optimal showed only mild cytotoxicity upon exposure to any of the wound care procedure which takes into account the type of three models. Therefore silver nitrate is safe to use when wound colonization and whether an autograft or HSS can be applying autograft or HSS and may be used to replace Beta- applied. dine or Furacine if an HSS is required. Dermacyn is used to clean burn wounds. As described Acknowledgments above for ulcer treatment, it is only mildly cytotoxic for auto- graft and the HSSs and therefore can be used to clean wounds The authors greatly appreciate and thank Dr E.M. de Boer and prior to application of autograft and HSS. Dr H.M. van den Hoogenband (both dermatologists at VU Antiseptic ointments which are used preoperatively to acute University Medical Centre, Amsterdam, Netherlands) for dis- surgical, trauma and burn wounds are Betadine , Fucidin , cussing patient care protocols concerning ulcer treatment; and Furacine , cerium-silver sulfadiazine cream and silver sulfadia- Dr A.F. Vloemans and Dr F.R. Tempelman (both burns sur- zine cream containing 1% acetic acid (Table 4). This study geons at the Red Cross Hospital, Beverwijk, Netherlands) for shows that Fucidin is only mildly cytotoxic to autograft and discussing patient care protocols in burns patients. This both HSSs and can therefore be used as required. Furacine research was financed by the Dutch Program for Tissue Engin- and Betadine both show moderate and severe cytotoxicity on eering, STW, NWO. autograft and HSS, respectively. Therefore their use before autograft transplantation is relatively safe. However, upon HSS References application care should be taken to remove all traces of Fura- cine and Betadine prior to placing the HSS. Fucidin is less 1 Janzekovic Z. A new concept in the early excision and immediate cytotoxic than Furacine and is used to treat the same wide grafting of burns. J Trauma 1970; 10:1103–8.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 40 Antiseptic cytotoxicity on skin substitutes, Q. le Duc et al.
2 Jones JE, Nelson EA. Skin grafting for venous leg ulcers. Cochrane 12 Smola H, Thiekotter G, Fusenig NE. Mutual induction of growth Database Syst Rev 2005; CD001737. factor gene expression by epidermal-dermal cell interaction. J Cell 3 Mekkes JR, Loots MA, van der Wal AC et al. Causes, inves- Biol 1993; 122:417–29. tigation and treatment of leg ulceration. Br J Dermatol 2003; 13 Ponec M, Kempenaar JA, De Kloet ER. Corticoids and cultured 148:388–401. human epidermal keratinocytes: specific intracellular binding and 4 Ehrenreich M, Ruszczak Z. Update on tissue-engineered biological clinical efficacy. J Invest Dermatol 1981; 76:211–14. dressings. Tissue Eng 2006; 12:2407–24. 14 Mosmann T. Rapid colorimetric assay for cellular growth and sur- 5 Pruitt BA Jr, McManus AT, Kim SH et al. Burn wound infections: vival: application to proliferation and cytotoxicity assays. J Immunol current status. World J Surg 1998; 22:135–45. Methods 1983; 65:55–63. 6 Jones SG, Edwards R, Thomas DW. Inflammation and wound 15 Spiekstra SW, Toebak MJ, Sampat-Sardjoepersad S et al. Induction healing: the role of bacteria in the immuno-regulation of wound of cytokine (interleukin-1alpha and tumor necrosis factor-alpha) healing. Int J Low Extrem Wounds 2004; 3:201–8. and chemokine (CCL20, CCL27, and CXCL8) alarm signals after 7 Gibbs S, van den Hoogenband HM, Kirtschig G et al. Autologous allergen and irritant exposure. Exp Dermatol 2005; 14:109–16. full-thickness skin substitute for healing chronic wounds. Br J 16 Boelsma E, Anderson C, Karlsson AM et al. Microdialysis technique Dermatol 2006; 155:267–74. as a method to study the percutaneous penetration of methyl nico- 8 Bell E, Ehrlich HP, Buttle DJ et al. Living tissue formed in vitro and tinate through excised human skin, reconstructed epidermis, and accepted as skin-equivalent tissue of full thickness. Science 1981; human skin in vivo. Pharm Res 2000; 17:141–7. 211:1052–4. 17 Newman GR, Walker M, Hobot JA et al. Visualisation of bacterial 9 Dinh TL, Veves A. The efficacy of Apligraf in the treatment of dia- sequestration and bactericidal activity within hydrating Hydrofiber betic foot ulcers. Plast Reconstr Surg 2006; 117:S152–7. wound dressings. Biomaterials 2006; 27:1129–39. 10 Hayes DW Jr, Webb GE, Mandracchia VJ et al. Full-thickness burn 18 Waymack P, Duff RG, Sabolinski M. The effect of a tissue engin- of the foot: successful treatment with Apligraf. A case report. eered bilayered living skin analog, over meshed split-thickness Clin Podiatr Med Surg 2001; 18:179–88. autografts on the healing of excised burn wounds. The Apligraf 11 Ponec M, Hasper I, Vianden GD et al. Effects of glucocorticosteroids Burn Study Group. Burns 2000; 26:609–19. on primary human skin fibroblasts. II. Effects on total protein and 19 Boelsma E, Gibbs S, Faller C et al. Characterization and comparison collagen biosynthesis by confluent cell cultures. Arch Dermatol Res of reconstructed skin models: morphological and immunohisto- 1977; 259:125–34. chemical evaluation. Acta Derm Venereol (Stockh) 2000; 80:82–8.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp33–40 CLINICAL AND LABORATORY INVESTIGATIONS DOI 10.1111/j.1365-2133.2007.07933.x Differences in survivin location and Bcl-2 expression in CD30+ lymphoproliferative disorders of the skin compared with systemic anaplastic large cell lymphomas: an immunohistochemical study G. Goteri, O. Simonetti,* S. Rupoli, G. Piccinini, C. Rubini, D. Stramazzotti, F. Fazioli, C. Capomagi, P. Leoni, A.M. Offidani* and L. Lomuzio§ Department of Neurosciences, Institute of Anatomic Pathology, Ancona, Italy *Clinic of Dermatology, Clinic of Hematology and Laboratory of Cellular and Molecular Biology, Institute of Clinical Medicine and Applied Biotechnologies, Polytechnic University of Marche Region, Via Conca 71, 60020 Torrette di Ancona, Ancona, Italy §Department of Surgical Sciences, University of Foggia, Foggia, Italy
Summary
Correspondence Background Cutaneous CD30+ lymphoproliferative disorders (LPDs) are a spectrum Oriana Simonetti. of disease associated with a favourable prognosis. Systemic anaplastic large cell E-mail: [email protected] lymphoma (ALCL), although morphologically and phenotypically similar, differs in clinical presentation and has a less favourable biological behaviour. Dysregula- Accepted for publication 17 January 2007 tion of apoptosis, the process regulating cell population by programmed death, can explain the differences among these disorders. Key words Objectives We investigated the expression of two inhibitors of apoptosis, survivin anaplastic large cell lymphoma, anaplastic and Bcl-2 protein, in serial skin lesion samples from CD30+ LPDs compared with lymphoma kinase, Bcl-2, CD30+ systemic ALCL. lymphoproliferative disorders, survivin Methods Immunohistochemical analysis with antibodies against anaplastic lymph- Conflicts of interest oma kinase (ALK)-1 protein, survivin and Bcl-2 protein was performed in 10 None declared. cutaneous CD30+ LPDs (five lymphomatoid papulosis, five ALCL) and 18 system- ic ALCLs. Reverse transcription–polymerase chain reaction studies for ALK and This work was presented at the 11th Congress of ALK/nucleophosmin were also performed. the European Hematology Association (15–18 Results Cutaneous CD30+ LPDs shared a heterogeneous expression of cytoplasmic June 2006, Amsterdam) and the Third Meeting of ) the European Association of Dermato-Oncology survivin with all systemic ALCLs, and of Bcl-2 with systemic ALK ALCLs; how- (23–25 June 2006, Rome). ever, they differ from systemic ALK) ALCLs because they lack nuclear survivin (P ¼ 0Æ045), and from systemic ALK+ ALCLs by a higher expression of Bcl-2 (P ¼ 0Æ045) and a lack of ALK-1. Overall, coexpression of Bcl-2 and nuclear survivin in CD30+ LPDs was associated with a less favourable disease survival. Conclusions The different patterns of expression of Bcl-2 and survivin in CD30+ LPDs might have an impact on their different biological and clinical behaviour. Moreover, nuclear localization of survivin, similarly to ALK, may be a useful marker for predicting a systemic form of ALCL with cutaneous presentation.
Cutaneous CD30+ lymphoproliferative disorders (LPDs) and course are used as decisive criteria for the definite diagnosis systemic anaplastic large cell lymphomas (ALCLs) are similar and choice of treatment, more than histological criteria.1 LyP morphologically and phenotypically, being composed of large presents as multiple disseminated papular-ulcerative lesions of atypical CD30+ cells, but differ in clinical presentation and the skin with spontaneous regression, not requiring treatment, biological behaviour. Cutaneous CD30+ LPDs, the second whereas PC-ALCL presents as growing tumours, partially most common group of cutaneous T-cell lymphomas (CTCLs), regressing, single or multiple and limited to the same skin accounting for approximately 30% of cases, are a spectrum of region, responsive to radiotherapy and not requiring aggres- disease ranging from lymphomatoid papulosis (LyP) to pri- sive therapy. In contrast, systemic ALCLs behave as aggressive mary cutaneous ALCL (PC-ALCL). The clinical appearance and tumours requiring chemotherapy, with a better outcome in
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 41 42 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al. the presence of 2p23 locus abnormalities that disrupt the ana- were prepared from exponentially growing cells fixed in acet- plastic lymphoma kinase (ALK) gene.2,3 one for 10 min at room temperature and immunohistochemi- Apoptosis is an active process regulating cell population by cal analysis was performed as described below. programmed death,4 and its deregulation plays an important role in normal and malignant lymphopoiesis.5 Although Immunohistochemical analysis extensively studied starting with Kikuchi and Nishikawa,6 it has not yet been completely clarified whether the involvement Five-micrometre tissue sections from each selected paraffin of different pathways of apoptosis might explain the striking block were submitted to immunohistochemical analysis using differences in biological course and prognosis of subtypes of the commercially available Dako Envision system (Dako Cor- cutaneous and systemic CD30+ LPDs. Several proteins with poration, Glostrup, Denmark). Briefly, slides were deparaffi- a proapoptotic or antiapoptotic function, such as CD30/ nized and rehydrated in xylene and in graded alcohol. CD30L,7 CD95/CD95L,8–11 caspase 3,12,13 FADD,13 Bax-L, Microwave treatment using citrate buffer, pH 6Æ0, was per- Bax-s, Mcl-1, Bad14,15 and PTEN,11 have been investigated in formed for antigenic retrieval. After blocking the endogenous these disorders, suggesting that the clinical behaviour of these peroxidase with 3% hydrogen peroxide, slides were incubated lymphomas is at least in part a reflection of the apoptotic ten- overnight at 4 C in moist chambers with the corresponding dencies of the neoplastic cells. Among proteins controlling primary antibody dilution. The following antibodies were apoptosis, we focused our attention on survivin and Bcl-2, used: rabbit monoclonal anti-CD3 (dilution 1 : 100; Dakocyto- two proteins with antiapoptotic function, both having a crit- mation, Milan, Italy), mouse monoclonal anti-CD4 (1F6, dilu- ical role in controlling the activation of caspases.16,17 Among tion 1 : 40; Novocastra Laboratories Ltd, Newcastle upon the 14 known members of the interleukin (IL)-1-converting Tyne, U.K.), mouse monoclonal anti-CD8 (4B11, dilution enzyme family of proteases, caspase 3 has been shown to be a 1 : 40; Novocastra Laboratories Ltd), rabbit polyclonal anti- key component of the apoptotic machinery.18 The aim of our survivin (FL-142, dilution 1 : 200; Santa Cruz Biotechnology study was to determine if differences in survivin cellular local- Inc., Santa Cruz, CA, U.S.A.); mouse monoclonal recognizing ization and Bcl-2 expression could be related to the different the intracellular domain of ALK2 (kindly provided by B. Falini, biological behaviour between cutaneous and systemic CD30+ University of Perugia, Italy and P.G. Pellicci, European Institute lymphomas. of Oncology, Milan, Italy); and mouse monoclonal anti-Bcl-2 (dilution 1 : 60; Dakocytomation). The peroxidase reaction Materials and methods was developed using 3,3¢-diaminobenzidine tetrahydrochloride (Sigma, St Louis, MO, U.S.A.) and 0Æ02% hydrogen peroxide. Specimens were then counterstained with Mayer’s haematoxy- Case selection lin. An appropriate positive control was used in each staining All the patients included in this study were examined at the run; negative control experiments included omission of the Clinics of Dermatology and Hematology, Azienda Ospedali primary antibody and/or replacement by nonspecific rabbit Riuniti Umberto I Hospital (Ancona, Italy) between 1995 polyclonal immunoglobulins or unrelated isotype-matched and 2003; only cases with complete clinicopathological data mouse monoclonal antibodies. Two investigators (G.G., C.R.) were considered eligible for the study. The histological mater- independently and in blinded fashion examined the immuno- ial was retrieved from the files of the Institute of Pathological histochemically stained sections. Anatomy and Histopathology, Polytechnic University of the A cell was considered immunoreactive for survivin in cases Marche Region (Ancona). Original slides were reviewed by showing deposition of fine brownish granules within the one of us (G.G.) in order to confirm the diagnosis according cytoplasm or the nucleus, irrespective of the staining intensity. to the recent World Health Organization classification system Immunostaining for Bcl-2 was considered positive in cases for haematopoietic and skin tumours1,19 and to select the with a strong brownish reactivity confined to the cytoplasm. most representative paraffin blocks for further immunohisto- Immunostaining for ALK was considered positive in cases with chemical and molecular investigations. brown reactivity in the cytoplasm and/or nucleus.
Cell lines RNA extraction
The human nucleophosmin (NPM)–ALK-positive Karpas 299 Total RNA was extracted using TRIzol (Invitrogen Ltd) accord- cell line (kindly provided by Prof. B. Falini, University of ing to the manufacturer’s instructions. For extraction of RNA Perugia, Italy) and the murine NPM/ALK-negative growth fac- from paraffin-embedded tissues the following modifications tor-dependent 32D cell line were used for the molecular were made: 20–25 sections of 5 lm thickness were placed in investigation of the t(2;5)(p23;q35) translocation, as positive Eppendorf tubes and deparaffinized in xylene. The deparaffi- and negative control, respectively. Both cell lines were main- nized tissues were washed in ethanol, pelleted, and then resus- tained in RPMI 1640 containing 10% fetal bovine serum pended in TRIzol. Following the extraction, total RNA was (Invitrogen Ltd, Paisley, U.K.) and (for 32D cell culture) 10% treated with DNAse as recommended. Reverse transcription WEHI conditioned medium as a source of IL-3. Cytospins (RT) of 1Æ7 lg of total RNA was performed in a mixture
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al. 43
) ) containing 50 mmol L 1 Tris-HCl, pH 8Æ3, 75 mmol L 1 KCl, sion of survivin and Bcl-2 among groups of cutaneous CD30+ )1 )1 )1 2 3 mmol L MgCl2, 10 mmol L dithiothreitol, 100 ng mL LPDs and systemic ALCL were assessed using the v test (or of Pd(N)6 random hexamer primers (Amersham Biosciences, Fisher’s exact test) for contingency tables. Overall survival ) Piscataway, NJ, U.S.A.) and 1 mmol L 1 deoxyribonucleoside (OS) was defined as the time from the start of treatment to triphosphates (dNTPs; Amersham Biosciences) for 1 h at death or to the last follow-up. OS was plotted using the 38 C in a volume of 25 lL. Successful cDNA synthesis was Kaplan–Meier method and differences between the curves checked by amplification of a 670-bp fragment of the b-actin were assessed using the log-rank test. Statistical significance housekeeping gene using the primers 5¢-CCTTCCTGGGCATG- was set at P <0Æ05. GAGTCCTG-3¢ (forward primer) and 5¢-GGAGCAATGATCTT- GATCTTC-3¢ (reverse primer). Results
Reverse transcription–polymerase chain reaction Clinicopathological features
Polymerase chain reaction (PCR) was performed with oligo- Clinical and pathological characteristics of the cases included nucleotides specific for both the chimeric NPM/ALK transcript in the study are summarized in Table 1. Ten cases were con- and for the ALK cytoplasmic portion. NPM/ALK fusion gene sidered primary cutaneous CD30+ LPDs because they lacked was evaluated by a nested PCR method using a 3¢ ALK-1 (5¢- extracutaneous involvement at the time of diagnosis: five cases CGA GGT GCG GAG CTT GCT CAG C-3¢) and the NPM/ALK were interpreted as LyP and five as PC-ALCL. The distinction junction oligonucleotide (5¢-TCC CTT GGG GGC TTT GAA between the two groups relied upon a combination of both ATA ACA CC-3¢) as primers for the first round (177-bp PCR clinical and morphological, immunohistochemical and mole- product) and the 3¢ ALK-1 primer in conjunction with the cular criteria. LyP cases showed the typical clinical picture of NPM-2 junction oligonucleotide (5¢-AGC ACT TAG TAG TGT multiple papular-ulcerative lesions with spontaneous regres- ACC GCC GGA-3¢) for the second run (98-bp PCR product), sion. Morphologically they all had the type A appearance according to the protocol described by Lamant et al.20 (‘histiocyte type’) with the typical histological wedge-shaped Two rounds of PCR were performed to detect transcripts pattern of the lymphoid infiltrate, which was distributed for the cytoplasmic portion of ALK using the 5¢-GCT GAG around vessels and was composed of medium-sized T lympho- CAA GCT CCG CAC CTC GAC GAC-3¢ (forward) and 5¢-CCC cytes admixed with neutrophilic and eosinophilic granulocytes GCC ATG AGC TCC AGC AGG ATG-3¢ (reverse) primers, and scattered or clustered large atypical Hodgkin-like large which amplify a 366-bp product. cells. A monoclonal rearrangement of the TCR c gene was Amplification was performed on total cDNA (130 ng) using detectable in two cases. PC-ALCL cases were all single tumoral an automated thermal cycler (Perkin Elmer, Foster City, CA, lesions. Histologically they exhibited a diffuse superficial and U.S.A.) in a final volume of 50 lL PCR buffer containing ) ) ) 1 lmol L 1 of each primer, 200 lmol L 1 dNTPs, 2 mmol L 1 Table 1 Clinicopathological features of the 28 cases of CD30+ MgCl2 and 1 U of Taq DNA polymerase (Finnzymes Oy, Espoo, lymphoproliferative disorders Finland). After an initial denaturation at 94 C for 3 min, 30 cycles of amplification were performed at the annealing tem- ) perature of 66 C. The second amplification was performed Variable LyP PC-ALCL ALK+ ALCL ALK ALCL with the same conditions using 2% of the first-round PCR prod- Age (years) uct. After amplification, PCR products were size fractionated by Median 68Æ060Æ031Æ061Æ8 Range 38–75 47–83 2–67 30–86 electrophoresis through a 2% agarose gel and visualized with Sex ethidium bromide staining. F313 4 M244 7 Polymerase chain reaction for T-cell clonality Phenotype T543 9 Data regarding T-cell clonality were obtained from the analysis ‘null’ 0 1 4 2 of T-cell receptor (TCR) c chain gene rearrangements using Status at last follow-up AW 5 4 6 3 PCR on paraffin sections as described previously.21 Results AWD 0 1 0 2 were considered as positive when one or two bands were DOD 0 0 1 6 observed in the expected range and the pattern was reprodu- Follow-up (months) cible in reduplicated analysis. Median 54 92 48 58 Range 23–56 36–216 1–145 3–216
Statistical analysis LyP, lymphomatoid papulosis; ALCL, anaplastic large cell lymph- oma; PC-ALCL, primary cutaneous ALCL; ALK, anaplastic lymph- Patients’ characteristics and descriptive data were expressed as oma kinase; AW, alive and well; AWD, alive with disease; DOD, median and range for continuous variables and by frequency dead of disease. tabulations for categorical factors. Differences in the expres-
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 44 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al.
deep dermal infiltrate of clearly atypical CD30+ T large atyp- 12 11 ical cells, mitotically active, which appeared to be of clonal LyP origin in PCR analysis in four of five cases. All 10 cutaneous 10 C-ALCL CD30+ LPDs were negative for ALK-1 immunostaining and for ALK RT-PCR. Systemic ALK- Eighteen cases were considered systemic ALCL based on the 8 positive ALCL 7 clinical presentation: immunostaining for ALK showed that Systemic ALK- seven cases were ALK+ and 11 were ALK). All the positive 6 negative ALCL cases showed both nuclear and cytoplasmic immunostaining. 55 5 Molecular analysis showed in all cases a 366-bp ALK transcript
No. of cases 4 by RT-PCR and the specific NPM/ALK fusion transcript of 4 98 bp, ruling out the presence of a different rearrangement. 3 Patients were followed up for a median of 55 months 2 2 (range 1–216). All the patients with LyP were alive and well 1 with complete resolution of the lesions: one patient experi- enced a relapse of the disease after 60 months which was suc- 000 0 cessfully treated by methotrexate. All patients with PC-ALCL Cytoplasmic Nuclear Bcl-2 had a complete remission after surgical excision of the lesions. survivin survivin However, three patients exhibited a disease relapse: in two cases new single cutaneous lesions appeared after 3 and Fig 1. Graph showing the distribution of immunoreactivity pattern 8 years, respectively, and were treated by surgical excision for survivin and Bcl-2 in the four groups of CD30+ lymphomas. and radiotherapy with success. One patient exhibited an extra- LyP, lymphomatoid papulosis; ALCL, anaplastic large cell lymphoma; C-ALCL, cutaneous ALCL; ALK, anaplastic lymphoma kinase. cutaneous relapse of a CD30+ ALCL involving the lymph
(a) (b)
(c) (d)
Fig 2. Immunostaining pattern for survivin in the four types of CD30+ lymphoproliferative disorders. Cases of lymphomatoid papulosis (a), primary cutaneous anaplastic large cell lymphoma (ALCL) (b) and nodal primary ALCL, anaplastic lymphoma kinase (ALK)-positive (c) show clusters of large atypical cells with similar cytoplasmic staining, whereas a case of nodal primary ALCL, ALK-negative (d) exhibits clusters of large atypical cells with predominant nuclear localization of immunostaining (immunoperoxidase; original magnification · 250).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al. 45 nodes and palatine tonsils after 88 months from initial diagno- 1·0 sis. He received chemotherapy with partial response and is alive with disease at 92 months. 0·9 All patients with systemic ALCL were treated with chemo- therapy except for two patients who died before starting treat- 0·8 ment and one patient who refused treatment: seven patients died of lymphoma and nine obtained a complete remission. 0·7 Two patients were alive with evidence of disease at last 0·6 follow-up. Patients were significantly younger in the group of ALK+ 0·5 lymphomas compared with all the other groups (ANOVA test, P <0Æ006), but no significant differences of distribution of 0·4 Survival probability sex and cell phenotype were detected. 0·3 As expected, disease-specific survival curves were signifi- cantly different between patients with primary cutaneous LPDs 0·2 Nuclear survivin negative and those with systemic lymphomas (P ¼ 0Æ0422): all the Nuclear survivin positive patients with primary cutaneous LPD were still alive at last fol- 0·1 low-up, whereas the median survival for those with systemic lymphoma was 99 months. Although the absence of ALK 0·0 expression was associated with a higher incidence of death, 0 50 100 150 200 250 the cases were too few to show any significant differences. Months
Fig 3. Disease-specific overall survival curve according to the nuclear 2 Survivin and Bcl- staining location of survivin. As shown, the median has not been reached for The distribution of survivin and Bcl-2 immunostaining in the cases without nuclear survivin and was 32 months in patients with nuclear survivin. At 99 months the probability to be alive tended to four subtypes of CD30+ LPDs is reported in Figure 1. All the be significantly higher in patients without vs. patients with nuclear 28 cases examined showed a clear cytoplasmic positivity for location of the marker (70% vs. 40%, log rank test: P ¼ 0Æ0612). survivin, independently from their clinicopathological group, in the large cells with atypical findings showing CD30 expression in serial sections. No double immunostaining was performed. Five cases of systemic ALCL, which were all ALK), 1·0 showed in addition a nuclear immunoreactivity for survivin in 0·9 the cells (Fig. 2). Nuclear expression of survivin was not observed in the other groups (P ¼ 0Æ045). Protein Bcl-2 cyto- 0·8 plasmic expression was found in 10 cases in the larger atypical cells of the different types of CD30+ lymphomas: systemic 0·7 ALK+ ALCL showed a lower frequency of Bcl-2 expression 0·6 (P ¼ 0Æ045). No association was found between Bcl-2 expres- sion or nuclear localization of survivin and the T or ‘null’ 0·5 phenotype. Instead, both the nuclear expression of survivin (P ¼ 0Æ082) and the presence of Bcl-2 expression (P ¼ 0·4 0Æ062) tended to be associated with an increased incidence of Survival probability 0·3 deaths from disease. Nuclear expression of survivin and/or Bcl-2 immunoreactivity were associated with a lower disease- 0·2 specific survival (Figs 3–5). Cytoplasmic Bcl-2 negative 0·1 Cytoplasmic Bcl-2 positive
Discussion 0·0 0 50 100 150 200 250 Our study adds interesting data on two inhibitors of apoptosis, Months Bcl-2 and survivin, in the groups of both cutaneous CD30+ LPDs and systemic ALCLs. As inhibitors of apoptosis, both pro- Fig 4. Disease-specific overall survival curve according to the teins have been implicated in promoting cancer, although the expression of Bcl-2. As shown, the median has not been reached for distribution of Bcl-2 and survivin expression varies widely in Bcl-2-negative cases and was 32 months in patients with Bcl-2- both normal and transformed cell types. Survivin is a recently positive lymphoproliferative disorders. At 99 months the probability described member of the inhibitors of apoptosis protein fam- to be alive was significantly higher in Bcl-2-negative vs. Bcl-2-positive ily, which participates in the complex network regulating cases (70% vs. 50%, log rank test: P ¼ 0Æ022).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 46 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al.
and differentiation and a step in the malignant progression.30 1·0 However, other investigators have reported conflicting results, 0·9 suggesting nuclear expression as a favourable prognostic factor in neoplasms at different locations, such as gastric, bladder and 0·8 breast carcinomas.35–37 Conflicting results have been reported in the same type of tumours: for instances, in melanocytic 0·7 tumours, Ding et al.38 detected nuclear survivin only in mela- 0·6 nomas and not in benign naevi using a polyclonal antibody, whereas Vetter et al.39 using a monoclonal antibody obtained 0·5 the opposite results with a higher positivity in nuclei of benign naevi compared with melanomas. We think that the use of dif- 0·4 ferent commercially available antibodies may have contributed Survival probability to the great variability of results. Li et al. have suggested com- 0·3 bining immunohistochemistry with Western blot for a precise 40 0·2 location of survivin. In our cases, however, no further analy- Nuclear survivin/Bcl-2 –/+ or –/– ses were possible to elucidate the finding, as only paraffin- Nuclear survivin/Bcl-2 +/+ 0·1 embedded, formalin-fixed tissue was available to us. However, based on our results, immunostaining for survivin with the 0·0 0 50 100 150 200 250 polyclonal antibody might have an application similar to ALK Time immunostaining, as cutaneous CD30+ lymphomas are highly unlikely to be primary if immunoreactive for ALK or for nuc- Fig 5. Disease-specific overall survival curve according to the lear survivin. expression of both nuclear survivin and cytoplasmic Bcl-2. At Bcl-2 protein, first identified in the t(14;18)(q32;q21) 99 months the probability to be alive was significantly higher in of B-cell follicular lymphoma, is a major negative regulator of Bcl-2-negative vs. Bcl-2-positive cases (73% vs. 0, log rank test: apoptosis41 that has been implicated in the pathogenesis of P <0Æ005). haematological and epithelial malignancies and has already been extensively studied in cutaneous CD30+ LPDs, with par- programmed cell death and also cell division.22,23 Overexpres- tially conflicting results. Paulli et al.42 found it expressed at sion of survivin has been associated with parameters of aggres- lower levels in regressing LyP compared with nonregressing siveness and poor prognosis in several tumours.24,25 This is, to cutaneous ALCL, suggesting that this protein may protect our knowledge, the first study analysing survivin expression in CD30+ cells from apoptosis in nonregressing lesions. Nevala cutaneous CD30+ LPDs, whereas systemic ALCLs have been the et al.11 also reported a lower expression of Bcl-2 (39%) in LyP subject of a previous study showing an association of survivin compared with the other CTCLs (46%). Mao et al.43 reported expression with a poor prognosis in ALK) ALCL.26 In our ser- absence of Bcl-2 expression (< 10%) in all of eight cutaneous ies of CD30+ LPDs, both the atypical Stembergoid cells of LyP ALCLs studied: in three of these cases they were able to and the anaplastic or pleomorphic large CD30+ cells of demonstrate also loss of Bcl-2 gene copy number by PCR. In PC-ALCL showed expression of cytoplasmic survivin, suggest- our series, Bcl-2 was expressed in half of our cases of LyP and ing that survivin is not an absolute marker of malignancy, PC-ALCL, suggesting that neither Bcl-2 nor cytoplasmic survi- being expressed also in indolent and potentially regressing vin expression helps in distinguishing in the spectrum of cuta- lesions. Survivin has indeed been demonstrated in many non- neous CD30+ LPDs. It might be postulated that despite Bcl-2 neoplastic cells of nonlymphoid nature, such as colonic epithe- and survivin cell expression LyP and a portion of PC-ALCL lium, uterine cervical mucosa and hepatocytes.27–29 Regarding lesions can undergo spontaneous regression through the acti- systemic ALCL, the most interesting and unexpected feature vation of other mechanisms still functioning that reduce cell was the observation of nuclear survivin immunostaining in growth and increase cell apoptosis, like the activation of the 45% of systemic ALK) ALCLs where it seemed to be associated CD30/CD30L and CD95/CD95L systems, the immune control with a less favourable prognosis, but without reaching a statis- mediated by activated cytotoxic T lymphocytes, and finally the tically significant value. This finding is in contrast with the necrosis related to vascular damage.44 results reported by Schlette et al.,26 who found survivin exclu- Moreover, our data confirm that Bcl-2 has a negative prog- sively located in the cytoplasm by immunohistochemistry and nostic value in systemic ALCL, particularly when it is com- by Western blotting. Survivin nuclear immunostaining has bined with nuclear localization of survivin. Bcl-2 expression already been described as an unfavourable marker in other has been described less frequently in ALK+ than in ALK) cases tumours, such as hepatocellular carcinoma,25,30,31 oesophageal of systemic ALCL.14,45 ALK+ systemic ALCLs bear a tendency carcinoma,32 lung carcinoma33 and mantle cell lymphoma,34 to cell apoptosis due to an imbalance between overexpressed and our results are in line with the hypothesis that transloca- Bax and absent Bcl-2.45 tion of survivin from the cytoplasm to the nucleus may consti- In conclusion, our results show that LyP and PC-ALCL share a tute an important regulatory mechanism for cell proliferation heterogeneous expression of cytoplasmic survivin with systemic
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al. 47
ALK+ ALCL, and of Bcl-2 with ALK) ALCL; however, cutaneous 15 Greisser J, Doebbeling U, Roos M et al. Apoptosis in CD30-positive CD30+ LPDs differ from systemic ALK) ALCLs because they lymphoproliferative disorders of the skin. Exp Dermatol 2005; lack nuclear survivin, and from systemic ALK+ ALCLs by a 14:380–5. 16 Yang J, Liu X, Bhalla K et al. Prevention of apoptosis by Bcl-2: higher Bcl-2 expression. Moreover, nuclear localization of survi- release of cytochrome c from mitochondria blocked. Science 1997; vin, similarly to ALK, may be a useful marker for predicting a 275:1129–31. systemic form of ALCL with cutaneous presentation. These 17 LaCasse EC, Baird S, Korneluk RG, MacKenzie AE. The inhibitors of results may have implications in their differing biological and apoptosis (IAPs) and their emerging role in cancer. Oncogene 1998; clinical aspects. Further studies on the expression and function 17:3247–59. of apoptosis-related proteins are needed, considering that the 18 Nunez G, Benedict MA, Hu Y, Inohara N. Caspases: the proteases factors involved in this process may provide novel targets for of the apoptotic pathway. Oncogene 1998; 17:3237–45. 19 Delsol G, Ralfkiaer E, Stein H et al. Anaplastic large cell lymphoma. experimental therapy in patients with ALCL. In: Pathology and Genetics of Tumors of Hemopoietic and Lymphoid Tissues (Jaffe ES, Harris NL, Stein H, Vardiman JW, eds). Lyon: IARC Acknowledgments Press, 2001; 230–5. 20 Lamant L, Dastugue N, Pulford K et al. A new fusion gene We thank Dr Alessandra Filosa for help in retrieving the archi- TPM3–ALK in anaplastic large cell lymphoma created by a val material and Mr Luca Balercia for the valuable technical (1;2)(q25;p23) translocation. Blood 1999; 93:3088–95. work of immunohistochemistry. 21 Ashton-Key M, Diss TC, Du MQ et al. The value of polymerase chain reaction in the diagnosis of cutaneous T-cell infiltrates. Am J Surg Pathol 1997; 21:743–7. References 22 Adida C, Crotty PL, McGrath J et al. Developmentally regulated expression of the novel cancer anti-apoptosis gene survivin in 1 Willemze R, Jaffe ES, Burg G et al. WHO–EORTC classification for human and mouse differentiation. Am J Pathol 1998; 152:43–9. cutaneous lymphomas. Blood 2005; 105:3768–85. 23 Ambrosini G, Adida C, Altieri DC. A novel anti-apoptotis gene, survi- 2 Pittaluga S, Wlodarska I, Pulford K et al. The monoclonal antibody vin, expressed in cancer and lymphoma. Nat Med 1997; 3:917–21. ALK1 identifies a distinct morphological subtype of anaplastic large 24 Li F, Ambrosini G, Chu EY et al. Control of apoptosis and mitotic cell lymphoma associated with 2p23/ALK rearrangements. Am J spindle checkpoint by survivin. Nature 1998; 396:580–4. Pathol 1997; 151:343–51. 25 Ito T, Shiraki K, Sugimoto K et al. Survivin promotes cell proli- 3 Falini B, Pileri S, Zinzani PL et al. ALK+ lymphoma: clinico-patho- feration in human hepatocellular carcinoma. Hepatology 2000; logical findings and outcome. Blood 1999; 93:2697–706. 31:1080–5. 4 Webb SJ, Harrison DJ, Wyllie AH. Apoptosis: an overview of 26 Schlette EJ, Meideros JL, Goy A et al. Survivin expression predicts the process and its relevance in disease. Adv Pharmacol 1997; 41:1–34. poorer prognosis in anaplastic large cell-lymphoma. J Clin Oncol 5 De Graaf AO, De Witte T, Jansen JH. Inhibitor of apoptosis pro- 2004; 22:1682–8. teins: new therapeutic targets in hematological cancer? Leukemia 27 Gianni R, Jarboe E, Orlicky D et al. Expression of survivin in nor- 2004; 18:1751–9. mal, hyperplastic, and neoplastic colonic mucosa. Hum Pathol 2001; 6 Kikuchi A, Nishikawa T. Apoptotic and proliferating cells in cutan- 32:119–25. eous lymphoproliferative diseases. Arch Dermatol 1997; 133:829–33. 28 Frost M, Jarboe EA, Orlicky D et al. Immunohistochemical localiza- 7 Mori M, Manuelli C, Pimpinelli N et al. CD30–CD30 ligand inter- tion of survivin in benign cervical mucosa, cervical dysplasia, and action in primary cutaneous CD30+ T-cell lymphomas: a clue to invasive squamous cell carcinoma. Am J Clin Pathol 2002; 117:738– the pathophysiology of clinical regression. Blood 1999; 94:3077– 44. 83. 29 Deguchi M, Shiraki K, Inoue H et al. Expression of survivin during 8 Zoi-Toli O, Vermeer MH, De Vries E et al. Expression of Fas and liver regeneration. Biochem Biophys Res Commun 2002; 297:59–64. Fas-ligand in primary cutaneous T-cell lymphoma (CTCL): associ- 30 Moon WS, Tarnawski AS. Nuclear translocation of survivin in ation between lack of Fas expression and aggressive types of CTCL. hepatocellular carcinoma: a key to cancer cell growth? Hum Pathol Br J Dermatol 2000; 143:313–19. 2003; 34:1119–26. 9 Levi E, Wang Z, Petrogiannis-Haliotis T et al. Distinct effects of 31 Fields AC, Cotsonis G, Sexton D et al. Survivin expression in CD30 and Fas signalling in cutaneous anaplastic lymphomas: a pos- hepatocellular carcinoma: correlation with proliferation, prognostic sible mechanism for disease progression. J Invest Dermatol 2000; parameters, and outcome. Mod Pathol 2004; 17:1378–85. 115:1034–40. 32 Grabowski P, Kuhnel T, Muhr-Wilkenshoff F et al. Prognostic 10 Sigel JE, His ED. Immunohistochemical analysis of CD30-positive value of nuclear survivin expression in oesophageal squamous cell lymphoproliferative disorders for expression of CD95 and CD95L. carcinoma. Br J Cancer 2003; 88:115–19. Mod Pathol 2000; 13:446–51. 33 Lu B, Gonzales A, Massion PP et al. Nuclear survivin as a biomarker 11 Nevala H, Karenko L, Vakeva L, Ranki A. Proapoptotic and for non-small-cell lung cancer. Br J Cancer 2004; 91:537–40. antiapoptotic markers in cutaneous T-cell lymphoma skin infiltrates 34 Martinez A, Bellosillo B, Bosh F et al. Nuclear survivin expression and lymphomatoid papulosis. Br J Dermatol 2001; 145:928–37. in mantle cell lymphoma is associated with cell proliferation and 12 Drakos E, Rassidakis GZ, Lai R et al. Caspase-3 activation in sys- survival. Am J Pathol 2004; 164:501–10. temic anaplastic large-cell lymphoma. Mod Pathol 2004; 17:109–16. 35 Okada E, Murai Y, Matsui K et al. Survivin expression in tumor cell 13 Clarke LE, Bayerl MG, Briggeman RD et al. Death receptor apoptosis nuclei is predictive of a favorable prognosis in gastric cancer signalling mediated by FADD in CD30-positive lymphoproliferative patients. Cancer Lett 2001; 163:109–16. disorders involving the skin. Am J Surg Pathol 2005; 29:452–9. 36 Lehner R, Lucia MS, Jarboe EA et al. Immunohistochemical localiza- 14 Rust R, Harms G, Blokzijl T et al. High expression of Mcl-1 in ALK tion of the IAP protein survivin in bladder mucosa and transitional positive and negative anaplastic large cell lymphoma. J Clin Pathol cell carcinoma. Appl Immunohistochem Mol Morphol 2002; 10:134–8. 2005; 58:520–4.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 48 Survivin and Bcl-2 in CD30+ lymphomas, G. Goteri et al.
37 Kennedy SM, O’Driscoll L, Purcell R et al. Prognostic importance of 42 Paulli M, Berti E, Boveri E et al. Cutaneous CD30+ lymphoprolifera- survivin in breast cancer. Br J Cancer 2003; 88:1077–83. tive disorders: expression of Bcl-2 and proteins of the tumor 38 Ding Y, Prieto VG, Zhang PS et al. Nuclear expression of the anti- necrosis factor receptor superfamily. Hum Pathol 1998; 29:1223–30. apoptotic protein survivin in malignant melanoma. Cancer 2006; 43 Mao X, Orchard G, Lillington DM et al. BCL2 and JUNB abnormal- 106:1123–9. ities in primary cutaneous lymphomas. Br J Dermatol 2004; 39 Vetter CS, Mu¨ller-Blech K, Schrama D et al. Cytoplasmic and nuclear 151:546–56. expression of survivin in melanocytic skin lesions. Arch Dermatol Res 44 Oudejans JJ, ten Berge RL, Meijer CJLM. Immune escape 2005; 297:26–30. mechanisms in ALCL. J Clin Pathol 2003; 56:423–5. 40 Li F, Yang J, Ramnath N et al. Nuclear or cytoplasmic expression of 45 Rassidakis GZ, Sarris AH, Herling M et al. Differential expression of survivin: what is the significance? Int J Cancer 2005; 114:509–12. BCL-2 family proteins in ALK-positive and ALK-negative anaplastic 41 Korsmeyer SJ. BCL-2 gene family and the regulation of pro- large cell lymphoma of T/null-cell lineage. Am J Pathol 2001; grammed cell death. Cancer Res 1999; 59:S1693–700. 159:527–35.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp41–48 CLINICAL AND LABORATORY INVESTIGATIONS DOI 10.1111/j.1365-2133.2007.07921.x Association of functional gene variants in the regulatory regions of COX-2 gene (PTGS2) with nonmelanoma skin cancer after organ transplantation M. Gomez Lira, S. Mazzola, G. Tessari,* G. Malerba, M. Ortombina, L. Naldi, G. Remuzzi, L. Boschiero,§ A. Forni,– C. Rugiu,** S. Piaserico, G. Girolomoni* and A. Turco Departments of Mother and Child, Section of Biology and Genetics, University of Verona, Strada Le Grazie 8, 37134 Verona, Italy *Department of Biomedical and Surgical Sciences, Section of Dermatology and Venereal Diseases, University of Verona, Verona, Italy Centro Studi Gruppo Italiano di Studi Epidemiologici in Dermatologia, Bergamo, Italy Department of Nephrology, Ospedali Riuniti di Bergamo, Bergamo, Italy §Kidney Transplantation Center, University of Verona, Verona, Italy –Division of Cardiac Surgery, University of Verona, Verona, Italy **Department of Nephrology, University of Verona, Verona, Italy Department of Dermatology, University of Padua, Padua, Italy
Summary
Correspondence Background Overexpression of cyclooxygenase-2 (COX-2), resulting in excessive Macarena Gomez Lira. prostaglandin production, has been observed in human epidermal keratinocytes E-mail: [email protected] after ultraviolet B injury, in squamous cell skin carcinoma (SCC), in actinic kera- toses, and in the early stages of carcinogenesis in a wide variety of tissues. The Accepted for publication 20 November 2006 dysregulation of COX-2 expression can in part be due to functional changes affecting regulatory elements in the promoter or 3¢ untranslated region (UTR) of Key words the gene. Two common polymorphisms (–765GfiC, and –1195AfiG) in the COX-2 gene, nonmelanoma skin cancer, promoter region of the COX-2 gene (now PTGS2), and one common polymorph- polymorphisms, PTGS2, transplantation ism in the 3¢ UTR (8473TfiC) have been described, and reported as associated Conflicts of interest with various malignancies. None declared. Objectives To determine if common known polymorphisms in the regulatory region of the COX-2 gene (PTGS2) can be associated with nonmelanoma skin cancer (NMSC) predisposition after organ transplantation, to evaluate if cancer risks are associated with specific COX-2 gene (PTGS2) haplotypes containing these polymorphisms, and to identify possible new genetic polymorphisms in the proximal 5¢ or 3¢ regulatory regions of the gene associated with disease. Methods The frequency of the three polymorphisms was determined in 240 Northern Italian transplant recipient patients (107 cases and 133 controls) with polymerase chain reaction–restriction fragment length polymorphism analysis. The proximal 5¢ and 3¢ regulatory regions of the gene were screened by hetero- duplex analysis. Results Stratification by age at transplant and type of tumours [SCC or basal cell car- cinoma (BCC)] demonstrated that allele –765C represented a protective factor in BCC cases undergoing transplantation before 50 years of age (CC + CG vs. GG, Fisher exact test P ¼ 0Æ003). One rare polymorphism, )62CfiG, was detected in the 5¢ flanking region. The allele frequency of –62G was 0Æ019, and no difference in genotype between cases and controls was observed. No other variants were found, suggesting that sequence variations in these regions are not likely to con- tribute to NMSC risk in this population. Haplotype analysis showed that the haplo- type containing all major alleles represents a protective factor in patients with SCC undergoing transplantation after 50 years of age [P ¼ 0Æ009; OR ¼ 0Æ37 (0Æ18– 0Æ79)] and that variant –1195AfiG may represent a risk factor in this subgroup of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 49 50 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al.
patients [P ¼ 0Æ01; OR ¼ 4Æ77 (1Æ47–16Æ41)]. Haplotype analysis in patients with BCC revealed that variant –765C might be a protective factor in patients undergoing transplantation before 50 years of age. Variant 8473TfiC, located in the 3¢ UTR region of the gene, showed no association with NMSC risk after transplantation. Conclusions COX-2 common variants –765GfiC and –1195AfiG appear to be asso- ciated with risk of NMSC, although in different ways in the SCC and BCC sub- groups, indicating that environmental and genetic risk factors may play different roles in the outcome leading to these two phenotypes.
Nonmelanoma skin cancer (NMSC) is the most commonly modulating the risk for various cancers. Two common poly- occurring cancer in humans, and its incidence is 4–20 times morphisms in the promoter region of the COX-2 gene higher in organ transplant recipients (OTRs) than in others, (PTGS2) (–765GfiC, and –1195AfiG) that affect promoter reaching the estimated incidence of 40–75% at 20 years activity,27,28 and one common polymorphism in the 3¢ un- after transplantation.1 Factors in OTRs that confer increased translated region (UTR) (8473TfiC), associated with colorectal susceptibility to NMSC are similar to those in the general adenoma, lung and breast cancer, have been described.29–31 population, but the tumours behave more aggressively, This study was designed to determine whether known com- occur at a younger age, are more numerous, grow more mon polymorphisms located in the regulatory region of the rapidly, and metastasize earlier in the transplant group.1–5 COX-2 gene (PTGS2) might be associated with a predisposition The most critical risk factor in NMSC is excessive ultraviolet to NMSC, to identify new genetic polymorphisms in the 5¢ or (UV) radiation exposure.6–8 Extensive documentation has 3¢ proximal regulatory regions of the gene that could be asso- validated the role of UVB irradiation (290–320 nm) as both ciated with disease, and to evaluate if cancer risks are associ- tumour initiator and promoter, inducing both squamous cell ated with specific COX-2 gene haplotypes. (SCC) and basal cell (BCC) carcinomas.8,9 UV radiation induces redox balance affecting the cell membrane, cell Materials and methods cycle, and rate of apoptosis in both human melanocytes and keratinocytes.10 Subjects Other factors including immune response, human papil- loma-virus (HPV) infection, and genetic predisposition, may Transplanted patients (kidney, heart and liver) undergoing also play a role in NMSC.11–14 Genetic susceptibility to skin immunosuppressive therapy with a functioning graft and a cancer may be particularly important in immunosuppressed minimum of 1 year of follow-up since transplantation, and individuals, who have the additional insults of long-term affected by at least one histologically proven NMSC, were con- immunosuppression and increased susceptibility to HPV infec- sidered cases. Controls were transplant patients free from any tions.15 One possible mechanism by which UV promotes car- skin cancer and matched for type of transplanted organ, sex cinogenesis is its ability to induce formation of prostaglandins and duration of transplantation. For each patient, basic demo- (PGs), which may then function as tumour promoters, or graphic data, date of transplantation and of the visit, long- may enhance initiation because of their ability to act as term immunosuppressive therapy, type of skin cancer, skin oxidants.16–19 The first step in the synthesis of PG from type, eye and hair colour were recorded. Long term mainten- arachidonic acid is catalysed by the enzyme cyclooxygenase ance immunosuppressive therapy consisted of a combination (COX). Two isoforms of COX have been described. COX-1 is of one or two immunosuppressive drugs (azathioprine, a housekeeping isoform constitutively expressed in most ciclosporin, mycophenolate mofetil and tacrolimus) with oral tissues, whereas COX-2 is induced by a variety of agents, methylprednisolone, according to current protocols.32 A total including proinflammatory agents and mitogens.20 An of 240 recipients of solid organ transplants, including 107 overexpression of COX-2, which results in excessive PG pro- cases and 133 controls, were enrolled in the study. duction, has been observed in human epidermal keratinocytes after UVB injury, in squamous cell skin cancer, and in early Molecular methods stages of carcinogenesis in a wide variety of tissues.21–23 Tran- scriptional regulation of COX-2 has been evaluated in multiple DNA was extracted from peripheral blood leucocytes by stand- cell lines.24 The promoter region of PTGS2, the COX-2 gene, ard salting out methods. Polymerase chain reaction (PCR) and contains a canonical TATA box and various putative transcrip- restriction fragment length polymorphism analyses were per- tional regulatory elements.25 COX-2 is also affected post-tran- formed to determine –765GfiC, –1195AfiG, and –62CfiG scriptionally, at the level of mRNA stability.26 Polymorphisms genotypes. The primers and restriction enzymes used are may either eliminate or create binding sites for various factors, shown in Table 1. Genotyping of polymorphism 8437TfiC potentially altering the expression of COX-2 and thereby was performed using restriction generating PCR, using a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al. 51
Table 1 Primers and endonuclease enzymes used for genotyping polymorphisms Primer forward Polymorphism Primer reverse Fragment Endonuclease –1195AfiG5¢CCCTGAGCACTACCCATGAT3¢ 273 bp PvuII 5¢GCCCTTCATAGGAGATACTGG3¢ –765GfiC5¢CCGCTTCCTTTGTCCATCAG3¢ 306 bp AciI 5¢GGCTGTATATCTGCTCTATATG3¢ –62CfiG5¢AAAGGCGGAAAGAAAGAG3¢ 108 bp HinfI 5¢TGCTCCTGACGCTCACTG3¢ 8473TfiC5¢GTTTGAAATTTTAAAGTACTTTTGAT3¢ 147 bp BclI 5¢TTTCAAATTATTGTTTCATTGGC3¢
modified forward primer that creates a BclI restriction site interval were also calculated. Haplotype frequencies were when allele 8473C is present (Table 1). derived using prediction of haplotype frequencies via the Screening for new polymorphisms in the proximal promo- Maximum Likelihood algorithm (http://www.bioinf.mdc- ter and 3¢ UTR regions was performed in 30 cases and 30 berlin.de/). Single locus and haplotype analyses were per- controls by PCR and heteroduplex analysis. A 407 bp fragment formed in the total group of NMSC. We also tested the of the proximal promoter region, and a 240 bp fragment hypothesis of association with SCC or BCC as the aetiology of of the proximal 3¢ UTR regions were amplified using SCC and BCC might be very different33,34 and by stratifying primers forward 5¢CGGTATCCCATCCAAGGC3¢ and reverse individuals by age < or > 50 years at transplantation, as age 5¢TGCTCCTGACGCTCACTG3¢, and amplification products is an important risk factor for NMSC.34–36 Adjusting by the were analysed for heteroduplex formation using SequaGel ND Bonferroni correction for multiple tests, a P-value 0Æ05/9 ¼ (National Diagnostics, Charlotte, NC, U.S.A.) gel electrophor- 0Æ005 was the threshold for acceptance of the hypothesis of esis. Fragments containing heteroduplex bands were directly association. sequenced in a 377 ABI PRISM automated DNA sequencer. Results Statistical methods Two hundred and forty Northern Italian transplant recipient Genotypes for all polymorphisms were measured and the Epi patients were enrolled in this study. Cases were represented by Info software package (version 2003) was used for assessing 49 patients presenting at least one BCC, 47 presenting at least association using a 2 · 2 · n contingency table. v2 tests, with one SCC, five with Bowen’s disease, two with keratoacan- Yates’ correction or Fisher’s exact test (two-tailed) were used thoma and four with other tumours (Table 2). Transplant for association. Odds ratios (ORs) and the OR 95% confidence patients without clinical or histological evidence of skin cancer
Table 2 Study group characteristics Cases Controls Total Sex M/F (total patients) 89/18 114/19 203/37 Type of transplantation (No.) Kidney 72 102 174 Heart 29 28 57 Liver 6 3 9 Mean ± SD age at transplantation (years) 48Æ9±1Æ142Æ2±12Æ945Æ2±12Æ5 (Range) (26Æ8–77Æ8) (28Æ5–74Æ6) (26Æ8–77Æ8) Cases vs. controls: t ¼ 4Æ73, P <0Æ00001 Immunosuppressive therapy (No. %) One immunosuppressive drug 10 (9Æ3) 9 (6Æ8) 19 (7Æ9) One immunosuppressive drug + 47 (43Æ9) 70 (52Æ6) 117 (48Æ8) methylprednisolone Two immunosuppressive drugs + 50 (46Æ7) 54 (40Æ6) 104 (43Æ3) methylprednisolone Cases vs. controls: v2 ¼ 1Æ93, P ¼ 0Æ38 Type of cancers (cases) Basal cell carcinoma 49 Squamous cell carcinoma 47 Bowen’s disease 5 Keratoacanthoma 2 Other 4
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 52 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al. acted as controls. The mean age at transplantation and at the plantation showed no difference between cases and controls consultation was significantly greater in cases than in controls (Table 3). Variant –1195G was over-represented in SCC (P <0Æ00001). The types of immunosuppressive drug did not patients undergoing transplantation after 50 years of age, differ significantly between cases and controls (P ¼ 0Æ38) but this difference did not reach significance (P ¼ 0Æ09) (Table 2). No difference in skin type, hair and eye colour (Table 3). between cases and controls was observed. In order to determine if other gene variants in the proximal regulatory regions of the gene could be associated with NMSC in the OTR population, we next amplified and screened by Single locus analysis heteroduplex analysis a 407 bp fragment of the proximal Three polymorphisms in the COX-2 gene (PTGS2), with rea- 5¢ flanking ()371/+ 36) region, and a 240 bp fragment of sonable frequency distribution in Caucasians and likely to have the proximal 3¢ UTR region (including 60 bp of the last functional relevance, were selected for analysis of association exon) of the gene, in 30 cases and in 30 control individuals. with the development of NMSC: two in the promoter region One control individual showed a heteroduplex band in the (–765GfiC, and –1195AfiG) and one in the 3¢ UTR region 5¢ promoter region. Sequence analysis identified a CfiG trans- (8473TfiC) of the gene. Observed genotype distributions of version at position –62. This rare variant –62CfiG had a total COX-2 polymorphisms are summarized in Table 3. Allelic fre- frequency of 0Æ02 in the study group, with no difference in quencies for the –765C, –1195G, and 8473C were 0Æ17, genotype distribution between cases and controls, and it was 0Æ15, and 0Æ33, respectively, in the overall population. Single not considered in further haplotype analysis because of its low locus analysis of polymorphisms showed no significant differ- frequency. This rare variant is described here for the first time, ences in genotype distributions between cases and controls and its frequency in the study group was not different com- when analysing NMSC as a total group of cases (Table 3). pared with that found by our laboratory in a control popula- Stratification by type of tumour demonstrated that allele tion of 100 Northern Italian individuals (frequency ¼ 0Æ025). –765C was under-represented in BCC cases [OR: 0Æ55 (0Æ25– No other variants were found in any of the regulatory regions 1Æ19)], although this difference did not reach significance analysed. (P ¼ 0Æ14) (Table 3). Further stratification of the BCC popula- tion by age at transplant showed that allele –765C was never Haplotype analysis present in BCC cases who underwent transplantation before 50 years of age (CC + CG vs. GG, P ¼ 0Æ003) (Table 3). We next calculated the haplotype frequencies of polymor- Although stratification leads to a relatively small number of phisms –765GfiC, –1195AfiG and 8473TfiC of COX-2 individuals by group (26 BCC cases and 90 control indivi- gene (PTGS2) using the ithap-cgi program37 (http://www. duals), these results suggest a significant association of this bioinf.mdc-berlin.de/). The three polymorphisms generated gene variant with a lower risk of BCC in this subgroup of four common haplotypes, which accounted for > 97% of the OTR patients, and this finding appears to be of great interest chromosomes observed. The global haplotype distribution did for further follow-up. Stratification of SCC by age at trans- not significantly differ between cases and controls when
Table 3 Genotype distribution of COX-2 polymorphisms in cases and controls as total NMSC and stratified by BCC, SCC and age at transplant
765GfiC 1195AfiG 8473TfiC
GG CG CC TotalP-value AA AG GG TotalP-value TT TC CC Total P-value Controls 88 36 5 129 96 33 2 131 64 51 15 130 NMSC 76 23 6 105 0Æ50 76 25 3 104 0Æ8 44 47 12 103 0Æ6 BCC 39 6 2 47 0Æ11a,b 35 10 2 47 0Æ5 23 18 5 46 1Æ0 SCC 29 15 3 47 0Æ64 32 14 1 47 0Æ8 17 24 6 47 0Æ3 < 50 years Controls 63 23 4 90 65 24 2 91 43 38 10 91 NMSC 38 8 1 18 0Æ43872470Æ3 20 23 4 47 0Æ7 BCC 26 0 0 26 0Æ005c,d 19 5 2 26 0Æ3 15 10 1 26 0Æ4 SCC 9 8 1 18 0Æ21620180Æ34122180Æ1 > 50 years Controls 25 13 1 39 31 9 0 40 21 13 5 39 NMSC 37 15 5 57 0Æ4 37 18 1 56 0Æ4 24 23 8 55 0Æ6 BCC 12 6 2 20 0Æ51550201Æ0 8 7 4 19 0Æ6 SCC 20 7 2 29 0Æ5 16 12 1 29 0Æ1e,f 13 12 4 29 0Æ8
aAllele frequency P ¼ 0Æ14, OR ¼ 0Æ55 (0Æ25–1Æ19). bCC + CG vs. GG, P ¼ 0Æ08, OR ¼ 0Æ44 (0Æ17–1Æ09). cAllele frequency P ¼ 0Æ003, OR ¼ 0Æ00 (0Æ00–0Æ47). dCC + CG vs. GG, P ¼ 0Æ003, OR ¼ 0Æ00 (0Æ000Æ48). eAllele frequency P ¼ 0Æ08, OR ¼ 2Æ5(0Æ9–6Æ9). fGG + AG vs. AA, P ¼ 0Æ1, OR ¼ 2Æ8(0Æ9–9Æ1).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al. 53 considered over all the NMSC cases. However, stratification of (0Æ24–0Æ83)] (Table 4). Compared with haplotype 111, the NMSC by age at transplantation, which is an important risk haplotype containing only variant –1195T (211) was over- factor for NMSC after transplantation, showed a significant represented in cases [P ¼ 0Æ01; OR ¼ 4Æ00 (1Æ28–13Æ27)]. global difference between NMSC and controls in the subgroup Further stratification by type of tumour demonstrated that the of patients who underwent transplantation after 50 years of risk conferred by haplotype 211 was particularly strong in age (P ¼ 0Æ04). This difference was mainly due to a higher SCC patients [OR ¼ 4Æ77 (1Æ47–16Æ41); P ¼ 0Æ01] (Table 5). frequency in controls than in NMSC patients of the haplotype Compared with haplotype 111, haplotype 211 conferred a carrying all the major variants (111) [P ¼ 0Æ01; OR ¼ 0Æ45 higher risk [OR ¼ 6Æ21 (1Æ79–22Æ70); P ¼ 0Æ002] which
Table 4 Distribution of COX-2 haplotype frequencies among cases and controls and Haplotype Controls NMSC OR (95% CI) P-value their association with nonmelanoma skin cancer (NMSC) Global test P ¼ n.s. (n ¼ 266) (n ¼ 214) A)1195-G)765-T8473 (111) 145 108 0Æ85 (0Æ58–1Æ24) 0Æ5 A)1195-G)765-C8473 (112) 38 39 1Æ34 (0Æ80–2Æ24) 0Æ3 G)1195-G)765-T8473 (211) 37 29 0Æ97 (0Æ54–1Æ69) 1Æ0 A)1195-C)765-C8473 (122) 46 33 0Æ87 (0Æ52–1Æ46) 0Æ7 Others 0 5 nd nd < 50 years Global test P ¼ n.s. (n ¼ 180) (n ¼ 96) 111 90 56 1Æ40 (0Æ82–2Æ38) 0Æ2 112 29 19 1Æ28 (0Æ64–2Æ55) 0Æ6 211 28 9 0Æ56 (0Æ23–1Æ32) 0Æ2 122 30 10 0Æ58 (0Æ25–1Æ31) 0Æ2 Others 3 2 nd nd > 50 years Global test P ¼ 0Æ04 (n ¼ 80) (n ¼ 116) 111 51 51 0Æ45 (0Æ24–0Æ83) 0Æ01 112 8 19 1Æ76 (0Æ68–4Æ67) 0Æ3 211 5 20 3Æ13 (1Æ04–10Æ01) 0Æ04 122 13 23 1Æ27 (0Æ57–2Æ88) 0Æ6 Others 3 3 nd nd
nd, not done.
Table 5 Distribution of COX-2 haplotype frequencies among cases and controls and P-value of their association with squamous cell independent carcinoma (SCC) Haplotype Controls SCC OR (95% CI) haplotype Global test P ¼ n.s. (n ¼ 266) (n ¼ 94)
A)1195-G)765-T8473 (111) 145 41 0Æ65 (0Æ39–1Æ06) 0Æ09 A)1195-G)765-C8473 (112) 38 16 1Æ23 (0Æ62–2Æ43) 0Æ64 G)1195-G)765-T8473 (211) 37 16 1Æ27 (0Æ64–2Æ51) 0Æ57 A)1195-C)765-C8473 (122) 46 20 1Æ29 (0Æ69–2Æ41) 0Æ48 Others 0 1 nd nd < 50 years Global test, P ¼ n.s. (n ¼ 180) (n ¼ 36) 111 90 18 1Æ00 (0Æ46–2Æ17) 0Æ9 112 29 6 1Æ04 (0Æ35–2Æ93) 0Æ9 211 28 2 0Æ32 (0Æ05–1Æ48) 0Æ2 122 30 10 1Æ92 (0Æ77–4Æ72) 0Æ2 Others 3 0 nd nd > 50 years Global test, P ¼ 0Æ017 (n ¼ 80) (n ¼ 58) 111 51 23 0Æ37 (0Æ18–0Æ79) 0Æ009 112 8 10 1Æ88 (0Æ63–5Æ68) 0Æ32 211 5 14 4Æ77 (1Æ47–16Æ41) 0Æ01 122 13 10 1Æ11 (0Æ40–2Æ89) 0Æ9 Others 3 1 nd nd
nd, not done.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 54 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al.
Table 6 Distribution of COX-2 haplotype P-value of frequencies among cases and controls independent and their association with basal cell carcinoma Haplotype Controls BCC OR (95% CI) haplotype (BCC) Global test P ¼ n.s. (n ¼ 266) (n ¼ 94)
A)1195-G)765-T8473 (111) 145 54 1Æ13 (0Æ68–1Æ86) 0Æ8 A)1195-G)765-C8473 (112) 38 17 1Æ32 (0Æ67–2Æ59) 0Æ5 G)1195-G)765-T8473 (211) 37 13 0Æ99 (0Æ47–2Æ05) 0Æ9 A)1195-C)765-C8473 (122) 46 9 0Æ51 (0Æ22–1Æ13) 0Æ1 Others 0 1 nd nd < 50 years Global test P ¼ 0Æ016 (n ¼ 180) (n ¼ 54) 111 90 35 2Æ06 (1Æ03–4Æ15) 0Æ04 112 29 10 1Æ18 (0Æ50–2Æ78) 0Æ8 211 28 7 1Æ24 (0Æ52–2Æ92) 0Æ8 122 30 0 0Æ00 (0Æ00–0Æ49) 0Æ004 Others 3 0 nd nd > 50 years Global test P ¼ n.s. (n ¼ 76) (n ¼ 42) 111 51 19 0Æ47 (0Æ2–1Æ07) 0Æ087 112 8 7 1Æ80 (0Æ53–6Æ05) 0Æ4 211 5 6 2Æ50 (0Æ62–10Æ27) 0Æ20 122 13 9 1Æ41 (0Æ49–3Æ98) 0Æ60 Others 3 1 nd nd
nd, not done. gives suggestive evidence of an association of this haplotype and different transcription factor binding sites might be import- with a higher risk of SCC development. ant in the regulation of COX-2 transcription in the development In BCC patients the at-risk haplotype was less frequent and of different kinds of tumours. In particular, the immediately did not reach significance. Stratification of patients by tumour proximal 5¢ flanking region of the COX-2 gene (PTGS2) has type (BCC and SCC), and age at transplant revealed a haplo- been shown to have the highest transcriptional activity in type conferring a lower risk of developing BCC in the sub- mouse skin carcinoma cells and in colon carcinoma cells,47 and group of patients undergoing transplantation before 50 years has also been shown to be a region inducible by liposaccharide of age. This was due to the absence of haplotype 122 and phorbol ester in vascular endothelial cells.48 Expression of
(A)1195-C)765-C8437) in this subgroup of patients [P ¼ 0Æ004; COX-2 can also be increased by stabilization of the COX-2 tran- OR ¼ 0Æ00 (0Æ00–0Æ49)] (Table 6). script.49 The 3¢ UTR regulatory region of the gene plays an im- portant role in mRNA stability and translation.26 Specifically, Discussion the immediately proximal 116 nucleotide of the 3¢ UTR region (CR1) contains six copies of the mRNA destabilizing motif AU- Several different mechanisms can potentially explain the link UUA (ARE), and mediates COX-2 mRNA degradation.50 ARE between COX-2 and malignancy, including NMSC. Increased sequences regulate mRNA via interaction with sequence-specific synthesis of COX-2-derived PGs can stimulate cell proli- RNA-binding proteins,51 and mutations in this region could feration,38 promote angiogenesis,39–42 increase invasiveness43 alter COX-2 levels by stabilization of the COX-2 mRNA or by and tumorigenesis,42 and inhibit apoptosis.44,45 dysregulation of translation. However, screening of the prox- This case-control study describes the analysis of COX-2 gen- imal 5¢ and 3¢ UTR regions of the gene revealed just one rare etic variability in relation to NMSC after transplantation. To our variant: –62CfiG. The frequency of this variant was 0Æ02 in the knowledge, this is the first study which investigates the effect overall population, and because of its low frequency, it was not of COX-2 genetic variants on the risk of NMSC after transplant- further considered for haplotype analysis. These results tend to ation, and where regulatory regions of the gene have been sys- exclude the hypothesis that sequence variations in the proximal tematically screened for polymorphisms in relation to this regulatory regions of the COX-2 gene (PTGS2) can contribute to phenotype. We have focused on the –1195GfiA and NMSC risk in our population. –765GfiC polymorphisms in the promoter region and the Single locus analysis of the three known polymorphisms 8473TfiC variant in the 3¢ UTR, as these common polymor- demonstrated that allele –765C was associated with a phisms located in the transcription or post-transcriptional regu- decreased risk of BCC in patients undergoing transplantation latory regions have been reported to be associated with various before 50 years of age, and allele –1195G was more frequent human diseases.27–31,46 Depending on the stimulus and the cell in patients with SCC who had transplantation after 50 years of type, transcriptional factors modulate the expression of COX-2, age, although this did not reach significance (Table 3).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al. 55
Haplotype analysis identified a haplotype that confers a monocytes was more than ten-fold higher in –765CC homozy- reduced risk of NMSC in the subgroup of patients undergoing gotes, who were over-represented in asthmatic patients com- 53,54 transplantation before 50 years of age. Haplotype A)1195- pared with controls, than in –765GG homozygotes. The
C)765-C8437 (122) was not observed in this subgroup of BCC COX-2 variant –765C has also been reported as an at-risk allele patients. As variant 8437C was also present in haplotype for oesophageal cancer.28 These contrasting results could be
A)1195-G)765-C8437 (112) which did not present any differ- explained because –765G can abolish an Sp1 binding site but ence in frequency between cases and controls, we inferred can create a binding site for E2F, a cyclin-dependent regulator that the putative protective effect of this haplotype might be of expression of several genes. In the same way, variant – due only to the presence of variant –765C. These results indi- 1195G may abolish a c-MYB binding site but, as assessed by cate that, irrespective of the other polymorphisms, allele – bioinformatic TESS Job, it may also create an interleukin (IL)-6 765C may indeed represent a protection factor against the responsive element, and could result in enhanced COX-2 tran- development of BCC tumours in individuals undergoing trans- scription in skin cancer cells after transplantation. It has been plantation before 50 years of age. We could speculate that reported that overexpression of IL-6 in BCC cell lines increases allele –765C, which may result in reduced activity of COX-2, expression of COX-2 and is accompanied by high vasculariza- represents a protection factor against the development of BCC tion.55 In particular, this new IL-6 response element created by tumours in individuals undergoing transplantation before –1195G may be specially important in OTR patients who are 50 years of age and that this effect is probably abolished in treated with immunosuppressive drugs in combination with elderly individuals where the additive effects of risk factor glucocorticoids, as a synergistic effect of IL-6 and glucocorti- accumulation and diminished DNA repair capacity exert stron- coids upon gene expression has been reported.56 As glucocorti- ger effects. coid treatment has been reported to be a risk factor for In the subgroup of patients who underwent transplantation NMSC,57 it could be worthwhile to determine if this risk varies after 50 years of age, the haplotype carrying all the major vari- depending on the status of the COX-2 –1195AfiG gene variant. ants (111) was more frequent in controls than in patients, con- Further functional analysis in an appropriate cell model and real ferring a lower risk of NMSC, especially SCC. The only time assessment of COX-2 expression directly in SCC specimens haplotype containing variant –1195G appeared to increase the from OTR patients are necessary to elucidate the possible bio- risk of developing the phenotype. Functional studies of the logical effects of this variant upon COX-2 expression in these effects of the –1195GfiA polymorphism have demonstrated tumours. that in HeLa cells culture, the reporter gene expression of con- In conclusion, the common COX-2 variants –765GfiC and structs carrying the –1195A variant was 4–6-fold greater than –1195AfiG appear to be associated with NMSC risks, those driven by the –1195G counterparts, and this effect was although in different ways in the SCC and BCC subgroups, postulated to be the result of the creation of a c-MYB binding indicating that environmental and genetic risk factors may play site by variation –1195A.28 Real-time PCR quantification of different roles in the outcome of these two types of lesions. COX-2 mRNA in individual oesophageal tissue demonstrated Variant –765C results, independent of other polymorphisms significantly higher COX-2 mRNA levels in individuals carry- analysed, were associated with a lower risk of developing BCC ing the –1195AA genotype than in individuals carrying in patients undergoing transplantation before 50 years of age. –1195GG.28 Allele –1195A has also been associated with a In SCC, the effect of variant –765C is not evident, while the higher risk of gastric cancer52,53 in Chinese populations. On the haplotype carrying all major variants (–1195A, –765G, and contrary, in this study population allele –1195A appears to con- 8473T) appears to represent a protective factor in patients fer a protective effect against the development of SCC. The undergoing transplantation at older age. The haplotype which frequency of allele –1195G is much lower in this study popula- appears to confer the highest risk in this subclass of patients is tion (0Æ15) compared with that reported in the Chinese popula- the one carrying –1195G which may increase COX-2 expres- tion (0Æ49), where allele –1195A is the major allele. Only five sion due to the creation of an IL-6 response element. individuals were homozygous for the –1195G, and 172 indi- viduals homozygous for the –1195A. Therefore, the presence Acknowledgments of the c-MYB binding site is the most common situation in this population. Our results cannot be easily explained from the The study was supported by MURST 60%, by grants from the functional role of –1195AfiG reported in HeLa culture cell Italian Ministry of University and Research, Italian Ministry of assays.28 This would indicate that probably other functional Health and by the Cariverona Foundation. We thank the Banca alleles may be in linkage disequilibrium with the –1195AfiG Popolare di Verona, Verona, Italy, for its generous financial alleles. However, the effects of polymorphisms can vary from support. tissue to tissue, especially for inducible genes such as COX-2 (PTGS2). In fact, the reduced expression of COX-2 has been References associated with variant –765C when compared with –765G, and this effect was postulated to be mediated by loss of an 1 Moloney FJ, Comber H, O’Lorcain P et al. A population based Sp1 transcription binding site.27 However, in contrast to this study of skin cancer incidence and prevalence in renal transplant observation, it has been reported that production of PG by recipients. Br J Dermatol 2006; 154:498–504.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 56 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al.
2 Liddington M, Richardson AJ, Higgins RM et al. Skin cancer in 23 Asgari M, White E, Chren MM. Nonsteroidal anti inflammatory renal transplant recipients. Br J Surg 1998; 76:1002–5. drug use in the prevention and treatment of squamous cell 3 Glover MT, Niranjan N, Kwan JT, Leigh IM. Non-melanoma skin carcinoma. Dermatol Surg 2004; 30:1335–42. cancer in renal transplant recipients: the extent of the problem and 24 Chun KS, Surh YJ. Signal transduction pathway regulating cyclo- a strategy for management. Br J Plast Surg 1994; 47:86–9. oxygenase 2 expression: potential molecular targets for chemo- 4 Martinez JC, Otley CC, Stasko T et al. Transplant skin cancer colla- prevention. Biochem Pharmacol 2004; 68:1089–110. borative. Defining the clinical course of metastatic skin cancer 25 Kosaka T, Miyata A, Ihara H et al. Characterization of the human in organ transplant recipients: a multicenter collaborative study. gene (PTGS2) encoding prostaglandin endoperoxide synthase 2. Eur Arch Dermatol 2003; 139:301–6. J Biochem 1994; 221:889–97. 5 Kovach BT, Sams HH, Stasko T. Systemic strategies for chemo- 26 Cok SJ, Morrison AR. The 3¢ untranslated region of murine cyclo- prevention of skin cancers in transplant patients. Clin Transplant oxygenase 2 contains multiple regulatory elements that alter mes- 2005; 19:726–34. sage stability and translational efficiency. J Biol Chem 2001; 6 Hall EJ, Astor M, Bedford J et al. Basic radiobiology. Am J Clin Oncol 276:23179–85. 1988; 11:220–52. 27 Papafili A, Hill MR, Brull D et al. Common promoter variant in 7 Black HS, deGruijl FR, Forbes PD et al. Photocarcinogenesis: an cyclooxygenase 2 represses gene expression: evidence of role in overview. J Photochem Photobiol B 1997; 40:29–47. acute phase inflammatory response. Arterioscler Thromb Vasc Biol 2002; 8 de Gruijl FR, van Kranen HJ, Mullenders LH. UV induced DNA 22:1631–6. damage, repair, mutations and oncogenic pathways in skin cancer. 28 Zhang X, Miao X, Tan W et al. Identification of functional genetic J Photochem Photobiol B 2001; 63:19–27. variants in cyclooxygenase 2 and their association with risk of eso- 9 Griffiths H, Mistry P, Herbert K, Lunec J. Molecular and cellular phageal cancer. Gastroenterology 2005; 129:565–76. effects of ultraviolet light induced genotoxicity. Crit Rev Clin Lab Sci 29 Park JM, Choi JE, Chae MH et al. Relationship between cyclooxyge- 1998; 35:189–237. nase 8473TfiC polymorphism and the risk of lung cancer: a case– 10 Larsson P, Andersson E, Johansson U et al. Ultraviolet A and B control study. BMC Cancer 2006; 6:70. affect human melanocytes and keratinocytes differently. A study 30 Langsenlehner U, Yazdani Biuki B, Eder T et al. The cyclooxygenase of oxidative alterations and apoptosis. Exp Dermatol 2005; 14:117– 2(PTGS2) 8473TfiC polymorphism is associated with breast can- 23. cer risk. Clin Cancer Res 2006; 12:1392–4. 11 Aubin F, Humbey O, Guerrini JS et al. Non melanoma skin 31 Ali IU, Luke BT, Dean M, Greenwald P. Allelic variants in regula- cancers and human papillomavirus. Ann Dermatol Venereol 2003; tory regions of cyclooxygenase 2: association with advanced colo- 130:1131–8. rectal adenoma. Br J Cancer 2005; 93:953–9. 12 Bouwes Bavinck JN, Feltkamp M, Struijk L, ter Schegget J. Human 32 Halloran PF. Immunosuppressive drugs for kidney transplantation. papillomavirus infection and skin cancer risk in organ transplant N Engl J Med 2004; 351:2715–29. recipients. J Invest Dermatol Symp Proc 2001; 6:207–11. 33 Boukamp P. Non melanoma skin cancer: what drives tumor devel- 13 Madan V, Hoban P, Strange RC, Fryer A et al. Genetics and risk factors opment and progession? Carcinogenesis 2005; 26:1657–67. for basal cell carcinoma. Br J Dermatol 2006; 154 (Suppl. 1): 5–7. 34 Lee JL, Kim A, Kopelovich L et al. Differential expression of E prost- 14 Tilli CM, van Steensel MA, Krekels GA et al. Molecular aetiology anoid receptors in murine and human non melanoma skin cancer. and pathogenesis of basal cell carcinoma. Br J Dermatol 2005; J Invest Dermatol 2005; 12:818–25. 152:1108–24. 35 Mithoefer AB, Supran S, Freeman RB. Risk factors associated with 15 Dang C, Koehler A, Forschner T et al. E6/E7 expression of human the development of skin cancer after liver transplantation. Liver papillomavirus types in cutaneous squamous cell dysplasia and Transpl 2002; 8:939–44. carcinoma in immunosuppressed organ transplant recipients. Br J 36 Herrero JI, Lorenzo M, Quiroga J et al. De novo neoplasia after Dermatol 2006; 155:129–36. liver transplantation: an analysis of risk factors and influence on 16 Vanderveen EE, Grekin RC, Swanson NA, Kragballe K. Arachidonic survival. Liver Transpl 2005; 11:89–97. acid metabolites in cutaneous carcinomas. Evidence suggesting that 37 Rohde K, Fuerst R. Haplotyping and estimation of haplotype elevated levels of prostaglandins in basal cell carcinomas are associ- frequencies for closely linked biallelic multilocus genetic pheno- ated with an aggressive growth pattern. Arch Dermatol 1986; types including nuclear family information. Hum Mutat 2001; 122:407–12. 17:289–95.
17 Kruza LL, Pentland AP. Mechanisms of UV induced inflammation. 38 Sheng H, Shao J, Washington MK, DuBois RN. Prostaglandin E2 J Invest Dermatol 1993; 100:355–415. increases growth and motility of colorectal carcinoma cells. J Biol 18 Grewe M, Trefzer U, Ballhorn A et al. Analysis of the mechanism Chem 2001; 276:18075–81. of ultraviolet (UV) B radiation induced prostaglandin E2 synthesis 39 Tsujii M, Kawano S, Tsuji S et al. Cyclooxygenase regulates angio- by human epidermoid carcinoma cells. J Invest Dermatol 1993; genesis induced by colon cancer cells. Cell 1998; 93:705–16. 101:528–31. 40 Nijsten T, Colpaert CG, Vermeulen PB et al. Cyclooxygenase 2 19 Vane JR, Mitchel JA, Appleton I et al. Inducible isoform of cyclo- expression and angiogenesis in squamous cell carcinoma of the oxygenase and nitric oxide synthase in inflammation. Proc Natl Acad skin and its precursors: a paired immunohistochemical study of 35 Sci USA 1994; 91:2046–50. cases. Br J Dermatol 2004; 151:837–45. 20 Chandrasekharan NV, Simmons DL. The cyclooxygenases. Genome 41 O’Grady A, O’Kelly P, Murphy GM et al. COX 2 expression corre- Biol 2004; 5:241. lates with microvessel density in non melanoma skin cancer from 21 Buckman SY, Gresham A, Hale P et al. COX 2 expression is induced renal transplant recipients and immunocompetent individuals. by UVB exposure in human skin: implications for the development Hum Pathol 2004; 35:1549–55. of skin cancer. Carcinogenesis 1998; 19:723–9. 42 Tjiu JW, Liao YH, Lin SJ et al. Cyclooxygenase 2 overexpression 22 Kim KH, Park EJ, Seo Y et al. Immunohistochemical study of cyclo- in human basal cell carcinoma cell line increases antiapoptosis, oxygenase 2 and p53 expression in skin tumors. J Dermatol 2006; angiogenesis, and tumorigenesis. J Invest Dermatol 2006; 126:1143– 33:319–25. 51.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 COX-2 gene variants in NMSC after transplantation, M. Gomez Lira et al. 57
43 Dohadwala M, Luo J, Zhu L et al. Non small cell lung cancer cyclo- 51 Dixon DA, Kaplan CD, McIntyre TM et al. Post transcriptional con- oxygenase 2 dependent invasion is mediated by CD44. J Biol Chem trol of cyclooxygenase 2 gene expression. J Biol Chem 2000; 2001; 276:20809–12. 275:11750–7. 44 Tsujii M, DuBois RN. Alterations in cellular adhesion and apoptosis 52 Liu F, Pan K, Zhang X et al. Genetic variants in cyclooxygenase 2: in epithelial cells overexpressing prostaglandin endoperoxide expression and risk of gastric cancer and its precursors in a Chinese synthase 2. Cell 1995; 83:493–501. population. Gastroentorology 2006; 130:1975–84. 45 Sheng H, Shao J, Morrow JD et al. Modulation of apoptosis and Bcl 53 Zhang XM, Miao XP, Tan W et al. Genetic polymorphisms in the 2 expression by prostaglandin E2 in human colon cancer cells. promoter region of cyclooxygenase 2 and their association with Cancer Res 1998; 58:362–6. risk of gastric cancer. Chung-Kuo i Hsueh Ko Hsueh Yuan Hsueh Pao Acta 46 Cipollone F, Toniato E, Martinoti S et al. A polymorphism in the Academiae Medicinae Sinicae 2006; 28:119–23. cyclooxygenese 2 gene as an inherited protective factor against 54 Szczeklik W, Sanak M, Szczeklik A. Functional effects and gender myocardial infarction and stroke. JAMA 2004; 291:2221–40. association of COX 2 gene polymorphism G765C in bronchial 47 Shao J, Sheng H, Inoue H et al. Regulation of constitutive cyclo- asthma. J Allergy Clin Immunol 2004; 114:248–53. oxygenase 2 expression in colon carcinoma cells. J Biol Chem 2000; 55 Jee SH, Chu CY, Chiu HC et al. Interleukin-6 induced basic fibro- 275:33951–6. blast growth factor-dependent angiogenesis in basal cell carcinoma 48 Inoue H, Yokoyama C, Hara S et al. Transcriptional regulation of cell line via JAK/STAT3 and PI3 kinase/Akt pathways. J Invest Dermatol human prostaglandin endoperoxide synthase 2 gene by lipopoly- 2004; 123:1169–75. saccharide and phorbol ester in vascular endothelial cells. Involve- 56 Takeda T, Kurachi H, Yamamoto T et al. Crosstalk between the ment of both nuclear factor for interleukin 6 expression site and interleukin-6 (IL-6)-JAK-STAT and the glucocorticoid-nuclear cAMP response element. J Biol Chem 1995; 270:24965–71. receptor pathway: synergistic activation of IL-6 response element 49 Skim J, Karin M. The control of mRNA stability in response to by IL-6 and glucocorticoid. J Endocrinol 1998; 159:323–30. extracellular stimuli. Mol Cell 2002; 14:323–31. 57 Karagas MR, Cushing GL Jr, Greenberg ER et al. Non mela- 50 Kim Y, Fischer SM. Transcriptional regulation of cyclooxygenase 2 noma skin cancers and glucocorticoid therapy. Br J Cancer 2001; in mouse skin carcinoma cells. J Biol Chem 1998; 273:27686–94. 85:683–6.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp49–57 CLINICAL AND LABORATORY INVESTIGATIONS DOI 10.1111/j.1365-2133.2007.07937.x Sentinel lymph node biopsy in melanoma: a micromorphometric study relating to prognosis and completion lymph node dissection S. Debarbieux, G. Duru,* S. Dalle, O. Be´atrix, B. Balme and L. Thomas Department of Dermatology, Hotel Dieu 69288, Lyon CEDEX 02, France *Health Systems Analysis Laboratory, FRE 2747, CNRS, Universite´ Claude Bernard Lyon 1, 69622 Villeurbanne CEDEX, France Department of Surgical Oncology, Centre Hospitalier Lyon Sud, Oullins, France
Summary
Correspondence Background Sentinel lymph node (SLN) positivity has been found to be strongly Luc Thomas. associated with a poor prognosis in melanoma. E-mail: [email protected] Objectives This large referral centre study was conducted: (i) to confirm the power- ful prognostic value of SLN biopsy (SLNB); (ii) to correlate patient prognosis to Accepted for publication 6 October 2006 the micromorphometric features of SLN metastasis in SLN-positive patients; and (iii) to correlate these micromorphometric features to the likelihood of positive Keywords completion lymph node dissection (CLND). melanoma, prognosis, sentinel lymph node Patients and methods SLNB was performed in 455 cases of primary melanoma between January 1999 and December 2004; for patients with positive SLN, the Conflicts of interest following micromorphometric features were registered: size of the largest meta- None declared. stasis (two diameters), depth of metastasis, number of millimetric slices involved, maximum number of metastases on a single section, presence of intracapsular lymphatic invasion and extracapsular spread. Kaplan–Meier survival curves were compared with the log-rank test; multivariate analysis was performed using a Cox regression model. Dependence of CLND status on micromorphometric fea- tures of SLN was assessed by the v2 test and predictive values of the different fea- tures were evaluated by multivariate analysis using a logistic regression model. Results A positive SLN was identified in 98 of our 455 cases. Survival was signifi- cantly shorter in SLN-positive patients than in SLN-negative patients. Extracapsu- lar invasion was found to be an independent prognostic factor of disease-free survival; ulceration of the primary and the maximum diameter of the largest metastasis were identified as independent predictive factors of disease-specific survival. Age and the lowest diameter of the largest metastasis were identified as independent predictive criteria of positive CLND, whereas depth of meta- stasis was not. Positivity of CLND was not significantly associated with a worse prognosis. Conclusions Our study confirms the previously demonstrated strong prognostic value of SLNB. It also confirms the relationship between tumour burden in the SLN (evaluated by the maximum diameter of the largest metastasis) and clinical outcome. We point out a new micromorphometric feature of SLN, which seems to be predictive of CLND status: the lowest diameter of the largest metastasis.
Sentinel lymph node biopsy (SLNB) is now offered in many metastasize in other nodes. Consequently, patients with neg- centres to patients with high-risk primary melanoma. The ative SLN are considered to be at lower risk of further procedure aims to identify patients whose melanoma has lymph node extension; in contrast, patients with positive already metastasized to regional lymph nodes.1 Thompson SLN are at high risk of subsequent lymph node metastasis et al.2 have demonstrated that in this situation it is rare for and usually undergo a completion lymph node dissection tumour cells to skip the sentinel lymph node (SLN) and (CLND).
2007 The Authors 58 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 Micromorphometric study of SLN in melanoma, S. Debarbieux et al. 59
Although the impact of the SLNB procedure on overall classified as cervical, supraclavicular, axillary, internal mam- survival is not known, it has been shown that positivity of the mary, inguinal and popliteal. SLN is a strong predictive factor of death related to melan- oma.3 For some authors, its predictive value is even stronger Sentinel node histological processing and evaluation than that of Breslow thickness of the primary tumour. It is likely that patients with positive SLN do not represent a After removal, sentinel nodes were fixed; the smallest ones homogeneous prognostic group. To date, no adjuvant therapy were cut into two slices whereas larger ones were cut into has definitively proven to be efficient for patients with lymph 2 to 3-mm slices before dehydration and paraffin embedding. node metastasis; therefore, it could be interesting to identify For each slice three roughly equidistant step-sections were homogeneous prognostic subgroups in order to include stained with haematoxylin–eosin–saffron (HES); in case of patients in clinical trials. negative or doubtful result, immunostaining with PS100 and In several studies, the rate of positive CLND among positive HMB45 or Mart-1 was performed. SLN patients varies from 7% to 30%; this could indicate that Histomorphometric features were assessed independently by in most cases, the objective lymphatic metastatic disease was two observers (S.D. and B.B.), a consensus statement being completely removed by the SLNB. Some authors have tried to obtained in case of different assessment. The following fea- identify factors predicting CLND status; few micromorphomet- tures were recorded: ric features of the SLN metastasis have been associated with 1 Type of metastasis, identified as: (i) exclusively peripheral a higher risk of positive CLND; however, so far, no study (in the subcapsular space or within medullar sinuses); has identified a subgroup of patients who could safely be (ii) mixed (peripheral and parenchymal); (iii) exclusively spared CLND. parenchymal; and (iv) subtotal (more than 80% of the lymph The present study was conducted to correlate patients’ clin- node). We were very stringent with the definition of periph- ical outcome and CLND status to micromorphometric features eral metastasis; in case of clusters of malignant cells sur- of positive SLN. rounded by lymphocytes, they were considered as partly parenchymal. Patients and methods 2 Size of the largest metastasis, evaluated by two diameters: diameter 1 is the maximum distance between two points of the metastasis (length of the metastasis); diameter 2 is the Inclusion criteria largest dimension measured perpendicularly to diameter 1 Between 1 January 1999 and 31 December 2004, patients (Fig. 1). with the following criteria were prospectively included in 3 Depth of metastasis, i.e. the distance between the capsule the SLN data bank of the Department of Dermatology, or the nearest medullary sinus and the deepest tumour cell in Hotel-Dieu Hospital, Lyons, France: (i) primary melanoma the nodal parenchyma. This deepest tumour cell did not of thickness > 1 mm, or < 1 mm with ulceration or evi- necessarily belong to the largest metastasis (Fig. 2). dence of regression on pathological examination or Clark 4 Number of step sections with metastatic invasion (sections level ‡ III; (ii) absence of in-transit metastasis (clinically or with isolated tumour cells were considered as positive). histologically) at > 2 cm from the primary lesion, absence 5 Maximum number of metastases on a single section: any of palpable lymph node or visceral metastasis detected by cluster of at least three cells was considered as a metastasis. initial clinical examination in the dermatology ward, nor- 6 Extracapsular invasion. mality of abdominal ultrasonography and chest X-ray; and (iii) no wide local excision of the primary lesion prior to inclusion. Patients who were not willing to undergo a CLND after positive SLN biopsy were excluded.
Sentinel lymph node biopsy
A lymphoscintigraphy with 99 m-Tc-labelled sulfur colloid was performed the day before SLNB to map the SLN(s). Intra- operative lymphatic mapping was performed with combined patent blue dye and the radioactive method as previously des- cribed.1 Blue and ⁄or hot nodes [(in vivo radioactivity ⁄back- ground) ‡ 3or(ex vivo radioactivity ⁄background) ‡ 10] were considered as sentinel nodes. Postexcision measurement of remnant radioactivity was used to confirm the complete exci- sion of all hot nodes. No preoperative pathological examina- Fig 1. Measurement of metastases size using two dimensions; tion was performed. The lymph node basins identified were haematoxylin, eosin and saffron staining, original magnification · 25.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 60 Micromorphometric study of SLN in melanoma, S. Debarbieux et al.
between CLND status and micromorphometric features of SLN was analysed by the v2 test; multivariate analysis was per- formed by logistic regression with Wald’s ascending stepwise method in order to determine the odds ratios (ORs) of inde- pendently relevant features. Survival curves were done by the Kaplan–Meier method and compared by the log-rank test. Multivariate Cox analysis of survival was performed by Wald’s ascending stepwise method.
Results
Between 1 January 1999 and 31 December 2004, 455 patients (mean age 51Æ7 years) underwent an SLNB [men: 240 (52Æ7%); women: 215 (47Æ3%)]; the mean Breslow thickness of the primary melanoma was 2Æ59 mm (range 0Æ5–20 mm). Fig 2. Depth of metastasis; haematoxylin, eosin and saffron staining, At least one SLN was identified in 446 patients (98Æ2%). The original magnification · 10. mean number of nodes identified was 1Æ5: a single node in 260 patients (58Æ3%), two nodes in 142 (31Æ8%), three nodes 7 Presence of tumour cells in lymphatic vessels within the in 33 (7Æ4%), and more than three nodes in 11 (2Æ5%). capsule: when several SLN were positive, the highest value The procedure identified 483 basins [axillary: 244 basins was registered for each criteria. (50Æ6%); groin: 173 basins (35Æ8%); cervical: 58 basins Nonsentinel nodes resected during the CLND in the case of (12%); popliteal: four basins (0Æ8%); supraclavicular: three positive SLNB were analysed in routine fashion, with one or basins (0Æ6%); internal mammary: one basin (0Æ2%)]. two sections stained with HES. Immunostainings were per- At least one SLN was positive in 99 basins of 98 patients. formed when HES sections were negative or doubtful. Among them, 88 had a single positive SLN, nine had two and one had three. Seventy-nine patients (80Æ6%) had mixed subcapsular and parenchymal metastasis; five patients Data concerning patient outcome (5Æ1%) had strictly peripheral involvement; eight (8Æ1%) Most patients were followed in the dermatology department had only parenchymal involvement and six (6Æ1%) had sub- and the data concerning their evolution were available in their total metastasis. medical records. For those who had interrupted their follow- For those 99 basins, completion of lymphadenectomy up in the ward, the dermatologist, oncologist or physician in found at least one positive non-SLN in 22 cases (22Æ2%). The charge of follow-up was interviewed systematically to get the precise status of CLND in one case could not be determined latest status. Cutaneous relapses or cutaneous metastasis were because of a preservation incident for one node among eight confirmed pathologically. In the case of subsequent onset of resected during surgical procedure. lymphatic metastatic disease and after confirmation of the The mean follow-up was 29 months. Only 10 patients were absence of visceral involvement, treatment systematically lost to follow-up. One patient was excluded from disease-free included whole basin lymph node dissection with pathological survival analysis because of coincidental evolutive urothelial examination of the surgical specimen. Surgical treatment or neoplasm. Ten patients were also withdrawn from disease- fine needle aspiration biopsy4 was systematically performed in specific survival calculation as the cause of their death was the case of apparently single visceral subsequent metastatic proven to be unrelated to melanoma. evolution. When several metastases were discovered together, they were attributed to melanoma either by cytological assess- Survival ment or by growth criteria on several close tomodensito- metries. Survival was calculated from the date of the initial Results are summarized in Tables 1 and 2. Survival curves complete excision of the primary tumour. of patients with negative and positive SLN are shown in Figure 3. For patients with positive SLN, log-rank comparison of sur- Statistical analysis vival curves showed that the presence of an ulceration of the Statistical evaluation was performed by an independent bio- primary melanoma (P =0Æ016 and 0Æ003), diameter 1 (cut- statistician (G.D.). The statistical unit was the patient for ana- off point: 2 mm; P =0Æ002 and < 0Æ001), depth of metastasis lysis of survival, and the basin (containing one or several (cut-off point: 1Æ5 mm; P =0Æ004 and 0Æ001) and extracapsu- positive SLNs) for analysis of CLND status. All data were com- lar spread (P <0Æ001) were significantly associated with puted in SPSS 13Æ0 statistical package (SPSS Inc., Chicago, IL, a shorter disease-free and disease-specific survival. Age U.S.A.). We relied on previous studies from the literature to (P =0Æ031), diameter 2 (cut-off point: 1 mm; P =0Æ002) determine cut-off points for continuous data. The relationship and lymphatic invasion within the capsule (P =0Æ016) were
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 Micromorphometric study of SLN in melanoma, S. Debarbieux et al. 61
Table 1 Comparison of survival with log-rank test (DFS and DSS)
Estimated mean Estimated mean
Variable DFS (months) 95% CI P(log-rank) DSS (months) 95% CI P(log-rank) Age 0Æ272 0Æ031 £ 60 (n = 71–72)a 38Æ531Æ9–45Æ152Æ3 47–57Æ5 >60(n = 26) 36Æ123Æ1–49 43Æ9 31–56Æ1 Sex 0Æ832 0Æ395 Male (n = 55–56)a 38Æ930Æ3–47Æ750Æ642Æ7–58Æ4 Female (n = 42) 40Æ830Æ8–50Æ857Æ249Æ1–65Æ4 Breslow thickness (1 data missing) 0Æ162 0Æ134 T1–T2 (n = 27) 42Æ433Æ3–51Æ553Æ745Æ7–61Æ6 T3–T4 (n = 69–70)a 38Æ330Æ1–46Æ550Æ743Æ2–58Æ2 Ulceration (2 data missing) 0Æ016 0Æ003 No (n = 48) 47Æ438Æ5–56Æ462Æ655Æ8–69Æ4 Yes (n = 47–48)a 33Æ524Æ5–42Æ545Æ336Æ9–53Æ6 Number of positive SLN 0Æ468 0Æ464 1(n = 87–88)a 38Æ631Æ8–45Æ353Æ247Æ2–59Æ1 >1(n = 10) 52Æ5 34–70Æ949Æ829Æ6–70 Diameter 1 0Æ002 < 0Æ001 £ 2mm(n = 60) 47Æ439Æ3–55Æ561Æ655Æ6–67Æ7 > 2 mm (n = 37–38)a 28 17Æ7–38Æ338Æ127Æ6–48Æ6 Diameter 2 0Æ08 0Æ002 £ 1mm(n = 61–62)a 44Æ335Æ9–52Æ759Æ953Æ2–66Æ5 > 1 mm (n = 36) 33Æ523Æ2–43Æ942Æ632Æ5–52Æ7 Depth of metastasis 0Æ004 0Æ001 £ 1Æ5mm(n = 63–64)a 46Æ738Æ5–54Æ960Æ253Æ7–66Æ8 >1Æ5 mm or subtotal metastases (n = 34) 27Æ117Æ1–37 40Æ129Æ8–50Æ4 Number of positive sections (1 data missing) 0Æ90Æ235 £ (n = 22) 33Æ424Æ5–42Æ348Æ63 40Æ16–57Æ1 >2(n = 74–75)a 41Æ533Æ8–49Æ252Æ42 45Æ53–59Æ31 Maximum number of metastases on 0Æ163 0Æ308 a single section (3 data missing) £ (n = 30) 27Æ319Æ4–35Æ139Æ831Æ3–48Æ2 > 2 and £ 5(n = 28) 34Æ626Æ1–43Æ14639Æ3–52Æ7 >5(n = 36–37)a 46Æ235Æ4–57 56Æ447Æ2–65Æ5 Extracapsular spread < 0Æ001 < 0Æ001 No (n = 79) 45Æ538Æ1–52Æ959Æ353Æ2–65Æ4 Yes (n = 18–19)a 18Æ57Æ7–29Æ433Æ921Æ6–46Æ2 Capsular lymphatic invasion (1 data missing) 0Æ106 0Æ016 No (n = 69–70)a 42Æ935Æ1–50Æ758Æ352Æ1–64Æ5 Yes (n = 27) 33 20Æ1–45Æ940Æ628Æ9–52Æ3 CLND status (1 data missing) 0Æ539 0Æ110 Negative (n = 75–76)a 42Æ6 35–50Æ256Æ549Æ7–63Æ2 Positive (n = 21) 36Æ723Æ8–49Æ444Æ632Æ8–56Æ3
aDFS = 97 patients; DSS = 98 patients; DFS, disease-free survival; DSS, disease-specific survival; CLND, completion lymph node dissection; SLN, sentinel lymph node; CI, confidence interval.
associated with a lower disease-specific survival but not with a Completion lymph node dissection status lower disease-free survival (Table 1). No significant correlation could be made between survival and the status of non-SLN. The results are summarized in Tables 3–5. By univariate analy- By multivariate analysis, only the presence of extracapsular sis, age (cut-off point: 60 years; P =0Æ001), diameter 2 (cut- invasion remained independently associated with a significantly off point: 1 mm; P =0Æ004) and depth of metastasis (cut-off shorter disease-free survival [OR = 3Æ808; 95% confidence point: 1Æ5 mm; P =0Æ009) were significantly associated with interval (CI) 1Æ936–7Æ489; P <0Æ001], whereas ulceration of a positive CLND. A trend towards statistical significance was the primary tumour and diameter 1 were independently asso- observed for the number of positive SLN (1 vs. ‡ 2; ciated with disease-specific survival (respectively: OR = 3Æ41; P =0Æ052) and the presence of an extracapsular spread 95% CI 1Æ322–8Æ791; P =0Æ011 and OR = 5Æ12; 95% CI (P =0Æ094). Histomorphometric features of the primary 2Æ119–12Æ39; P <0Æ001) (Fig. 4, Table 2). melanoma, such as Breslow thickness and the presence of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 62 Micromorphometric study of SLN in melanoma, S. Debarbieux et al.
Table 2 Multivariate analysis of survival by Disease-specific survival Cox model (DFS and DSS) Disease-free survival (DFS) (DSS)
Variable P OR 95% CI P OR 95% CI Age 0Æ698 0Æ203 £ 60 (n = 71) >60(n = 26) Sex 0Æ922 0Æ226 Male (n = 55) Female (n = 42) Breslow thickness 0Æ936 0Æ267 (1 data missing) T1–T2 (n = 27) T3–T4 (n = 69) Ulceration (2 data missing) 0Æ077 0Æ011 3Æ41 1Æ322–8Æ791 Yes (n = 47) No (n = 48) Number of positive SLN 0Æ587 0Æ186 1(n = 87) >1(n = 10) Diameter 1 0Æ633 < 0Æ001 5Æ12 2Æ119–12Æ39 £ 2mm(n = 60) > 2 mm (n = 37) Diameter 2 0Æ543 0Æ508 £ 1mm(n = 61) > 1 mm (n = 36) Depth of metastasis 0Æ531 0Æ714 £ 1Æ5mm(n = 63) >1Æ5 mm or subtotal metastasis (n = 34) Extracapsular spread < 0Æ001 3Æ808 1Æ936–7Æ489 0Æ348 Yes (n = 18) No (n = 79) Capsular lymphatic invasion 0Æ648 0Æ213 (1 data missing) Yes (n = 27) No (n = 69) CLND status (1 data missing) 0Æ670 0Æ558 Positive (n = 21) Negative (n = 75)
CLND, completion lymph node dissection; SLN, sentinel lymph node; OR, odds ratio; CI, confidence interval.
ulceration, were not associated with the CLND status; neither 40) for patients > 60 years old or diameter 2 > 1 mm, and was the maximum diameter of the largest metastasis. 54Æ5% (six of 11) for patients with both criteria. By multivariate analysis (including sex, age, Breslow thickness, ulceration, number of positive SLN, diameters 1 and 2, depth Discussion of metastasis, extracapsular spread and lymphatic invasion within the capsule), age (OR = 4Æ434; 95% CI 1Æ481–13Æ273; Studies conducted during the early 1990s to evaluate the P =0Æ008) and diameter 2 (OR = 4Æ453; 95% CI 1Æ504– therapeutic impact of elective lymph node dissection (ELND) 13Æ188; P =0Æ007) remained independently associated with a have failed to show a benefit for all melanoma patients;5 how- higher risk of positive CLND whereas depth of metastasis did ever, it is likely that subgroups of patients may benefit from not (Table 4). When diameter and depth only were associated the procedure. Although the setting is different from the SLNB in the logistic regression model, depth was not significantly procedure, Cascinelli et al.6 have shown that patients for whom independently prognostic either (Table 5). ELND identified occult metastasis have a longer survival than When patients were subclassified by age and diameter 2, patients who underwent a lymph node dissection for patent the percentages of positive CLND were 6Æ4% (three of 47) for metastasis. Balch et al.5 have shown that ELND worsens the patients £ 60 years old and diameter 2 £ 1 mm, 30% (12 of prognosis of elderly patients whereas it could confer a benefit
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 Micromorphometric study of SLN in melanoma, S. Debarbieux et al. 63
(a) (b) 1·0 1·0 Negative SLN
Negative SLN 0·8 0·8
0·6 0·6 Positive SLN
0·4 0·4 Positive SLN Disease-free survival Disease-specific survival 0·2 0·2
P log-rank < 0·001 0·0 0·0
0 20 40 60 80 0204060 80 Length of follow-up (months) Length of follow-up (months)
Fig 3. Disease-free (a) and disease-specific (b) survival curves of negative and positive sentinel lymph node patients.
1·0 1·0 1·0
0·8 0·8 0·8
0·6 0·6 0·6 No extracapsular invasion ≤ No ulceration Diameter 1 2 mm 0·4 0·4 0·4 Ulceration Disease-free survival 0·2 Disease-free survival 0·2 Diameter 1 > 2 mm 0·2 Disease-free survival Extracapsular invasion 0·0 P log-rank = 0·016 0·0 P log-rank = 0·002 0·0 P log-rank < 0·001 0 20 40 60 80 0204060 80 0204060 80 Length of follow-up (months) Length of follow-up (months) Length of follow-up (months)
1·0 1·0 1·0 Diameter 1 ≤ 2 mm 0·8 No ulceration 0·8 0·8 No extracapsular invasion 0·6 0·6 0·6 Ulceration 0·4 0·4 0·4 Diameter 1 > 2 mm
0·2 0·2 Extracapsular invasion
Disease-specific survival 0·2 Disease-specific survival Disease-specific survival 0·0 P log-rank = 0·003 0·0 P log-rank < 0·001 0·0 P log-rank < 0·001 0204060 80 0204060 80 Length of follow-up (months) Length of follow-up (months) 0204060 80 Length of follow-up (months)
Fig 4. Disease-free (a–c) and disease-specific (d–f) survival curves of patients according to ulceration of primary (a and d), maximum diameter of the largest metastasis (b and e) and extracapsular invasion (c and f). to patients under 60 years, especially with an intermediate node metastases do not usually bypass the SLN.2 SLNB is a thickness melanoma (1–2 mm). very powerful predictor of survival in melanoma. As a consequence, SLNB was developed as a ‘minimally However, most patients who undergo a CLND following a invasive procedure’, with a low complication rate,7 aimed at positive SLN do not bear additional intralymphatic metastatic identifying patients at risk for subsequent lymph node inva- disease, as such involvement is found in about 20% of cases.9 sion. The identification rate of the procedure is excellent, Within the last few years, studies have been conducted to between 98% and 99%,8 and it has been shown that lymph determine whether some patients could be classified as ‘very
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 64 Micromorphometric study of SLN in melanoma, S. Debarbieux et al.
Table 3 Univariate analysis of CLND status by v2 test (98 basins) Table 4 Multivariate analysis of CLND status with a logistic regression model (98 basins) Variable % of CLND+ P (v2) Age 0Æ001 Variable P OR 95% CI £ 60 (n = 71) 14Æ1% Age 0Æ008 4Æ434 1Æ481–13Æ273 >60(n = 27) 44Æ4% £ 60 (n = 71) Sex 0Æ750 >60(n = 27) Male (n = 55) 23Æ6% Sex 0Æ755 Female (n = 43) 20Æ9% Male (n = 55) Breslow thickness (1 data missing) 0Æ251 Female (n = 43) T1–T2 (n = 27) 14Æ8% Breslow thickness 0Æ866 T3–T4 (n = 70) 25Æ7% (1 data missing) Ulceration (2 data missing) 0Æ260 T1–T2 (n = 27) No (n = 47) 17% T3–T4 (n = 70) Yes (n = 49) 26Æ5% Ulceration (2 data missing) 0Æ778 Number of positive SLN 0Æ052 No (n = 47) 1(n = 87) 19Æ5% Yes (n = 49) >1(n = 11) 45Æ5% Number of positive SLN 0Æ328 Diameter 1 0Æ220 1(n = 87) £ 2mm(n = 60) 18Æ33% >1(n = 11) > 2 mm (n = 38) 28Æ9% Diameter 1 0Æ302 Diameter 2 0Æ004 £ 2mm(n = 60) £ 1mm(n = 61) 13Æ1% > 2 mm (n = 38) > 1 mm (n = 37) 37Æ8% Diameter 2 0Æ007 4Æ453 1Æ504–13Æ188 Depth of metastasis 0Æ009 £ 1mm(n = 61) £ 1Æ5mm(n = 63) 14Æ2% > 1 mm (n = 37) >1Æ5mm(n = 35) or 37Æ1% Depth of metastasis 0Æ685 subtotal metastasis £ 1Æ5mm(n = 63) Number of positive sections 0Æ567 >1Æ5mm(n = 35) (1 data missing) or subtotal metastasis £ 2(n = 22) 18Æ2% Extracapsular spread 0Æ821 >2(n = 75) 24% No (n = 79) Maximum number of metastases 0Æ518 Yes (n = 19) on a single section (3 data missing) Capsular lymphatic invasion 0Æ613 £ 2(n = 30) 16Æ7% (1 data missing) > 2 and £ 5(n = 28) 17Æ8% No (n = 69) >5(n = 37) 27% Yes (n = 28) Extracapsular spread 0Æ094 No (n = 79) 18Æ9% CLND, completion lymph node dissection; SLN, sentinel lymph Yes (n = 19) 36Æ8% node; OR, odds ratio; CI, confidence interval. Capsular lymphatic invasion 0Æ156 (1 data missing) No (n = 69) 18Æ8% Table 5 Multivariate analysis of CLND status with a logistic regression Yes (n = 28) 32Æ1% model including only diameter 2 and depth of metastasis (98 basins) CLND, completion lymph node dissection; SLN, sentinel lymph node. Variable P Odds ratio 95% CI Diameter 2 0Æ006 4Æ033 1Æ488–10Æ929 £ 1mm(n = 61) > 1 mm (n = 37) Depth of metastasis 0Æ557 low risk’ of further nodal metastasis according to the type of £ 1Æ5mm(n = 63) involvement of the SLN. Such patients could be spared the >1Æ5mm(n = 35) morbidity of a CLND. Several studies have aimed at classifying positive SLN CLND, completion lymph node dissection; CI, confidence interval. patients with regard to prognosis on the basis of micromorph- ometric features of the SLN combined with demographic and ⁄or pathological criteria of the primary tumour. cases, it was logical to try to predict the outcome according to In 1989, Cochran et al.10 showed that the number of the features of SLN invasion. Gershenwald et al.3 have demon- involved lymph nodes and the ratio (diameter of metastasis ⁄ strated clearly that SLN positivity has a significant impact on diameter of lymph node) was a good predictor of patient out- disease-free and overall survival. The result of our study is come. As the SLN is the only involved node in about 80% of concordant (see Fig. 3). With their S-classification including
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 Micromorphometric study of SLN in melanoma, S. Debarbieux et al. 65 the number of millimetric slices involved by metastasis and correlation between this diameter (using the cut-off point of the depth of tumour cells from the inner margin of the cap- 2 mm) and the involvement of non-SLN. Minimal involve- sule (or ‘centripetal depth’), Starz et al.11,12 underlined that ment of subcapsular sinuses of the SLN was not associated patients with a positive SLN could have different prognoses. with a lower risk of positive CLND either (21%). S3-classified (> two sections involved and centripetal depth Reeves et al.18 reported a series of 98 patients with positive > 1 mm) patients had a significantly higher rate of distant SLNs; they determined a size–ulceration score, which was the metastasis during follow-up and shorter overall survival than only significant predictive parameter of non-SLN involvement S1 (£ two sections involved and centripetal depth £ 1 mm), by multivariate analysis (OR = 2Æ9). No patient with a non- S2 (> two sections involved and centripetal depth £ 1 mm) ulcerated primary melanoma and SLN metastasis < 2 mm and negative SLN patients. (size–ulceration score = 0; 21 patients) had any additional Ranieri et al.13 drew the conclusion that the maximum positive node on CLND. In their study, the thickness of the diameter of the largest SLN tumour deposit was associated primary melanoma and the location of the metastasis in with outcome: the cut-off point of 3 mm appeared to be par- the lymph node (nonsubcapsular) were not predictive of the ticularly relevant with regard to 5-year disease-free and overall status of non-SLN. Our results do not confirm the impact of survival by multivariate analysis. the ulceration of the primary melanoma on the involvement For Cochran et al.,14 both Breslow thickness of the primary of non-SLN. melanoma and the relative area of the metastasis were predic- Salti and Das Gupta19 described sentinel node metastasis tive of the clinical outcome. of 56 patients and classified them as follows: positive on According to our study, although several parameters were immunohistochemistry only; one cluster < 2 mm; several associated with shorter disease-free and disease-specific survi- ‘small’ clusters; ‘large’ clusters; and extracapsular invasion. By vals, few remained independently significant by multivariate univariate analysis, this classification tended to be significant analysis: the presence of an extracapsular invasion for disease- (P =0Æ09). However, multivariate analysis did not find any free survival and both the maximum diameter of the largest significant predictive value. metastasis and the presence of an ulceration of the primary In the series reported by Cochran et al.,14 which included 90 melanoma for disease-specific survival. These results are con- patients, both the Breslow thickness and the relative area of cordant with the conclusions of Ranieri et al.13 and Cochran the metastasis (metastasis area ⁄lymph node area) were signifi- et al.14 that the tumour burden was predictive of survival. How- cantly associated with the result of CLND. Although analysed ever, Ranieri et al.13 found that ulceration had no impact on the as a discrete criteria in our study (Cochran et al.14 analysed it prognosis. Our study is the first to point out such a correlation. as a continuous variable), the Breslow thickness was not found Our series also confirms, both by univariate and multivari- to be predictive of CLND status. ate analyses, that in the case of positive SLN, micromorpho- In a series of 146 positive SLN patients, Dewar et al.20 found metric features are more accurate in predicting the outcome that non-SLN status was related to the location and depth of than Breslow thickness of the primary tumour. metastasis but not to the Breslow thickness of the primary On the whole, the outcome has been less thoroughly melanoma. Subcapsular location was associated with fewer studied than the risk of invasion of non-SLN, probably positive CLND. Multivariate analysis was not performed. In because it requires a longer observation phase; however, our experience, differentiation of strictly subcapsular location there is a consensus on the predictive value of the tumour from early parenchymal invasion is often difficult and prone burden, although variously measured in previously pub- to be differently assessed by distinct observers. This may lished works. explain the difference between our data and those of Dewar Regarding the involvement of non-SLN, the first study, et al.20 (5% vs. 26% of subcapsular metastasis). conducted by Wagner et al.,15 did not find any correlation by Lee et al.9 concluded from 191 positive SLN patients that the univariate analysis between Breslow thickness of the primary cut-off points of 3 mm for Breslow thickness of the primary melanoma, number of positive SLN, number of metastatic melanoma and 2 mm for the size of SLN metastasis were rele- deposits within the SLN and the result of the CLND. They vant with regard to the status of non-SLNs. However, even in found no correlation by univariate analysis; however, the the lower risk group, 12Æ3% of CLND were positive. number of patients was quite low. The latest study was conducted by Scolyer et al.,21 in According to Starz et al.,11,12,16 the risk of non-SLN involve- which the SLN metastases from 140 patients were analysed ment was significantly correlated with S-classification (OR by univariate analysis. Penetrative depth of metastasis (cut-off 3Æ31; 95% CI 1Æ325–8Æ267). Among 65 positive SLN patients, point: 2 mm), deposit area (cut-off 10 mm2), effacement only one of their 12 patients belonging to the S1 group of nodal architecture by tumour and perinodal lymphatic (£ two involved sections and centripetal depth £ 1 mm) had involvement were significantly associated with positive an invasion of non-SLN. They suggested that CLND could be CLND, whereas the number of positive SLN, the number of avoided for such patients. In our study, the number of sec- deposits, the deposit diameter, the location of metastasis tions involved was not associated with the status of CLND. (subcapsular ⁄intraparenchymal and spread of cells within the Carlson et al.17 measured the diameter of the largest metasta- node) and extracapsular invasion were not. No multivariate sis within the sentinel node of 104 patients. They found no analysis was performed.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 66 Micromorphometric study of SLN in melanoma, S. Debarbieux et al.
On the whole, the results from these different studies are ment of non-SLN. Unlike descriptions in previous studies, strikingly heterogeneous. The first reason is probably that cri- the maximum diameter of the largest metastasis was not teria have been evaluated differently. The second reason could predictive; however, the second diameter of the same meta- be a poor reproducibility of the evaluation of some of the his- stasis is strongly indicative of positive CLND. Actually, most tomorphometric criteria; Scolyer et al.21 described some diffi- metastases are not spheric; they often have a maximum culties that we have encountered ourselves. The size of the diameter (diameter 1) which is roughly parallel to the cap- largest deposit was often determined; however, clusters are sule, whereas their width reflects the centripetal invasion sometimes close to each other, leading to the question of into the parenchyma. However, this parameter should not whether one should measure the whole size or consider two be confused with the depth of the deepest cluster of meta- separate clusters. static cells, which is not significant by multivariate analysis Another point which is not specified in the literature is in our study. Scolyer et al.21 underlined the difference; how one should consider clusters localized along medullary according to them, the term ‘centripetal thickness’ used by sinuses; we considered that clusters in medullary sinuses were Starz et al.11,12 implies a continuous thickness from the peripheral and in such cases we measured the depth of meta- inner margin of the capsule whereas the term ‘penetrative stasis from the sinuses (Fig. 5). depth’ does not. The second diameter of the largest Moreover, one must be aware of the technical pitfalls; the ini- metastasis could be in our view more relevant than tial hypothesis of most studies is that very small and ⁄or per- measurement of either ‘centripetal’ or ‘penetrative depth’ ipheral metastasis could be less susceptible to being associated (Fig. 6). We also believe that this measure is more repro- with non-SLN involvement. However, these metastases are also ducible because of variations of the ‘centripetal’ or ‘penetrative more susceptible to being missed by sections, as suggested depth’ that could be induced by different sectioning of the by molecular biology studies with reverse transcriptase- specimen. polymerase chain reaction.22 It is obvious that the efficiency Our study did not confirm the predictive value of Breslow in detecting micrometastasis depends on the sampling method thickness, ulceration of the primary tumour, number of posit- and that no sampling method will be able to detect all micro- ive sections and maximum diameter of the largest metastasis metastases. A consensus was established a few years ago to for CLND status. standardize the different steps of the SLNB procedure and par- In conclusion, our study confirms the negative impact on ticularly of lymph node tissue sampling and the use of immu- survival of positivity of SLN in melanoma patients. nohistochemistry techniques;1 however, several studies were On the basis of previously published works and of the pre- conducted before this consensus statement was produced and sent study, there is some relationship between the depth of the sampling methods were quite heterogeneous, which partly invasion of the lymph node by melanoma cells and the posi- explains the heterogeneous findings. tivity of CLND. However, the methods used to evaluate this According to our study, the two parameters that were invasion were extremely heterogeneous and conflicting data associated with a higher probability of non-SLN invasion by have been published. On the basis of our experience and of multivariate analysis are age (OR = 4Æ434; 95% CI = 1Æ481– difficulties of assessment reproducibility, we propose to 13Æ273; P =0Æ008) and diameter 2 (longest distance per- simplify the micromorphometric evaluation by measuring the pendicular to the maximum diameter) of the metastasis second diameter (shortest diameter) of the largest metastasis (OR = 4Æ453; 95% CI 1Æ504–13Æ188; P =0Æ007). Our study observed in serial sections. Our study demonstrates a strong is the first to find a correlation between age and involve-
Medullary sinus
Depth
Fig 5. Measurement of metastasis depth in a case of sinusal location; Fig 6. Penetrative depth according to Scolyer et al.21; staining: Mart-1, staining: Mart-1, original magnification · 10. original magnification · 10.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 Micromorphometric study of SLN in melanoma, S. Debarbieux et al. 67 predictive value of this new criterion for positive CLND 9 Lee JH, Essner R, Torisu-Itakura H et al. Factors predictive of (OR = 4Æ453). tumor-positive nonsentinel lymph nodes after tumor-positive senti- Our study confirms the predictive value for poor outcome nel lymph node dissection for melanoma. J Clin Oncol 2004; 22:3677–84. of the maximum diameter of the largest metastasis. For the 10 Cochran AJ, Lana AM, Wen DR. Histomorphometry in the assess- first time, we have also demonstrated that in positive SLN ment of prognosis in stage II malignant melanoma. Am J Surg Pathol patients, ulceration of the primary lesion was strongly correlated 1989; 13:600–4. to an increased risk of death. Further multicentre investiga- 11 Starz H, Balda BR, Kramer KU et al. A micromorphometry-based tions are needed to confirm the reliability of our newly intro- concept for routine classification of sentinel lymph node metastases duced criteria. and its clinical relevance for patients with melanoma. Cancer 2001; 91:2110–21. 12 Starz H, Siedlecki K, Balda BR. Sentinel lymphonodectomy and Acknowledgments S-classification: a successful strategy for better prediction and improvement of outcome of melanoma. Ann Surg Oncol 2004; This work has been supported in part by Hospices Civils de 11(Suppl. 3):162S–8S. Lyon grant (to L.T.), Ligue de´partementale du Rhoˆne contre 13 Ranieri JM, Wagner JD, Azuaje R et al. Prognostic importance of le cancer grant (to L.T.) APiCiL foundation grant (to L.T.), lymph node tumor burden in melanoma patients staged by sentinel and University Claude Bernard E.A 37-32 grant (to L.T.). node biopsy. Ann Surg Oncol 2002; 9:975–81. 14 Cochran AJ, Wen DR, Huang RR et al. Prediction of metastatic mel- anoma in nonsentinel nodes and clinical outcome based on the References primary melanoma and the sentinel node. Mod Pathol 2004; 17:747–55. 1 Cochran AJ, Balda BR, Starz H et al. The Augsburg Consensus. 15 Wagner JD, Davidson D, Coleman JJ III et al. Lymph node tumor Techniques of lymphatic mapping, sentinel lymphadenectomy, and volumes in patients undergoing sentinel lymph node biopsy for completion lymphadenectomy in cutaneous malignancies. Cancer cutaneous melanoma. Ann Surg Oncol 1999; 6:398–404. 2000; 89:236–41. 16 Starz H, De Donno A, Balda BR. The Augsburg experience: 2 Thompson JF, McCarthy WH, Bosch CM et al. Sentinel lymph node histological aspects and patient outcomes. Ann Surg Oncol 2001; status as an indicator of the presence of metastatic melanoma in 8(Suppl. 9):48S–51S. regional lymph nodes. Melanoma Res 1995; 5:255–60. 17 Carlson GW, Murray DR, Lyles RH et al. The amount of meta- 3 Gershenwald JE, Thompson W, Mansfield PF et al. Multi-institu- static melanoma in a sentinel lymph node: does it have prognostic tional melanoma lymphatic mapping experience: the prognostic significance? Ann Surg Oncol 2003; 10:575–81. value of sentinel lymph node status in 612 stage I or II melanoma 18 Reeves ME, Delgado R, Busam KJ et al. Prediction of nonsentinel patients. J Clin Oncol 1999; 17:976–83. lymph node status in melanoma. Ann Surg Oncol 2003; 10:27– 4 Dalle S, Paulin C, Lapras V et al. Fine-needle aspiration with ultra- 31. sound guidance in patients with malignant melanoma and palpable 19 Salti GI, Das Gupta TK. Predicting residual lymph node basin dis- lymph nodes. Br J Dermatol 2006; 155:552–6. ease in melanoma patients with sentinel lymph node metastases. 5 Balch CM, Soong SJ, Bartolucci AA et al. Efficacy of an elective Am J Surg 2003; 186:98–101. regional lymph node dissection of 1 to 4 mm thick melanomas for 20 Dewar DJ, Newell B, Green MA et al. The microanatomic location patients 60 years of age and younger. Ann Surg 1996; 224:255–63; of metastatic melanoma in sentinel lymph nodes predicts non- discussion 263–6. sentinel lymph node involvement. J Clin Oncol 2004; 22:3345– 6 Cascinelli N, Morabito A, Santinami M et al. Immediate or delayed 9. dissection of regional nodes in patients with melanoma of the 21 Scolyer RA, Li LX, McCarthy SW et al. Micromorphometric features trunk: a randomised trial. WHO Melanoma Programme. Lancet of positive sentinel lymph nodes predict involvement of nonsenti- 1998; 351:793–6. nel nodes in patients with melanoma. Am J Clin Pathol 2004; 7 Wrightson WR, Wong SL, Edwards MJ et al. Complications associ- 122:532–9. ated with sentinel lymph node biopsy for melanoma. Ann Surg Oncol 22 Ulrich J, Bonnekoh B, Bockelmann R et al. Prognostic significance 2003; 10:676–80. of detecting micrometastases by tyrosinase RT ⁄PCR in sentinel 8 Morton DL, Wen DR, Wong JH et al. Technical details of intraoper- lymph node biopsies: lessons from 322 consecutive melanoma ative lymphatic mapping for early stage melanoma. Arch Surg 1992; patients. Eur J Cancer 2004; 40:2812–19. 127:392–9.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp58–67 CLINICAL AND LABORATORY INVESTIGATIONS DOI 10.1111/j.1365-2133.2007.07986.x Prevalence of metabolic syndrome in patients with psoriasis: a hospital-based case–control study P. Gisondi, G. Tessari, A. Conti,* S. Piaserico, S. Schianchi,* A. Peserico, A. Giannetti* and G. Girolomoni Section of Dermatology, Department of Biomedical and Surgical Science, Section of Dermatology and Venereology, University of Verona, Piazzale A. Stefani 1, I-37126 Verona, Italy *Department of Dermatology, University of Modena and Reggio Emilia, Modena, Italy Department of Dermatology, University of Padua, Padua, Italy
Summary
Correspondence Background Psoriasis is a chronic inflammatory disease associated with an increased Paolo Gisondi. cardiovascular risk. Metabolic syndrome is a significant predictor of cardiovascu- E-mail: [email protected] lar events. Objective To investigate the prevalence of metabolic syndrome in patients with Accepted for publication 28 February 2007 psoriasis. Methods We performed a hospital-based case–control study on 338 adult patients Key words with chronic plaque psoriasis and 334 patients with skin diseases other than cardiovascular risk, chronic plaque psoriasis, psoriasis. metabolic syndrome Results Metabolic syndrome was significantly more common in psoriatic patients Æ Æ Æ Conflicts of interest than in controls (30 1% vs. 20 6%, odds ratio 1 65, 95% confidence interval None declared. 1Æ16–2Æ35; P=0Æ005) after the age of 40 years. Psoriatic patients also had a higher prevalence of hypertriglyceridaemia and abdominal obesity, whereas hyperglycaemia, arterial hypertension and high-density lipoprotein cholesterol plasma levels were similar. Although psoriasis patients were more frequently smokers, the association of psoriasis with metabolic syndrome was independent from smoking. There was no correlation between severity of psoriasis and preva- lence of metabolic syndrome. Psoriatic patients with metabolic syndrome were older and had a longer disease duration compared with psoriatic patients without metabolic syndrome. Conclusion Psoriatic patients have a higher prevalence of metabolic syndrome, which can favour cardiovascular events. We suggest psoriatic patients should be encouraged to correct aggressively their modifiable cardiovascular risk factors.
Psoriasis is a chronic inflammatory skin disease that affects factors such as hyperlipidaemia, hypertension and hyper- about 3% of the population.1,2 Psoriasis can be associated homocysteinaemia.13 with other diseases, which may have a major impact on Metabolic syndrome is a cluster of risk factors including patients. The more common comorbidities include psoriatic central obesity, atherogenic dyslipidaemia, hypertension and arthritis and anxiety ⁄depression disorder.3–5 More recently, glucose intolerance, and is a strong predictor of cardiovascular psoriasis has also been reported to be associated with diseases, diabetes and stroke.14–16 The importance of meta- metabolic disorders including obesity, dyslipidaemia and bolic syndrome is that it may confer a cardiovascular risk diabetes.6–8 Moreover, an increased mortality from cardio- higher than the individual components. Men with metabolic vascular disease in patients with severe psoriasis has been syndrome are almost three times more likely to die of coron- documented, and psoriasis may confer an independent risk ary artery disease after adjustment for conventional cardiovas- of myocardial infarction especially in young patients.9–11 cular risk factors.17 The aim of this study was to investigate Major factors that may contribute to this unfavourable car- the prevalence of metabolic syndrome in patients with psoria- diovascular risk profile include cigarette smoking, obesity, sis. The results indicate that psoriatic patients have a higher physical inactivity, hyperhomocysteinaemia and psychological prevalence of metabolic syndrome. The association between stress, which have a higher prevalence among patients with psoriasis and metabolic syndrome is also true for mild severity psoriasis.12 In addition, many traditional systemic thera- psoriasis and it is independent from the tendency of psoriatic pies for psoriasis may also worsen cardiovascular risk patients to be obese.
2007 The Authors 68 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 Psoriasis and metabolic syndrome, P. Gisondi et al. 69
Materials and methods computed. Associations between the presence of psoriasis and various covariates were tested by using the Fisher’s exact test for categorical variables and t-test for continuous variables. Study population The proportion of patients with metabolic syndrome in cases This was a hospital-based case–control study involving a series was compared with that in controls stratifying data by sex and of 338 psoriatic patients (cases) and 334 controls consecutive- age according to the Mantel–Haenszel v2 test. Linear correl- ly admitted to the outpatient clinics of three university hospi- ation between covariates was analysed according to the Spear- tals in north Italy. Inclusion criteria for cases were age more man test. To account simultaneously for the effect of age, sex than 18 years and clinical diagnosis of chronic plaque psoriasis and smoking habit unconditional multiple logistic regression (diagnosis of psoriasis lasting at least 6 months). Patients was performed. All P-values are two-sided and P <0Æ05 was receiving any systemic treatment for psoriasis including acitre- considered statistically significant. Sample size of the study tin, ciclosporin, methotrexate, phototherapy or biologics for at population was determined as follows: prevalence of metabolic least 1 month before enrolment were not included in the syndrome in the control group was expected to be about study. Psoriatic arthritis (PsA) was diagnosed according to 20%,22 and we posited a prevalence difference of about 5% in standard criteria.18 Controls were enrolled among patients patients with psoriasis. For a =0Æ05 and 1 ) b =0Æ90, it was referred for dermatological conditions other than psoriasis. necessary to enrol at least 302 patients per group in order to The source population for cases and controls was the same. guarantee the above-significance level and power. After signed informed consent, all subjects were visited by a dermatologist who registered demographic, biometric and the Results other relevant data on a case report form. Relevant data collec- ted included age, gender, weight, height, body mass index The study included 338 cases and 334 controls. Descriptive (BMI), waist circumference, blood pressure, smoking habit, characteristics of the study population are reported in Table 1. age of psoriasis onset, type and severity of psoriasis, presence There was no significant interhospital variability within the and distribution of psoriatic arthropathy and concomitant dataset collected. Moreover, the three hospitals contributed to medications. BMI was calculated as weight (kg) ⁄height (cm2). the study with similar numbers of cases and controls. Patients To determine waist circumference, we located the upper hip had mild to severe psoriasis with a PASI score ranging from bone and placed a measuring tape at the level of the upper- 1Æ3to60Æ2 with a median 7Æ9 [95% confidence interval (CI) most part of the hipbone around the abdomen (ensuring that 7Æ4–9Æ6]; 186 (57Æ3%) patients had a PASI score < 10, the tape measure was horizontal). The tape measure was snug whereas 152 (42Æ7%) had a PASI score ‡ 10. BSA affected ran- but did not cause compressions on the skin. Blood pressure ged from 1% to 95%, with a median of 10Æ6 (95% CI 14Æ9– was recorded as the average of two measurements after sub- 19Æ1); 152 (45Æ3%) patients had a BSA < 10%, whereas 183 jects had been sitting for 5 min. Severity of psoriasis was (54Æ7%) had a BSA ‡ 10%. Static PGA mean value was 3 assessed according to Psoriasis Area and Severity Index (95% CI 2Æ8–3Æ0). Chronic plaque psoriasis was the more (PASI),19 body surface area (BSA) measurement and static common clinical type accounting for 96Æ3% of cases. PsA was Physician’s Global Assessment (PGA).20 Chronic plaque psoria- present in 71 patients (21%). Prevalence of dermatological sis was considered localized or disseminated when it covered diagnoses in the control group were as follows: 38Æ2% had less or more than 10% of the BSA. Metabolic syndrome was melanoma and nonmelanoma skin cancers (mainly actinic ker- diagnosed in the presence of three or more criteria of the atosis and basal cell carcinoma), 24Æ9% eczema, 14Æ3% viti- National Cholesterol Education Program’s Adult Panel III (ATP ligo, 12Æ9% infective skin diseases (mainly viral warts), 9Æ7% III): waist circumference > 102 cm in men or > 88 cm in autoimmune bullous diseases (mainly bullous pemphigoid). ) women; hypertriglyceridaemia > 1Æ7 mmol L 1; high-density Psoriatic patients had a higher mean BMI and were more fre- ) lipoprotein (HDL) cholesterol < 1Æ0 mmol L 1 in men or quently smokers compared with controls. We found a higher ) <1Æ3 mmol L 1 in women; blood pressure > 135 ⁄85 mmHg; prevalence of metabolic syndrome in cases than in controls ) fasting plasma glucose > 6Æ1 mmol L 1.21 Venous samples [30Æ1% vs. 20Æ6%, odds ratio (OR) 1Æ65, 95% CI 1Æ16–2Æ35; were taken at the enrolment visit after the subjects had fasted P =0Æ005] after controlling for sex and age. Multiple logistic overnight (at least 8 h). Serum cholesterol and triglycerides regression analysis revealed that metabolic syndrome was asso- were measured with enzymatic procedures. Plasma glucose ciated with psoriasis independently of age and smoking habit. was measured using a glucose oxidase method. The higher prevalence of metabolic syndrome in psoriatic patients was confirmed in all age classes of 40 years and older (Fig. 1). Metabolic syndrome was present in 35Æ2% of patients Statistical analysis with PsA (P =0Æ003). Prevalence of metabolic syndrome was Analysis was carried out using the STATA (version 6.0 Stata- not correlated to severity of psoriasis. In particular, there was Corp LP, College Station, TX, U.S.A.) and Graphpad (version no difference in the prevalence of metabolic syndrome in 4.0 GraphPad Software, El Camino Real, San Diego, CA, patients with a PASI score lower or higher than 10 (30Æ1% vs. U.S.A.) software packages. Standard descriptive statistics such 29Æ4%, respectively; P =0Æ9), or in patients with BSA involve- as mean, standard deviation and prevalence proportion were ment lower or greater than 10% (32Æ2% vs. 28Æ4%, respectively;
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 70 Psoriasis and metabolic syndrome, P. Gisondi et al.
Table 1 Study population. Descriptive Cases Controls characteristic of cases and controls (n=338) (n = 334) P-value Sex M ⁄F 160 ⁄178 150 ⁄184 0Æ8 Age at enrolment (years), mean ± SD 62Æ1±15Æ163Æ8±20Æ40Æ7 Body mass index, mean ± SD 27Æ7±4Æ825Æ4±4Æ90Æ0001 Smoker, n (%) 121 (36Æ2) 72 (21Æ0) 0Æ0001 Metabolic syndrome, n (%) 102 (30Æ1) 69 (20Æ6) 0Æ005 Waist circumference > 102 cm (M) or 193 (57Æ1) 159 (47Æ6) 0Æ01 > 88 cm (F), n (%) ) Triglyceridaemia > 1Æ7 mmol L 1, n (%) 128 (37Æ8) 78 (23Æ3) 0Æ001 ) HDL cholesterol < 1Æ0 mmol L 1 (M) or 61 (18Æ0) 72 (21Æ5) 0Æ2 ) <1Æ3 mmol L 1 (F), n (%) Blood pressure > 135 ⁄85 mmHg, n (%) 138 (40Æ8) 132 (39Æ5) 0Æ7 ) Fasting plasma glucose > 6Æ1 mmol L 1, 65 (19Æ2) 70 (20Æ9) 0Æ6 n (%)
HDL, high-density lipoprotein.
syndrome had a higher mean age, an earlier age of onset of 35 psoriasis and a longer disease duration. There were no differ- Controls 30 ences regarding gender, clinical type of psoriasis, psoriasis Psoriasis severity and prevalence of smokers. 25
20 Discussion 15 Psoriasis is a chronic inflammatory skin disease, which confers 10 an unfavourable cardiovascular risk profile. A higher mortality risk for arterial and venous thrombosis,9 and a higher risk of
Metabolic syndrome (%) 5 myocardial infarction especially in young patients with severe 0 psoriasis10 have been reported. Major factors that may con- 18–30 31–40 41–50 51–60 61–70 >71 tribute to this increased cardiovascular risk include cigarette Classes of age smoking, dyslipidaemia, obesity, physical inactivity, hyper- homocysteinaemia and psychological stress, all of which have Fig 1. Prevalence of metabolic syndrome in different age classes. a higher occurrence among patients with psoriasis.11 A direct correlation between severity of psoriasis and the prevalence of P =0Æ4). There were no differences in the prevalence of obesity, dyslipidaemia and hyperhomocysteinaemia has been metabolic syndrome between men and women, but it was reported in psoriatic patients,12,23,24 suggesting that skin more frequent after the age of 40 in both cases and controls. changes (inflammation) caused by psoriasis have a direct role Waist circumference > 102 cm in men or > 88 cm in in determining these risk factors. Also PsA has been found to ) women, and triglyceridaemia > 1Æ7 mmol L 1 were signifi- be associated with relevant cardiovascular risk factors.25 Other cantly more prevalent in cases than in controls. In contrast, T-helper 1-mediated autoimmune diseases, such as rheuma- there were no significant differences regarding the prevalence toid arthritis and systemic lupus erythematosus are character- of low HDL cholesterol, hypertension and fasting plasma ized by accelerated atherosclerosis and consequently higher glucose between cases and controls. We found a mild but cardiovascular morbidity and mortality rates.26,27 The con- significant correlation between severity of psoriasis and comitant occurrence of dyslipidaemia, glucose intolerance, triglyceridaemia in patients with psoriasis (r =0Æ16; insulin resistance, obesity and hypertension constitutes the P =0Æ04) and between plasma fasting glucose and triglycerid- metabolic syndrome, which has been similarly defined by the aemia in subjects with metabolic syndrome (r =0Æ23; World Health Organization, the National Cholesterol Educa- P =0Æ0006). In contrast, we found no significant correlation tion Program’s ATP III and the European Group on Insulin between psoriasis severity with waist circumference, blood Resistance.21,28,29 Each panel group agrees on the essential pressure, fasting plasma glucose and body weight. There was components, but they differ in the details and criteria. Meta- a direct correlation between PASI score and disease duration bolic syndrome is a strong predictor of cardiovascular diseases, (r =0Æ21; P =0Æ02). diabetes and stroke and significantly increases the risk of Comparing psoriatic patients with and without metabolic cardiovascular mortality compared with the individual syndrome (Table 2), we observed that patients with metabolic factors.16,30 Metabolic syndrome also increases the risk of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 Psoriasis and metabolic syndrome, P. Gisondi et al. 71
Table 2 Descriptive characteristics of psoriatic patients with and without Psoriatic patients Psoriatic patients metabolic syndrome with metabolic without metabolic syndrome (n = 102) syndrome (n = 236) P-value Sex M ⁄F65⁄37 161 ⁄75 0Æ4 Age at enrolment (years), 63Æ1±10Æ951Æ7±15Æ30Æ0001 mean ± SD Age at onset of psoriasis (years), 49Æ4±18Æ836Æ6±15Æ50Æ0001 mean ± SD Duration of psoriasis (years), 18Æ1±16Æ113Æ3±12Æ00Æ03 mean ± SD Smoker, n (%) 32 (31) 89 (37) 0Æ2 Type of psoriasis, n (%) Chronic plaque localized 46 (45) 104 (44) 0Æ2 (BSA < 10%) Chronic plaque disseminated 53 (51) 123 (52) (BSA > 10%) Pustular 3 (4) 3 (1) Other (inverse, erythrodermic) 0 6 (2) PASI, mean ± SD 11Æ1±10Æ411Æ2±9Æ40Æ9 PASI > 10, n (%) 41 (42Æ6) 98 (42Æ8) 0Æ9 BSA, mean ± SD 17Æ3±18Æ018Æ5±20Æ80Æ6
BSA, body surface area; PASI, Psoriasis Area and Severity Index.
all-cause and colon cancer mortality.31,32 Comparison of pub- secretion or sensitivity compared with control patients.38 Total lished prevalence for different populations is difficult because plasma triglycerides directly correlated to psoriasis severity of the different diagnostic criteria used. Just as the prevalence and, as expected, to plasma glucose levels. An atherogenic lipid of the individual components of the syndrome varies among profile at the onset of psoriasis has been observed in a well- populations, so does the prevalence of the metabolic syn- designed study, which excluded the possible role of confound- drome itself. Differences in genetic background, diet, levels of ing factors such as obesity, hypertension, cigarette smoking physical activity, population age and sex, levels of over- and and physical activity.39 We did not find differences between undernutrition, and body habits all influence its prevalence. cases and controls regarding prevalence of the other cardiovas- According to the ATP III definition, almost 25% of U.S. adults cular risk factors including diabetes, hypertension and reduc- have metabolic syndrome.33 In Western Europe the prevalence tion of HDL cholesterol. In contrast to our findings, a recent of metabolic syndrome is similar to the U.S.A. ranging from hospital-based case–control study from Northern Europe 15% to 35%;34 in developing countries the prevalence of showed that diabetes mellitus type 2 and arterial hypertension metabolic syndrome is lower, but recent epidemiological stud- as well as coronary artery disease occurred significantly more ies are registering a rapid increase.35 A very consistent finding frequently in patients with psoriasis than in controls.40 The among different studies is that prevalence of metabolic syn- same authors observed a very low prevalence of metabolic syn- drome is strictly age dependent, increasing sharply after the drome in both patients (4Æ3%) and controls (1Æ1%), and there age of 60.36 was a consistent trend for an enhanced risk of metabolic syn- We have found that the prevalence of metabolic syndrome is drome when comparing patients with severe psoriasis to those significantly higher in psoriatic patients compared with con- with moderate psoriasis. trols after the age of 40 years, and it directly correlates to psor- In our study, we found that psoriasis is associated with iasis duration. The prevalence of metabolic syndrome in our metabolic syndrome independently of its severity, estimated control population was similar to that recently estimated with both PASI and BSA scoring and smoking habit. The study among Italian adults.22 As expected, we have found that psori- was cross-sectional and therefore the directionality of the asso- atic patients have a BMI and a waist circumference significantly ciation between psoriasis and metabolic syndrome could not higher compared with controls patients. Overwhelming evi- be determined. We cannot determine at the moment what dence points towards the primary role of insulin resistance in comes first, but the absence of correlation between psoriasis pathogenesis of metabolic syndrome. More recently, chronic severity and metabolic syndrome may suggest that it is obesity inflammation has been suggested as a part of the insulin resist- that favours psoriasis. Some evidence indicates that psoriasis ance syndrome. Therapeutic intervention based on methotrexate comes first. It is possible that depression, eating habits, phys- and tumour necrosis factor (TNF)-a antagonists seems to ical inactivity, alcohol consumption, stress and inflammatory diminish the insulin resistance state.37 However, it has been mediators associated with psoriasis favour obesity in pre- reported that patients with psoriasis have no deficit in insulin disposed individuals. On the other hand, obesity may in turn
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 72 Psoriasis and metabolic syndrome, P. Gisondi et al. favour psoriasis. Indeed obesity is a proinflammatory state and 12 Malerba M, Gisondi P, Radaeli A et al. Plasma homocysteine and the adipose tissue is a rich source of inflammatory mediators folate levels in patients with chronic plaque psoriasis. Br J Dermatol know as adipocytokines. These include adiponectin, leptin, 2006; 155:1165–9. 13 Naldi L, Griffiths CE. Traditional therapies in the management of resistin and visfatin, which are thought to provide an import- moderate to severe chronic plaque psoriasis: an assessment of the ant link between obesity, insulin resistance and related inflam- benefits and risks. Br J Dermatol 2005; 152:597–615. matory disorders. Other products of adipose tissue that have 14 Wilson PW, D’Agostino RB, Parise H et al. Metabolic syndrome as been characterized include TNF-a, interleukin 6 and monocyte a precursor of cardiovascular disease and type 2 diabetes mellitus. chemoattractant protein 1. These products have well-known Circulation 2005; 112:3066–72. roles in the pathogenesis of psoriasis and at the interface 15 Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet between the immune and metabolic systems. In line with this 2005; 365:1415–28. 16 Wannamethee SG, Shaper AG, Lennon L, Morris RW. Metabolic hypothesis is the evidence that obesity is a risk factor for other 41,42 syndrome vs. Framingham Risk Score for prediction of coronary inflammatory diseases such as atherosclerosis and asthma. heart disease, stroke, and type 2 diabetes mellitus. Arch Intern Med Our study presents several limitations. Firstly, it was a 2005; 165:2644–50. cross-sectional study that does not allow the directionality of 17 Lakka HM, Laaksonen DE, Lakka TA et al. The metabolic syndrome the association to be ascertained. Secondly, the study was con- and total and cardiovascular disease mortality in middle-aged men. ducted in tertiary care centres, and thus patients are biased JAMA 2002; 288:2709–16. toward having more severe psoriasis; and, thirdly, the study 18 Taylor W, Gladman D, Helliwell P et al.; CASPAR Study Group. Classification criteria for psoriatic arthritis: development of new was conduced in northeast Italy and the population analysed criteria from a large international study. Arthritis Rheum 2006; may not be representative of the entire country. 54:2665–73. In conclusion, we have found a higher prevalence of meta- 19 Fredriksson T, Pettersson U. Severe psoriasis – oral therapy with a bolic syndrome in patients with psoriasis, which could play a new retinoid. Dermatologica 1978; 157:238–44. relevant role in accelerating atherosclerosis. The association is 20 Langley RG, Ellis CN. Evaluating psoriasis with Psoriasis Area and not limited to severe cases but also occurs with mild cases. Severity Index, Psoriasis Global Assessment, and Lattice System The hypothesis that obesity can favour psoriasis needs to be Physician’s Global Assessment. J Am Acad Dermatol 2004; 51:563–9. 21 National Institute of Health. Third report of the National Cholesterol Educa- urgently addressed in prospective studies. We suggest that all tion Program Export Panel on Detection, Evaluation, and Treatment of High Blood patients with psoriasis should be encouraged to correct aggres- Cholesterol in Adults (Adult Treatment Panel III). Executive Summary. NIH sively their modifiable cardiovascular risk factors, in particular, publ. no. 01-3670. Bethesda, MD: National Institute of Health, metabolic syndrome. National Heart, Lung and Blood Institute, 2001. 22 Miccoli R, Bianchi C, Odoguardi L et al. Prevalence of the metabolic syndrome among Italian adults according to ATP III definition. References Nutr Metab Cardiovasc Dis 2005; 15:250–4. 1 Schon MP, Boehncke WH. Psoriasis. N Engl J Med 2005; 23 Naldi L, Chatenoud L, Linder D et al. Cigarette smoking, body mass 352:1899–912. index, and stressful life events as risk factors for psoriasis: results 2 Krueger G, Ellis CN. Psoriasis – recent advances in understanding its from an Italian case–control study. J Invest Dermatol 2005; 125:61– pathogenesis and treatment. J Am Acad Dermatol 2005; 53:S94–100. 7. 3 Gisondi P, Girolomoni G, Sampogna F et al. Prevalence of psoriatic 24 Herron MD, Hinckley M, Hoffman MS et al. Impact of obesity and arthritis and joint complaints in a large population of Italian smoking on psoriasis presentation and management. Arch Dermatol patients hospitalized for psoriasis. Eur J Dermatol 2005; 15:279–83. 2005; 141:1527–34. 4 Gelfand JM, Gladman DD, Mease PJ et al. Epidemiology of psoriatic 25 Kimhi O, Caspi D, Bornstein NM et al. Prevalence and risk factors arthritis in the population of the United States. J Am Acad Dermatol of atherosclerosis in patients with psoriatic arthritis. Semin Arthritis 2005; 53:573. Rheum 2007; 36:203–9. 5 Kimball AB, Jacobson C, Weiss S et al. The psychosocial burden of 26 Chung CP, Oeser A, Raggi P et al. Increased coronary-artery psoriasis. Am J Clin Dermatol 2005; 6:383–92. atherosclerosis in rheumatoid arthritis: relationship to disease 6 Henseler T, Christophers E. Disease concomitance in psoriasis. JAm duration and cardiovascular risk factors. Arthritis Rheum 2005; Acad Dermatol 1995; 32:982–6. 52:3045–53. 7 Neimann AL, Shin DB, Wang X et al. Prevalence of cardiovascular 27 Sherer Y, Shoenfeld Y. Mechanisms of disease: atherosclerosis in risk factors in patients with psoriasis. J Am Acad Dermatol 2006; autoimmune diseases. Nat Clin Pract Rheumatol 2006; 2:99–106. 55:829–35. 28 Alberti KG, Zimmet PZ. Definition, diagnosis and classification of 8 Mallbris L, Ritchlin CT, Stahle M. Metabolic disorders in patients diabetes mellitus and its complications. Part 1: diagnosis and classi- with psoriasis and psoriatic arthritis. Curr Rheumatol Rep 2006; fication of diabetes mellitus provisional report of a WHO consult- 8:355–63. ation. Diabet Med 1998; 15:539–53. 9 Mallbris L, Akre O, Granath F et al. Increased risk for cardiovascular 29 Balkau B, Charles MA. Comment on the provisional report from mortality in psoriasis inpatients but not in outpatients. Eur J Epidemiol the WHO consultation. European Group for the Study of Insulin 2004; 19:225–30. Resistance. Diabet Med 1999; 16:442–3. 10 Gelfand JM, Neimann AL, Shin DB et al. Risk of myocardial infarc- 30 Meigs JB, Wilson PW, Fox CS et al. Body mass index, metabolic tion in patients with psoriasis. JAMA 2006; 296:1735–41. syndrome, and risk of type 2 diabetes or cardiovascular disease. 11 Wakkee M, Thio HB, Prens EP et al. Unfavorable cardiovascular risk J Clin Endocrinol Metab 2006; 91:2906–12. profiles in untreated and treated psoriasis patients. Atherosclerosis 31 Cowey S, Hardy RW. The metabolic syndrome. A high risk state 2007; 190:1–9. for cancer? Am J Pathol 2006; 169:1505–22.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 Psoriasis and metabolic syndrome, P. Gisondi et al. 73
32 Lakka HM, Laaksonen DE, Lakka TA. The metabolic syndrome and 37 Kiortsis DN, Mavridis AK, Vasakos S et al. Effects of infliximab total and cardiovascular disease mortality in middle-aged men. treatment on insulin resistance in patients with rheumatoid arthritis JAMA 2002; 228:2709–16. and ankylosing spondylitis. Ann Rheum Dis 2005; 64:765–6. 33 Moller DE, Kaufman KD. Metabolic syndrome: a clinical and 38 Reynoso-von Drateln C, Martinez-Abundis E, Balcazar-Munoz BR molecular perspective. Annu Rev Med 2005; 56:45–62. et al. Lipid profile, insulin secretion and insulin sensitivity in 34 Cameron AJ, Shaw JE, Zimmet PZ. The metabolic syndrome: preva- psoriasis. J Am Acad Dermatol 2003; 48:882–5. lence in worldwide populations. Endocrinol Metab Clin North Am 2004; 39 Mallbris L, Granath F, Hamsten A, Stahle M. Psoriasis is associated 33:351–75. with lipid abnormalities at the onset of skin disease. J Am Acad 35 Du G, Reynolds K, Wu X et al. Prevalence of the metabolic syn- Dermatol 2006; 54:614–21. drome and overweight among adults in China. Lancet 2005; 40 Sommer DM, Jenisch S, Suchan M et al. Increased prevalence of the 365:1398–405. metabolic syndrome in patients with moderate to severe psoriasis. 36 American Heart Association; National Heart, Lung, and Blood Insti- Arch Dermatol Res 2006; 298:321–8. tute, Grundy SM, Cleeman JI et al. Diagnosis and management of 41 Tilg H, Moschen AR. Adipocytokines: mediators linking adipose tis- the metabolic syndrome. An American Heart Association ⁄National sue, inflammation and immunity. Nat Rev Immunol 2006; 6:772–83. Heart, Lung, and Blood Institute Scientific Statement. Executive 42 Weiss ST. Obesity: insight into the origins of asthma. Nat Immunol summary. Cardiol Rev 2005; 13:322–7. 2005; 6:537–9.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp68–73 CONTACT DERMATITIS AND ALLERGY DOI 10.1111/j.1365-2133.2007.07944.x How irritant is alcohol? H. Lo¨ffler,* G. Kampf, § D. Schmermund and H.I. Maibach* *Department of Dermatology, University of California, San Francisco, CA, U.S.A. Department of Dermatology, Philipp-University, Deutschhausstr. 9, Marburg, Germany Bode Chemie GmbH & Co. KG, Scientific Affairs, Hamburg, Germany §Institute for Hygiene and Environmental Medicine, Ernst-Moritz-Arndt University, Greifswald, Germany
Summary
Correspondence Background Alcohol-based hand rubs are used worldwide to prevent transmission Harald Lo¨ffler. of nosocomial pathogens. E-mail: harald.loeffl[email protected] Objectives To investigate skin irritation caused by alcohols alone and in combin- ation with detergent washing. Accepted for publication 8 November 2006 Methods Single and repetitive patch testing with 60–100% alcohols [ethanol, 1-propanol, 2-propanol (synonyms: isopropyl alcohol, isopropanol)], a positive Key words control [0Æ5% sodium lauryl sulphate (SLS)] and negative controls (empty cham- alcohol-based hand rubs, irritant contact ber and water) were performed. Wash tests were performed with 80% ethanol dermatitis, patch test, sodium lauryl sulphate, and 0Æ5% SLS on the forearms with each agent alone and with both agents in a tandem application, wash test tandem design. Skin hydration, erythema and barrier disruption [measured as Conflicts of interest transepidermal water loss (TEWL)] were evaluated (always 15 volunteers). None declared. Results We found no significant change in skin barrier or erythema induced by the alcohols in the patch tests, whereas skin hydration decreased significantly. Appli- cation of alcohols to previously irritated skin did not show a stronger skin barrier disruption than application of SLS alone. Wash tests demonstrated that alcohol application caused significantly less skin irritation than washing with a detergent (TEWL, P <0Æ001; skin hydration, P <0Æ05; erythema, P <0Æ05). Even on pre- viously irritated skin, ethanol did not enhance irritation. By contrast, a protective effect of ethanol used after skin washing was observed (TEWL, P <0Æ05; skin hydration, P <0Æ05; erythema, P <0Æ05). Conclusions Alcohol-based hand rubs cause less skin irritation than hand washing and are therefore preferred for hand hygiene from the dermatological point of view. An alcohol-based hand rub may even decrease rather than increase skin irritation after a hand wash due to a mechanical partial elimination of the detergent.
Millions of healthcare workers perform hand disinfection with wet work and contact with detergents.9 Nevertheless, ‘hand alcohol-based hand rubs several times daily. Their efficacy for hygiene’ procedures are often quoted as important patho- control of nosocomial infection led to widespread U.S. gov- genetic factors for the development of hand dermatitis.10–12 ernment endorsement.1 The assumption of poor skin toler- When the irritant effect of alcohol on the skin has been ance, however, remains a major reason for low compliance evaluated, most authors found low toxicity.13–16 By contrast, rates.2 ‘Hand hygiene’ may lead to irritation and hand eczema. many healthcare workers complain about unacceptable skin The prevalence of eczematous hand lesions in medical staff irritation caused by alcohol-based hand rubs. Even in the remains between 20% and 40%.3,4 The nursing and related Guideline for Hand Hygiene in Healthcare Settings of the Centers for professions (employees in the healthcare system) have up to a Disease Control,1 skin tolerability of alcohol-based hand rubs six times increased risk for occupational dermatitis.5,6 Irritant is stated as potentially problematic: ‘Although alcohols are contact dermatitis is frequently observed in these occupations among the safest antiseptics available, they can cause dryness and is widely accepted as unavoidable. Even a mild interdigital and irritation’. eczema can be an important sign for future hand eczema on The assumed irritation due to alcohol-based hand antiseptics which microbes can grow more easily.7,8 Irritant skin changes may hold up their wide use, especially in the U.S.A.7 This in healthcare employees are undoubtedly caused by frequent study, therefore, evaluates the irritant potencies of the relevant
2007 The Authors 74 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 How irritant is alcohol?, H. Lo¨ffler et al. 75 types of alcohols alone and in sequence with use of a deter- was subsequently applied on the same area for another 24 h. gent in a highly standardized test design (patch test) as well The removal of the second patch was performed at 48 h; as in a more realistic standardized wash test. evaluation was conducted before the first application (0 h) and after 72 h (hence, substance-nonspecific alterations 17 Materials and methods of measurements due to occlusion could be minimized ). Each group consisted of 15 volunteers, and an empty chamber and a distilled water chamber were used as controls. Alcohols Study population and design (> 99% purity) were supplied by Bode Chemie GmbH & Co. A total of 105 healthy volunteers without skin diseases (49 KG (Hamburg, Germany); sodium lauryl sulphate (SLS, women and 56 men, age range 18–52 years, mean ± SD > 99% purity) was supplied by Sigma Chemicals (St Louis, 32 ± 8Æ2) participated. Atopic individuals were excluded. MO, U.S.A. and Munich, Germany). Written informed consent was obtained, and the study was Part 1 consisted of repetitive applications of the same patch: approved by the ethical committee of the University of Cali- ethanol, 1-propanol or 2-propanol (synonyms: isopropyl alco- fornia San Francisco (CA, U.S.A.) and the University of Mar- hol, isopropanol), each in the concentration of 60%, 70%, burg (Germany). The study was performed in a multicentre 80%, 90% and 100% (ethanol 99%). design in San Francisco, U.S.A., and Marburg, Germany. Sub- Part 2 consisted of repetitive applications of alcohol patches jects were instructed not to apply topical ‘leave-on’ products (ethanol 80%, 1-propanol 60% and 2-propanol 70%, analog- such as lotions or creams to the test sites for 1 week prior to ous to the concentrations used in commonly used alcohol- study. Before the measurements began, the volunteers rested based hand rubs,18 followed by SLS 0Æ5% patches and vice in the test room for acclimatization for at least 30 min. versa. SLS 0Æ5% and an empty chamber served as controls.
Test procedure Wash test
Two different tests were performed: the repetitive occlusive Procedure 1: SLS wash test. This standardized wash test was per- patch test (Fig. 1) and the wash test. formed on the forearm: a foam roller (Lehnartz, Remscheid, Germany) was soaked with SLS 0Æ5% and moved 10 times within 1 min up and down on the volar forearm. Then, the Repetitive occlusive patch test roller was soaked again with test solution and the whole pro- The repetitive occlusive patch test consists of two occlusive ap- cedure was repeated for altogether five times. At the end, the plications of substances on the same test area (tandem applica- forearm was rinsed with clear tap water and dried carefully tion), each lasting 24 h. Sixty microlitres of the test solution with a paper towel. For each washing procedure 50 mL of test was pipetted in large Finn Chambers (Epitest, Helsinki, Fin- solution was used. In two groups (see below) tap water was land; inner diameter 12 mm) and applied for 24 h on the used instead of SLS. back. After removal of the patch, the test area was marked and a further patch (with the same agent or with another agent) Procedure 2: disinfection test. For simulation of hand disinfection with an alcohol-based hand rub, the procedure was basically identical to the SLS wash test, except that alcohol was allowed to air dry on the skin between each treatment.
Procedure 3: combination of hand washing and disinfection. In this pro- cedure the forearm was washed and dried first and disinfec- ted thereafter. Therefore, a surgical hand disinfection was mimicked.
The described procedures were performed twice daily for 7 days with evaluation on day 0 (baseline: before the first procedure), on day 8 and after 2 days of skin recovery on day 10. In each test group, a comparison between two different procedures (each of them on one forearm) was performed by a randomized assignment of a procedure to one forearm. The following groups were tested: (i) ethanol 80% alone (procedure 2) vs. SLS 0Æ5% alone (procedure 1); (ii) ethanol Fig 1. Test design of the repetitive patch tests (each group consists of 80% alone (procedure 2) vs. SLS 0Æ5% followed by ethanol 15 volunteers): group 1 with twice 24-h application of the same 80% (procedure 3); (iii) SLS 0Æ5% (procedure 1) alone vs. SLS alcohol patch; group 2 with alternating application of alcohol and 0Æ5% followed by ethanol 80% (procedure 3); (iv) SLS 0Æ5% detergent. SLS, sodium lauryl sulphate. (procedure 1) followed by tap water (procedure 1) vs. SLS
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 76 How irritant is alcohol?, H. Lo¨ffler et al.
0Æ5% followed by ethanol 80% (procedure 3); and (v) ethanol 80% alone (procedure 2) vs. tap water alone (procedure 1).
Biometrics
Bioengineering measurements of transepidermal water loss (TEWL), skin hydration and erythema were performed. TEWL was evaluated with a Tewameter TM210 (Courage & Khazaka, Cologne, Germany; Acaderm, Menlo Park, CA, U.S.A.). During the TEWL measurements, the probe was hand held by use of an insulating glove until a stable TEWL value was established (~1 min). Air convection was prevented by reducing move- ments and talking in the test room. Temperature and humidity were recorded (20–22 C, 40–55% relative humidity). The results were evaluated according to the guidelines for TEWL measurement by the Standardization Group of the European Society of Contact Dermatitis.19 Each TEWL test value consis- ted of the mean of two single measurements. Fig 2. Mean ± SD changes in skin hydration compared with basal Skin hydration was evaluated with a Corneometer CM 920 value after repeated application (twice, each for 24 h) of ethanol (Courage & Khazaka; Acaderm); each test value was attained (red), 1-propanol (green) and 2-propanol (blue) at different 20 by taking the mean of five single measurements. Erythema concentrations and two control chambers in a patch test design. was measured with a Chromameter CR 300 (Minolta, Osaka, Differences between the alcohols and the water or empty chamber Japan); during measurement the light in the test room was were significant at P <0Æ05. dimmed, and each test value was attained by taking the mean of five single measurements.21 chamber (Fig. 2). This decrease tended to be stronger with ethanol and 1-propanol than with 2-propanol. Remarkably, decreased hydration seemed less pronounced at higher alcohol Subjective sensations concentrations. No significant change from the negative con- After the washing procedure each volunteer estimated his ⁄her trols was seen regarding erythema (chromameter values) and subjective sensations of dryness, itching and burning on a vis- skin barrier (TEWL values) at all patches (data not shown). ual analogue scale ranging from 0 to 10. The full length of the scale was defined as the subjective maximum sensation; Alcohols and sodium lauryl sulphate each participant marked the degree of his ⁄her individual sen- sation on the scale. We measured the length between 0 and When SLS 0Æ5% was applied in a tandem test design with the the marked point for each sensation. alcohols or controls, no significant change was found for Dskin hydration (D = difference between SLS + alcohol and SLS + empty chamber) for any sequence. A different picture was Statistics observed for TEWL and erythema. The highest DTEWL and Statistical analysis was performed using SPSS software version Derythema were found with a tandem application of SLS 11.5 (SPSS, Chicago, IL, U.S.A.). The results of the bioengin- (Table 1). A single SLS application followed by an empty cham- eering measurements were calculated regarding their symmet- ber (control) significantly increased TEWL. Replacement of the rical distribution with the Kolmogorov–Smirnov test. Because empty chamber with any of the three alcohols revealed a similar they showed a symmetrical distribution, the values are shown DTEWL. When SLS was applied in the second patch after a pre- as mean ± SD. Differences between the bioengineering values ceding empty chamber (control) or a preceding patch with any of each test procedure were calculated regarding significance of the three alcohols, DTEWL was also in a similar range. by means of the paired, two-sided Student’s t-test. Statistical significance was accepted when P £ 0Æ05. Wash tests
Results Ethanol 80% vs. sodium lauryl sulphate 0Æ5% As shown in Table 2, there was a significant increase of barrier Patch tests disruption, a lower skin hydration and a greater erythema at the forearm washed with the detergent compared with the dis- Alcohols alone infected forearm at the end of the treatment period (day 8). We noticed a significantly decreased skin hydration induced Although a stabilization of the skin physiology was seen by alcohols compared with the empty chamber and the water 2 days later, the SLS-treated side was still more affected
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 How irritant is alcohol?, H. Lo¨ffler et al. 77
Table 1 Changes in transepidermal water loss (TEWL), skin hydration and erythema Type of treatment DTEWL DSkin hydration DErythema (mean ± SD) after tandem application of ) various hand hygiene agents in a repetitive SLS 0Æ5% followed by empty chamber 23Æ0±8Æ6 5Æ6±9Æ05Æ2±2Æ3 occlusive patch test design (control) SLS 0Æ5% followed by SLS 0Æ5% 40Æ9±12Æ5** )1Æ6±10Æ77Æ4±3Æ2* SLS 0Æ5% followed by ethanol 80% 22Æ6±8Æ6 )4Æ8±11Æ74Æ7±2Æ5 SLS 0Æ5% followed by 1-propanol 60% 24Æ5±11Æ3 )0Æ4±18Æ94Æ9±2Æ5 SLS 0Æ5% followed by 2-propanol 70% 18Æ2±8Æ9* )0Æ4±10Æ13Æ6±1Æ4*
Empty chamber followed by SLS 0Æ5% 16Æ2±9Æ3 )3Æ2±15Æ84Æ9±2Æ1 (control) Ethanol 80% followed by SLS 0Æ5% 17Æ3±11Æ1 )3Æ6±7Æ83Æ1±2Æ2 1-propanol 60% followed by SLS 0Æ5% 15Æ1±7Æ5 )4Æ4±10Æ13Æ6±2Æ6 2-propanol 70% followed by SLS 0Æ5% 14Æ3±8Æ5 )0Æ1±11Æ82Æ8±1Æ9*
SLS, sodium lauryl sulphate. Difference from control is significant at *P <0Æ05 or **P <0Æ001.
Table 2 Changes in transepidermal water loss (TEWL), skin hydration and erythema (mean ± SD) after tandem wash tests with hand hygiene agents; procedures were performed on the forearms (side A and side B), and D-values (difference from basal values on day 0) are shown
DTEWL DSkin hydration DErythema
Substance ⁄procedure Day 8 Day 10 Day 8 Day 10 Day 8 Day 10 Side A Ethanol 80% 3Æ9±4Æ0** 1Æ5±1Æ1** )6Æ8±5Æ8* )3Æ2±6Æ10Æ6±0Æ6* 0Æ1±0Æ4* Side B SLS 0Æ5% 9Æ7±5Æ6** 6Æ7±2Æ1** )10Æ1±6Æ1* )5Æ8±7Æ21Æ8±1Æ6* 0Æ9±1Æ1* Side A Ethanol 80% 2Æ4±1Æ0** 1Æ1±3Æ4** )7Æ7±3Æ5* )3Æ5±3Æ6* 0Æ6±0Æ3* 0Æ1±0Æ4 Side B SLS 0Æ5% followed by ethanol 80% 8Æ0±4Æ0** 4Æ1±2Æ0** )11Æ7±5Æ2* )6Æ4±6Æ1* 1Æ0±6Æ5* 0Æ4±0Æ5 Side A SLS 0Æ5% 9Æ9±5Æ1* 6Æ8±3Æ4* )10Æ6±4Æ6* )5Æ9±5Æ6* 2Æ1±1Æ2** 1Æ0±0Æ9* Side B SLS 0Æ5% followed by ethanol 80% 6Æ9±3Æ3* 4Æ7±2Æ3* )6Æ6±5Æ8* )2Æ8±6Æ3* 1Æ2±0Æ9** 0Æ5±0Æ6* Side A SLS 0Æ5% followed by water 7Æ9±9Æ14Æ9±5Æ8 )7Æ2±6Æ3 )5Æ4±5Æ11Æ7±1Æ50Æ8±0Æ5 Side B SLS 0Æ5% followed by ethanol 80% 8Æ6±6Æ45Æ2±4Æ3 )9Æ6±5Æ8 )4Æ5±5Æ71Æ6±1Æ40Æ6±0Æ6 Side A Ethanol 80% 1Æ6±2Æ10Æ6±1Æ3 )3Æ5±8Æ2 )1Æ9±8Æ50Æ1±0Æ6 )0Æ2±0Æ9 Side B Water 1Æ5±2Æ01Æ2±1Æ7 )1Æ1±7Æ5 )0Æ1±9Æ70Æ2±0Æ8 )0Æ2±1Æ5
SLS, sodium lauryl sulphate. The difference between side A and side B is significant at *P <0Æ05 or **P <0Æ001.
(Fig. 3a). Subjective sensation was overall slight but on day 8 less influenced by the combination than by the washing pro- was greater regarding dryness, itching and burning at the SLS- cedure alone. Therefore, the skin was less irritated by washing treated side. and disinfection compared with washing alone (Fig. 3c). By contrast, there were moderate subjective sensations, especially on the SLS ⁄ethanol side, regarding burning and dryness. Ethanol 80% vs. sodium lauryl sulphate 0Æ5% followed by ethanol 80% Sodium lauryl sulphate (SLS) 0Æ5% followed by water vs. The combination of washing and disinfection caused signifi- SLS 0Æ5% followed by ethanol 80% cantly greater impairment of all evaluated physiological skin parameters (increased barrier disruption, decreased skin hydra- Both procedures led to comparable skin physiology changes. A tion and increased erythema) than disinfection alone. This tendency towards a pronounced decrease of skin hydration effect was also apparent on day 10 but with diminished values was seen with the combination of SLS with ethanol but this (Fig. 3b). Subjective sensations were stronger on the SLS ⁄ did not reach the level of significance. However, there were ethanol side, especially regarding dryness and itching. no significant differences between both forearms at 8 or 10 days (Table 2). Although there were only moderate sub- jective sensations, the number of burning sensations at the Sodium lauryl sulphate (SLS) 0Æ5% vs. SLS 0Æ5% followed SLS ⁄ethanol side was higher. by ethanol 80%
The skin physiology evaluated by water barrier disruption, Ethanol 80% vs. water skin hydration and erythema was clearly affected by washing with SLS. When the skin was disinfected with ethanol after Forearm disinfection with ethanol led to similar changes in washing, all skin physiological parameters were significantly skin physiology as did washing with water alone. There was a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 78 How irritant is alcohol?, H. Lo¨ffler et al.
Fig 3. Bioengineering changes (mean ± SD) caused by wash procedure with (a) ethanol 80% vs. sodium lauryl sulphate (SLS) 0Æ5%, (b) ethanol 80% vs. SLS 0Æ5% followed by ethanol 80%, or (c) SLS 0Æ5% vs. SLS 0Æ5% followed by ethanol 80%. Differences between both procedures were significant at *P <0Æ05 or **P <0Æ001, with SLS inducing a stronger irritation (a), SLS ⁄ethanol inducing a stronger irritation (b) and SLS alone inducing a stronger irritation (c). tendency on day 8 and 10 towards a pronounced decrease of decrease with 1-propanol, followed by ethanol; the smallest skin hydration caused by ethanol, but these changes were not decrease was observed with 2-propanol. With a single 24-h significant. At day 10 (3 days after completion of the wash- application, decreased skin hydration was not detected (data ing) the values were nearly normal. Subjective sensations were not shown). There is a tendency towards a greater decrease of not different between the two sides. skin hydration perceivable at the lower concentrations of eth- anol and 1-propanol. These differences are small; we currently Discussion have no explanation for this phenomenon. However, our results demonstrate the ability of epicutaneously applied Ethanol, 1-propanol and 2-propanol lead to only minor skin alcoholic substances to reduce skin hydration, as described barrier changes (comparable with those with water or the earlier.23–26 It was not possible to induce detectable irritation empty chamber) and no changes in erythema independent of (barrier disruption or erythema). the concentration tested. This is a first indication that these Because this test design is exaggerated (long-term occlusive substances are not important irritants when the contact dur- testing) and its relevance for the daily work of healthcare ation is limited, and supports previous findings.14,15 In employees unclear, the washing ⁄disinfection study was per- contrast, detergents can (depending on their type and formed. When ethanol 80% was repetitively applied in a more concentration) induce relevant barrier disruption and inflam- realistic use test design over 1 week, only a minimal barrier mation even after a single patch test.22 A decrease of skin disruption was induced, comparable with that induced by tap hydration can also be observed after the application of alco- water. This barrier disruption might be due to scrubbing hols even in the occlusive patch test design, with the strongest (moving the roller over the skin) rather than to an inherent
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 How irritant is alcohol?, H. Lo¨ffler et al. 79 feature of ethanol. We did not detect an increase in erythema, to different amounts of detergents applied, or to the glycerol in accordance with results from studies in which no or which was added to the disinfection solution.40 In our study, only minor irritant potency is described after application of an we demonstrated that the combination of SLS washing with alcohol-based hand rub,13–16,27–32 which can be further subsequent ethanol treatment induced similar skin irritation as reduced by addition of emollients.33,34 In our model skin the combination of SLS washing with subsequent water treat- hydration was moderately reduced by ethanol treatment, ment. Hence, the protective effect is most likely to be caused slightly more than by tap water. By contrast to ethanol, the by a washout of detergent molecules which are on and in the detergent SLS induced a much stronger barrier disruption and stratum corneum and which may lead to a prolonged skin a pronounced skin hydration decrease, despite the fact that the irritation.41 This washout can be achieved with similar results concentration of the detergent was low (0Æ5%), which was by ethanol or water treatment. The important finding of this emphasized by the moderate increase of erythema. In our study is that alcohols used in hand rubs did not induce further view ethanol has only a low irritancy which is clinically rele- skin irritation. Contrarily, disinfection may even reduce irrita- vant only regarding its potential for reducing skin hydration. tion caused by detergents, which might be an important ele- The most interesting question remains the ability of alco- ment in the concept of early prevention.42 hol-based hand rubs to induce skin damage when applied on Why do so many healthcare employees believe that alcohol- previously irritated skin as described previously.15 The hands based hand rinses are strong irritants and that these substances of healthcare workers are often previously irritated because of lead to hand dermatitis? One reason is the fact that alcoholic wet work and the occlusive milieu of gloves.4,5 Because alco- solutions induce burning sensations (sensory irritation) at pre- hol-based hand rubs are sometimes used after hand washing viously irritated skin. This was observed by our volunteers (classic procedure of surgical hand antisepsis), the combin- especially in the groups in which SLS treatment was followed ation of washing with a detergent and disinfection may be a by ethanol. In these groups, irritation was induced by SLS and crucial point in occupationally induced hand dermatitis. the subsequently applied ethanol could possibly penetrate First, we investigated the influence of a repetitive tandem more easily to sensory nerve endings. However, despite the application on skin physiology with patch tests. Consequently, fact that the application caused sensory discomfort, the physi- ethanol, 1-propanol and 2-propanol were applied before or ology of the skin was not altered. Healthcare employees, how- after an SLS patch. Remarkably, no increase of irritation was ever, will blame the alcohol-based product for this discomfort induced when the alcohol (regardless of which) was applied and not the underlying disturbed barrier function.12,43 The after the SLS patch. This demonstrates that the detergent- consequences are obvious: the probably harmless disinfection induced skin irritation, detected by barrier disruption, procedure will be neglected and increasingly replaced by fur- decrease of skin hydration and increase of erythema, was not ther hand washing. This does not lead to immediate discom- exacerbated by these alcohols. Even after the combined appli- fort, but will exacerbate the skin condition in a vicious cation, skin hydration was not different from that following circle.44 Burning after the use of alcohol-based hand rubs use of the detergent alone. When the application order was should probably be recognized as a sign for an already dis- reversed (first an alcohol patch then the SLS patch), the degree turbed skin barrier and for an impending hand dermatitis. of irritation remained the same. Hence, no alcohol impaired Chew and Maibach summarize the extensive recent literature the skin physiology such that the following detergent induced on irritation that may be beyond this mechanism.45 Based on a greater irritation. In this feature, all three types of alcohols our results, the influence of longer and exaggerated (20 times are different from several other (including physical) irritants, every day) use of alcohols and different aspects of tandem which enhanced skin reaction induced by a detergent in a effects, as well as the effect on individuals with an increased tandem patch test design as shown previously.35–38 We con- risk of hand dermatitis (e.g. atopic individuals) should be firmed the finding of Kappes et al., who demonstrated that investigated in detail. 1-propanol did not enhance the cumulative skin irritation Our study is of relevance for all medical staff, because it when used after SLS,39 and we can furthermore extend their demonstrates the good skin compatibility of three alcohols fre- results to 2-propanol and ethanol. quently used in alcohol-based hand rubs and underlines the The relevant question remains, however, if the detergent- known problematic skin compatibility of detergents and hand induced irritation can be exacerbated by a subsequent applica- washing. Promotion and education (most effectively during tion of ethanol. Surprisingly, all skin physiological parameters clinical training) may individually be necessary to encourage evaluated were less impaired by the combination of SLS with the use of alcohol-based hand rubs.25,42,46 Moreover, the vis- ethanol compared with SLS alone. This suggests that applica- ibly improved skin condition after the regular use of alcoholic tion of ethanol after hand washing may reduce irritant skin hand disinfection may encourage healthcare employees to changes caused by washing. Similar results were shown in a change their behaviour.2,15,25,27,31 short test protocol by Pedersen et al.40 They used commercial The advantages of alcohol-based hand disinfection com- products in an intensive repetitive application test over 2 days pared with hand washing regarding its bactericidal efficacy are and also detected diminished irritation induced by the com- obvious. Because there is evidence that hand washing before bination of washing and disinfection compared with washing disinfection may decrease rather than increase bactericidal effi- alone. It was uncertain whether this protective effect was due cacy of the combined procedure, reduction of hand washing
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 80 How irritant is alcohol?, H. Lo¨ffler et al. may be recommended not only from the skin physiological 17 Friebe K, Effendy I, Lo¨ffler H. Effects of skin occlusion in patch point of view but also for hygienic reasons.47,48 testing with sodium lauryl sulphate. Br J Dermatol 2003; 148:65–9. We do not wish to overgeneralize the results and under- 18 Kampf G, Kramer A. Epidemiologic background of hand hygiene and evaluation of the most important agents for scrubs and rubs. stand the difficulties of extrapolating from 100 volunteers to Clin Microbiol Rev 2004; 17:863–93. millions of users. Yet the data provided by these highly con- 19 Pinnagoda J, Tupker RA, Agner T et al. Guidelines for transepider- trolled observations offer the foundation for epidemiological mal water loss (TEWL) measurement: a report from the Standard- investigations and further investigations such as methods of ization Group of the European Society of Contact Dermatitis. decreasing sensory irritation.49,50 Contact Dermatitis 1990; 22:164–78. 20 Courage W. Hardware and measuring principle: Corneometer. In: Bioengineering and the Skin: Water and the Stratum Corneum (Elsner P, Acknowledgments Berardesca E, Maibach HI, eds). Boca Raton, FL: CRC Press, 1994. 21 Fullerton A, Fischer T, Lahti A et al. Guidelines for measurement of This work was generously supported by a grant from Bode skin colour and erythema. A report from the Standardization Group Chemie GmbH & Co. KG, Hamburg, Germany. H.L. per- of the European Society of Contact Dermatitis. Contact Dermatitis formed the investigations, had full access to the data in this 1996; 35:1–10. study and takes responsibility for the integrity of the data and 22 Lo¨ffler H, Happle R. Profile of irritant patch testing with deter- the accuracy of the data analysis. gents: sodium lauryl sulfate, sodium laureth sulfate and alkyl polyglucoside. Contact Dermatitis 2003; 48:26–32. 23 Rotter ML, Koller W, Neumann R. The influence of cosmetic addi- References tives on the acceptability of alcohol-based hand disinfectants. J Hosp Infect 1991; 18 (Suppl. B):S57–63. 1 Boyce JM, Pittet D. Guideline for hand hygiene in healthcare set- 24 Walter CW. Disinfection of hands. Am J Surg 1965; 109:691–3. tings. Recommendations of the Healthcare Infection Control Prac- 25 Pittet D. Compliance with hand disinfection and its impact on tices Advisory Committee and the HIPAC ⁄SHEA ⁄APIC ⁄IDSA Hand hospital-acquired infections. J Hosp Infect 2001; 48 (Suppl. A):S40– Hygiene Task Force. Am J Infect Control 2002; 30:S1–46. 6. 2 Kampf G, Lo¨ffler H. Dermatological aspects of a successful intro- 26 Widmer AF. Replace hand washing with use of a waterless alcohol duction and continuation of alcohol-based hand rubs for hygienic hand rub? Clin Infect Dis 2000; 31:136–43. hand disinfection. J Hosp Infect 2003; 55:1–7. 27 Rotter ML. Arguments for alcoholic hand disinfection. J Hosp Infect 3 Schnuch A, Uter W, Geier J et al. Contact allergies in healthcare 2001; 48 (Suppl. A):S4–8. workers. Results from the IVDK. Acta Derm Venereol (Stockh) 1998; 28 Rotter ML. Hand washing and hand disinfection. In: Hospital Epidemi- 78:358–63. ology and Infection Control (Mayhall CG, ed.), 2nd edn. Philadelphia: 4 Kavli G, Angell E, Moseng D. Hospital employees and skin prob- Lippincott Williams & Wilkins, 1999; 1339–55. lems. Contact Dermatitis 1987; 17:156–8. 29 Newman JL, Seitz JC. Intermittent use of an antimicrobial hand gel 5 Smit HA, Burdorf A, Coenraads PJ. Prevalence of hand dermatitis for reducing soap-induced irritation of health care personnel. Am J in different occupations. Int J Epidemiol 1993; 22:288–93. Infect Control 1990; 18:194–200. 6 Stingeni L, Lapomarda V, Lisi P. Occupational hand dermatitis in 30 Larson EL, Butz AM, Gullette DL et al. Alcohol for surgical hospital environments. Contact Dermatitis 1995; 33:172–6. scrubbing? Infect Control Hosp Epidemiol 1990; 11:139–43. 7 Steere AC, Mallison GF. Handwashing practices for the prevention 31 Kampf G, Muscatiello M. Dermal tolerance of Sterillium, a propanol- of nosocomial infections. Ann Intern Med 1975; 83:683–90. based hand rub. J Hosp Infect 2003; 55:295–8. 8 Larson E, Friedman C, Cohran J et al. Prevalence and correlates 32 Jungbauer FH, van der Harst JJ, Groothoff JW et al. Skin protection of skin damage on the hands of nurses. Heart Lung 1997; in nursing work: promoting the use of gloves and hand alcohol. 26:404–12. Contact Dermatitis 2004; 51:135–40. 9 Seitz JC, Newman JL. Factors affecting skin condition in two nur- 33 Kampf G, Wigger-Alberti W, Schoder V et al. Emollients in a pro- sing populations: implications for current handwashing protocols. panol-based hand rub can significantly decrease irritant contact Am J Infect Control 1988; 16:46–53. dermatitis. Contact Dermatitis 2005; 53:344–9. 10 Pittet D. Improving compliance with hand hygiene in hospitals. 34 Kampf G, Ennen J. Regular use of a hand cream can attenuate Infect Control Hosp Epidemiol 2000; 21:381–6. skin dryness and roughness caused by frequent hand washing. BMC 11 Pittet D. Improving adherence to hand hygiene practice: a multi- Dermatol 2006; 6:1. disciplinary approach. Emerg Infect Dis 2001; 7:234–40. 35 Ale SI, Laugier JP, Maibach HI. Differential irritant skin responses 12 Larson E, Killien M. Factors influencing handwashing behavior of to tandem application of topical retinoic acid and sodium lauryl patient care personnel. Am J Infect Control 1982; 10:93–9. sulphate: II. Effect of time between first and second exposure. Br J 13 Boyce JM, Kelliher S, Vallande N. Skin irritation and dryness associ- Dermatol 1997; 137:226–33. ated with two hand-hygiene regimens: soap-and-water hand wash- 36 Effendy I, Weltfriend S, Patil S et al. Differential irritant skin ing versus hand antisepsis with an alcoholic hand gel. Infect Control responses to topical retinoic acid and sodium lauryl sulphate: alone Hosp Epidemiol 2000; 21:442–8. and in crossover design. Br J Dermatol 1996; 134:424–30. 14 de Haan P, Meester HHM, Bruynzeel DP. Irritancy of alcohol. In: 37 Fluhr JW, Akengin A, Bornkessel A et al. Additive impairment of The Irritant Contact Dermatitis Syndrome (van der Valk PGM, Maibach HI, the barrier function by mechanical irritation, occlusion and sodium eds). New York: CRC Press, 1996; 65–70. lauryl sulphate in vivo. Br J Dermatol 2005; 153:125–31. 15 Lu¨bbe J, Ruffieux C, van Melle G et al. Irritancy of the skin dis- 38 Wigger-Alberti W, Krebs A, Elsner P. Experimental irritant contact infectant n-propanol. Contact Dermatitis 2001; 45:226–31. dermatitis due to cumulative epicutaneous exposure to sodium 16 Winnefeld M, Richard MA, Drancourt M et al. Skin tolerance and lauryl sulphate and toluene: single and concurrent application. effectiveness of two hand decontamination procedures in everyday Br J Dermatol 2000; 143:551–6. hospital use. Br J Dermatol 2000; 143:546–50.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 How irritant is alcohol?, H. Lo¨ffler et al. 81
39 Kappes UP, Goritz N, Wigger-Alberti W et al. Tandem application 46 Creedon SA. Healthcare workers’ hand decontamination practices: of sodium lauryl sulfate and n-propanol does not lead to enhance- compliance with recommended guidelines. J Adv Nurs 2005; ment of cumulative skin irritation. Acta Derm Venereol (Stockh) 2001; 51:208–16. 81:403–5. 47 Hu¨bner NO, Kampf G, Kamp P et al. Does a preceding hand 40 Pedersen LK, Held E, Johansen JD et al. Short-term effects of wash and drying time after surgical hand disinfection influence alcohol-based disinfectant and detergent on skin irritation. Contact the efficacy of a propanol-based hand rub? BMC Microbiol 2006; Dermatitis 2005; 52:82–7. 6:57. 41 Rhein LD. Review of properties of surfactants that determine their 48 Hu¨bner NO, Kampf G, Lo¨ffler H et al. Effect of a 1 min hand wash interactions with stratum corneum. J Soc Cosmet Chem 1997; 48: on the bactericidal efficacy of consecutive surgical hand disinfec- 253–74. tion with standard alcohols and on skin hydration. Int J Hyg Environ 42 Lo¨ffler H, Effendy I. Prevention of irritant contact dermatitis. Eur J Health 2006; 209:285–91. Dermatol 2002; 12:4–9. 49 Zhai H, Hannon W, Hahn GS et al. Strontium nitrate decreased 43 Lu¨bbe J, Ruffieux C, Perrenoud D. A stinging cause for preventive histamine-induced itch magnitude and duration in man. Dermatology skin care. Lancet 2000; 356:768–9. 2000; 200:244–6. 44 Huynh NT, Commens CA. Scrubbing for cutaneous procedures can 50 Zhai H, Hannon W, Hahn GS et al. Strontium nitrate suppresses be hazardous. Australas J Dermatol 2002; 43:102–4. chemically-induced sensory irritation in humans. Contact Dermatitis 45 Chew AL, Maibach HI (eds). Handbook of Irritant Dermatitis. Berlin: 2000; 42:98–100. Springer-Verlag, 2006.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp74–81 CONTACT DERMATITIS AND ALLERGY DOI 10.1111/j.1365-2133.2007.07965.x Artificial reduction in transepidermal water loss improves skin barrier function I. Buraczewska, U. Brostro¨m and M. Lode´n ACO HUD NORDIC AB, Research & Development, Box 622, SE-194 26 Upplands Va¨sby, Sweden
Summary
Correspondence Background Artificial reduction of abnormal transepidermal water loss (TEWL) is Marie Lode´n. considered to improve skin diseases associated with a defective barrier function. E-mail: [email protected] Treatment of the skin with moisturizers is also known to influence skin barrier function. Whether or not differences in occlusion between creams contribute to Accepted for publication 31 January 2007 their effects on the skin barrier function is unknown. Objectives To investigate the long-term effects of a semipermeable membrane on Key words the skin barrier function in normal skin. In addition, the occlusive properties of gloves, irritation, occlusion, silicone membrane, two creams were studied. sodium lauryl sulphate, susceptibility, Methods The study was randomized, controlled and evaluator-blind using measure- transepidermal water loss ment of TEWL and skin susceptibility to sodium lauryl sulphate as indicators of Conflicts of interest skin barrier function. I.B., U.B. and M.L. are employees of ACO HUD Results Coating of the skin with a silicone membrane for 23 h per day for 3 weeks AB, the manufacturer of Canoderm . improved skin barrier function, whereas no significant changes were found after using the membrane for 8 h per day. Conclusions Differences between creams in terms of their effect on skin barrier function cannot be solely explained by their occlusive properties.
Clinical studies in patients with xerosis support the benefit of ingredients in affecting skin barrier function. Furthermore, the treatment with moisturizers on visible dryness symptoms, complete ingredient listing, fat content, pH and other product whereas more divergent effects are reported with regard to characteristics are usually not given with the test results. The the functional changes in skin barrier function. No changes in role of possible occlusion from fats in the applied formula- transepidermal water loss (TEWL),1,2 decreased TEWL3–6 and tions is also unknown. increased TEWL7,8 have all been reported from moisturizer The aim of this study was to investigate the effect of studies on skin barrier disorders. Differences in skin suscepti- occlusion on skin barrier function without introducing any bility to external stressors have also been reported.7,9 potential influence on skin permeability from ingredients in Treatment of normal skin with moisturizers may influence the creams. We reduced TEWL by application of a very thin its barrier properties. Repeated applications of certain urea- sheet of soft semipermeable silicone membrane to the skin containing moisturizers have been shown to reduce TEWL and for 8 or 23 h per day for 3 weeks. Since the membrane make skin less susceptible to sodium lauryl sulphate (SLS)- should mimic the occlusion induced by creams, we also induced irritation,10–12 whereas a lipid-rich cream without measured the reduction in TEWL (i.e. degree of occlusion) any humectant has been found to increase skin susceptibility from two creams which have been reported to have opposite to SLS irritation.13 Increased sensitivity to nickel has also been effects on skin barrier function: urea cream11 and lipid-rich reported after treatment of nickel-sensitive patients with mois- cream.13,17 Skin barrier function was assessed by measure- turizers without humectant14,15 compared with treatment with ments of TEWL and by measurement of skin susceptibility to moisturizers containing humectants.15,16 In addition, differ- the irritant SLS. ences in skin reactivity to the vasodilator benzyl nicotinate have been reported between treatments with a lipid-rich Materials and methods cream and a moisturizer containing urea.17 The role of the proposed ‘active’ ingredients in the formu- Materials lations (e.g. certain lipids or humectants) has not been conclu- sively confirmed for the reported effects in the studies. Only a TEWL was reduced by application of a self-adherent semiper- limited number of vehicle-controlled studies have been carried meable membrane, which is marketed for topical use to out, reducing the possibility of understanding the role of the reduce hypertrophic scars and keloids (Oleeva Clear ; BioMed
2007 The Authors 82 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp82–86 TEWL and skin barrier function, I. Buraczewska et al. 83
Sciences Inc., Allentown, PA, U.S.A.). The membrane is wash- After 10, 22 and 24 days TEWL was measured. After the able and is formed from a blend of silicone and polytetra- measurements on day 22, the skin on the arms was patch tes- fluoroethylene (PTFE). PTFE provides an internal reinforcing ted with 1% SLS in a large aluminium chamber (diameter mechanism, thereby creating very thin sheets of soft silicone 12 mm, Finn Chambers, Epitest Oy, Finland). Fifty microlitres with enhanced physical strength. The membrane adheres well of the aqueous solution of 1% SLS was pipetted onto a layer to skin, but in order to secure a close fit an elastic tube ban- of filter paper placed in the chamber which was fixed to the dage (Tubifast ; Scholl, Oldham, U.K.) was applied. skin for 24 h using adhesive tape (Scanpor; Norgeplaster Facil- The urea cream Canoderm (ACO HUD AB, Stockholm, ity, Oslo, Norway). After removal of the patch, the skin was Sweden) contains aqua, caprylic ⁄capric triglyceride, canola, gently rinsed with water and allowed to dry. After another urea, propylene glycol, cetearyl alcohol, glyceryl polymetha- 24 h the degree of irritation was evaluated visually, as well as crylate, dimethicone, paraffin, sodium lactate, carbomer, poly- by measurements of TEWL and blood flow. In addition, skin sorbate 60, glyceryl stearate, PEG-100 stearate, methylparaben, capacitance was measured on an adjacent area of skin. propylparaben, lactic acid and citric acid, while the lipid- rich cream Locobase (Yamanouchi Europe, Leiderdorp, the Evaluation Netherlands) contains petrolatum, aqua, paraffinum liquidum, cetearyl alcohol, ceteareth-25, methylparaben, citric acid and The subjective assessment of the degree of irritation was made sodium citrate. The creams were studied with respect to their in accordance with the guidelines on simple scoring of acute occlusive properties. SLS irritation as suggested by the European Society of Contact Dermatitis:18 where 0 is no reaction; 0.5 is doubtful, very weak erythema or minute scaling; 1 is weak erythema, slight Study design oedema, slight scaling and ⁄or slight roughness; 2 is moderate The study was divided into two parts. The first part addressed degree of erythema, oedema, scaling and ⁄or roughness, or the degree of occlusion, i.e. reduction in water loss induced minor degree of erosions, vesicles, crusting and ⁄or fissuring; by the membrane and the creams, and the second part focused 3 is strong with marked degree of erythema, oedema, scaling, on the influence on skin barrier function induced by the roughness, erosions, vesicles, bullae, crusting and ⁄or fissuring; membrane. Informed consent was obtained from all partici- and 4 is very strong ⁄caustic as in 3, with necrotic areas. pants and the local ethics committee approved the study. After the visual assessment, measurements of TEWL were Treatments were randomized. performed using DermaLab (Cortex Technology, Hadsund, The reduction in water evaporation was measured after Denmark) (open chamber). After the probe was applied to the application of the membrane and the creams to defined areas on skin, TEWL values were automatically transferred to a compu- the volar forearm in 10 healthy individuals (aged 24–42 years, ter during the subsequent 60 s and the mean value from the mean 32.5; eight women and two men). On the morning last 20 s was recorded and used for further calculations. before attending the clinic the individuals were asked to wash Blood flow was measured with a laser Doppler flowmeter their forearms with soap and water. Four areas were marked on (Periflux PF1d; Perimed, Stockholm, Sweden) equipped with a the volar site of the forearm, and baseline TEWL at these areas special multifibre probe (PF 113 integrating probe, Perimed) was measured. One area was left untreated and the other three which has seven fibre triplets, one in the middle and six ) were treated with the two creams (1 mg cm 2) or covered with around, forming a small circle. The measured value is there- the membrane for 7 h. TEWL was measured again at 0.5 and fore a mean from the seven fibres, which helps in reducing 7 h, after which a second application of the respective cream variation due to spotty erythema. The probe was attached to was made. On the morning of the following day, TEWL in the the skin, without pressure, with a double adhesive tape and areas was again measured. No washing of the areas was allowed the values were registered on a chart strip recorder from and care was taken not to contact the cream-treated area during which a number for the calculations was taken after the the first 30 min after application. In three other individuals recordings had reached equilibrium. (aged 22–29 years, mean 26.7; women only), the membrane For the capacitance measurements a corneometer (Corneo- was kept in place for 24 h and measurement of TEWL was made meter CM-820; Courage & Khazaka GmbH, Cologne, Ger- after 0.3, 1 and 24 h on the membrane and then 5, 30 and many) (which is considered to reflect skin hydration) was 120 min after removal of the membrane. used. The clean corneometer probe was pressed gently onto In the long-term study, 29 healthy individuals (aged 20– the skin and results were obtained after 1 s; the median from 59 years; 25 women and 4 men), with no visible signs of three measurements within the same area was used for the skin disease or allergies, were included. The study was ran- calculations. Both the laser Doppler and the corneometer give domized and blind with respect to the evaluation. The subjects their results in arbitrary units. were divided into two groups: one group used the membrane for 8 h per day (during the night), and the other group used Data treatment and statistics it for 23 h per day (it was removed for 1 h when showering or when washing the treated arm). Both groups used the The Wilcoxon signed rank test on paired differences within membranes for 21 days. the same individual was used to detect differences between
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp82–86 84 TEWL and skin barrier function, I. Buraczewska et al. treatments. The Mann–Whitney U-test was used to detect 20 differences between individuals. The statistical program used Vol 1 Vol 2 Vol 3 was Minitab 14; P < 0.05 was considered significant.
) 15 –1 h –2 Results 10
TEWL in the forearm skin of the healthy volunteers was 5 )2 )1 3–8 g m h (mean 4.9, median 4.1). Application of the TEWL (g m Δ membrane to the skin reduced water evaporation by approxi- 0 ) ) mately 1–2 g m 2 h 1, measured on top of the membrane 0.5 –5 and 7 h after coverage of the area (P < 0.05 compared with 0 0·3 1 24 0·1 0·5 2 the untreated area for both time points) (Fig. 1). The creams Application of membrane Removal of membrane also reduced evaporation, but to a lesser extent (Fig. 1). Signi- Time (h) ficant reduction in TEWL compared to baseline was obtained Fig 2. Changes in transepidermal water loss (TEWL) from baseline in after all three time points with the lipid-rich cream, but only three volunteers (Vol) after application and removal of the membrane. 0.5 and 7 h after application of the urea cream. A slight reduc- tion in TEWL was also noted in the untreated area, and only the lipid-rich cream showed significantly lower TEWL values 10 compared with the untreated area at 7 and 24 h (P = 0.03). 23-h group Examples of the time course for the changes in TEWL after 8-h group application and removal of the membrane in three individuals are given in Fig. 2. After removal of the membrane signifi- cantly higher levels of TEWL were observed for about 30 min, 0 after which the levels subsided to the levels found before application of the membrane. TEWL of treated– After application of the membrane for 10 days, TEWL ten- Δ ded to be lower in the area that had been covered by the untreated undamaged skin membrane for 23 h per day (P = 0.06), whereas no change –10 P = 0·06 P = 1 P = 0·002 P = 0·625 P = 0·003 P = 0·315 in TEWL was observed in the area that had been covered by Day 10 Day 22 Day 24 the membrane for 8 h per day (P = 0.95) (Fig. 3). After a further 12 days, TEWL was significantly lower on the area ) ) Fig 3. Transepidermal water loss (TEWL) (g m 2 h 1) after 10, 22 covered for 23 h per day (P = 0.002), whereas no change and 24 days in skin to which the membrane was applied for 23 or 8 h per day, shown as the difference between membrane-treated skin and the untreated arm (n = 29). The results are presented as box plots 1 with the median value represented by a line across the box and the 0 first quartile value at the bottom and the third quartile value at the )
–1 –1 top. The whiskers are the lines that extend from the top and bottom h
–2 of the box to the lowest and highest observation within a defined –2 region. Outliers outside the lower and upper limits are plotted as –3 asterisks. –4 TEWL (g m Δ –5 was found in the area covered for 8 h per day (P = 0.625) –6 (Fig. 3). In some of the membrane-coated areas, slight scaling - - was observed, but no significant changes in skin capacitance Urea Lipid Urea Lipid Urea Lipid were found (Table 1). Membrane Untreated Membrane Untreated Untreated 0·5 h 7 h 24 h Application of SLS to the skin showed a significant reduc- tion in susceptibility, measured as TEWL (P = 0.003) and skin Fig 1. Changes in transepidermal water loss (TEWL) from baseline blood flow (P = 0.017), in skin that had been covered by the in each of the areas of forearm skin following application of the membrane for 23 h per day (Fig. 4). No significant changes membrane, urea cream and lipid-rich cream and in the area left were found visually, or in the skin which had been covered untreated. TEWL was measured above the areas, i.e. on top of the by the membrane for 8 h per day (Fig. 4). membrane. The results are presented as box plots with the median value represented by a line across the box and the first quartile value at the bottom and the third quartile value at the top. The whiskers are Discussion the lines that extend from the top and bottom of the box to the lowest and highest observation within a defined region. *P < 0.05 The suppression of TEWL by topical formulations is known to compared with baseline (n = 10). be related to the amount applied, fat content and types of fat
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp82–86 TEWL and skin barrier function, I. Buraczewska et al. 85
Table 1 Skin capacitance after 22 and 24 days in the areas of skin Artificial reduction of abnormal TEWL is considered to pro- coated with the membrane for 8 or 23 h per day and in the vide benefit in a large number of diseases associated with a corresponding untreated control areas. The results are median values defective barrier function (e.g. psoriasis, atopic dermatitis and with 25% and 75% percentiles in parentheses. There were no hypertrophic scars or keloids),23 whereas complete occlusion significant differences between the membrane-coated skin and the control areas (n = 29) of normal skin with, for example, aluminium chambers has been found to induce clinical and histopathological inflamma- tion and impairments in TEWL.24–26 The use of gloves has 8-h group 23-h group therefore been suggested as an essential factor in the patho- Membrane- Membrane- genesis of cumulative irritant contact dermatitis.24 Surpris- Day coated skin Control coated skin Control ingly, in this long-term study, we found an improved barrier 22 48 (44–50) 48 (43–54) 48 (44–50) 47 (42–50) function with the chronic use (23 h per day) of the silicone 24 48 (46–56) 51 (44–56) 52 (46–54) 50 (46–52) membrane, while the intermittent use (8 h per day) did not affect skin permeability. In contrast to the wearing of gloves, and possibly also aluminium chambers, the membrane in our study allowed for perspiration of the skin. The membrane also adhered well to the skin and prevented water accumulation 100 TEWL between the plastic edge and the skin surface. The good Blood flow adherence to the skin was achieved without any use of glue on the surface, which could have interfered with the barrier 50 function of the skin. There are several possible mechanisms behind the unex- 0 pected improvement of the skin barrier function. The molecu- TEWL/blood flow of SLS-damaged skin lar events that are involved in the process of cornification and Δ barrier lipid assembly allow for a number of intriguing specu- 27 –50 lations. One obvious explanation is a possible increase in P = 0·003 P = 0·017 P = 0·379 P = 0·925 thickness due to the mechanical protection of the stratum cor- 23-h group 8-h group neum by the membrane. Protection of the surface may have reduced the rate of desquamation in those wearing the mem- Fig 4. Degree of irritation after exposure to sodium lauryl sulphate brane for 23 h a day. This theory is compatible with the visual (SLS) shown as the difference in transepidermal water loss (TEWL) appearance of slight scaly skin in some individuals after ) ) (g m 2 h 1) and blood flow (arbitrary units) between membrane- removal of the membrane, although this was not confirmed by coated skin and the untreated area (n = 29). The results are presented a reduced capacitance. However, increased hydration usually as box plots with the median value represented by a line across the favours the activity of desquamatory enzymes, thereby counter- box and the first quartile value at the bottom and the third quartile acting increased thickness and scaliness. Any excess desquamat- value at the top. The whiskers are the lines that extend from the top ing cells should therefore have been removed during the daily and bottom of the box to the lowest and highest observation within a cleansing procedure and exposure to ambient conditions. defined region. Outliers outside the lower and upper limits are plotted Another conceivable explanation for the improved barrier func- as asterisks. tion is increased differentiation of the stratum corneum layer, which may have resulted in a larger projected size of the cor- in the formulation.19,20 In published studies, the amounts neocytes. The long-term reduction in TEWL may have induced applied are usually exaggerated in order to be able to detect a changed pattern of cytokines, which could have positively differences between formulations. In this study we aimed to influenced the stratum corneum differentiation.28 Larger size ) mimic normal application rates, i.e. 1 mg cm 2, and repeated of corneocytes makes the pathway for penetration longer.29,30 the applications two times to facilitate detection of possible Changes in the stratum corneum thickness and in proliferation occlusion. In doing so we could detect a slight reduction in might well be detected by histology and immunohistology, TEWL from the creams, but found no obvious differences in preferably also addressing the projected size of the corneocytes. occlusion between the formulations. Application of the mem- The third suggested explanation for improved barrier function brane to the skin induced a more pronounced reduction in relates to the composition and structure of the bilayer lipids, ) ) TEWL (1–2 g m 2 h 1). Reduction in skin water loss by which are known to be essential factors in determining skin creams19 as well as by plastic membranes,21 is known to permeability. The fraction of lipids forming a fluid phase may increase skin hydration. This can be observed as a release of have been changed, either due to compositional or organiza- excess water when the occlusion is removed and the skin is tional changes.31 The long-term increase in hydration may exposed to the ambient atmosphere,19,21,22 as shown in have influenced the bilayer lipids in a similar way to the ingre- Fig. 2. Our data, as well as previous findings, suggest a rela- dients in the formulations. Absence of pronounced differences tionship between the degree of occlusion induced by the in occlusive properties between the two tested formulations applied material and the subsequent rate of evaporation. suggests that they exert deeper effects due to their content of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp82–86 86 TEWL and skin barrier function, I. Buraczewska et al.
‘active’ ingredients rather than differences in occlusive capabil- 12 Serup J. A double-blind comparison of two creams containing urea ity. ‘Active’ ingredients or optimal physiological lipid mixtures as the active ingredient. Assessment of efficacy and side-effects by have been suggested to be responsible for barrier-improving non-invasive techniques and a clinical scoring scheme. Acta Derm Venereol Suppl (Stockh) 1992; 177:34–43. effects.3,32–34 The lipid-rich cream in the present study has also 13 Held E, Sveinsdottir S, Agner T. Effect of long-term use of moistur- shown deleterious effects on skin barrier function, with a izers on skin hydration, barrier function and susceptibility to resulting increase in susceptibility to irritants and aller- irritants. Acta Derm Venereol (Stockh) 1999; 79:49–51. 7,8,13,14 gens, whereas the urea cream has repeatedly been 14 Zachariae C, Held E, Johansen JD et al. Effect of a moisturizer on skin shown to reinforce barrier function.9,11,16 susceptibility to NiCl2. Acta Derm Venereol (Stockh) 2003; 83:93–7. In this study we were able to confirm previous findings of 15 Hachem JP, De Paepe K, Vanpe´eEet al. The effect of two mois- marginal reductions in skin water loss by application of for- turisers on skin barrier damage in allergic contact dermatitis. Eur J Dermatol 2002; 12:136–8. mulations,19 but failed to link previously reported differences 16 Lode´n M, Kuzmina N, Nyre´nMet al. Nickel susceptibility and skin on skin barrier function with differences in occlusion from barrier function to water after treatment with a urea-containing the two studied emulsions. Furthermore, we were able to moisturizer. Exog Dermatol 2004; 3:99–105. demonstrate an improved skin barrier function by long-term 17 Duval D, Lindberg M, Boman A et al. Differences among moisturiz- increase in skin hydration induced by a semipermeable sili- ers in affecting skin susceptibility to hexyl nicotinate, measured as cone membrane. The findings in the present study emphasize time to increase skin blood flow. Skin Res Technol 2002; 8:1–5. the importance of artificial reduction of TEWL as a measure to 18 Tupker RA, Willis C, Berardesca E et al. Guidelines on sodium lau- ryl sulfate (SLS) exposure tests. A report from the Standardization influence skin susceptibility to environmental stimuli. Previous Group of the European Society of Contact Dermatitis. Contact Derma- studies have shown that urea improves normal as well as titis 1997; 37:53–69. defective skin barrier function in humans. The use of other 19 Lode´n M. The increase in skin hydration after application of emol- substances in formulations to retard water loss may further lients with different amounts of lipids. Acta Derm Venereol (Stockh) enhance the clinical effect of moisturizers. 1992; 72:327–30. 20 Tsutsumi H, Utsugi T, Hayashi S. Study on the occlusivity of oil films. J Soc Cosmet Chem 1979; 30:345–56. References 21 Berardesca E, Vignoli GP, Fideli D et al. Effect of occlusive dressings on the stratum corneum water holding capacity. Am J Med Sci 1992; 1 Halkier-Sorensen L, Thestrup-Pedersen K. The efficacy of a mois- 304:25–8. turizer (Locobase) among cleaners and kitchen assistants during 22 Suetak T, Sasai S, Zhen YX et al. Effects of silicone gel sheet on the everyday exposure to water and detergents. Contact Dermatitis 1993; stratum corneum hydration. Br J Plast Surg 2000; 53:503–7. 29:266–71. 23 Tree S, Marks R. An explanation for the ‘placebo’ effect of bland 2 Vilaplana J, Coll J, Trulla´sCet al. Clinical and non-invasive evalu- ointment bases. Br J Dermatol 1975; 92:195–8. ation of 12% ammonium lactate emulsion for the treatment of dry 24 Ramsing DW, Agner T. Effect of glove occlusion on human skin skin in atopic and non-atopic subjects. Acta Derm Venereol (Stockh) (II). Long-term experimental exposure. Contact Dermatitis 1996; 1992; 72:28–33. 34:258–62. 3 Chamlin SL, Kao J, Frieden IJ et al. Ceramide-dominant barrier 25 Tsai TF, Maibach HI. How irritant is water? An overview. Contact repair lipids alleviate childhood atopic dermatitis: changes in bar- Dermatitis 1999; 41:311–14. rier function provide a sensitive indicator of disease activity. JAm 26 Ramsing DW, Agner T. Effect of glove occlusion on human skin. Acad Dermatol 2002; 47:198–208. (I). Short-term experimental exposure. Contact Dermatitis 1996; 4 Andersson A-C, Lindberg M, Lode´n M. The effect of two urea- 34:1–5. containing creams on dry, eczematous skin in atopic patients. I. 27 Candi E, Schmidt R, Melino G. The cornified envelope: a model of Expert, patient and instrumental evaluation. J Dermatolog Treat 1999; cell death in the skin. Nat Rev Mol Cell Biol 2005; 6:328–40. 10:165–9. 28 Elias PM, Wood LC, Feingold KR. Epidermal pathogenesis of 5 Grice K, Sattar H, Baker H. Urea and retinoic acid in ichthyosis and inflammatory dermatoses. Am J Contact Dermatol 1999; 10:119–26. their effect on transepidermal water loss and water holding capacity 29 Potts RO, Francoeur ML. The influence of stratum corneum mor- of stratum corneum. Acta Derm Venereol (Stockh) 1973; 53:114–18. phology on water permeability. J Invest Dermatol 1991; 96:495–9. 6 Pope FM, Rees JK, Wells RS et al. Out-patient treatment of ichthyo- 30 Rougier A, Lotte C, Corcuff P et al. Relationship between skin per- sis: a double-blind trial of ointments. Br J Dermatol 1972; 86:291–6. meability and corneocyte size according to anatomic site, age, and 7 Kolbe L, Kligman AM, Stoudemayer T. Objective bioengineering sex in man. J Soc Cosmet Chem 1988; 39:15–26. methods to assess the effects of moisturizers on xerotic leg of 31 Lindberg M, Forslind B. The skin as a barrier. In: Dry Skin and Mois- elderly people. J Dermatolog Treat 2000; 11:241–5. turizers. Chemistry and Function (Lode´n M, Maibach HI, eds), Vol. 2. 8Ga˚nemo A, Virtanen M, Vahlquist A. Improved topical treatment of Boca Raton, FL: Taylor & Francis Group, 2005; 9–21. lamellar ichthyosis: a double-blind study of four different cream 32 Lode´n M, Barany E, Mandahl P et al. The influence of urea treat- formulations. Br J Dermatol 1999; 141:1027–32. ment on skin susceptibility to surfactant-induced irritation. A pla- 9 Lode´n M, Andersson A-C, Lindberg M. Improvement in skin bar- cebo-controlled and randomized study. Exog Dermatol 2004; 3:1–6. rier function in patients with atopic dermatitis after treatment with 33 Man MM, Feingold KR, Thornfeldt CR et al. Optimization of a moisturizing cream (Canoderm ). Br J Dermatol 1999; 140:264–7. physiological lipid mixtures for barrier repair. J Invest Dermatol 1996; 10 Lode´n M. Urea-containing moisturizers influence barrier properties 106:1096–101. of normal skin. Arch Dermatol Res 1996; 288:103–7. 34 Fluhr JW, Gloor M, Lehmann L et al. Glycerol accelerates recovery 11 Lode´n M. Barrier recovery and influence of irritant stimuli in skin of barrier function in vivo. Acta Derm Venereol (Stockh) 1999; treated with a moisturizing cream. Contact Dermatitis 1997; 36:256– 79:418–21. 60.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp82–86 DERMATOLOGICAL SURGERY AND LASERS DOI 10.1111/j.1365-2133.2007.07946.x Randomized, double-blind, prospective study to compare topical 5-aminolaevulinic acid methylester with topical 5-aminolaevulinic acid photodynamic therapy for extensive scalp actinic keratosis F.J. Moloney* and P. Collins* *City of Dublin Skin and Cancer Hospital, Dublin 2, and St Vincent’s University Hospital, Dublin 4, Ireland
Summary
Correspondence Background 5-aminolaevlinic acid methylester (MAL) and 5-aminolaevulinic acid Fergal Moloney and Paul Collins. (ALA) photodynamic therapy (PDT) are both effective treatment options for acti- E-mail: [email protected]; nic keratosis (AK). While MAL is significantly more expensive than ALA, no stud- [email protected] ies have directly compared their efficacy in the treatment of extensive scalp AK. Accepted for publication Objectives To compare the efficacy and adverse effects of MAL-PDT with ALA-PDT 13 November 2006 in the treatment of scalp AK. Methods Sixteen male patients aged 59–87 years with extensive scalp AK were ran- Key words domized into a double-blind, split-scalp prospective study. Two treatment fields actinic keratoses, aminolaevulinic acid, were defined (right and left frontoparietal scalp) and treated 2 weeks apart. These aminolaevulinic acid methylester, pain, fields were randomized to receive either MAL or ALA as first or second treatment. photodynamic therapy, randomized controlled trial MAL cream was applied for 3 h; 20% ALA cream was applied for 5 h. A blinded Conflicts of interest observer assessed efficacy comparing AK counts before and 1 month after treat- None declared. ment. Pain was assessed using a visual analogue scale at 3, 6, 12 and 16 min. Results Fifteen patients completed treatment to both fields. There was a mean reduction from baseline in AK counts with the use of ALA-PDT of 6Æ2±1Æ9 compared with 5Æ6±3Æ2 with MAL-PDT (P =0Æ588). All patients experienced pain which was of greater intensity in the ALA-treated side at all time points: 3 min (P =0Æ151), 6 min (P =0Æ085), 12 min (P =0Æ012) and 16 min (P =0Æ029). Similarly, duration of discomfort post-procedure persisted for lon- ger following treatment with ALA when compared with MAL-PDT (P =0Æ044). Conclusions This study demonstrates that both ALA-PDT and MAL-PDT result in a significant reduction in scalp AK. There is no significant difference in efficacy. However, ALA-PDT is more painful than MAL-PDT in the treatment of extensive scalp AK.
The safety and efficacy of both topical 5-aminolaevulinic acid The main adverse event with both treatments is pain, which (ALA) and 5-aminolaevlinic acid methylester (MAL) photody- is described by most patients. The severity of pain varies from namic therapy (PDT) have been demonstrated in the treatment a transient discomfort to severe pain.6 Our aim was to com- of scalp actinic keratosis (AK).1–3 MAL is currently approved pare both efficacy and adverse effects of MAL with ALA-PDT in Europe for the treatment of AK and basal cell carcinoma of scalp AK in a split-scalp randomized, controlled double- (BCC) whereas ALA is unlicensed. MAL demonstrates blind study. improved lipophilicity when compared with ALA,4 allowing for enhanced lesion penetration and also a greater specificity Patients and methods for neoplastic cells.5 As a result it has been suggested that MAL-PDT may be more effective than ALA-PDT. In Ireland, Patients MAL (Metvix ; Photocure ASA, Oslo, Norway) is significantly more expensive than ALA (Porphin ; Mandeville Medicines, This study received local ethics committee approval and Aylesbury, Bucks, U.K.). all patients gave informed consent to participate. Sixteen
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp87–91 87 88 MAL-PDT compared with ALA-PDT for extensive scalp AK, F.J. Moloney and P. Collins caucasian male patients with bald scalps and extensive scalp Patients were assessed for erythema and erosions 14 days AK were selected for PDT. The mean age of patients was after their first and second treatment by one investigator 71Æ2±8Æ78 (range 59–87) years. One patient, following (F.M.) and asked to indicate at what time following treatment treatment to one side of the scalp, requested that the same were they no longer aware of discomfort in the treatment site treatment be applied to the contralateral scalp. This patient and whether analgesia was required following treatment. was withdrawn from the study and his data was not included Patients were further assessed for clinical response 1 month in the primary outcome analysis. after their last treatment by a second investigator (P.C.), who recorded patient preference and side-effect profile. He categor- ized response to treatment as clear, improved or no response, Treatment procedure and recorded the number of residual palpable AKs. Patients Prior to treatment, patients were assessed and equal treat- were asked if they had a preference for one treatment modal- ment fields were measured on the right and left frontopari- ity. Both patients and investigators remained blinded until etal scalp. The number of palpable AKs was recorded. The study completion. AKs were classified according to the most severe type pre- sent: grade 1 (not easily seen, slightly palpable), grade 2 Data analysis (well developed, easily palpable) or grade 3 (hyperkeratotic) AKs. Hyperkeratotic AKs were treated with white paraffin gel Baseline AK counts, reduction in AK counts and pain scores to remove any keratotic debris. Patients were randomized so were normally distributed as confirmed using the Shapiro– that half would receive ALA and half MAL as their first split- Wilk test (Stata Corporation, College Station, TX, U.S.A.); scalp treatment. This allowed patients to act as their own therefore, differences between treatment groups were tested by controls, and thus reduce confounding factors. The second the paired Student’s t-test. Nonparametric two-sided Wilcoxon field was treated separately, 2 weeks later, with the other signed ranks tests were employed to assess differences in the treatment modality. Care was taken to avoid any overlap AK count and duration of discomfort post-treatment. Statistical between treatment fields, as evidenced by the absence of any analyses were performed with JMP IN version 5.1 for local reaction extending beyond the treatment margins at Windows (SAS Institute, Cary, NC, U.S.A.). P-values < 0Æ05 fortnightly review. were deemed statistically significant. ) A visible layer of 20% ALA (30 g 100 cm 2 Porphin )or )2 MAL (30 g 100 cm Metvix ), was applied to either right or Results left frontal scalp for a 5-h or a 3-h period, respectively. The area was occluded with Tegaderm (3M, Loughborough, Baseline characteristics U.K.) and covered with a gauze dressing. Five hours after application of ALA and 3 h after application of MAL, the All patients had diffuse AKs; 31Æ3% (five patients) had pre- excess was wiped off. dominantly grade 1 AKs, 56Æ3% (nine patients) had predom- Fluorescence was graded on a scale of 1–3, using a Wood’s inantly grade 2 AKs and 12Æ5% (two patients) had scattered light after 5 h for ALA and after 3 h for MAL; 1 was light ⁄ grade 3 AKs on a grade 2 background. There was no differ- ) pale; 2, moderate; and 3, strong. Patients received 50 J cm 2 ence in the baseline number of AKs in the MAL and ALA treat- ) at 50 mW cm 2 using a Waldmann PDT lamp MSR 1200 ment fields (P =0Æ159) (Table 1). Fluorescence was noted as (580–740 nm) to the treatment field. Treatment time was grade 2 in seven treatment areas and grade 3 in 24 treatment 16 min 40 s. Irradiance was measured using a calibrated areas. No patient had taken pretreatment analgesia on either hand-held meter (International light 1400A with Selo33 ⁄ the first or second treatment. F ⁄W ⁄QND52 detector with spectral shaping and neutral density filters, calibrated by D. Taylor, Gloucester, U.K.). All Efficacy patients were cooled with a fan and refrigerated cold-water spray during treatment. Fifteen patients received treatment to both sides of the scalp. The median number of AKs on the ALA-treated fields 1 month post-treatment was 1 [interquartile range (IQR), 0–2] com- Evaluation pared with a median of 2 (IQR, 0–4) in MAL-treated fields Pain was assessed using a visual analogue scale (VAS) (1– (Table 1). This represented a mean reduction from baseline in 100 mm) at 3, 6, 12 and 16 min. Patients moved a counter AK counts with the use of ALA of 6Æ2±1Æ9 compared with along a 100-mm scale from ‘no pain’ to ‘worst pain ever’. 5Æ6±3Æ2 with MAL (P =0Æ588). One month following the The flip-side of the scale indicated the scores, 0 being no pain last treatment seven of 15 fields (46Æ7%) treated with and 100 the worst pain ever as previously described.7 Nurses MAL-PDT were clear compared with six of 15 fields (40%) recorded the numerical pain score and patients were not treated with ALA-PDT. This represented a clearing of 71% of aware of this value. If treatment had to be discontinued AKs treated with MAL-PDT compared with 87% treated with because of pain, the timing of this was recorded. Adverse ALA-PDT. Treatment failed, with no clinical response, in one effects were documented. field, which had received MAL-PDT. Blinded dermatologist
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp87–91 MAL-PDT compared with ALA-PDT for extensive scalp AK, F.J. Moloney and P. Collins 89
Table 1 Assessment of split-scalp treatment efficacy and patient cantly longer following ALA-PDT [median, 480 min (IQR preference for ALA-PDT and MAL-PDT 330–600)] when compared with MAL-PDT [median, 120 min (IQR 0–495)], P =0Æ044. Three patients required oral anal- ALA-PDT MAL-PDT gesia following ALA-PDT compared with one patient follow- (n = 15) (n = 15) P-value ing MAL-PDT. Efficacy, n (%) Clear 6 (40) 7 (46Æ7) Improved 9 (60) 7 (46Æ7) Adverse events No response 0 ( 0 ) 1 (0Æ7) One patient reported burning pain localized to the half-scalp AK count baseline Mean ± SD (CI) 7Æ3±1Æ6 8Æ8±1Æ5 0Æ159 treated with ALA-PDT, induced by natural or artificial light (5Æ8–8Æ7) (7Æ0–10Æ6) exposure. This was persisting at the 1-month assessment. He Median 8 8 was not on phototoxic therapy. There were no other adverse Range 2–11 2–16 events recorded apart from mild erythema in all treated sites AK count 1 month post-treatment and superficial erosions in two patients from both treatment Mean ± SD (CI) 1Æ1±1Æ2 2Æ7±3Æ4 0Æ227 groups noted at the 2-week assessment. (0Æ4–1Æ7) (0Æ9–4Æ6) Median 1 2 IQR 0–2 0–2 Patient preference Reduction in AK count from baseline Mean ± SD (CI) 6Æ2±1Æ9 5Æ6±3Æ2 0Æ588 When asked which treatment they would prefer to receive if (5Æ2–7Æ3) (3Æ9–7Æ5) further treatment was required, ten of 15 patients favoured Median 7 6 MAL-PDT because it was associated with less pain. Three Range 2–9 0–12 patients expressed no preference while two patients chose Patient preference, n (%)a 2 (13Æ3) 10 (66Æ7) ALA-PDT on the grounds that they perceived it to be more ALA, 5-aminolaevulinic acid; MAL, 5-aminolaevlinic acid methyl- effective than MAL-PDT. ester; PDT, photodynamic therapy; AK, actinic keratoses; SD, standard deviation; CI, 95% confidence intervals; IQR, interquar- tile range. Discussion a Three patients expressed no preference. Chronically sun-damaged scalps are commonly seen in elderly bald males. In a fair-skinned Irish population, such scalps often demonstrate diffuse visible and palpable AKs on a back- assessment indicated a better clinical response with ALA in six ground of actinically damaged skin, so-called field canceriza- patients (40%), MAL in one patient (6Æ7%) and no difference tion. Our study has shown that ALA-PDT and MAL-PDT are in the remaining eight patients (53Æ3%). both effective in the treatment of such scalps. A similar reduc- tion in the number of AKs and clearance rate was achieved with both treatments. All patients experienced pain during Pain scores treatment, the intensity of which varied. We demonstrated Pain scores were higher for ALA-PDT when compared with that MAL-PDT proved less painful for patients both during and MAL-PDT during treatment of scalp AKs (Table 2). This following MAL-PDT treatment when compared with ALA-PDT. held true at all time points: 3 min (P =0Æ151), 6 min To date, there have been many open studies, which have (P =0Æ085), 12 min (P =0Æ012) and 16 min (P =0Æ029). demonstrated the efficacy of both ALA-PDT and MAL-PDT in The duration of discomfort following treatment was signifi- the treatment of AKs.8–10 MAL-PDT is also effective in both
Table 2 Comparison of pain scores during ALA-PDT and MAL-PDT and duration of ALA-PDT MAL-PDT P-value discomfort post-treatment Pain score, mean ± SD (CI) 3 min 38Æ3±28Æ2 (22Æ7–53Æ9) 25Æ9±22Æ1 (13Æ7–38Æ2) 0Æ151 6 min 33Æ2±23Æ5 (20Æ3–46Æ3) 24Æ4±16Æ2 (15Æ4–33Æ4) 0Æ085 12 min 38Æ6±24Æ4 (25Æ1–52Æ1) 21Æ6±15Æ1 (13Æ3–29Æ9) 0Æ012a 16 min 37Æ3±25Æ6 (23Æ2–51Æ5) 22Æ3±15Æ7 (13Æ6–30Æ9) 0Æ029a Duration of discomfort 480 (330–600) 120 (0–495) 0Æ044a in min, median (IQR)
ALA, 5-aminolaevulinic acid; MAL, 5-aminolaevlinic acid methylester; PDT, photodynamic therapy; SD, standard deviation; CI, 95% confidence intervals; IQR, interquartile range aResults that are significant at 5% level.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp87–91 90 MAL-PDT compared with ALA-PDT for extensive scalp AK, F.J. Moloney and P. Collins the treatment and prevention of AKs in immunosuppressed corneum, which in theory should reduce the penetration of populations.11,12 Further studies have compared MAL-PDT ALA and MAL into target lesions. This was in part counteract- treatment of AKs with other treatment modalities such as ed by removal of keratotic debris from the AK surface. It has cryotherapy13–15 and topical 5-fluorouracil,16 demonstrating been postulated that ALA but not MAL is transported by similar response rates and cosmetic outcomes. Those studies, c-aminobutyric acid (GABA) receptors present in peripheral which employed a randomized controlled intraindividual nerve endings, thereby causing more pain.20 However, a design, have a greater power to demonstrate differences recent study has demonstrated that there is no significant dif- between the modalities being compared. ference in the penetration depths of ALA and MAL in normal To our knowledge, no study has directly compared the effi- skin but that ALA is more effective at inducing protoporphyrin cacy of ALA-PDT with MAL-PDT in the treatment of AKs, IX (PpIX) relative to an equivalent concentration of MAL despite the significant cost difference between the two prepa- applied to normal skin.21 These findings have yet to be repro- rations. Eight weeks after a single treatment using ALA-PDT, duced in diseased skin. Jeffes et al.8 recorded total clearing of 91% of lesions on the The exact level of PpIX required in tissue to generate a clin- face and scalp. Tarstedt et al.10 assessed patients 3 months after ically significant photodynamic effect is unknown. There is no their last treatment and documented a complete response rate agreed standard application time for ALA. Work which meas- of 93% for thin lesions but only 70% for thicker AKs after a ured levels of PpIX in plaques of psoriasis after application of single MAL-PDT treatment. Drawing comparisons between the ALA showed that levels peaked at 5 or 6 h, reaching 80–90% findings of such studies is obviously limited by differing study at 4 h.22 While it is possible that the longer duration of appli- methodologies, treatment regimens and lesion selection. It is cation for ALA, compared with MAL, could in itself influence also important not to compare clearance rates of individual pain severity and efficacy, there is no evidence to support lesions treated with clearance rates for a treatment field. In this. An ideal comparison of efficacy would also compare our study, for example, the total AK count in patients treated PpIX distribution within the epidermis prior to irradiation. with ALA was reduced by 87% at 1 month. This corresponds Robinson et al.23 highlighted the importance of diffuse distri- directly to 40% of scalp sites treated with ALA achieving total bution of PpIX in the epidermis compared with localized hot clearance at 1 month, while appearing to present a more spots of PpIX in the stratum corneum as a key determinant of favourable response. Follow-up studies will be required to efficacy. assess whether response rates are maintained over longer time Our statistically lower pain scores for MAL-PDT are mir- periods. rored clinically in the patient preferences should further treat- A number of approaches have been employed to reduce the ment be required, a majority selecting MAL-PDT on the basis pain associated with PDT. Two prospective, randomized stud- of its pain profile. Our findings support those of Kasche ies, using tetracaine gel (Ametop ; Smith and Nephew, Hull, et al.24 who asked patients with scalp AKs receiving either U.K.) in the first and a eutetic mixture of lignocaine 2Æ5% and MAL-PDT or ALA PDT to rate their pain. They stopped treat- prilocaine 2Æ5% (EMLA cream; Astra Pharmaceuticals, King’s ment in 54% of patients treated with ALA-PDT due to un- Langley, U.K.) in the second, concluded that topical anaesthe- bearable pain, in comparison with 14% of those receiving sia conferred no beneficial pain relief during PDT treat- MAL-PDT. No patients in our study required treatment to be ment.7,17 We standardized our treatment using a fan and discontinued. cold-water spray during treatment. We identified treatment The persisting photoaggravated burning pain experienced fields with comparable areas and lesion size in keeping with by one of our patients in the area of scalp treated with the work of Grapengiesser et al.18 showing that site and size of ALA-PDT is not a documented side-effect, nor had we experi- the lesion and individual patient characteristics were the main enced other patients with the same problem. In theory, all determinants of PDT-associated pain. To counteract the possi- PpIX should be metabolized to the photodynamically inactive bility of patient anxiety and unfamiliarity with the proposed haem within 48 h;25 however, PpIX has been demonstrated in treatment influencing pain scores, half received MAL-PDT and plaques of psoriasis for up to 14 days following a single ALA half ALA-PDT as their first treatment. In practice there was no application.26 difference in scores relative to which treatment was first The results of this study allow for the application of data- received. based rationale in choosing a mode of delivering PDT to sun- The pattern of pain intensity was similar for both ALA-PDT damaged skin. The absence of a demonstrable difference in and MAL-PDT, with both treatments inducing pain, which the efficacy of ALA-PDT and MAL-PDT treatment of scalp AKs gradually intensified during the first minute of treatment, places greater emphasis on the side-effect profile and patient reaching a maximum prior to the first recorded pain score at acceptability of these treatments. Despite optimizing pain-relief 3 min, and easing immediately on discontinuing treatment. techniques during PDT treatment, pain remains a major para- Less pain with MAL-PDT when compared with ALA-PDT had meter in our experience, often dissuading patients from fur- been documented previously in normal tape-stripped skin.19 ther treatment, when severe. The results of this study have All fields treated in this study had evidence of chronic ultra- encouraged us to employ MAL-PDT in the treatment of scalp violet (UV) damage. The field changes in skin that has been AKs, while reserving ALA-PDT treatment for those scalps that chronically UV damaged includes thickening of the stratum fail to respond.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp87–91 MAL-PDT compared with ALA-PDT for extensive scalp AK, F.J. Moloney and P. Collins 91
Acknowledgments 13 Freeman M, Vinciullo C, Francis D et al. A comparison of photo- dynamic therapy using topical methyl aminolevulinate (Metvix) We thank Sister Nora Tracey and the nurses in the day care with single cycle cryotherapy in patients with actinic keratosis: a unit for their help with this study. prospective, randomized study. J Dermatolog Treat 2003; 14:99–106. 14 Morton C, Horn M, Leman J et al. Comparison of topical methyl aminolevulinate photodynamic therapy with cryotherapy or fluor- References ouracil for treatment of squamous cell carcinoma in situ: results of a multicenter randomized trial. Arch Dermatol 2006; 142:729–35. 1 Pariser DM, Lowe NJ, Stewart DM et al. Photodynamic therapy with 15 Szeimies RM, Karrer S, Radakovic-Fijan S et al. Photodynamic ther- topical methyl aminolevulinate for actinic keratosis: results of a apy using topical methyl 5-aminolevulinate compared with cryo- prospective randomized multicenter trial. J Am Acad Dermatol 2003; therapy for actinic keratosis: a prospective, randomized study. JAm 48:227–32. Acad Dermatol 2002; 47:258–62. 2 Markham T, Collins P. Topical 5-aminolaevulinic acid photo- 16 Kurwa HA, Yong-Gee SA, Seed PT et al. A randomized paired com- dynamic therapy for extensive scalp actinic keratoses. Br J Dermatol parison of photodynamic therapy and topical 5-fluorouracil in the 2001; 145:502–4. treatment of actinic keratoses. J Am Acad Dermatol 1999; 41:414–18. 3 Szeimies RM, Karrer S, Sauerwald A et al. Photodynamic therapy 17 Holmes MV, Dawe RS, Ferguson J et al. A randomized, double- with topical application of 5-aminolevulinic acid in the treatment blind, placebo-controlled study of the efficacy of tetracaine gel of actinic keratoses: an initial clinical study. Dermatology 1996; (Ametop) for pain relief during topical photodynamic therapy. Br J 192:246–51. Dermatol 2004; 150:337–40. 4 Kloek J, van Beijersbergen H. Prodrugs of 5-aminolevulinic acid 18 Grapengiesser S, Ericson M, Gudmundsson F et al. Pain caused by for photodynamic therapy. Photochem Photobiol 1996; 64:994–1000. photodynamic therapy of skin cancer. Clin Exp Dermatol 2002; 5 Fritsch C, Homey B, Stahl W et al. Preferential relative porphyrin 27:493–7. enrichment in solar keratoses upon topical application of delta- 19 Wiegell SR, Stender IM, Na R et al. Pain associated with photo- aminolevulinic acid methylester. Photochem Photobiol 1998; 68:218– dynamic therapy using 5-aminolevulinic acid or 5-aminolevulinic 21. acid methylester on tape-stripped normal skin. Arch Dermatol 2003; 6 Fink-Puches R, Hofer A, Smolle J et al. Primary clinical response 139:1173–7. and long-term follow-up of solar keratoses treated with topically 20 Rud E, Gederaas O, Hogset A et al. 5-aminolevulinic acid, but not applied 5-aminolevulinic acid and irradiation by different wave 5-aminolevulinic acid esters, is transported into adenocarcinoma bands of light. J Photochem Photobiol B 1997; 41:145–51. cells by system BETA transporters. Photochem Photobiol 2000; 71:640– 7 Langan SM, Collins P. Randomized, double-blind, placebo- 7. controlled prospective study of the efficacy of topical anaesthesia 21 Juzeniene A, Juzenas P, Ma LW et al. Topical application of 5-amino- with a eutetic mixture of lignocaine 2.5% and prilocaine 2.5% for laevulinic acid, methyl 5-aminolaevulinate and hexyl 5-aminolaevu- topical 5-aminolaevulinic acid-photodynamic therapy for extensive linate on normal human skin. Br J Dermatol 2006; 155:791–9. scalp actinic keratoses. Br J Dermatol 2006; 154:146–9. 22 Stringer MR, Collins P, Robinson DJ et al. The accumulation of 8 Jeffes EW, McCullough JL, Weinstein GD et al. Photodynamic ther- protoporphyrin IX in plaque psoriasis after topical application of apy of actinic keratosis with topical 5-aminolevulinic acid. A pilot 5-aminolevulinic acid indicates a potential for superficial photo- dose-ranging study. Arch Dermatol 1997; 133:727–32. dynamic therapy. J Invest Dermatol 1996; 107:76–81. 9 Morton CA, Brown SB, Collins S et al. Guidelines for topical photo- 23 Robinson DJ, Collins P, Stringer MR et al. Improved response of dynamic therapy: report of a workshop of the British Photoderma- plaque psoriasis after multiple treatments with topical 5-aminolaev- tology Group. Br J Dermatol 2002; 146:552–67. ulinic acid photodynamic therapy. Acta Derm Venereol 1999; 79:451– 10 Tarstedt M, Rosdahl I, Berne B et al. A randomized multicenter 5. study to compare two treatment regimens of topical methyl ami- 24 Kasche A, Luderschmidt S, Ring J et al. Photodynamic therapy induces nolevulinate (Metvix)-PDT in actinic keratosis of the face and scalp. less pain in patients treated with methyl aminolevulinate compared Acta Derm Venereol 2005; 85:424–8. to aminolevulinic acid. J Drugs Dermatol 2006; 5:353–6. 11 Dragieva G, Prinz BM, Hafner J et al. A randomized controlled clin- 25 Marmur ES, Schmults CD, Goldberg DJ. A review of laser and ical trial of topical photodynamic therapy with methyl aminolaevu- photodynamic therapy for the treatment of nonmelanoma skin linate in the treatment of actinic keratoses in transplant recipients. cancer. Dermatol Surg 2004; 30:264–71. Br J Dermatol 2004; 151:196–200. 26 Collins P, Robinson DJ, Stringer MR et al. The variable response of 12 Wulf HC, Pavel S, Stender I et al. Topical photodynamic therapy plaque psoriasis after a single treatment with topical 5-aminolaevu- for prevention of new skin lesions in renal transplant recipients. linic acid photodynamic therapy. Br J Dermatol 1997; 137:743–9. Acta Derm Venereol 2006; 86:25–8.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp87–91 DERMATOLOGICAL SURGERY AND LASERS DOI 10.1111/j.1365-2133.2007.07970.x Lipomas after blunt soft tissue trauma: are they real? Analysis of 31 cases M.C. Aust, M. Spies, S. Kall, A. Gohritz, P. Boorboor, P. Kolokythas* and P.M. Vogt Department of Plastic, Hand and Reconstructive Surgery, Medizinische Hochschule Hannover, Carl-Neuberg-Str. 1, 30659 Hannover, Germany *Department of Hand-, Plastic and Microsurgery, BG-Trauma Hospital, Bergedorfer Straße 10, 21033 Hamburg, Germany
Summary
Correspondence Background Soft tissue trauma and lipomas are common occurrences in surgical Matthias C. Aust. practice. Lipomas are defined as benign tumours of adipose tissue with so far E-mail: [email protected] unexplained pathogenesis and aetiology. A link between preceding blunt soft tissue trauma at the site of the tumour and the formation of lipomas has been Accepted for publication 14 February 2007 described earlier. These soft tissue tumours have been named ‘post-traumatic lipomas’. Key words Objectives In a retrospective review, to analyse all patients with benign adipose haematoma, lipomas, post-traumatic, tissue tumours treated at our institution between August 2001 and January 2007. pseudolipomas, soft tissue tumour Methods All cases were reviewed regarding medical history, magnetic resonance Conflicts of interest imaging findings, intraoperative findings, clinical chemistry and histology. None declared. Results In 170 patients presenting with lipomas, 34 lipomas in 31 patients were identified as post-traumatic. The mean ± SD age of the patients with post- traumatic lipomas was 52 ± 14Æ5 years. The mean time elapsed between soft tissue trauma and lipoma formation was 2Æ0 years (range 0Æ5–5). Twenty-five of the 31 patients reported an extensive and slowly resolving haematoma after blunt tissue trauma at the site of lipoma formation. The mean ± SD body mass index ) was 29Æ0±7Æ6kgm 2. Fourteen of 31 patients presented with an elevated par- tial thromboplastin time. Eleven of 34 lipomas were found on the upper extrem- ities, five on the lower extremities, 13 on the trunk, and two on the face. All tumours were located subcutaneously, superficial to the musculofascial system. Thirty-three lipomas were removed by surgical excision and one by liposuction following an incisional biopsy. Histological examination revealed capsulated and noncapsulated benign adipose tissue in all 34 tumours. Conclusions The existence of a pathogenic link between blunt soft tissue trauma and the formation of post-traumatic lipomas is still controversial. Two potential mech- anisms are discussed. Firstly, the formation of so-called post-traumatic ‘pseudo- lipomas’ may result from a prolapse of adipose tissue through fascia induced by direct impact. Alternatively, lipoma formation may be explained as a result of preadipocyte differentiation and proliferation mediated by cytokine release follow- ing soft tissue damage after blunt trauma and haematoma formation.
Lipomas are benign soft tissue tumours composed of mature mediastinal, gastrointestinal or intraneural location.1,3,4 adipocytes. With an incidence of 1%, lipomas represent the Although intramuscular lipomas are described as benign most frequent benign mesenchymal tumours.1,2 The aetiology lesions in the literature, there is a current consensus to con- and pathogenesis are still unknown; however, they occur sider subfascial lipomas as low-grade liposarcomas.5 Arising more frequently in obese patients. Lipomas are composed of from subcutaneous fat tissue, lipomas usually present as single mature adipocytes and connective tissue that may create a or multiple well-defined, slowly growing, painless soft tissue multilobular appearance by forming fibrous septa. Predomin- tumours, that are nonadherent to the overlying skin and may antly located in the subcutaneous tissue layer, lipomas rarely be easily removed. Lipomas may become symptomatic due to present in an intermuscular, subfascial, retroperitoneal, tumour growth and subsequent neural irritation by mechanical
2007 The Authors 92 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 Post-traumatic lipomas, M.C. Aust et al. 93 pressure. The clinical diagnosis is in most cases apparent, following a blunt soft tissue trauma of the tumour site. Patient although ultrasound or magnetic resonance imaging (MRI) charts and photographic documentation were reviewed retro- may be used on special occasions. Therapy mostly consists of spectively. Medical history, diagnostic examinations including surgical excision of the lipoma including its capsule or aspir- MRI and ultrasound, clinical chemistry and histopathology ation lipectomy of the tumour (Figs 1 and 2).6–8 were analysed. Post-traumatic lipomas are adipose tissue tumours forming at the location of a preceding blunt soft tissue trauma at Diagnostics and therapy between 5 months and 6 years after trauma. In contrast to idiopathic lipomas, a marked gender predominance towards The presence of a post-traumatic lipoma was diagnosed women (12 : 1) has been described. Post-traumatic lipomas with the clinical finding of a clearly defined, soft, mobile, are predominantly localized at the lower extremity, and at the nonadherent mass in subcutaneous tissue and the history of trochanteric and gluteal regions.6,9,10 Although there is no a preceding blunt soft tissue trauma at exactly the same proof of a specific hypothesis of origin, several mechanisms location. Diagnostic tests included ultrasound and MRI if are discussed. Post-traumatic lipoma may develop due to a the tumour had a subfascial location, if malignancy was prolapse of adipose tissue following blunt soft tissue trauma. suspected, or if clinical signs of nerve compression were This has also been called pseudolipoma, as there is no real present. Preoperative tests included routine clinical labora- de novo growth of adipose tissue. This theory of pseudolipoma tory tests for changes in fat metabolism, glucose metabolism formation was first proposed by Brooke and MacGregor in and coagulation. As the charts were reviewed retrospectively 1969.6 They reported several cases with post-traumatic pro- only available clinical laboratory data could be extracted and lapse of adipose tissue through Scarpa’s fascia resulting in analysed. pseudolipomas at the site of blunt soft tissue trauma. Meggitt In 30 patients, the lipomas were removed by surgical exci- and Wilson reported 12 patients with pseudolipomas follow- sion; in one case liposuction was performed after an incisional ing trauma.10 biopsy. All lipomas were completely removed and the tissue The second theory is based on the effect of cytokines and submitted for histopathology. growth factors released by platelets, macrophages and fibro- blasts in the haematoma and the local inflammation after blunt Statistics soft tissue trauma. These effects may lead to differentiation of preadipocytes into mature adipocytes forming new adipose Statistical analysis was performed using the v2 test where tissue.1–3,6,11 appropriate. P £ 0Æ05 was considered statistically significant. In an adult organism new adipocytes originate from pre- cursor cells, so-called preadipocytes. This differentiation is Results promoted by several inflammatory mediators and growth factors released by cells and platelets in post-traumatic Demographics haematoma.9,12 Potential cellular and molecular signal cas- cades leading to terminal differentiation of preadipocytes The demographic data of all patients are shown in Tables 1–3. have previously been described.13 Signorini and Campiglio There were 21 women and 10 men (gender ratio of 2Æ1:1) first presented the hypothesis of de novo formation of adipose with post-traumatic lipomas as compared with 100 women tissue by stimulation of preadipocytes.9 In addition, sex and 39 men (2Æ6:1) in the group with idiopathic lipomas. hormones may influence the proliferation and differentiation The mean ± SD age of 52 ± 14Æ5 years was not different of adipocytes and thus the formation of adipose tissue. The compared with 44 ± 15Æ5 years in the idiopathic group. stimulation of female subdermal preadipocytes with estradiol The time between the preceding soft tissue trauma and pre- in vitro led to a higher proliferation rate when compared sentation with the tumour mass ranged between 6 months with male preadipocytes. Hence, the local estradiol concen- and 5 years with a mean ± SD interval of 2Æ0±2Æ8 years after tration in adipose tissue may present an effective stimulus the trauma (see Table 2). Twenty-five of 31 patients (81%) in to the proliferation of preadipocytes.12,14 The purpose of the study group reported the presence of an extensive post- our retrospective study was to analyse patients presenting traumatic haematoma at the later lipoma site. In 30 patients with a lipoma and a concomitant preceding blunt soft tissue (33 lipomas) the tumour was surgically excised with primary trauma. closure and in one case the tumour was removed by liposuc- tion after incisional biopsy. None of the patients presented Patients and methods with postoperative recurrence of the lipomas.
Patients Histology
Between August 2001 and January 2007, 170 patients present- Histology revealed no signs of malignancy in any tumour. ing with a lipoma were treated at our institution. Thirty-one Twenty-eight of the excised lipomas (82%) had a fibromatous of these patients (with 34 lipomas) presented with lipomas capsule and were designated as real lipomas. In six tumours
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 94 Post-traumatic lipomas, M.C. Aust et al.
(a) (b)
(c) (d)
(e) (f)
Fig 1. Post-traumatic lipoma of the sacral area 3 years after a car accident (d–f). (a–c) Pseudolipoma of the lower abdomen.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 Post-traumatic lipomas, M.C. Aust et al. 95
(a) (b)
(c) (d)
(e) (f)
Fig 2. (a, b) Post-traumatic lipoma of the left cheek. (c, d) Magnetic resonance imaging scan demonstrates signal-intensified soft tissue mass. (e, f) Clinical appearance at 12 months.
(18%) no fibrous capsule was present, and thus the tumours Clinical chemistry were defined as pseudolipomas (Fig. 3). Hypercholesterolaemia was present in five patients (16%) with post-traumatic lipomas compared with 11 patients with lipo- Body mass index mas of idiopathic origin (8%, P £ 0Æ05). Fourteen patients The mean ± SD body mass index (BMI) of patients with (45%) with post-traumatic lipoma had an elevated spontan- ) post-traumatic lipomas (n = 31) was 29Æ0±7Æ6kgm 2 eous partial thromboplastin time (PTT) exceeding the standard compared with the idiopathic control group with a BMI laboratory value of 26–36 s compared with 30 of 139 patients ) of 28Æ3±7Æ6kgm 2. This difference was not statistically (22%) with idiopathic lipomas (P £ 0Æ05). Other coagulation significant. tests showed no significant differences.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 96 Post-traumatic lipomas, M.C. Aust et al.
Table 1 Demographic data of patients with idiopathic and post- Table 3 Clinical chemistry and histology traumatic lipoma Post- Idiopathic Idiopathic traumatic P-value group Post-traumatic group group (n = 139) group (n = 31) P-value Clinical chemistry n = 139 n =31 Gender ratio (female ⁄male) 2Æ6:1 2Æ1:1 NS Hypercholesterolaemia 11 (8%) 5 (16%) £ 0Æ05 )1 Age (years), mean ± SD 44 ± 15Æ552±14Æ5NS (> 230 mg dL ) ) Body mass index (kg m 2), 28Æ3±7Æ629Æ0±7Æ6NS Elevated partial 30 (22%) 14 (45%) £ 0Æ05 mean ± SD thromboplastin time (> 36 s) Prothrombin time > 130% 4 (3%) 2 (6%) NS 9 )1 NS, not significant. Platelet count < 150 · 10 L 6 (4%) 3 (10%) NS Histology n = 139 n =34 Encapsulated––lipoma 120 (86%) 28 (82%) NS Nonencapsulated––pseudolipoma 19 (14%) 6 (18%) NS Discussion NS, not significant. The aetiology and pathogenesis of lipomas remain unclear despite their frequent occurrence in clinical practice. Mechan- ical, endocrine and inflammatory mechanisms have been dis- trauma may be distinguished from spontaneous formation of cussed. Recently, a genetic defect with translocation or partial lipomas, both rarely cause functional impairment but may loss of chromosome 12 has also been suggested as a factor cause an aesthetic embarrassment. contributing to the formation of lipomas.1–3,11 Although pre- In the present study we report 31 patients with 34 post- sentation of lipomas at the site of a preceding blunt soft tissue traumatic lipomas. There are only few reports connecting soft
Table 2 Patient data in 31 cases of post-traumatic lipomas
Time from Age trauma to Patient Sex (years) tumour Type of trauma Localization Size (cm) Treatment Histology 1 F 64 1 year Blunt trauma Toe 2 · 1 Excision Lipoma 2 F 39 6 months Car accident Neck 11 · 7 Excision Lipoma 3 F 74 4 years Blunt trauma Groin 1 · 1 Excision Lipoma 4 F 56 4 years Blunt trauma Forearm 9 · 8 Excision Pseudolipoma 5 F 18 6 months Hit during fight Forehead 3 · 3 Excision Lipoma 6 F 51 2 years Blunt trauma Neck 15 · 20 Liposuction Pseudolipoma 7 M 38 1 year Blunt trauma Neck 6 · 5 Excision Lipoma 8 F 64 3 years Blunt trauma Back 6 · 5 Excision Lipoma 9 F 44 2 years Blunt trauma Neck 12 · 14 Excision Lipoma 10 F 32 1 year Blunt trauma Thigh 8 · 7 Excision Lipoma 11 F 64 4 years Car accident Chest 4 · 3 Excision Pseudolipoma 12 F 47 5 years Blunt trauma Upper arm 7 · 6 Excision Lipoma 13 F 33 1 year Blunt trauma Face 2 · 2 Excision Lipoma 14 M 64 2 years Car accident Arm (multiple) 4 · 4 Excision Lipoma 15 F 73 2 years Car accident Back 16 · 7 Excision Pseudolipoma 16 F 41 1 year Car accident Forearm 3 · 2 Excision Lipoma 17 M 62 1 year Blunt trauma Hand 5 · 4 Excision Lipoma 18 M 40 3 years Blunt trauma Back 10 · 10 Excision Lipoma 19 F 52 2 years Car accident Forearm 6 · 4 Excision Lipoma 20 F 55 6 months Car accident Chest 7 · 7 Excision Lipoma 21 M 55 1 year Car accident Lower leg 4 · 3 Excision Lipoma 22 F 51 1 year Blunt trauma Lower leg 3 · 6 Excision Lipoma 23 F 60 6 months Blunt trauma Forearm 2 · 2 Excision Lipoma 24 F 57 2 years Car accident Hand 2 · 4 Excision Pseudolipoma 25 F 60 1 year Blunt trauma Chest 4 · 6 Excision Lipoma 26 M 46 3 years Car accident Forearm 2 · 2 Excision Lipoma 27 M 51 6 months Blunt trauma Chest 4 · 2 Excision Pseudolipoma 28 F 57 1 year Blunt trauma Chest 8 · 8 Excision Lipoma 29 M 53 4 years Car accident Lower leg 4 · 3 Excision Lipoma 30 M 59 6 months Blunt trauma Back 7 · 6 Excision Lipoma 31 M 52 4 years Hit during fight Lower leg 3 · 2 Excision Lipoma
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 Post-traumatic lipomas, M.C. Aust et al. 97
(a) (b)
(c)
(d)
Fig 3. Post-traumatic lipoma of the first webspace of the right hand after blunt tissue trauma. (a, b) Macroscopic presentation of the tumour demonstrating a defined capsule. (c) Wound closure by Z-plasty to avoid web contracture. (d) Histology demonstrates mature adipocytes and connective tissue (haematoxylin and eosin).
tissue trauma to the development of fatty soft tissue tumours. true lipomas, these pseudolipomas have no surrounding cap- Adair et al.15 first suggested a link between trauma and benign sule. In pseudolipomas only the fat lobules of the deep subcu- fatty tissue tumours in patients after severe trauma as early as taneous fat layer prolapse through Scarpa’s fascia as described 1932. Soft tissue trauma leading to a prolapse of adipose tis- by several authors.6,10,16 In addition, chronic local compres- sue through Scarpa’s fascia forming pseudolipomas was first sion causing anatomical defects in the fascia may also result in described by Brooke and MacGregor in 1969.6 In contrast to pseudolipoma formation. Herbert and DeGeus reported a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 98 Post-traumatic lipomas, M.C. Aust et al. patient with an abdominal lipoma induced by wearing tight localization. In these cases MRI scans may give valuable infor- trousers.17 An extensive study on the origin of post-traumatic mation to discriminate between lipomas and liposarcomas. lipomas postulated that subcutaneous scar formation and con- The removal of post-traumatic lipomas is usually performed tracture after a soft tissue trauma may also lead to the devel- by surgical excision or by liposuction.6–22 However, prior to opment of pseudolipomas.7 liposuction an excisional biopsy is recommended to exclude On the other hand, 28 of our patients presented with malignant liposarcoma.9 In our series 33 post-traumatic lipo- encapsulated lipomas without any lesions in Scarpa’s fascia. mas were treated by excision and one by liposuction. The sur- Thus the mechanism explained above appears unlikely. gical excision was chosen because of the small size of these Another explanation for formation of lipomas is the induction lipomas. In large extensive lipomas liposuction may be a less of the differentiation of preadipocytes to mature adipocytes by invasive and safe method of tumour removal, although it has cytokine–cell interaction at the trauma site.9 This mechanism the disadvantage of leaving the fibrous capsule behind. As was first discussed by Signorini and Campiglio, who hypothes- none of our patients presented with a local recurrence of lipo- ized that post-traumatic lipomas may be formed by terminal mas we cannot determine if the latter approach increases the proliferation of adipocyte precursor cells and de novo formation rate of recurrence. of adipose tissue. They further investigated the effect of local In conclusion, the mechanisms leading to the formation of and systemic cytokines and growth factors on the proliferation post-traumatic lipomas remain unclear. Although the prolapse of preadipocytes. This hypothesis is very well illustrated by of adipose tissue through a post-traumatic lesion in Scarpa’s the so-called ‘lipoma arborescens’. This subsynovial adipose fascia has been discussed by several authors, it only accounts tissue proliferation is induced by chronic inflammation.18 for a small number of post-traumatic lipomas (18%). Most of Inflammatory cytokines and mediators are released by dam- the post-traumatic lipomas in our series were true encapsulated aged and necrotic cells after blunt soft tissue trauma.5,8 This lipomas. Thus it appears more likely that the formation of the triggers a cascade of further secretion of other pro- and latter must be considered a true proliferative tumour growth. anti-inflammatory cytokines and growth factors, which stimu- Most likely this proliferation and de novo growth of adipose tis- late the proliferation of preadipocytes. After blunt soft tissue sue is triggered by release of local and systemic cytokines and injury, localized adipose tissue necrosis at the site of a persist- growth factors from haematoma and local tissue necrosis. ing extensive haematoma followed by local formation of lipo- Further clinical and experimental studies are necessary to mas has been described by Copcu and Sivrioglu.19 Other clarify the pathogenesis and the mechanism of post-traumatic authors have previously reported a correlation between soft lipoma formation. tissue trauma covering an area of > 100 cm2, prolonged haematoma, adipose soft tissue necrosis, and the occurrence References of lipomas.12,20 Adipose tissue necrosis and severe haemato- mas have been described to increase the risk of post-traumatic 1 Allen PW. Tumors and Proliferations of Adipose Tissue. New York: Masson, lipomas by stimulation of preadipocyte proliferation.9,12,20 1981; 318–24. Interestingly, we found an elevated spontaneous PTT in 14 2 Enzinger FM, Weiss SW. Benign lipomatous tumor. In: Soft Tissue Tumors (Enzinger FM, Weiss SW, eds), 3rd edn. St Louis: Mosby, of 31 patients with post-traumatic lipomas. A link between 1995; 381–92. manifest or subclinical coagulation disorders, such as elevated 3 Anders KH, Ackermann AB. Neoplasms of the subcutaneous fat. In: coagulation tests like an elevated PTT and the development of Fitzpatrick‘s Dermatology in General Medicine (Freedberg IM, Eisen AZ, post-traumatic lipomas, has not been previously described. Wolff K, Austen KF, Goldsmith LA, Katz SI et al., eds), 5th edn. In our population, patients with post-traumatic lipomas pre- New York: McGraw-Hill, 1999; 1292–3. ) sented with a mean ± SD BMI of 29Æ0±7Æ6kgm 2 com- 4 Austin M, Mack GR, Townsend CM. Infiltrating (intramuscular) ) pared with 28Æ3±7Æ6kgm 2 in patients with idiopathic lipomas and angiolipomas: a clinicopathologic study of six cases. Arch Surg 1980; 115:281–4. lipomas. Although a correlation between obesity and the inci- 4 5 McPherson T. Benign tumors of fibrous tissue and adipose tissue dence of lipomas has been previously suggested, we could of the hand. J Hand Ther 2005; 18:53–4. not demonstrate this correlation. Only one patient in our 6 Brooke RI, MacGregor AJ. Traumatic pseudolipoma of the buccal study had four lipomas, all of them encapsulated lipomas mucosa. Oral Surg Oral Med Oral Pathol 1969; 28:223–5. localized on the arm. This appears to be exceptional as only 7 Rozneri L, Isaacs GW. The traumatic pseudolipoma. Aust N Z J Surg one single case with multiple post-traumatic lesions has previ- 1977; 47:779–82. ously been reported.9,17 8 Penoff JH. Traumatic lipomas ⁄pseudolipomas. J Trauma 1982; 22:63–5. At initial presentation, four patients had had a previous 9 Signorini M, Campiglio GL. Posttraumatic lipomas: where do they ultrasound examination of the lipomas and seven patients an come from? Plast Reconstr Surg 1998; 101:699–705. MRI scan. If there is any clinical doubt about the type of 10 Meggitt BF, Wilson JN. The battered buttock syndrome: fat frac- lesion encountered, we consider ultrasound the method of tures: a report on a group of traumatic lipomata. Br J Surg 1972; choice as it is inexpensive, noninvasive, highly sensitive, spe- 59:165–9. cific and reliable.9,21 However, MRI may be necessary for a 11 Turc Carel C, Dal Cin P, Boghosian L. Breakpoints in benign lip- more precise diagnosis and differentiation of the tumour, with oma may be at 12q13 or 12q14. Cancer Genet Cytogenet 1988; 36:131–4. recurring lesions or with suspected subfascial or intramuscular
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 Post-traumatic lipomas, M.C. Aust et al. 99
12 Salaus IB, Cushman SW, Weismann RE. Studies of human adipose 18 Tiao WM, Yeh LR, Lu YC et al. Lipoma arborescens of the knee: tissue; adipose cell size and number of nonobese and obese a case report. J Formos Med Assoc 2001; 100:412–15. patients. J Clin Invest 1973; 52:929–32. 19 Copcu E, Sivrioglu NS. Posttraumatic lipoma: analysis of 10 cases 13 Ailhaud G, Grimaldi P, Negrel R. Cellular and molecular aspects of and explanation of possible mechanisms. Dermatol Surg 2003; adipose tissue development. Annu Rev Nutr 1992; 12:207–33. 29:215–20. 14 Gregoire FM. Adipocyte differentiation: from fibroblast to endo- 20 Weiss SW, Goldblum JR (eds). Enzinger and Weiss’s Soft Tissue Tumors, crine cell. Exp Biol Med 2001; 226:997–1002. 4th edn. St Louis: Mosby, 2001; 578–80. 15 Adair FE, Pack GT, Parrior JH. Lipoma. Am J Cancer 1932; 16:1104–6. 21 David LR, DeFranzo A, Marks M. Posttraumatic pseudolipoma. 16 Elsahy NI. Post-traumatic fatty deformities. Eur J Plast Surg 1989; J Trauma 1996; 40:396–400. 12:208–11. 22 Dodenhoff TT. Trauma induced saddle-bag: case-report. Lipoplasty 17 Herbert DC, DeGeus J. Post-traumatic lipomas of the abdominal Newsletter 1988; 5:55–7. wall. Br J Plast Surg 1975; 28:303–6.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp92–99 EPIDEMIOLOGY AND HEALTH SERVICES RESEARCH DOI 10.1111/j.1365-2133.2007.07969.x Impetigo in epidemic and nonepidemic phases: an incidence study over 4½ years in a general population S. Rørtveit and G. Rortveit* Municipal Health Services of Austevoll Kommune, 5399 Bekkjarvik, Norway *The Section for General Practice, Department of Public Health and Primary Health Care, University of Bergen, Kalfarv. 31, 5018 Bergen, Norway The Research Unit for General Practice, Unifob Health, Bergen, Norway
Summary
Correspondence Background Little is known about incidence and natural variation of impetigo in Sverre Rørtveit. general populations. E-mail: [email protected] Objectives To investigate the natural course of impetigo in a well-defined popula- tion, and to study the resistance pattern of the causal bacteria over time. Accepted for publication 13 February 2007 Methods This is a population-based incidence study in Austevoll, an island commu- nity of 4457 inhabitants in Norway, in the years 2001–2005. Incidence rates are Key words given as events per person-year. Epidemic periods were identified by statistical epidemic, fusidic acid, impetigo, population based, process-control analyses. skin diseases, staphylococcal infections Results The incidence rate of impetigo for the whole study period was 0Æ017 Conflicts of interest events per person-year, corresponding to a total of 334 cases. The incidence rates None declared. were 0Æ009, 0Æ026, 0Æ019, 0Æ016 and 0Æ009 in the years 2001, 2002, 2003, 2004 and 2005, respectively. Three epidemics were identified, starting in August of 2002, 2003 and 2004, lasting for 11, 11 and 5 weeks, respectively. Incidence rates in these epidemic periods were 0Æ099, 0Æ045 and 0Æ074, respectively. In epidemic periods, Staphylococcus aureus was the causal bacterium in 89% (117 ⁄132) of cases, while this proportion was 68% (84 ⁄123) in nonepidemic periods (P <0Æ01). Staphylococcus aureus was resistant to fusidic acid in 84% (98 ⁄117) and 64% (54 ⁄84) of impetigo cases in epidemic and nonepidemic periods, respect- ively (P <0Æ01). When investigating all types of infections caused by S. aureus in the study period, the proportion of fusidic acid resistance in impetigo cases (152 ⁄201, 76%) differed significantly from fusidic acid resistance in other infec- tions (18 ⁄116, 16%) (P <0Æ01). Conclusions Distinctive epidemic outbreaks occurred during the summer of three of the five follow-up years. In outbreaks, S. aureus was more frequently the causal agent and the sensitivity to fusidic acid decreased significantly.
There are limited data on the natural incidence and fluctuations In the summer of 2002, an epidemic outbreak of impetigo of impetigo in populations, even though it is the most common was observed in Norway. The causal bacterium for the epi- skin infection in children.1 A study published from the Nether- demic was shown to be Staphylococcus aureus, and most isolates lands reported the incidence rate in a random selection of gen- of the bacterium were shown to be resistant to fusidic eral practices to be 0Æ017 events per person-year in 1987 and acid.7,11 We have previously reported on this outbreak in 0Æ021 in 2001.2 A study from Britain, giving incidence data Austevoll, an island community on the Western coast.12 The from sentinel general practices over the years 1999–2003, gives aim of the present study was to describe the natural course of the total incidence in the whole study period as 18Æ7 events per primary impetigo in a well-defined population and the resist- week per 100 000 inhabitants, corresponding to 0Æ01 events ance pattern of the causal bacteria over time. The population per person-year.3 Apart from studies of selected populations in Austevoll was followed over a 4½-year period. from general practice or specialist clinics,4–6 most studies of the epidemiology of impetigo have been based on bacterial Materials and methods samples received at microbiological laboratories.7–10 As far as we know, no previous study has investigated continuous vari- The study was conducted as a population-based investigation ation of the incidence of impetigo in a general population. in Austevoll, an island municipality in Western Norway. Eight
2007 The Authors 100 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 Incidence of impetigo, S. Rørtveit and G. Rortveit 101 of the islands are habitated, some of them interconnected by In order to compare resistance patterns, we also identified bridges, but ferries and boats are the only connection with two other groups of patients: (i) those with superficial skin in- the mainland. There are four general practitioners in the fections other than impetigo, and (ii) those where S. aureus had municipality. Norway has a registered list system, and few been cultured from infections more invasive than impetigo and inhabitants see a general practitioner outside Austevoll. other superficial skin infections. Cases of superficial skin infec- For study purposes, we made an operational definition of tions other than impetigo were identified using the four search impetigo as being a superficial skin infection with spontan- terms mentioned above when cases were found not to conform eous erosions and ⁄or crusts and ⁄or bullae. This definition was to the definition of impetigo given in this study. Cases with based on literature study. The basis for inclusion of impetigo cultures resulting in growth of S. aureus from more invasive cases was a diagnosis of impetigo made by a doctor. Specific infections were identified by applying the name of the bacter- criteria for the diagnosis were not, however, discussed among ium as a search term in the laboratory section, when cases with the general practitioners. If the doctor’s description of the impetigo or other superficial skin infections were excluded. eruption in the patient record explicitly deviated from our Examples of such infections are deeper skin infections, infected definition, the patient was not included as an impetigo case traumatic wounds, paronychia and chronic leg ulcers. Extensive (n = 32). Likewise, patients where the diagnosis of impetigo attempts to identify all cases of nonimpetigo superficial skin had been revised in a consecutive consultation were not inclu- infections and invasive infections potentially caused by S. aureus ded (n = 11), nor were cases with secondary infection of were not made; however, the application of the name of the other skin diseases (n = 1). Each impetigo event was counted. bacterium as one of the search terms enabled the finding of all However, patients who consulted their doctor again for impe- cases where the bacterium was cultured. tigo within 6 months of an event were regarded as having Throughout the study period all the bacteriological specimens recurrence of the primary episode. These recurrent episodes were investigated at the Department of Microbiology and were not counted again. Immunology, Haukeland University Hospital, Norway. Fusidic An electronic medical journal (EMJ) was introduced in acid susceptibility was determined by disc diffusion. Before Austevoll in June 2001. The EMJ system used is Infodoc and January 2005, the laboratory used equipment supplied by the all general practitioners of the municipality share information AB Biodisk company (Solna, Sweden). The discs contained fusi- ) on patients’ medical records. The same system is used for day- dic acid 50 lgL 1 and were applied to PDM agar. An inhibitory time appointments and after-hours care. The EMJ can be zone of > 32 mm was recorded as sensitivity to fusidic acid, searched for individual search terms in any given record sec- and an inhibitory zone £ 32 mm as resistance to fusidic acid. tion (history, clinical findings, laboratory results, treatment From January 2005, the laboratory has been using antibiotic and diagnosis, and the doctors’ patient schedule). discs and agar plates supplied by Becton Dickinson (Trond- ) Identification of impetigo cases was done in three different heim, Norway). Discs of 10 lgL1 are applied to Muller ways according to time period. In the period week 27 of Hinton II agar. An inhibitory zone of > 24 mm is recorded as 2001–week 28 of 2002, cases were identified by applying the sensitivity to fusidic acid, whereas an inhibitory zone search terms ‘impetigo’, ‘brennkopper’ (the Norwegian term £ 24 mm is recorded as resistance to fusidic acid. for impetigo), ‘skin infection’ and ‘pyoderma’ in electronic These procedural evaluations both correspond to a mini- searching of the diagnosis section, and ‘Staphylococcus aureus’in mum inhibitory concentration breakpoint at 0Æ5. The methods electronic searching of the laboratory results section. In the were compared to ensure that there would be no difference in period week 29–52 of 2002, cases were registered by reading susceptibility categorization as a consequence of the change of all sections of all patient records from all consultations with method, and were found to perform equally. Apart from this, all the doctors in the community. Cases evidently evaluated by there was no change in the standard operating procedure for the doctor as representing impetigo were included when they culture ⁄sensitivity testing. Laboratory results received electron- did not deviate from the operational definition, irrespective of ically ensured the precision of bacteriological data. which diagnostic term was given. For validation purposes, all In 2003, the Norwegian Medicines Agency published diagnostic terms used in the records were registered as part of national guidelines for the treatment of impetigo taking this hand search. Four diagnoses, identifying all cases except account of the increased frequency of S. aureus being resistant two, were subsequently applied in an electronic search in to fusidic acid.11 Local treatment with bacitracin or dibrom- order to find all cases. This resulted in the identification of propamidine ointment was recommended and, in cases where seven additional cases. These diagnoses are identical to the oral treatment would be preferable, dicloxacillin. All general search terms described above, which were applied in the elec- practitioners in Austevoll have been aware of these guidelines tronic search for the first time period. From 2003 onwards, during the study period. the general practitioners agreed on attributing the term ‘impe- All general practitioners agreed on sending specimens for tigo’, or the corresponding Norwegian term, to cases of impe- bacteriological culture in as many impetigo cases as possible tigo, for ease of identification. Accordingly, cases of impetigo from week 32 of 2002 onwards. Data on the pathogen and occurring from week 1 of 2003 were identified by an elec- resistance pattern were recorded for each case of impetigo, tronic search in the diagnosis section for these two terms at along with age, sex, week of first consultation, year, localiza- the end of each week. tion of the rash, and initial treatment. In order to calculate
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 102 Incidence of impetigo, S. Rørtveit and G. Rortveit
Fig 1. Cases of impetigo per week. UCL, upper control limit. age-specific incidence rates, age was categorized in Statistics Ethical approval Norway age categories.13 When appropriate, age categories were merged to simplify the presentation. According to the Regional Committees for Research Ethics in Incidence rates were calculated as (n ⁄N) ⁄(w ⁄52) where Norway, ethical approval was not required because all data n = number of new impetigo events, N = number of people were collected from medical records obtained as part of regu- in the population, w = number of weeks in the observation lar visits by patients, and treatment procedures were in period and 52 being the number of weeks in a full year. accordance with usual Norwegian practice. The first author Accordingly, incidence rates are given per person-year. The had access to these records due to his position as a general number of people in the population changed slightly in the practitioner in the municipality. The second author had access study period, from 4439 to 4487, and we used the mean to anonymous data only. number to calculate incidence rate (N = 4457).13 To explore variation in incidence over time, modified statis- Results tical process-control (SPC) analyses were performed (EpiData Analysis 1.1).14 Four periods of reliable endemic phases The incidence rate of impetigo for the total study period was were established by visual inspection of the curve shown in 0Æ017 events per person-year, corresponding to 334 cases. Fig. 1 (week 27 ⁄2001–week 31 ⁄2002, week 43 ⁄2002–week Table 1 shows the variation in incidence by year and 6-month 26 ⁄2003, week 42 ⁄2003–week 29 ⁄2004, week 44 ⁄2004– week 52 ⁄2005). Running I-charts for each of these endemic Table 1 Incidence rates (per person-year) in the total population (N = 4457) according to year and half yeara periods showed that the mean weekly frequency for each of them was 0Æ63, 0Æ71, 0Æ78 and 0Æ81 cases, respectively. In Year Incidence n Half year Incidence n SPC, incidence levels superior to mean + 3 SD of endemic a phase are commonly viewed as beginning epidemics and are 2001 0Æ009 20 Second 0Æ009 20 Æ Æ given the denomination ‘upper control limit’.15 I-chart analy- 2002 0 026 115 First 0 006 13 Second 0Æ046 102 ses yielded upper control limits of the established endemic 2003 0Æ019 85 First 0Æ008 18 phases as 2Æ85, 3Æ61, 2Æ96 and 3Æ38 cases per week, respect- Second 0Æ030 67 ively. We defined the start of an epidemic as the first of three 2004 0Æ016 72 First 0Æ006 13 or more consecutive weeks exceeding the upper control limit, Second 0Æ026 59 and the end as the week before the start of three consecutive 2005 0Æ009 42 First 0Æ009 19 weeks below the upper control limit. Second 0Æ010 23 Æ v2 tests were performed to test differences between propor- All 0 017 334 tions using SPSS software version 13.0 (SPSS, Chicago, IL, aNo data from first half of 2001. U.S.A.). Statistical significance was accepted at the 0Æ05 level.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 Incidence of impetigo, S. Rørtveit and G. Rortveit 103 period. The SPC analysis revealed three epidemics of impetigo fusidic acid in 76% (152 ⁄201). However, there was consid- starting in August of 2002, 2003 and 2004, as shown in erable variation, with S. aureus being the causal agent signifi- Fig. 1. The epidemic lasted 11 weeks in years 2002 and 2003 cantly more frequently in epidemic as compared with and 5 weeks in 2004, and the incidence rates in these periods nonepidemic periods (Table 3). Also, when S. aureus was the were 0Æ099, 0Æ045 and 0Æ074 events per person-year, respect- causal bacterium, resistance to fusidic acid was significantly ively. In total, the epidemics comprised about one half of all higher in epidemic as compared with nonepidemic periods. the cases of impetigo (166 of 334). Table 2 shows the inci- Group A streptococci were the causal agent in nonepidemic dence rates by year in different age groups. The incidence was periods only (Table 3). by far the highest among children aged 15 years and younger. Compared with impetigo, resistance to fusidic acid was sig- This was the case for both nonepidemic and epidemic periods nificantly lower both in other superficial skin infections and in (data not shown). more invasive infections in cases where S. aureus was the causal Swabs were taken in 255 of 334 (76%) impetigo cases agent (Table 4). Resistance to erythromycin was low in all (Table 3). Staphylococcus aureus was isolated from 79% three infection groups, whereas resistance to penicillin G was (201 ⁄255) of these cases, and the bacterium was resistant to generally high in all groups. The resistance of S. aureus to
Table 2 Incidence rates (per person-year) of impetigo by age group and year
Incidence (n) Age group (years) Population All years 2001a 2002 2003 2004 2005 0–6 467 0Æ059 (124) 0Æ026 (6) 0Æ088 (41) 0Æ081 (38) 0Æ047 (22) 0Æ036 (17) 7–15 651 0Æ039 (115) 0Æ028 (9) 0Æ060 (39) 0Æ041 (27) 0Æ043 (28) 0Æ018 (12) 16–24 594 0Æ014 (37) 0Æ003 (1) 0Æ029 (17) 0Æ010 (6) 0Æ017 (10) 0Æ005 (3) 25–66 2158 0Æ005 (49) 0Æ004 (4) 0Æ008 (17) 0Æ006 (10) 0Æ005 (11) 0Æ003 (7) 67+ 587 0Æ003 (9) 0 (0) 0Æ002 (1) 0Æ007 (4) 0Æ002 (1) 0Æ005 (3)
aNo data from first half of 2001.
Table 3 Incidence rates (per person-year) and microbial characteristics of impetigo in epidemic and nonepidemic periods
Fusidic acid Group A Swabs takena S. aureusb resistance (S. aureus)c streptococcib
Period Incidence Cases, n (%) n (%) P-valued n (%) P-valued n (%) P-valued n (%) P-valued Epidemic 0Æ071 166 (49Æ7) 132 (80) 0Æ18 117 (89) < 0Æ01 98 (84) < 0Æ01 0 (0) < 0Æ01 Nonepidemic 0Æ009 168 (50Æ3) 123 (73) 84 (68) 54 (64) 25 (100) All 0Æ017 334 (100) 255 (76) 201 (79) 152 (76) 25 (100)
S. aureus, Staphylococcus aureus. aProportion of swabs taken per number of incident cases. bProportion of swabs with S. aureus (or group A strepto- cocci, last column) per number of swabs taken. cProportion of fusidic acid resistance per number of swabs with S. aureus. dP-value for v2 test (comparison between epidemic and nonepidemic periods).
Table 4 Proportion of resistance to fusidic acid, erythromycin and penicillin G related to type of infection, in cases where Staphylococcus aureus was the causal pathogen
Fusidic acid resistance Erythromycin resistance Penicillin G resistance Type of infection n (%) P-valuea n (%) P-valuea n (%) P-valuea Impetigo (n = 201) 152 (76) 13 (7) 190 (96) Other superficial skin infections (n = 34) 6 (18) < 0Æ001 1 (3) 0Æ43 25 (76) < 0Æ001 Invasive infections (n = 82) 12 (15) < 0Æ001 2 (2) 0Æ17 60 (73) < 0Æ001
Two cases of impetigo with intermediate sensitivity for erythromycin, and one invasive infection case with intermediate sensitivity for peni- cillin G, were defined as sensitive in these analyses. Resistance to erythromycin was not tested in three cases of impetigo, one other super- ficial skin infection and one invasive infection. Resistance to penicillin G was not tested in four cases of impetigo and one case of other superficial infection. aP-value for v2 test (pairwise comparison of resistance pattern in impetigo vs. other superficial skin infections or inva- sive infections).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 104 Incidence of impetigo, S. Rørtveit and G. Rortveit penicillin G was significantly higher in impetigo than in other cial skin infections in children declined from almost 50% in infections (Table 4). Among the isolates of S. aureus, one was 2002 to about 20% in 2004Æ10 Their results indicate that a methicillin resistant. specific clone of S. aureus was responsible for the epidemic of impetigo in Scandinavia from the late 1990s, that the epi- Discussion demic reached its peak in 2002, and that it is now reced- ing.8,10 One cannot know whether the isolates of S. aureus As far as we know, this is the first study to report variations from Austevoll belong to this assumed clone, as no molecular in incidence of impetigo in a well-defined general population typing was done in our study. However, the seasonal variation over a time-span of several years. Well-defined epidemics of incidence of resistance to fusidic acid is the same as the were demonstrated in late summer of three consecutive years variation found by the Swedish researchers. during the study period of 4½ years. Staphylococcus aureus was The data reported by O¨ sterlund et al., being collected retro- the predominant causal agent, and was resistant to fusidic acid spectively from microbiological laboratories, preclude the in the majority of cases. This varied significantly, however, detection of differences between diagnostic categories.10 O¨ sterl- between epidemic and nonepidemic periods. und et al. reported a maximum of almost 50% of resistance to The resistance pattern also differed significantly between fusidic acid of superficial skin infections caused by S. aureus. For S. aureus isolates causing impetigo and S. aureus isolates causing the whole period of our study we found a total of 76% of resist- other infections. The magnitude of this difference is strikingly ance to fusidic acid among impetigo-causing S. aureus. In our high. There was no general recommendation to send bacterio- study, there was an inverse relationship between impetigo and logical cultures for infections other than impetigo. However, other superficial skin infections with respect to resistance to fus- it is very unlikely that this would lead to a bias in the direc- idic acid. The advantages offered by our data being collected in tion of a lower rate of resistance to fusidic acid. real time and with fairly good control of the diagnostic preci- The distance from the island community to the mainland sion give us reason to conclude that the lower proportion of poses considerable restraints to consulting doctors in other resistance to fusidic acid in the Swedish study might result from communities. Accordingly, our data give a unique opportunity homogeneous analysis of heterogeneous diseases. Also, the to observe variations in incidence and microbiological change Swedish study reported less resistance to fusidic acid among with high validity. We presume that some patients with minor patients older than 13 years than in patients of younger age. degrees of the disease did not consult a doctor, resulting in a However, most patients with superficial skin infections other slight underestimation of incidence. than impetigo belong to higher age groups. The incidence in a Dutch national survey was 0Æ017 events The current study adds considerable data to the scarce per person-year in 1987 and 0Æ021 in 20012 which is close to knowledge about impetigo, causal agents and natural fluctua- the incidence of 0Æ017 events per person-year reported for the tions in a general population. Similar studies of well-defined total period in the current study, while in a British study the populations should be conducted in other countries to see if incidence was estimated to 0Æ01 events per person-year,3 even our findings are generalizable to other populations and geo- though the study covered a time period partly overlapping graphical areas. Further epidemiological studies should also that of the present study, and Norway and Britain are fairly provide more detailed information on genetic properties of close geographically. The reason for the lower incidence bacteria causing impetigo. reported from Britain might be that a substantial proportion of patients with impetigo in Britain is seen at specialist clinics Acknowledgments rather than by general practitioners.5 In both the Dutch and the British study there was a seasonal variation, with higher We thank Janecke Thesen, MD, for her support in the work incidence in summer and autumn, which is in line with our with the SPC tool, Haima Mylvaganam, MD, Department of findings. We continuously monitored the level of the disease Microbiology and Immunology, Haukeland University Hospi- in a general population and were therefore able to demon- tal, for permitting the use of microbiological data in this art- strate real-time variability and the existence of epidemics, icle, and the general practitioners of Austevoll who have which exerted a strong influence on the incidence. treated the patients in this study. We have been unable to find a definition of epidemic out- 15 break of impetigo in the literature. Both SPC analysis and References non-SPC types of analysis16 of epidemics establish baseline levels of endemic phase, and use magnitude of deviation from 1 Koning S, van der Wouden JC. Treatment for impetigo. BMJ 2004; the baseline in order to define an epidemic. A statistical tool 329:695–6. for establishing an epidemic was very valuable, and allowed 2 Koning S, Mohammedamin RS, van der Wouden JC et al. Impetigo: incidence and treatment in Dutch general practice in 1987 and us to describe important variations in microbial patterns 2001 – results from two national surveys. Br J Dermatol 2006; between epidemic and nonepidemic phases. 154:239–43. O¨ sterlund et al. reviewed national data from microbiological 3 Elliot AJ, Cross KW, Smith GE et al. The association between impe- laboratories in Sweden and showed that the proportion of tigo, insect bites and air temperature: a retrospective 5-year study resistance to fusidic acid among S. aureus isolates from superfi-
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 Incidence of impetigo, S. Rørtveit and G. Rortveit 105
(1999–2003) using morbidity data collected from a sentinel gen- 10 O¨ sterlund A, Kahlmeter G, Haeggman S, Olsson-Liljequist B. Staphy- eral practice network database. Fam Pract 2006; 23:490–6. lococcus aureus resistant to fusidic acid among Swedish children: a 4 Rogers M, Dorman D, Gapes M, Ly J. A three-year study of impe- follow-up study. Scand J Infect Dis 2006; 38:332–4. tigo in Sydney. Med J Aust 1987; 147:63–5. 11 Statens Legemiddelverk. Terapianbefalinger ved impetigo (brenn- 5 Loffeld A, Davies P, Lewis A, Moss C. Seasonal occurrence of impe- kopper). MSIS-Rapport 2003; 31:1B. tigo: a retrospective 8-year review (1996–2003). Clin Exp Dermatol 12 Rørtveit S, Rortveit G. An epidemic of bullous impetigo in the 2005; 30:512–14. municipality of Austevoll in the year 2002. Tidsskr Nor Laegeforen 6 Koning S, van Suijlekom-Smit LW, Nouwen JL et al. Fusidic acid 2003; 123:2557–60. cream in the treatment of impetigo in general practice: double 13 Statistics Norway. Population Numbers, 1 January 2005. Available at: blind randomised placebo controlled trial. BMJ 2002; 324:203–6. http://www.ssb.no/emner/03/hjulet/tabell-01.html (last accessed 7 Tveten Y, Jenkins A, Kristiansen BE. A fusidic acid-resistant clone 16 May 2007). of Staphylococcus aureus associated with impetigo bullosa is spreading 14 Carey R. Improving Healthcare with Control Charts: Basic and Advanced SPC in Norway. J Antimicrob Chemother 2002; 50:873–6. Methods and Case Studies. Milwaukee: ASQ Quality Press, 2003. 8O¨ sterlund A, Eden T, Olsson-Liljequist B et al. Clonal spread among 15 Arantes A, Carvalho Eda S, Medeiros EA et al. Use of statistical pro- Swedish children of a Staphylococcus aureus strain resistant to fusidic cess control charts in the epidemiological surveillance of nosoco- acid. Scand J Infect Dis 2002; 34:729–34. mial infections. Rev Saude Publica 2003; 37:768–74. 9 Ravenscroft JC, Layton A, Barnham M. Observations on high levels 16 Naumova EN, O’Neil E, MacNeill I. INFERNO: a system for early of fusidic acid resistant Staphylococcus aureus in Harrogate, North outbreak detection and signature forecasting. MMWR Morb Mortal Yorkshire, UK. Clin Exp Dermatol 2000; 25:327–30. Wkly Rep 2005; 54 (Suppl.):77–83.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp100–105 EPIDEMIOLOGY AND HEALTH SERVICES RESEARCH DOI 10.1111/j.1365-2133.2007.07987.x Hairdressing is associated with scalp disease in African schoolchildren N.P. Khumalo, S. Jessop, F. Gumedze* and R. Ehrlich Division of Dermatology, *Department of Statistical Sciences and School of Public Health and Family Medicine, Groote Schuur Hospital and the University of Cape Town, Observatory 7925, South Africa
Summary
Correspondence Background Anecdotal reports suggest that certain disorders are common in African N.P. Khumalo. hair and may be associated with hairstyles. E-mail: [email protected]; Objectives A cross-sectional study of 1042 schoolchildren was performed to test [email protected] this hypothesis. Accepted for publication Methods A questionnaire was administered and scalp examinations performed, after 19 October 2006 ethics approval. Results Participants included 45% boys and 55% girls. The majority of boys, Key words 72Æ8%, kept natural hair with frequent haircuts (within 4 weeks). The prevalence acne keloidalis nuchae, African hair, braids, hair of acne (folliculitis) keloidalis nuchae (AKN) was 0Æ67% in the whole group and grooming, relaxed hair, traction alopecia highest (4Æ7%) in boys in the final year of high school, all of whom had fre- Conflicts of interest quent haircuts. The majority of girls (78Æ4%) had chemically relaxed hair, which None declared. was usually combed back or tied in ponytails, vs. 8Æ6% of boys. Traction alopecia (TA) was significantly more common with relaxed than natural hair, with an overall prevalence of 9Æ4% (98 of 1042) and of 17Æ1% in girls, in whom it increased with age from 8Æ6% in the first year of school to 21Æ7% in the last year of high school. The proportion with TA in participants with a history of braids on natural hair was lower (22Æ9%), but not significantly, than among those with a history of braids on relaxed hair (32Æ1%). No cases of central centrifugal cicatricial alopecia were identified. Conclusions We found associations between hairstyle and disease in our population of schoolchildren. AKN appears to be associated with frequently cut natural hair and TA with relaxed hair. These associations need further study for purposes of disease prevention.
For the purpose of this article, African hair is defined as if the hair is kept short (Fig. 2) or a ‘blow out’ if allowed to tightly curly (spiral) black hair typical of sub-Saharan Africans. grow longer. The latter allows a larger ‘Afro’ than is usually In South Africa, although natural (chemically unaltered) hair- possible, which is popular with some girls. Relaxed hair in styles such as twists, coils and dreadlocks are becoming more this study is inclusive of all the above groups. Some of these popular, chemical straighteners or relaxers are commonly used hairstyles may result in complications such as chemical burns, by females (80% of age group 15–50 years; personal commu- allergic reactions and hair loss.2–4 nication, cosmetic company in-house study). Chemical hair Anecdotal reports suggest that certain scalp disorders, which relaxers are used on African hair and contain alkaline chemi- may be associated with hairdressing, are more common in cals such as sodium hydroxide and guanidine hydroxide.1 individuals of African descent. These are acne (folliculitis) Girls tend to use relaxers for length; the hair is usually keloidalis nuchae (AKN)4 – a scarring alopecia that has a pre- combed back and ⁄or tied in a ponytail to which artificial dilection for, but is not limited to, the nuchal scalp; central extensions may be attached (Fig. 1). Hair braiding of virgin centrifugal cicatricial alopecia (CCCA)4 – previously called or chemically treated hair, with or without hair extensions, is ‘hot comb alopecia’ and ‘follicular degeneration syndrome’; also popular. The same chemicals used as relaxers are occa- and traction alopecia (TA).4,5 We performed a systematic sionally applied to the hair for shorter periods (< 5 vs. 10– review of the evidence for this excess prevalence and found a 20 min) than when relaxing, so as to produce a loose curl paucity of population studies on these conditions.6 We were (wavy) effect popular with some boys, and called the ‘s-curl’ also unable to identify studies that report hairstyle trends and
2007 The Authors 106 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp106–110 Hairdressing and scalp disease in African schoolchildren, N.P. Khumalo et al. 107
Materials and methods
Objectives
There were three primary objectives of this study: (i) to deter- mine the prevalence of various hairstyles in African children attending school in Langa Township in Cape Town; (ii) to determine the prevalence of specific scalp ⁄hair disorders (TA, AKN, CCCA) and the variation with age and gender; and (iii) to describe possible associations between scalp disorders and hairstyle. A secondary objective was to document the presence of any other clinical disorders of hair ⁄scalp in study participants.
Pilot study
In view of the previously mentioned lack of data, a pilot study which included all age groups was conducted at a primary Fig 1. Relaxed hair tied into a ponytail to which an additional straight healthcare centre in Cape Town in order to get estimates to or curly extension can be attached to increase volume. use for sample size calculation for at least some of the condi- tions. The study included all ages. Consent was received from the relevant administrators and the participants. Clinical exam- ination of each participant’s scalp was performed by a derma- tologist and the following data were collected: age, gender, current hairstyle and clinical diagnosis of pathology (if any). All 104 people who were asked agreed to participate in the study: 80 females and 24 males, mean age 27Æ9 years (range 1–78). No participants were found to have CCCA. The preva- lence of TA in females was 46% (37 of 80) vs. 0% in males, and of AKN 8% in males (two of 24) vs. 0% in females. No other disorders were diagnosed.
Main study
Ethical approval was received from our institution’s Research Ethics Committee and permission was granted by the Western Cape Provincial Government’s Ministry of Education. Informed consent was obtained from parents or from participants if older than 18 years. The study design was a cross-sectional survey. Participants Fig 2. The ‘s-curl’ – a loose curl (wavy) effect, prominent when hair were in their first and last years in two primary schools and is kept short, is produced by applying the same chemicals used as two high schools in Langa Township in Cape Town (randomly relaxers for much shorter periods than when relaxing. When the selected out of six primary schools and four high schools, hair is allowed to grow longer the same procedure produces an respectively). A sample of 352 girls was required to obtain a effect called the ‘blow out’ allowing a larger ‘Afro’ than is usually 95% confidence interval of ± 5% around a prevalence estimate possible. of 46% for TA. Similarly, a sample of 706 boys was required to obtain a 95% confidence interval of ± 2% around a preva- possible association with scalp (and hair) disease in Africans. lence estimate of 8% for AKN, based on the pilot study. The However, a subsequent Nigerian study has suggested an asso- plan was to look for the disorders of interest (AKN ⁄TA ⁄CCCA) ciation between frequent hair relaxer use and scarring alope- and to document any other clinical diagnoses of scalp dis- cia.7 Although TA and CCCA are usually seen in adult women orders, in all participants irrespective of gender. However, and AKN in adult men, the age at onset is not certain. There during consultation with the principals it became apparent are also no prevalence data available in our population for any that examining different numbers of girls and boys would hair ⁄scalp disease. We thus undertook the first population prove too disruptive. A decision was thus taken to examine all study of this subject in children from the first to last year of consenting pupils in each of the chosen grades irrespective of school in Cape Town. gender, and 1060 pupils in total were invited to participate.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp106–110 108 Hairdressing and scalp disease in African schoolchildren, N.P. Khumalo et al.
A previously piloted questionnaire was administered by a trained assistant and included demographic data as well as questions on grooming methods, reasons for the choice of hairstyle and current and past hairstyles. All subjects were examined by the same dermatologist to record the current hairstyle and the condition of the scalp. Digital photography was used to record interesting hairstyles and disease findings. Statistical analyses were performed using Stata version 9.0 (Stata Corporation, College Station, TX, U.S.A.). Prevalence figures were calculated as percentages and associations were compared using Fisher’s exact test and the v2 test. All signifi- cance tests were two tailed and significance was defined at the 5% alpha level.
Results
Of 1060 invited scholars, consent was received from 1042 Fig 3. High school boy with acne (folliculitis) keloidalis nuchae – (age range 6–21 years). There were two refusals and 16 were typically mild in this age group. not at school on study days. Participants included 467 (45%) boys and 574 (55%) girls (Table 1). Children in the first year of school were not asked to fill in the questionnaire because (The latter difference is partly explained by the fact that the of their age (6 years: n = 177) and this accounts for most of examination included the previously mentioned 6 year olds). the difference in totals of the results in this study. The prevalence of AKN in the whole group was seven of On examination, natural hair was more common in boys 1042 (0Æ67%), all in boys [seven of 466 (1Æ5%)]. In high than girls [407 of 459 (88Æ7%) vs. 153 of 561 (27Æ3%)]. A school boys the prevalence was seven of 321 (2Æ2%) and was much higher proportion of boys than girls had frequent hair- highest [six of 129 (4Æ7%)] in boys in the last year of high cuts (within 4 weeks) [268 of 368 (72Æ8%) vs. 65 of 480 school. Disease severity in all cases was mild (Fig. 3). There (13Æ5%), P <0Æ0001]. In contrast, a greater proportion of was a significant association, in all participants with natural girls than boys recorded their hair as relaxed [382 of 487 hair, between AKN and frequently cut hair (hair cuts (78Æ4%) vs. 31 of 362 (8Æ6%), P <0Æ0001]. This difference < 4 weeks) vs. participants with hair cuts > 4 weeks (seven was similar to that found on examination of the whole group of 261 vs. none of 199, P =0Æ02). [407 of 561 (72Æ5%) vs. 52 of 459 (11Æ3%), P <0Æ0001]. In girls, most relaxed hair was worn in a ‘pushed back style’ or ponytail whereas in boys it was kept short. Heat (hairdryers, tongs, etc.) was not used daily ⁄weekly by any Table 1 Characteristics of study population participants, but rather rarely on special occasions, e.g. wed- dings. In addition, no participants used hot combs. n (%)a The prevalence of TA was 98 of 1042 (9Æ4%) in all partici- Æ Characteristic Total Girls Boys P-value pants, and 17 1% in girls, among whom it increased with age from 8Æ6% in the first to 21Æ7% in the last year of school Age group (years) < 10 176 (17Æ6) 81 (14Æ6) 95 (21Æ2) (Fig. 4, Table 2). There was a larger proportion of partici- 10–15 215 (21Æ5) 126 (22Æ7) 89 (20Æ0) pants with relaxed hair who had TA than those with natural > 15 610 (60Æ9) 347 (62Æ6) 263 (58Æ8) hair (90 of 459 vs. eight of 560, P <0Æ0001), a finding that Total 1001 554 447 0Æ021b was maintained when the analysis was limited to girls (90 Hairstyle on examination of 407 with relaxed vs. eight of 153 with natural hair, Natural 560 (54Æ9) 153 (27Æ3) 407 (88Æ7) Relaxed 459 (45) 407 (72Æ5) 52 (11Æ3) P <0Æ0001). Other 1 (0Æ1) 1 (0Æ2) 0 (0Æ0) Although braids and extensions were not allowed with Total 1020 561 459 < 0Æ0001b school uniform, these were worn during weekends and school History of braids holidays. The history of previous hairstyles included braids on Braids 205 (66Æ6) 170 (65Æ6) 35 (71Æ4) natural hair, braids on relaxed hair and hair worn as dread- (natural hair) Braids 89 (28Æ9) 84 (32Æ4) 5 (10Æ2) locks. The prevalence of TA in the latter groups of participants (relaxed hair) was, respectively, 39 of 205 (19Æ0%), 27 of 89 (30Æ3%) and Dreadlocks 14 (45Æ5) 5 (2Æ0) 9 (18Æ4) none of 14 (0%) in all participants, and 39 of 170 (22Æ9%), Total 308 259 49 < 0Æ0001b 27 of 84 (32Æ1%) and none of five (0%) in girls. Although the proportion of girls with TA with a history of braids on aDenominators differ due to missing values. bP-value based on v2 test of association for difference between boys and girls. natural hair was lower than that of girls with braids on relaxed hair, the difference was not significant (P =0Æ116).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp106–110 Hairdressing and scalp disease in African schoolchildren, N.P. Khumalo et al. 109
bacterial infection. Boys had a higher disease prevalence of clinical tinea capitis, accounting for two thirds, than did girls (28 of 39 vs. 11 of 39, P =0Æ025) but the proportion with severe disease was similar in both groups (three of 28 vs. one of 11, respectively, P =0Æ692).
Discussion
Neither TA nor AKN have previously been reported in general or school population studies.6 We found a significantly higher disease occurrence of TA than that reported in clinic-based studies (1%) in London8 and recently (7Æ7% of 39 women) from Nigeria,7 and closer to that reported among adult female African volunteers (33%) in an industry study.9 In this study the prevalence of TA was significantly higher in girls with Fig 4. Young girl in grade 1 class with relaxed hair, demonstrating relaxed hair than in those with natural hair. The high preva- prominent traction alopecia. lence of TA in relaxed hair, which is mostly combed back and worn in ponytails, is consistent with the hypothesis that trac- tion is causal. However, the extent of the contribution of Table 2 Variation in the prevalence of traction alopecia in girls chemical relaxers to increased hair fragility remains uncertain. It has been suggested that relaxers ‘weaken the hair struc- n ⁄N (%) ture’10 and increase fibre fragility.11 Thus relaxed hair may be Total 98 ⁄574 (17Æ1) less resistant to traction, predisposing it to a higher risk of School level developing TA. Although school rules forbade our participants First year of school 7 ⁄81 (8Æ6) from wearing braids, these were worn on weekends and dur- Last year of primary and 40 ⁄258 (15Æ6) ing holidays. The prevalence of TA was higher, although not first year of high school significantly, in participants with a past history of braids on Last year of high school 51 ⁄235 (21Æ7) Hairstyle on examination relaxed hair than on natural hair. The additional traction from Natural hair 8 ⁄153 (5Æ2) intermittent braids is likely to have contributed to the preva- Relaxed hair 90 ⁄407 (22Æ1) lence of TA in our population. Adult studies that include par- Hairstyle on history ticipants with braids as usual hairstyles are likely to give more Braids on natural hair 39 ⁄170 (22Æ9) accurate estimates of TA in braids. Braids on relaxed hair 27 ⁄84 (32Æ1) The prevalence of AKN in our population was similar to the case frequency (1Æ3%)12 reported among patients in a Nige- rian skin clinic, and lower than the 13Æ7%, which included scalp folliculitis, reported from a London skin clinic. It was There was a significant difference in the prevalence of TA also lower than that reported among American football players between girls with natural hair and those with a history of who all wear helmets, of 13Æ6% vs. 0% in blacks and whites, braids on natural hair [eight of 153 (5Æ2%) vs. 39 of 170 respectively.13 In the latter study the prevalence in high school (22Æ9%), P <0Æ0001], suggesting that a history of braids on vs. older players was 5Æ2% (8Æ1% in blacks, 0% in whites) vs. natural hair may increase the prevalence of TA. However, this 9Æ4% (16Æ1% in blacks, 0% in whites). The prevalence in high difference is confounded by currently relaxed hair in the school players in the latter study is closer to that in our popu- majority of these participants (36 of 170). No cases of CCCA lation of boys in the last year of school (4Æ7%). were identified. Of the three conditions (TA, AKN, CCCA) thought to be Other findings included three cases of folliculitis and one common in Africans and to have an association with hairstyle case of alopecia areata. Surprisingly, no participants were choice, we previously failed to identify any studies estimating diagnosed with head lice infestation or androgenetic alopecia. the frequency of CCCA in any population.6 The recent Nigerian The overall prevalence of scaly scalp was 61 of 1042: 22 cases clinic study7 reported CCCA in six (15%) of 39 adult women of scaly scalp only and 39 cases diagnosed as clinical tinea presenting with hair loss and found an association between capitis (i.e. scaly scalp and at least one patch of alopecia, prolonged and frequent use of relaxers (mean 23Æ2± broken hairs or black dots). Most children with possible tinea 9Æ3 years) and ‘scarred alopecia’. We did not identify a single capitis were in the 6-year-old group [34 of 177 (19Æ2%), participant with alopecia consistent with CCCA. including four severe cases, all in the first year of school], Hair grooming is not thought to play a role in the pathogene- vs. five of 865 (0Æ6%, none severe) in the older group sis of certain other scalp diseases, but because they are import- (P <0Æ0001). The severe cases included one with kerion and ant causes of morbidity it was important to document their three with impetiginized lesions suggestive of superimposed prevalence. It was thus interesting that only one case of alopecia
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp106–110 110 Hairdressing and scalp disease in African schoolchildren, N.P. Khumalo et al. areata was identified, giving a prevalence of 0Æ1%. A prevalence References of androgenetic alopecia of 5Æ6% has been reported in adoles- 1 Bolduc C, Shapiro J. Hair care products: waving, straightening, cents at a Turkish boarding school14 and of 32% at ages conditioning, and coloring. Clin Dermatol 2001; 19:431–6. 15 17–26 years in Singapore. The predominance of very short 2 Swee W, Klontz KC, Lambert LA. A nationwide outbreak of hairstyles in boys in this study could have concealed this latter alopecia associated with the use of a hair-relaxing formulation. Arch condition, which even if present was likely to be mild in this Dermatol 2000; 136:1104–8. age group. 3 Nicholson AG, Harland CC, Bull RH et al. Chemically induced cos- Transmission of tinea capitis and head lice can occur metic alopecia. Br J Dermatol 1993; 128:537–41. through hairdressing implements in addition to direct contact. 4 Quinn CR, Quinn TM, Kelly AP. Hair care practices in African American women. Cutis 2003; 72:280–2. The overall prevalence of scaly scalp was 61 of 1042 (22 scale 5 Hantash BM, Schwartz RA. Traction alopecia in children. Cutis alone; 39 scale plus alopecia or broken hairs or black dots). 2003; 71:18–20. The predominance of short hair in boys diagnosed with clinical 6 Khumalo NP, Jessop S, Ehrlich R. The prevalence of cutaneous tinea capitis in our study could suggest possible transmis- adverse effects of hair dressing. A systematic review. Arch Dermatol sion during frequent hair cuts. The latter would be consistent 2006; 142:377–83. with the low prevalence among 6-year-old girls with relaxed 7 Nnoruka EN. Hair loss: is there a relationship with hair hair, which is infrequently cut. However, clinical diagnosis of care practices in Nigeria? Int J Dermatol 2005; 44 (Suppl. 1):13– 17. tinea capitis is likely to be an overestimation of the true preva- 8 Child FJ, Fuller LC, Higgins EM et al. A study of the spectrum of lence, as demonstrated in a recent study that reported that skin disease occurring in a black population in south-east London. although 66 of 300 prepubertal children had scaly scalps, only Br J Dermatol 1999; 141:512–17. nine (14%) had culture-positive tinea infection.10 Even though 9 Loussouarn G, El-Rawadi C, Genain G. Diversity of hair growth our clinical diagnosis was stricter, the suggested association profiles. Int J Dermatol 2005; 44 (Suppl. 1):6–9. can only be confirmed in studies which include fungal 10 Williams JV, Eichenfield LF, Burke BL et al. Prevalence of scalp sca- cultures. ling in prepubertal children. Pediatrics 2005; 115:e1–6. 11 Buczek A, Markowska-Gosik D, Widomska D et al. Pediculosis capi- Not a single case of head lice infestation was identified. The tis among schoolchildren in urban and rural areas of eastern latter has been reported at a prevalence ranging from 0 to Poland. Eur J Epidemiol 2004; 19:491–5. 11,16–19 28% in primary schools in various countries. In African 12 George AO, Akanji AO, Nduka EU et al. Clinical, biochemical and study participants with similar hair to ours, the prevalence morphologic features of acne keloidalis in a black population. Int J was 17Æ1% in Kenya,20 6Æ8% in Sierra Leone,21 and 3Æ1% in Dermatol 1993; 32:714–16. urban and 0Æ1% in rural Nigerian children.22 13 Knable AL Jr, Hanke CW, Gonin R. Prevalence of acne keloi- Finally, the prevalence of scaly scalp had an inverse relation- dalis nuchae in football players. J Am Acad Dermatol 1997; 37:570–4. 14 Tuncel AA, Erbagci Z. Prevalence of skin diseases among male ship with age. Alopecia areata was infrequent, while lice adolescent and post-adolescent boarding school students in Turkey. infestation, androgenetic alopecia and CCCA were not diag- J Dermatol 2005; 32:557–64. nosed in our school population. This study has found an asso- 15 Tang PH, Chia HP, Cheong LL et al. A community study of male ciation between hairstyle choice and the prevalence of specific androgenetic alopecia in Bishan, Singapore. Singapore Med J 2000; scalp disorders. TA is common in girls, increases with age and 41:202–5. is significantly associated with relaxed hair. The effect of a his- 16 Ciftci IH, Karaca S, Dogru O et al. Prevalence of pediculosis and tory of braids on disease prevalence is difficult to interpret. scabies in preschool nursery children of Afyon, Turkey. Korean J Parasitol 2006; 44:95–8. AKN is more common in older boys who have frequent hair- 17 Al-Shawa RM. Head louse infestations in Gaza governorates. J Med cuts. Studies among adults, who have a wider hairstyle choice Entomol 2006; 43:505–7. and have manipulated their hair for longer, are likely to yield 18 Khokhar A. A study of pediculosis capitis among primary school more information about the association between hair groom- children in Delhi. Indian J Med Sci 2002; 56:449–52. ing and disease. Estimation of the contribution of individual 19 Counahan M, Andrews R, Buttner P et al. Head lice prevalence in variables to disease development will help in the formulation primary schools in Victoria, Australia. J Paediatr Child Health 2004; of effective preventive and treatment strategies. 40:616–19. 20 Chunge RN. A study of head lice among primary schoolchildren in Kenya. Trans R Soc Trop Med Hyg 1986; 80:42–6. Acknowledgments 21 Gbakima AA, Lebbie AR. The head louse in Sierra Leone: an epidemiological study among school children, in the Njala area. We thank the South African Medical Research Council for West Afr J Med 1992; 11:165–71. partial funding of the study, and are grateful to the school 22 Ebomoyi E. Pediculosis capitis among primary schoolchildren in administrators, parents, pupils and to Mr S. Nkepu, the urban and rural areas of Kwara State, Nigeria. J Sch Health 1988; research assistant. 58:101–3.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp106–110 PHOTOBIOLOGY DOI 10.1111/j.1365-2133.2007.07959.x Variable pulsed light is less painful than light-emitting diodes for topical photodynamic therapy of actinic keratosis: a prospective randomized controlled trial P. Babilas, R. Knobler,* S. Hummel, C. Gottschaller, T. Maisch, M. Koller, M. Landthaler and R.-M. Szeimies Department of Dermatology and Centre of Clinical Studies, University of Regensburg, 93042 Regensburg, Germany *Department of Special and Environmental Dermatology, Medical University of Vienna, 1090 Vienna, Austria
Summary
Correspondence Background Photodynamic therapy (PDT) of actinic keratosis (AK) using methyl- Rolf-Markus Szeimies. aminolaevulinate (MAL) is an effective and safe treatment option, but the proce- E-mail: Rolf-Markus.Szeimies@klinik. dure is painful. uni-regensburg.de Objectives To evaluate the efficacy and pain associated with variable pulsed light Accepted for publication (VPL), a prospective, randomized, controlled split-face study was performed. 5 February 2007 Methods Topical MAL-PDT was conducted in 25 patients with AK (n ¼ 238) who were suitable for two-sided comparison. After incubation with MAL, irradiation ) ) Key words was performed with a light-emitting diode (LED) (50 mW cm 2;37Jcm 2) vs. ) ) 5-aminolaevulinic acid, actinic keratosis, VPL (80 J cm 2, double pulsed at 40 J cm 2, pulse train of 15 impulses each photodynamic therapy, protoporphyrin IX, with a duration of 5 ms, 610–950 nm filtered hand piece) followed by re-evalu- randomized controlled trial ation up to 3 months. Conflicts of interest Results The pain during and after therapy was significantly lower with VPL irradi- None declared. ation [t (d.f. ¼ 24) ¼ 4Æ42, P <0Æ001]. The overall mean ± SD infiltration and keratosis score at 3 months after treatment was 0Æ86 ± 0Æ71 (LED system) vs. 1Æ05 ± 0Æ74 (VPL device) (no statistically significant difference; P ¼ 0Æ292). Patient satisfaction following both treatment modalities did not significantly differ at the 3-month follow up (P ¼ 0Æ425). Conclusions VPL used for MAL-PDT is an efficient alternative for the treatment of AK that results in complete remission and cosmesis equivalent to LED irradiation but causes significantly less pain.
Photodynamic therapy (PDT) using 5-aminolaevulinic acid in PDT,3,10–13 and light-emitting diodes (LEDs), lasers and (ALA) is based on the topical application of ALA to a skin incoherent light sources such as halogen lamps all give ade- lesion and the subsequent illumination of the lesion with light quate complete remission rates.3,14,15 of an appropriate wavelength. In situ, ALA is enzymatically Incoherent polychromatic filtered flashlamp (intense pulsed converted into the active endogenous photosensitizer proto- light, IPL) devices developed in the early 1990s have been porphyrin IX (PpIX), which selectively sensitizes diseased cells. standard light sources for the treatment of several skin condi- Excitation of the photosensitizer by light of a certain wave- tions such as vascular lesions16 or photoageing,17–19 or for length and fluence results in the generation of reactive oxygen hair removal.20,21 The emission spectrum of IPL devices ran- species (ROS), particularly singlet oxygen. These ROS mediate ges from 500 to 1300 nm. With the aid of convertible cut-off cellular effects, e.g. lipid peroxidation, and vascular effects filters, the IPL device can easily be adapted to the desired result in a direct or indirect cytotoxic impact on treated cells.1 wavelengths, which underlines its high versatility. Additional Usually ALA-PDT shows excellent cosmetic results.1,2 When advantages as compared to lasers are higher skin coverage rates multiple lesions are treated simultaneously, pain is a major and lower costs. The adaptation to the absorption spectrum of side-effect of PDT. The sense of pain depends on the extent of PpIX allows the use of IPLs for PDT. Recently, several authors the lesions3–6 and – according to our clinical experience – on have reported successful treatments of acne vulgaris,22–24 the type of lesion, the site of the lesions and the light source AK,25,26 photoageing19,27,28 and sebaceous gland hyperplasia29 used. The efficacy of topical ALA-PDT in the treatment of acti- with IPL devices for PDT. However, up to the present no con- nic keratosis (AK) has been frequently documented.3,7–9 Var- trolled studies providing evidence-based data have been pub- ious light sources have been successfully used for irradiation lished regarding the effectiveness and pain associated with the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 111 112 VPL for photodynamic therapy, P. Babilas et al. use of IPL devices for ALA-PDT in the treatment of epidermal Germany). After a 3-h incubation time the cream was washed neoplasias such as AK. off with saline. Immediately after removing dressing and This prospective, controlled and randomized study com- cream, irradiation was performed. Light source and treatment pares both the treatment-associated pain and the effectiveness side were assigned randomly in each patient. The allocation of methylaminolaevulinate (MAL)-PDT using an LED system schedule was assigned by lot prior to study initiation by a vs. a variable pulsed light (VPL) system for the treatment of study nurse and remained concealed until each individual was AK (n ¼ 238). The VPL system is an IPL device in which it is included. One side was irradiated with the VPL device (Ener- possible to change the number, width and delay of the micro- gist Ultra VPLTM; Energist Ltd, Swansea, U.K.) using a total ) pulses used. light dose of 80 J cm 2 (610–950 nm cut-off filter, double ) pulsed at 40 J cm 2, pulse train of 15 impulses, each pulse Materials and methods duration 5 ms, 20 ms delay time between pulses). Prior to irradiation a contact gel was applied. The contralateral side was irradiated with the LED system (AktiLite ; Galderma, Study design ) La De´fense, France) using a total light dose of 37 J cm 2 )2 This comparative study was designed as a prospective, ran- (kem ¼ 635 ± 3 nm, light intensity 50 mW cm , treatment domized split-face study according to the revised CONSORT time 12 min, size 80–180 mm). For both light sources, dosi- statement.30 Each patient served as his ⁄her own control and metry was chosen based on the manufacturers’ recommenda- received the treatment of interest on one side of the face and tions. Irradiation was performed in a single treatment setting the control treatment on the other side. Treatments were with the right side being irradiated first. No analgesia or local assigned randomly. A simple randomization scheme was anaesthesia was given. However, a cooling system was applied and prepared prior to the study. The study included required by all patients during LED irradiation (Cryo5; three points of assessment: prior to treatment, 2 weeks follow Zimmer Medizinsysteme, Neu-Ulm, Germany), whereas VPL up, and 3 months follow up. All patients were given verbal was applied without cooling. Stopping rules such as intoler- and written information on the nature of the study. Signed able pain were defined prospectively. informed consent was obtained prior to study initiation. The Patients were instructed strictly to avoid sunlight and to use study was approved by the Independent Ethic Committee of sunscreen for the following 6 weeks. Treatment was per- the University of Regensburg. The clinical investigations were formed only once. conducted according to the Declaration of Helsinki principles. Assessments and response evaluation Patients: inclusion and exclusion criteria Prior to treatment and at each follow-up visit each lesion was Twenty-five patients aged at least 18 years (eight women and classified by two independent and blinded investigators. If the 17 men, mean ± SD age 71 ± 16 years) with a clinical diag- investigators documented different values, the mean was cal- nosis of mild to moderate AK (forehead and scalp, n ¼ 238) culated. Each lesion was rated using a previously established were recruited for this clinical study. To be considered eli- score31, in which the overall infiltration of each lesion was gible, a patient had to have a minimum of two AKs in a sym- graded as 0, none (no palpable infiltration); 1, weak (slightly metrical distribution suitable for a two-sided comparison and palpable AK, more easily felt than seen); 2, medium (moder- a previous treatment-free period of at least 4 months. Diag- ately thick AK, easy to feel); 3, strong infiltration; or 4, very nosis was based on clinical assessment and, if necessary, strong infiltration. In addition, the grade of keratosis was clas- histology. All patients had Fitzpatrick skin types II (84%) or sified as 0, no visible keratosis; 1, just visible; 2, moderate; III (16%). Patients with pigmented lesions in the target area 3, thick keratosis; or 4, very thick keratosis.31 The mean of or with porphyria were excluded. Additional exclusion criteria both score values was calculated; lesions with a mean value included pregnancy, lactation and allergy to ALA. up to 1Æ5 were classified as mild AKs, from 1Æ51 to 2Æ5as moderate and from 2Æ51 to 4 as severe AKs. A cosmetic rating of the perilesional area was performed regarding wrinkling Treatment procedure (0, none; 1, normal skin texture; 2, weak wrinkling; 3, strong The study was conducted from July 2005 to June 2006 at the wrinkling), hyperpigmentation and hypopigmentation (each Department of Dermatology, University Hospital Regensburg, 0, none, i.e. normal; 1, weak; 2, medium; 3, strong), hairi- Germany. ness (0, hairless; 1, normal body part hairiness; 2, slightly Prior to PDT, loose crusts and debris were removed from increased hairiness; 3, strong hairiness), and redness and des- the lesion using a small curette and the surface was gently quamation (each 0, none; 1, weak; 2, medium; 3, strong). roughened. MAL (Metvix ; Galderma, France; 16% MAL) was The sum of the score value (0–18) was calculated. The overall applied to the lesion as a 1-mm thick layer including 5 mm cosmesis was evaluated by both the patient and the physicians of the surrounding normal tissue. The size of treatment site (two independent and blinded investigators) on a score from was exactly equal in both treatment arms. The area was 1 (very bad) to 10 (excellent). The lesions were photographed covered with occlusive dressing (Tegaderm ; 3M, Borken, at each visit.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 VPL for photodynamic therapy, P. Babilas et al. 113
As a primary outcome measure patients were asked to score P <0Æ05 and the beta error at P <0Æ20. Assuming that the pain separately for each site directly after treatment. Pain was test group has an advantage over the control group of 2 score assessed using a visual analogue scale (VAS; 0, none–10, points, a minimum of n ¼ 15 pairwise observations (i.e. insufferable pain). The VAS is a widely accepted technique for number of patients in the context of a split-face design) is measuring pain.32 A VAS typically consists of a horizontal line needed to detect this effect. In order to compensate for drop- anchored with labels at opposite ends representing no pain outs and to increase the overall clinical relevance and the gen- and worst imaginable pain, respectively.33 The patient is asked eralizability of the results, we decided to enrol 25 patients in to regard the VAS as a continuum and to mark the point along the study. Before ⁄after differences were calculated with the the line corresponding to his ⁄her current level of pain.33 In dependent t-test (interval-scaled, normally distributed varia- addition, immediate side-effects such as burning sensation, bles), the Wilcoxon test (rank-ordered measures or non-nor- pruritus, redness or swelling were documented on a scale mal distributions) or the McNemar test (proportions). All from 1 (weak) to 10 (unbearably strong). Adverse events such analyses were based on the intention-to-treat approach. Data as allergic reactions were documented. analyses were carried out using SPSS for Windows version Sense of pain (primary outcome measure), lesion response, 12.02 (SPSS, Chicago, IL, U.S.A.), and results are reported as cosmesis and patient satisfaction (secondary outcome meas- mean ± SD. ures) were assessed at 3 months following therapy; side- effects (secondary outcome measure) were assessed at 2 weeks Results and 3 months following therapy by two independent and blinded investigators. Side-effects such as pain, swelling, red- Twenty-five patients with a total of 238 AK lesions completed ness, crusting, local infection, erosion, hyperpigmentation or the study. All patients showed a lesion distribution suitable for hypopigmentation were documented on a scale from 1 (weak) a two-sided comparison (Fig. 1). All lesions were located in to 10 (unbearably strong). The overall outcome was classified the face (68%) or on the scalp (32%). The treatment groups as complete remission if no sign of infiltration or keratosis were very similar with respect to number of lesions and lesion was detectable. The complete remission rate at 3 months fol- grades. 120 lesions were treated with the LED system and 118 lowing treatment is based on all treated AKs. Patient satisfac- lesions with the VPL device. The initial overall infiltration of tion was rated on a score from 1 (very bad) to 10 (excellent). the lesions (1Æ33 ± 0Æ70 vs. 1Æ32 ± 0Æ69) and the initial over- Adverse events such as local phototoxicity reactions that all keratosis (1Æ42 ± 0Æ68 vs. 1Æ53 ± 0Æ69) were equal in both normally occur after PDT were recorded. treatment arms. At baseline, 70% of the lesions showed mild overall infiltration and keratosis, 27Æ5% showed moderate overall infiltration and keratosis and 2Æ5% showed severe over- Statistical methods all infiltration and keratosis. The sum of the cosmetic rating of The primary dependent variable was pain as assessed by the the perilesional area prior to treatment was 5Æ56 ± 1Æ81 for VAS. The VAS ranges from 0 (no pain) to 10 (highest level of the LED system side vs. 5Æ48 ± 1Æ84 for the VPL device pain) with an assumed SD of 2Æ50. The alpha error was set at side (no significant difference, see Table 1). The patients
39 patients with actinic keratosis in a symmetrical dissemination screened 4 excluded 10 refused
25 randomized Score: infiltration/keratosis cosmesis Incubation with MAL
Score: sense of pain, One side irradiated with LED side-effects, One side irradiated with VPL adverse events
Score: 2 weeks follow-up cosmesis, side-effects, 2 weeks follow-up adverse events
Score: Fig 1. Trial profile: prospective, randomized, infiltration/keratosis, 3 months follow-up 3 months follow-up controlled split-face study. MAL, cosmesis, side-effects, adverse events methylaminolaevulinate; VPL, variable pulsed light; LED, light-emitting diode.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 114 VPL for photodynamic therapy, P. Babilas et al.
Table 1 Cosmesis prior to and following Prior to PDT At 3 months photodynamic therapy (PDT). Each factor was assessed on a 0–3 scale (sum: 0–18) and VPL LED VPL LED physician and patient scores on a scale from 1 (very bad) to 10 (excellent). Scores are Wrinkling 1Æ72 ± 0Æ72 1Æ72 ± 0Æ72 1Æ72 ± 0Æ72 1Æ72 ± 0Æ72 given as mean ± SD Hyperpigmentation 1Æ28 ± 0Æ83 1Æ36 ± 0Æ79 1Æ28 ± 0Æ83 1Æ36 ± 0Æ79 Hypopigmentation 0Æ72 ± 0Æ60 0Æ72 ± 0Æ60 0Æ72 ± 0Æ60 0Æ72 ± 0Æ60 Hairiness 0Æ52 ± 0Æ50 0Æ52 ± 0Æ50 0Æ52 ± 0Æ50 0Æ52 ± 0Æ50 Redness 0Æ76 ± 0Æ71 0Æ76 ± 0Æ71 0Æ60 ± 0Æ69 0Æ64 ± 0Æ69 Desquamation 0Æ48 ± 0Æ75 0Æ48 ± 0Æ75 0Æ32 ± 0Æ55 0Æ32 ± 0Æ55 Sum 5Æ48 ± 1Æ84 5Æ56 ± 1Æ81 5Æ16 ± 1Æ67 5Æ28 ± 1Æ66 Overall evaluation Physician 6Æ64 ± 1Æ79 6Æ64 ± 1Æ79 8Æ20 ± 1Æ06 8Æ16 ± 0Æ97 Patient 5Æ52 ± 2Æ06 5Æ52 ± 2Æ06 7Æ76 ± 1Æ39 7Æ92 ± 1Æ29
VPL, variable pulsed light; LED, light-emitting diode.
(5Æ52 ± 2Æ06) and physicians (6Æ64 ± 1Æ79) graded the over- Table 2 Comparison between visible pulsed light (VPL) and light- all cosmesis prior to PDT equally for both sides (Table 1). emitting diode (LED) regarding pain during treatment (measured on Lesion preparation was performed before PDT in all individual a 0–10 visual analogue scale) and the mean of the infiltration and treatments. The time between lesion preparation and irradi- keratosis scores (each assessed on a 0–4 scale) and patient satisfaction (0–10 scale) at 3 months following photodynamic therapy (PDT). ation was equal in both treatment arms (192 ± 7 min) as Scores are given as mean ± SD every lesion of each patient was incubated simultaneously. The impression of pain as evaluated by VAS differed sig- VPL LED P-value nificantly between the treatment arms. For the LED system, Æ Æ Æ Æ Æ the patients documented a score value of 6Æ40 ± 2Æ63, in Pain during treatment 4 34 ± 2 46 6 40 ± 2 63 < 0 001 Overall infiltration and 1Æ05 ± 0Æ74 0Æ86 ± 0Æ71 0Æ292 comparison with 4Æ34 ± 2Æ46 for the VPL device [t (d.f. ¼ keratosis score at 24) ¼ 4Æ42, P <0Æ001, see Fig.2,Table 2]. No patient dis- 3 months following PDT continued treatment because of severe discomfort. Two to Patient satisfaction at 7Æ76 ± 1Æ39 7Æ92 ± 1Æ29 0Æ425 four days after irradiation, other side-effects such as redness 3 months following PDT and crusting were documented for all AKs in both treatment sides. These lasted for about 8–10 days. At the follow-up periods of 14 days and 3 months after treatment, all patients 7Æ76 ± 1Æ39 (VPL device); corresponding physician-assessed showed minor or no side-effects (Table 3). scores were 8Æ16 ± 0Æ97 vs. 8Æ20 ± 1Æ06 (no significant The sum of the cosmetic rating of the perilesional area difference in each case, see Table 2). There was no significant 3 months after treatment was 5Æ28 ± 1Æ66 for the LED system difference in either rating prior to PDT as compared with side vs. 5Æ16 ± 1Æ67 for the VPL device side (no significant 3 months following PDT. Patient satisfaction was high at difference, see Table 1). The patients graded the overall cos- 3 months after therapy, without any statistically significant mesis 3 months after PDT as 7Æ92 ± 1Æ29 (LED system) vs. difference between the treatment sides (P ¼ 0Æ425). The overall lesion complete response rates at 3 months after treatment were 56Æ7% (68 of 120 AKs) following PDT with the LED system and 46Æ6% (55 of 118 AKs) following PDT 10 9 with the VPL device (Fig. 3). The overall infiltration and kera- 8 tosis score at 3 months after treatment was 0Æ86 ± 0Æ71 (LED 7 6 system) vs. 1Æ05 ± 0Æ74 (VPL device). Statistical analysis 5 revealed no statistically significant difference between the 4 treatment regimens (P ¼ 0Æ292, Table 2). Pain intensity Pain 3 2 1 0 Discussion LED VPL As AK has the potential to progress into invasive squamous Fig 2. Pain visual analogue scale (VAS) values (0, none–10, cell carcinoma, timely and appropriate treatment is of import- insufferable pain) documented immediately after illumination of each ance. Treatment options include cryotherapy, topical 5-fluoro- side. Corresponding values are connected with bars: blue, lower VAS uracil, curettage, electrosurgery, laser surgery and excision. for variable pulsed light (VPL); red, lower VAS for light-emitting Most treatments are hampered by considerable discomfort diode (LED). during the treatment sessions or periods, and often lead to
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 VPL for photodynamic therapy, P. Babilas et al. 115
Table 3 Side-effects after photodynamic therapy. Each side-effect was assessed on a scale from 1 (weak) to 10 (unbearably strong) and scores are given as mean ± SD
Immediately after irradiation At 2 weeks At 3 months
VPL LED VPL LED VPL LED Redness 4Æ20 ± 2Æ08 5Æ28 ± 1Æ97 1Æ44 ± 1Æ60 1Æ44 ± 1Æ60 1Æ44 ± 1Æ60 1Æ44 ± 1Æ60 Swelling 0Æ60 ± 1Æ50 0Æ80 ± 1Æ50 0Æ04 ± 0Æ20 0Æ00 ± 0Æ00 0Æ04 ± 0Æ20 0Æ00 ± 0Æ00 Burning sensation 2Æ16 ± 1Æ95 5Æ08 ± 2Æ780000 Pruritus 0Æ04 ± 0Æ20 0Æ04 ± 0Æ200000 Crusting – – 0Æ88 ± 1Æ37 0Æ76 ± 1Æ30 0Æ08 ± 0Æ27 0Æ08 ± 0Æ27 Local infection – – 0Æ00 ± 0Æ00 0Æ00 ± 0Æ00 0Æ00 ± 0Æ00 0Æ00 ± 0Æ00 Erosion 0 0 0Æ04 ± 0Æ20 0Æ00 ± 0Æ00 0Æ04 ± 0Æ20 0Æ00 ± 0Æ00 Hyperpigmentation – – 1Æ21 ± 0Æ82 1Æ20 ± 0Æ80 1Æ20 ± 0Æ85 1Æ28 ± 0Æ83 Hypopigmentation – – 0Æ76 ± 0Æ65 0Æ76 ± 0Æ65 0Æ72 ± 0Æ60 0Æ72 ± 0Æ60
VPL, variable pulsed light; LED, light-emitting diode.
(a) (b)
Fig 3. Complete remission of the actinic keratosis area at the left side of the nose as early as 2 weeks following photodynamic therapy (PDT) with variable pulsed light. (a) Before treatment; (b) 2 weeks after PDT. scars or less than optimal cosmetic end results. Especially excludes all extraneous wavelengths. Thus, photobleaching of regarding field cancerization, PDT using topical photosensitiz- chromophores and tissue hyperthermia are avoided which ers such as MAL represents an excellent treatment alternative. might occur in high-intensity light sources.34,35 In contrast to PDT can potentially improve the cosmetic outcome, as it is lasers, both light systems easily allow the simultaneous treat- less destructive and more selective than other treatment ment of apparent AK and perilesional sun-damaged skin. options. Besides lasers and the widely used incoherent light Therefore, this therapy focuses on subclinical lesions, which is sources, VPL devices have recently been introduced for topical of special interest in field cancerization.36 A major advantage PDT. However, no controlled and randomized comparisons of of IPL devices is the high skin coverage rate. The use of large the efficacy of VPL devices as compared with established light footprints allows a rapid treatment of most anatomical sources in the treatment of AK have been published up to the regions.37 In contrast, irradiation with LED systems or, for present. example, halogen lamps takes at least 8 min per irradiation In the VPL system, a flashlamp is pulsed under computer field. However, light application with an IPL device to uneven control. Recently, the use of high-energy flashlamps for PDT anatomical regions such as around the nose or ear is difficult has become more popular for a variety of reasons. They emit due to the lack of manoeuvrability and large spot sizes. An light ranging in wavelengths from blue to infrared, which additional advantage of the VPL device as compared with the covers all absorption peaks of PpIX including Q and Soret LED system may be the fact that photodynamically active bands (410, 504, 538, 576 and 630 nm). In contrast to the products can also be activated at wavelengths beyond the acti- broadband emission of IPL devices, the LED system closely vation spectrum of the key porphyrin PpIX, which makes matches the activation spectra of key chromophores and broadband illumination more effective for MAL-PDT.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 116 VPL for photodynamic therapy, P. Babilas et al.
A disadvantage of IPLs is the possible inconsistency of the application time) in 28 patients with 251 AKs. The primary emitted light spectrum from pulse to pulse caused by an complete response rate of AKs to ALA-PDT was 64% after a instantaneous discharge of the capacitors.17 As IPLs may be single treatment, but was 85% when responses to a second used for photoepilation, the risk of unwanted hair reduction treatment were included.39 In an uncontrolled study of 17 in the treated area is evident but may not be relevant for the patients treated twice with ALA-PDT (20% ALA ⁄oil-in-water indication studied. emulsion; 4 h application time) using an IPL device (kem ¼ ) To show the possible advantage of using a VPL device as 615–1200 nm, 40 J cm 2, double pulse mode of 4Æ0 ms, compared with the established LED system as light source for delay time 20 ms), complete remission in 33 of 38 AKs was MAL-PDT, 25 patients with AK (n ¼ 238) were treated in a achieved 3 months following treatment (87%). Crusting was split-face study. As the sense of pain was the primary outcome observed as a side-effect, which took 1 week to resolve.40 measure, a single treatment setting was chosen. Besides, a dou- Unfortunately, no study systematically documented and ana- ble treatment of PDT (which is recommended for the treatment lysed the sense of pain as a marker for the quality of treatment of AK) possibly leads to overtreatment and thus is likely to felt by patients. mask minor differences in the efficacy of the light sources. When the impact on cosmetic outcome was evaluated, no According to the VAS, the sense of pain during irradiation difference comparing both light sources (Table 1) was noted. with the LED system was 6Æ40 ± 2Æ63, which agrees with our The excellent cosmetic outcome of PDT documented in this previous study.3 However, during irradiation with the VPL study is in accordance with previous results.3,7,39 Due to the device the sense of pain was 4Æ34 ± 2Æ46, which is signifi- selectivity of ALA for diseased cells, PpIX has been shown to cantly lower as compared with the LED system. This represents concentrate in AKs as compared with the normal surrounding a remarkable result, as a cooling device was used during LED tissue.34 Hence, optimal destruction of lesions is achieved with irradiation to reduce the burning sensation, whereas no treat- the surrounding skin left intact. However, an improvement of ment for pain reduction was required during VPL irradiation. cosmesis of the perilesional area, as described by Alster et al.,28 ) In addition, the VPL device uses a higher dose (80 J cm 2)as Goldman et al.19 and Avram and Goldman,25 was not detected. ) compared with the LED system (37 J cm 2). Some authors In conclusion, our study was able to show significantly suggest that the sense of pain increases with increasing light lower pain for AKs treated with MAL and the VPL device dose.15,38 The lower pain level induced by the VPL is when compared with the LED system. Treatment of AKs with most possibly due to the short pulse duration of 5 ms (total a single MAL-PDT treatment produced an equivalent clinical pulse time 355 ms) as compared with the significantly lon- outcome with both light sources. Residual lesions can be ger treatment time of the LED system. Karrer et al.11 com- re-treated to achieve higher response rates. The overall cosmetic pared pain and efficacy of ALA-PDT using a long-pulse tunable outcome with both light sources was assessed to be excellent flashlamp-pumped pulsed dye laser (LPDL; 1Æ5 ms; 585 nm; by both investigators and patients. According to this study, ) ) 18 J cm 2) and an incoherent lamp (160 mW cm 2; 60– the VPL device is associated with a lower side-effect profile ) 160 J cm 2). Topical PDT was performed on 24 patients with regarding pain during treatment and is an effective and useful AKs on the head (n ¼ 200). Results showed that pain during alternative for illumination in MAL-PDT of AK. light treatment was significantly reduced by using the LPDL, with equal complete remission rates following both light regi- Acknowledgments mens. Sandberg et al.5 investigated pain during ALA-PDT in 91 patients with AK using a VAS (0–10). They found out that the The Energist Ultra VPLTM system was provided by Energist size and the redness of the actinic lesions as well as the effi- Ltd, Swansea, U.K. The work was supported by a grant of cacy of treatment were related to PDT-induced pain. the Dr-Heinz-Maurer-Stiftung, Boppard, Germany. The editorial Three months after therapy, no clinically significant differ- assistance of Mrs Monika Scho¨ll is gratefully acknowledged. ence in therapeutic outcome was detected between both light regimens. The 3-month follow-up revealed a complete remis- References sion rate of 46Æ6% (VPL device) vs. 56Æ7% (LED system). These results correspond to other findings published in the 1 Babilas P, Landthaler M, Szeimies RM. Photodynamic therapy in literature: Kim et al. investigated the clinical and histological dermatology. Eur J Dermatol 2006; 16:340–8. response of 12 AKs (in more pigmented Asian skin) 12 weeks 2 Szeimies RM, Karrer S, Radakovic-Fijan S et al. Photodynamic ther- apy using topical methyl 5-aminolevulinate compared with cryo- following ALA-PDT (20% ALA ⁄oil-in-water emulsion; 4 h therapy for actinic keratosis: a prospective, randomized study. JAm application time) using an IPL device (kem ¼ 555–950 nm, Acad Dermatol 2002; 47:258–62. )2 12–16 J cm , pulse duration 20–30 ms, delay time 20 ms). 3 Babilas P, Kohl E, Maisch T et al. In vitro and in vivo comparison of The study revealed 50% clinical clearance and 42% histological two different light sources for topical photodynamic therapy. Br J clearance of individual AKs in one single treatment setting.26 Dermatol 2006; 154:712–18. The unexpectedly low complete remission rate obtained in the 4 Wiegell SR, Stender IM, Na R et al. Pain associated with photo- present study as well as in the cited work may be explained dynamic therapy using 5-aminolevulinic acid or 5-aminolevulinic acid methylester on tape-stripped normal skin. Arch Dermatol 2003; by the single treatment setting. This suggestion is confirmed 139:1173–7. by Fink-Puches et al. who performed ALA-PDT (20% ALA, 4 h
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 VPL for photodynamic therapy, P. Babilas et al. 117
5 Sandberg C, Stenquist B, Rosdahl I et al. Important factors for pain 23 Melnick S. Cystic acne improved by photodynamic therapy with during photodynamic therapy for actinic keratosis. Acta Derm Venereol short-contact 5-aminolevulinic acid and sequential combination of (Stockh) 2006; 86:404–8. intense pulsed light and blue light activation. J Drugs Dermatol 2005; 6 Kasche A, Luderschmidt S, Ring J et al. Photodynamic therapy indu- 4:742–5. ces less pain in patients treated with methyl aminolevulinate com- 24 Gold MH, Bradshaw VL, Boring MM et al. The use of a novel pared to aminolevulinic acid. J Drugs Dermatol 2006; 5:353–6. intense pulsed light and heat source and ALA-PDT in the treatment 7 Szeimies RM, Karrer S, Sauerwald A et al. Photodynamic therapy of moderate to severe inflammatory acne vulgaris. J Drugs Dermatol with topical application of 5-aminolevulinic acid in the treatment 2004; 3:S15–19. of actinic keratoses: an initial clinical study. Dermatology 1996; 25 Avram DK, Goldman MP. Effectiveness and safety of ALA-IPL in 192:246–51. treating actinic keratoses and photodamage. J Drugs Dermatol 2004; 8 Braathen L, Szeimies RM, Basset-Seguin N et al. Guidelines on the use 3:S36–9. of photodynamic therapy (PDT) for non-melanoma skin cancer – 26 Kim HS, Yoo JY, Cho KH et al. Topical photodynamic therapy an international consensus. J Am Acad Dermatol 2007; 56:125–43. using intense pulsed light for treatment of actinic keratosis: clinical 9 Jeffes EW, McCullough JL, Weinstein GD et al. Photodynamic ther- and histopathologic evaluation. Dermatol Surg 2005; 31:33–6. apy of actinic keratosis with topical 5-aminolevulinic acid. A pilot 27 Dover JS, Bhatia AC, Stewart B et al. Topical 5-aminolevulinic acid dose-ranging study. Arch Dermatol 1997; 133:727–32. combined with intense pulsed light in the treatment of photoaging. 10 Szeimies RM, Hein R, Baumler W et al. A possible new incoherent Arch Dermatol 2005; 141:1247–52. lamp for photodynamic treatment of superficial skin lesions. Acta 28 Alster TS, Tanzi EL, Welsh EC. Photorejuvenation of facial skin Derm Venereol (Stockh) 1994; 74:117–19. with topical 20% 5-aminolevulinic acid and intense pulsed light 11 Karrer S, Baumler W, Abels C et al. Long-pulse dye laser for photo- treatment: a split-face comparison study. J Drugs Dermatol 2005; dynamic therapy: investigations in vitro and in vivo. Lasers Surg Med 4:35–8. 1999; 25:51–9. 29 Gold MH, Bradshaw VL, Boring MM et al. Treatment of sebaceous 12 Brancaleon L, Moseley H. Laser and non-laser light sources for gland hyperplasia by photodynamic therapy with 5-aminolevulinic photodynamic therapy. Lasers Med Sci 2002; 17:173–86. acid and a blue light source or intense pulsed light source. J Drugs 13 Juzeniene A, Juzenas P, Ma LW et al. Effectiveness of different light Dermatol 2004; 3:S6–9. sources for 5-aminolevulinic acid photodynamic therapy. Lasers Med 30 Moher D, Schulz KF, Altman DG. The CONSORT statement: revised Sci 2004; 19:139–49. recommendations for improving the quality of reports of parallel- 14 Soler AM, Angell-Petersen E, Warloe T et al. Photodynamic therapy group randomised trials. Lancet 2001; 357:1191–4. of superficial basal cell carcinoma with 5-aminolevulinic acid with 31 Olsen EA, Abernethy ML, Kulp-Shorten C et al. A double-blind, dimethylsulfoxide and ethylendiaminetetraacetic acid: a comparison vehicle-controlled study evaluating masoprocol cream in the treat- of two light sources. Photochem Photobiol 2000; 71:724–9. ment of actinic keratoses on the head and neck. J Am Acad Dermatol 15 Clark C, Bryden A, Dawe R et al. Topical 5-aminolaevulinic 1991; 24:738–43. acid photodynamic therapy for cutaneous lesions: outcome and 32 Kumar S, Tandon OP, Mathur R. Pain measurement: a formidable comparison of light sources. Photodermatol Photoimmunol Photomed task. Indian J Physiol Pharmacol 2002; 46:396–406. 2003; 19:134–41. 33 Ohnhaus EE, Adler R. Methodological problems in the measure- 16 Bjerring P, Christiansen K, Troilius A. Intense pulsed light source ment of pain: a comparison between the verbal rating scale and for the treatment of dye laser resistant port-wine stains. J Cosmet the visual analogue scale. Pain 1975; 1:379–84. Laser Ther 2003; 5:7–13. 34 Fritsch C, Homey B, Stahl W et al. Preferential relative porphyrin 17 Ross EV, Smirnov M, Pankratov M, Altshuler G. Intense pulsed enrichment in solar keratoses upon topical application of delta- light and laser treatment of facial telangiectasias and dyspigmenta- aminolevulinic acid methylester. Photochem Photobiol 1998; 68:218– tion: some theoretical and practical comparisons. Dermatol Surg 21. 2005; 31:1188–98. 35 Pogue BW, Momma T, Wu HC et al. Transient absorption changes 18 Weiss RA, Goldman MP, Weiss MA. Treatment of poikiloderma of in vivo during photodynamic therapy with pulsed-laser light. Br J Civatte with an intense pulsed light source. Dermatol Surg 2000; Cancer 1999; 80:344–51. 26:823–7. 36 Grant WE, Hopper C, Speight PM et al. Photodynamic therapy of 19 Goldman MP, Weiss RA, Weiss MA. Intense pulsed light as a non- malignant and premalignant lesions in patients with ‘field canceri- ablative approach to photoaging. Dermatol Surg 2005; 31:1179–87. zation’ of the oral cavity. J Laryngol Otol 1993; 107:1140–5. 20 Hee Lee J, Huh CH, Yoon HJ et al. Photo-epilation results of axil- 37 Ross EV. Laser versus intense pulsed light: competing technologies lary hair in dark-skinned patients by intense pulsed light: compari- in dermatology. Lasers Surg Med 2006; 38:261–72. son between different wavelengths and pulse width. Dermatol Surg 38 Fijan S, Honigsmann H, Ortel B. Photodynamic therapy of epithe- 2006; 32:234–40. lial skin tumours using delta-aminolaevulinic acid and desferriox- 21 Toosi P, Sadighha A, Sharifian A et al. A comparison study of the amine. Br J Dermatol 1995; 133:282–8. efficacy and side effects of different light sources in hair removal. 39 Fink-Puches R, Hofer A, Smolle J et al. Primary clinical response Lasers Med Sci 2006; 21:1–4. and long-term follow-up of solar keratoses treated with topically 22 Santos MA, Belo VG, Santos G. Effectiveness of photodynamic ther- applied 5-aminolevulinic acid and irradiation by different wave apy with topical 5-aminolevulinic acid and intense pulsed light bands of light. J Photochem Photobiol B 1997; 41:145–51. versus intense pulsed light alone in the treatment of acne vulgaris: 40 Rodriguez R, Sanchez T, Cordoba S. Photodynamic photorejuvenation. comparative study. Dermatol Surg 2005; 31:910–15. Dermatol Surg 2002; 28:742–4.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp111–117 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07884.x Use of oral glycopyrronium bromide in hyperhidrosis V. Bajaj and J.A.A. Langtry* Department of Dermatology, University Hospital of North Durham, Durham DH1 5TW, U.K. *Department of Dermatology, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, U.K.
Summary
Correspondence Background Idiopathic hyperhidrosis may be a disabling condition causing emo- Vrinda Bajaj. tional stress and negative impact on a patient’s quality of life. Oral anticholiner- E-mail: [email protected] gics are one of the treatments available. There are few published data on the use of the anticholinergic drug glycopyrronium bromide (glycopyrrolate) given Accepted for publication 20 December 2006 orally in the treatment of hyperhidrosis. Objectives To report a retrospective analysis describing the treatment responses, Key words doses and side-effects of oral glycopyrrolate in the treatment of idiopathic hyper- glycopyrronium bromide, hyperhidrosis hidrosis. Methods Review of case notes in a series of 24 patients, nine with generalized and Conflicts of interest 15 with localized hyperhidrosis. None declared. Results Fifteen of 19 evaluable patients (79%) responded to oral glycopyrrolate. This study was presented as a poster at the British However, treatment was limited by side-effects in around one third of patients. Association of Dermatologists Annual Conference in Conclusions A prospective clinical study to compare the efficacy and side-effects of Manchester, July 2006. oral anticholinergics is warranted.
Idiopathic hyperhidrosis may be a disabling condition causing undertaken. These did not reveal a cause for the hyper- emotional stress and negative impact on a patient’s quality of hidrosis in any of the patients. life. Oral anticholinergics are one of the treatments available. Systemic unwanted effects may limit their use, although gly- Results copyrronium bromide (glycopyrrolate) is thought to have fewer adverse effects than other anticholinergics.1 Topical Over the 4-year period 24 patients were treated with oral glyco- glycopyrrolate may be effective in treating gustatory hyper- pyrrolate for idiopathic hyperhidrosis. The patients ranged in hidrosis.2 Glycopyrrolate–tap water iontophoresis has shown age from 19 to 62 years (mean 33). In this series 70% were efficacy in treating palmoplantar hyperhidrosis.3,4 There is a women and 30% men. The duration of hyperhidrosis ranged dearth of published reports on the use of oral glycopyrrolate from 18 months to a patient-stated time of ‘life-long’. The in idiopathic hyperhidrosis.5 We report a retrospective analysis mean age at onset was 25 years but did not include the two describing the treatment responses, doses and side-effects patients whose stated duration was ‘life-long’. of oral glycopyrrolate in the treatment of idiopathic hyper- hidrosis. Sites of hyperhidrosis
Patients and methods Figure 2 shows the frequencies for the sites of sweating, with some patients having a combination of sites involved. The Case notes of 24 patients with hyperhidrosis who had been most commonly involved areas were the axillae in nine treated with oral glycopyrrolate between 2001 and 2004 at patients, and palms or soles in six; hyperhidrosis was general- Sunderland Royal Hospital were retrospectively reviewed. The ized in nine patients. following information was obtained: demographic details, site of sweating, previous treatments, dose of and response Other treatments to oral glycopyrrolate, and side-effects. A linear analogue scale (LAS; Fig. 1) was scored by the patient for severity of Three patients had not had any previous treatments. Medical hyperhidrosis and quality of life, before and during treat- treatments used other than oral glycopyrrolate included topical ment. An underlying cause for the excessive sweating was aluminium chloride (n ¼ 17), beta-blockers [atenolol (n ¼ sought in 17 patients (including all those with generalized 2), propanolol (n ¼ 4), not specified (n ¼ 1)], diltiazem hyperhidrosis), in whom at least a drug history and prelim- (n ¼ 4), clonidine (n ¼ 3) and propantheline (n ¼ 2) inary biochemical and haematological investigations were (Table 1). These were either ineffective or not tolerated. One
2007 The Authors 118 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp118–121 Use of oral glycopyrronium bromide in hyperhidrosis, V. Bajaj and J.A.A. Langtry 119
and four patients varied between 2 mg twice and thrice daily No treatment depending on their symptoms. 0 Severity left palm 10 X 0 Severity right palm 10 X Treatment response 010Affects quality of life X Objective evaluation including Minor’s iodine starch test or After glycopyrrolate 2 mg bd gravimetric testing of treatment response had not been con- 0 Severity left palm 10 ducted and was therefore not available in this retrospective X 10 0 Severity right palm study. The degree of treatment response could not be accur- 010X Affects quality of life ately assessed from the patients’ case notes, so the absolute X responses were recorded. There was no follow-up on five patients. One patient was discharged from the clinic and four did not attend follow-up. Fig 1. Example of a linear analogue scale for a patient with palmoplantar hyperhidrosis, showing improvement in severity and Of the remaining 19 patients with hyperhidrosis, 15 (79%) quality of life. bd, twice daily. responded to glycopyrrolate. A patient-assessed LAS (0–10 points) of the severity of hyperhidrosis and quality of life before and after treatment with glycopyrrolate was completed in only eight of the 19 9 patients. The response was rated as excellent (‡9 points) in 8 three patients, improvement (up to 8 points) in four patients and no improvement in one patient. 7 The three patients who showed an excellent improvement 6 were a 41-year-old woman with generalized hyperhidrosis, a 34-year-old woman with axillary/palmoplantar hyperhidrosis 5 and a 19-year-old woman with palmar hyperhidrosis. Of the 4 four patients who demonstrated some improvement, two (a
No. of patients No. 23-year-old woman with axillary/palmoplantar hyperhidrosis 3 and a 20-year-old woman with palmar hyperhidrosis) repor- ted a 5-point LAS improvement and the other two (a 31-year- 2 old woman with truncal hyperhidrosis and a 26-year-old man 1 with axillary/palmoplantar hyperhidrosis) reported a 3-point LAS improvement in the severity of hyperhidrosis and quality 0 of life. A 44-year-old woman with genitocrural hyperhidrosis
Axilla Tr unk did not find any improvement after taking glycopyrrolate Palmar Axilla/head 2 mg three times daily and subsequently failed to respond to Generalised Genitocrural Head andFace trunk and scalp localized injections of botulinum toxin type A. Axilla/palmoplantar Glycopyrrolate was effective and tolerated well with no Generalised/axilla/soles side-effects in a patient in whom propantheline had been in- Site effective (a 62-year-old man with generalized hyperhidrosis). This patient is now in remission and takes the glycopyrrolate Fig 2. Sites of hyperhidrosis. on an as-required basis. However, the second patient on pro- pantheline (a 44-year-old woman with genitocrural hyper- patient had undergone sympathectomy which had helped the hidrosis), who developed xerostomia, also had to stop palmar but not the axillary hyperhidrosis. None of the patients glycopyrrolate for the same reason. had iontophoresis. Botulinum toxin type A intradermal injec- Of the three patients who were started on glycopyrrolate as tions to the axillae of one patient were successful, after treat- first-line treatment for their hyperhidrosis, two responded well ment with oral glycopyrrolate failed. (a 53-year-old man with generalized hyperhidrosis and a 57- year-old woman with generalized hyperhidrosis) and the third (a 36-year-old woman with face and scalp hyperhidrosis) takes Dose glycopyrrolate on an intermittent basis due to side-effects. Treatment was started with glycopyrrolate 2 mg twice daily and the dose was increased according to control of sweating. Side-effects Eight patients were treated with glycopyrrolate 2 mg twice daily and 10 tolerated 2 mg three times daily. One patient Our data show that 15 of 19 (79%) patients developed side- with generalized hyperhidrosis required up to 4 mg twice effects. Dry mouth was the commonest adverse effect (12 daily for symptom control. One patient took 2 mg once daily patients) and occurred in eight patients whose doses were
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp118–121 120 Use of oral glycopyrronium bromide in hyperhidrosis, V. Bajaj and J.A.A. Langtry
Table 1 Doses and duration of treatment of other agents used and reason for stopping
Topical aluminium chloride Atenolol Propanolol Diltiazem Clonidine Propantheline Dose NA 50 mg daily 80 mg daily 60 mg three 50 lg once daily 15 mg three times times daily increasing to twice daily and 30 mg daily after 4 weeks four times daily then three times daily after a further 4 weeks Duration 1 day–4 months 2–4 months 2 years 3–4 months 4 months 4 months No. of patients using the agent 17 2 4 4 3a 2 before glycopyrronium bromide Side-effects (no. of patients) 12b 1c 001d 1e Ineffective (no. of patients) 5 1 4 4 1 1
aThe reason for stopping was not specified in one patient. bMany patients found this too irritant including three patients who used this with topical steroids. cNausea; dlethargy; exerostomia. NA, not applicable. higher than 2 mg twice daily. It resulted in three patients thought to be the mechanism for the fewer central nervous stopping this treatment. However, nine patients were able to system side-effects of glycopyrrolate. It may also have less tolerate this symptom and continued with glycopyrrolate. effect on the heart rate at lower doses. However, all anticho- Around one third of patients had to stop glycopyrrolate due linergics result in inhibition of salivation. to side-effects (xerostomia, n ¼ 3; erectile dysfunction, n ¼ 1; The intravenous use of glycopyrrolate in general anaesthesia headaches and urinary retention, n ¼ 1). There was no appar- has been well described; however, information regarding its ent correlation between higher dose and the development of oral use is limited. Its safety and efficacy have been outlined these additional side-effects in this small group of patients. in a number of case series for the control of drooling in Four patients stopped glycopyrrolate due to ineffectiveness. In patients with disabilities.9,10 Most of such reports have been total, around one half discontinued treatment because of side- in children; there has been only one adult case reported. In effects or lack of efficacy. the latter, control of drooling was achieved while the patient was on oral glycopyrrolate 3–4 mg once daily. Its use in Discussion hyperhidrosis is limited, with one case of diabetic gustatory sweating being treated with glycopyrrolate at up to 1 mg Hyperhidrosis affects around 1% of the U.K. population.6 twice daily.11 Control of sweating was achieved with minimal Treatment options depend on the site of hyperhidrosis and side-effects of xerostomia at doses of 1 mg once daily after a include a variety of medical and surgical methods as well as year of follow-up. Klaber and Catterall advocate the use of gly- iontophoresis and localized intradermal injections of botuli- copyrrolate at doses of up to 2 mg three times daily, particu- num toxin type A.7 Although it is not licensed for use in larly in young women with generalized hyperhidrosis.5 It hyperhidrosis, our interest in the use of oral glycopyrrolate should, however, be used with caution in those with closed started after the successful outcome in a patient who had angle glaucoma, bladder outflow obstruction, gastro-oesopha- proved difficult to control. geal reflux disease and cardiac insufficiency. It is contraindi- The mechanism of action of anticholinergic drugs is the cated in those with myasthenia gravis, paralytic ileus and competitive antagonism of acetylcholine at the muscarinic pyloric stenosis. receptor. The pharmacological action of anticholinergics is not Although hyperhidrosis affects men and women equally, limited to sweat ducts as muscarinic receptors are present 70% of our patient group was female. It has been suggested throughout the central and autonomic nervous system. At least that women are intolerant of sweating levels that men tend to five subtypes of muscarinic receptors (M1–M5) have been ignore and may be more likely to approach the medical pro- identified.8 M1 and M4 are found predominantly in neuronal fession for treatment.12 Palmoplantar hyperhidrosis tends to tissue, M2 has been located to the heart, M3 is found in glan- present in childhood or around puberty, whereas axillary dular tissue and the M5 receptor role is to regulate cerebral hyperhidrosis is uncommon before puberty. The variety of blood flow. The highly polar quaternary ammonium group of body sites involved in this case series may account for the glycopyrrolate limits its passage across lipid membranes, such mean age at onset being 25 years. This is likely to be an over- as the blood-brain barrier, in contrast to atropine sulphate estimate as it did not include those two patients who stated and scopolamine hydrobromide, which are highly nonpolar that their hyperhidrosis was ‘life-long’ and were unable to tertiary amines that penetrate lipid barriers easily.1 This is estimate the age at onset.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp118–121 Use of oral glycopyrronium bromide in hyperhidrosis, V. Bajaj and J.A.A. Langtry 121
Propantheline, like glycopyrrolate, is also a quaternary 3 Abell E, Morgan K. The treatment of idiopathic hyperhidrosis by amine compound and the only anticholinergic licensed for use glycopyrronium and tap water iontophoresis. Br J Dermatol 1974; in gustatory hyperhidrosis, although the evidence for its oral 91:87–91. 4 Dolianitis C, Scarff CE, Kelly J et al. Iontophoresis with glycopyrro- use is limited to anecdotal reports.13,14 In our case series two late for the treatment of palmoplantar hyperhidrosis. Australas patients had tried propantheline before glycopyrrolate. One J Dermatol 2004; 45:208–12. patient stopped propantheline (15 mg three times daily) as it 5 Klaber M, Catterall M. Treating hyperhidrosis. Anticholinergic was ineffective, and found glycopyrrolate to be effective at the drugs were not mentioned. BMJ 2000; 321:703 (Letter). dose of 2 mg three times daily without developing adverse 6 Collin J, Whatling P. Treating hyperhidrosis. Surgery and botuli- effects. The second patient found propantheline effective at num toxin are treatments of choice in severe cases. BMJ 2000; high doses of 30 mg four times daily but stopped this due to 320:1221–2. 7 Langtry JAA. Hyperhidrosis. In: Treatment of Skin Disease (Lebwohl xerostomia and also was unable to tolerate glycopyrrolate MG, Heymann WR, Berth-Jones J, Coulson I, eds), 2nd edn. Phil- 2 mg three times daily for the same reason. There are no adelphia: Elsevier Mosby, 2006; 288–91. studies comparing the efficacy and side-effects of glycopyrro- 8 Matsui M, Yamada S, Oki T et al. Functional analysis of muscarinic late and propantheline. Glycopyrrolate tablets are not widely acetylcholine receptors using knockout mice. Life Sci 2004; available15 and are significantly more expensive than propan- 75:2971–81. theline. Glycopyrrolate may be made up as an oral solution 9 Blasco PA, Stansbury JC. Glycopyrrolate treatment of chronic drooling. but it has a short shelf life. Arch Pediatr Adolesc Med 1996; 150:932–5. 10 Stern LM. Preliminary study of glycopyrrolate in the management In conclusion, glycopyrrolate was effective in 75% of patients of drooling. J Paediatr Child Health 1997; 33:52–4. with both generalized and localized hyperhidrosis; however, 11 Edick CM. Oral glycopyrrolate for treatment of diabetic gustatory treatment was limited by side-effects in around one third of sweating. Ann Pharmacother 2005; 39:1760. those treated. A prospective clinical study to compare the effi- 12 Moschella SL, Hurley HJ. Diseases of the eccrine sweat glands. In: cacy and side-effects of oral anticholinergics is warranted. Dermatology (Moschella SL, Hurley HJ, eds), 3rd edn. Philadelphia: WB Saunders, 1992; 1518. 13 Cunliffe WJ, Johnson CE. Gustatory hyperhidrosis. A complication References of thyroidectomy. Br J Dermatol 1967; 79:519–26. 14 Canaday BR, Stanford RH. Propantheline bromide in the manage- 1 Ali-Melkkila T, Kanto J, Lisalo E. Pharmacokinetics and related ment of hyperhidrosis associated with spinal cord injury. Ann Phar- pharmacodynamics of anticholinergic drugs. Acta Anaesthesiol Scand macother 1995; 29:489–92. 1993; 37:633–42. 15 Kavanagh GM, Burns C, Aldridge RD. Topical glycopyrrolate 2 Kim WO, Kil HK, Yoon DM, Cho MJ. Treatment of compensatory should not be overlooked in treatment of focal hyperhidrosis. Br J gustatory hyperhidrosis with topical glycopyrrolate. Yonsei Med J Dermatol 2006; 155:487 (Letter). 2003; 44:579–82.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp118–121 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07896.x Oral R115866 in the treatment of moderate to severe facial acne vulgaris: an exploratory study C.J. Verfaille,* M. Coel, I.H. Boersma,§ J. Mertens, M. Borgers* and D. Roseeuw– *Departments of Dermatology and Molecular Cell Biology, GROW, Maastricht University, Maastricht, The Netherlands Barrier Therapeutics NV, Cipalstraat 3, B-2440 Geel, Belgium Private Practice in Dermatology, Nijverheidsstraat 75, B-1800 Vilvoorde, Belgium §Department of Dermatology, Albert Schweitzer Hospital, NL-3331 LZ Zwijndrecht, The Netherlands –Department of Dermatology, Free University Brussels (AZ-VUB), B-1090 Jette, Belgium
Summary
Correspondence Background R115866 (RambazoleTM; Barrier Therapeutics NV, Geel, Belgium), a Christel Verfaille. new-generation retinoic acid metabolism-blocking agent, is a nonretinoid com- E-mail: [email protected] pound enhancing intracellularly the endogenous levels of all-trans-retinoic acid by blocking its catabolism. By virtue of this property, and the proven positive effects Accepted for publication 6 January 2007 of retinoids in the treatment of acne, R115866 could potentially be a useful drug for acne. Key words Objectives To explore the efficacy, safety and tolerability of systemic R115866 in acne, R115866, RambazoleTM, retinoic acid, male patients with moderate to severe facial acne vulgaris (at least 15 papules retinoic acid metabolism-blocking agent and/or pustules and at least two nodulocystic lesions). Conflicts of interest Methods In this exploratory trial, 17 patients were treated with oral R115866 C.J.V and J.M. are employees of, and own stock 1 mg once daily for 12 weeks, followed by a 4-week treatment-free period. options in, Barrier Therapeutics NV. M.B. owns Results At the end of treatment (week 12, n ¼ 16) a mean reduction in inflamma- stock options in Barrier Therapeutics NV. D.R. tory lesion count of 77Æ4% (P <0Æ001), in noninflammatory lesion count of has received a fee for speaking from Barrier 58Æ3% (P <0Æ001) and in total lesion count of 76Æ0% (P <0Æ001) was observed Therapeutics NV. as compared with baseline. All lesion counts were significantly reduced from
Data from this study were presented at the 14th week 4 onwards. Mild side-effects were reported occasionally. Congress of the European Academy of Dermatology Conclusions The current data indicate that treatment with oral R115866 1 mg once and Venereology, London, 2005. daily for 12 weeks in patients with moderate to severe facial acne vulgaris is effi- cacious and well tolerated and merits further investigation.
R115866 (RambazoleTM; Barrier Therapeutics NV, Geel, Bel- dynamic property and the well-known modulating effects of all- gium) is a triazole identified as a new-generation retinoic acid trans-RA on epidermal and sebocyte growth and differentiation, metabolism-blocking agent (RAMBA). At nanomolar concen- R115866 is potentially of use for disorders associated with trations, R115866 selectively inhibits cytochrome P-450 either aberrant keratinization/desquamation, such as psoriasis, (CYP) dependent hydroxylases involved in the catabolism of or when combined with aberrant sebum production/excre- all-trans-retinoic acid (all-trans-RA) whereas micromolar con- tion, such as acne vulgaris. centrations are needed to inhibit other CYPs.1 By blocking the The aim of this exploratory trial was to assess efficacy, tol- breakdown of endogenous all-trans-RA, RAMBAs transiently erability and safety of oral R115866 administered at 1 mg increase endogenous cellular all-trans-RA levels. Animal experi- once daily for 12 weeks in patients with moderate to severe ments in rats confirmed the in vivo suppressive effects of facial acne vulgaris. R115866 on all-trans-RA metabolism. Similar to all-trans-RA, R115866 exhibited a modulating effect on epithelial growth Patients and methods and differentiation in various animal models of keratinization.2 In addition, topical application of R115866 revealed a reduc- Study design tion in utriculus size in rhino mouse skin and a reduction of mRNA of proinflammatory cytokines, such as interleukin-1a, Patients were enrolled in three study centres in Belgium and the in skin biopsies of healthy volunteers, suggesting a possible Netherlands. The protocol was approved by the independent anti-inflammatory activity.3,4 By virtue of this pharmaco- ethics committee and patients provided written informed
2007 The Authors 122 Journal Compilation 2007 Barrier Therapeutics NV • British Journal of Dermatology 2007 157, pp122–126 Oral R115866 in facial acne vulgaris, C.J. Verfaille et al. 123 consent prior to study enrolment. Clinical investigations were Safety and tolerability evaluation conducted according to the Declaration of Helsinki principles. Adverse events (AEs) were recorded throughout the study. Blood samples for haematology, biochemistry and endocrinol- Patient population ogy were taken at every visit. Urinalysis was performed at Healthy male patients aged between 16 and 50 years (with at screening. Cardiovascular safety was monitored at each visit. least 15 papules and/or pustules and at least two nodulocystic At all visits, the severity of all-trans-RA-related cutaneous side- lesions) were eligible. Washout periods before entering the effects (xerosis, pruritus, cheilitis, epistaxis, periungual gra- study were 2 weeks for topical or ultraviolet treatment and nulomatous reaction, hair loss and fragile skin) was evaluated 4 weeks for previous systemic treatment of acne. Main exclu- (none, mild, moderate or severe). sion criteria included other types of acne, concurrent inflam- matory disease, coexisting serious disease, hyperlipidaemia, Statistics clinically relevant electrocardiogram (ECG) abnormalities, heart disorders, use of drugs that could interfere with either acne or The intent-to-treat population consisted of all patients assigned the study drug, use of vitamin A (>1000 lg daily) or oral to treatment (¼ safety population). The per-protocol popula- retinoids in the 6 months prior to screening, known intoler- tion consisted of all enrolled patients who used trial medica- ance to retinoids and the suspicion of alcohol or drug abuse. tion and no prohibited therapy for the specified period and who were evaluable at end of treatment (¼ primary popula- tion for analysis of efficacy). Statistical significance of changes Treatment in lesion counts and overall acne severity compared with base- Patients took two 0Æ5-mg capsules of R115866 (daily dose of line values was assessed with the Wilcoxon signed rank 1 mg) daily for 12 weeks followed by a 4-week treatment- test. Global assessments of severity were compared with free period. ‘no change’ (¼ score 1) as hypothesized outcome using the Wilcoxon signed rank test. Two-sided P-values £ 0Æ05 were considered statistically significant.6 Clinical evaluations
Clinical evaluations were performed by the same investigator Results per site over the course of the study at baseline, at week 4, 8 and 12 of treatment, and at end of follow-up. Photographs of Patients the facial area were taken under standardized conditions. Twenty patients were screened, and 17 (male, white-skinned) enrolled and treated with R115866. There were three screen- Facial acne-related lesion count ing failures due to an insufficient number of lesions. Mean Acne-related lesions (open/closed comedones, papules, pustules age (range) was 22Æ3 (13–36) years and mean weight (range) and nodules) were counted by the investigator on the forehead, was 71Æ3 (55–97) kg. One patient discontinued the study (lost left/right cheek, chin and nose, resulting in a separate count to follow-up) after 4 weeks of treatment. Sixteen patients per lesion type, a total noninflammatory lesion count (open completed the entire study. and closed comedones), a total inflammatory lesion count (papules, pustules and nodules) and a total acne lesion count. Clinical evaluations
Overall acne severity Facial acne-related lesion count
Both investigator and patient evaluated the overall facial acne Detailed information on the acne lesion counts is given in severity on a five-point scale in which 0 ¼ none (clear) and Tables 1 and 2. 4 ¼ severe (numerous or extensive papules or pustules, many At baseline the mean (range) for the total lesion count was nodules).5 81Æ3 (22–199), for the total inflammatory lesion count it was 56Æ2 (15–170) and for the total noninflammatory lesion count it was 25Æ1 (0–60). Global assessment of effectiveness Compared with baseline, a statistically significant reduction A global evaluation of the patient’s overall disease severity was seen in both total noninflammatory (P ¼ 0Æ016) and total relative to baseline was made by investigator and patient on a inflammatory lesion count (P ¼ 0Æ018) from week 4 of the six-point scale: 0, worse; 1, unchanged (<10% improve- treatment onwards, and a statistically significant reduction was ment); 2, slight improvement (approximately 25%); 3, mod- seen at all time points for the total lesion count (P £ 0Æ001). erate improvement (approximately 50%); 4, marked The mean percentage reduction in total lesion count was improvement (approximately 75%); and 5, complete or 31Æ5% after 4 weeks of treatment, 76Æ0% at the end of treat- almost complete clearance (90–100%). ment (week 12; P <0Æ001) and 77Æ2% at the end of follow-up
2007 The Authors Journal Compilation 2007 Barrier Therapeutics NV • British Journal of Dermatology 2007 157, pp122–126 124 Oral R115866 in facial acne vulgaris, C.J. Verfaille et al.
Table 1 Summary of total noninflammatory Baseline Treatment Follow-up lesion count (N-ILC) and total inflammatory lesion count (ILC) per patient at baseline, at Subject Week 0 Week 4 Week 8 Week 12 Week 16 week 4, 8 and 12 of treatment with R115866 and at follow-up N-ILC/ILC N-ILC/ILC N-ILC/ILC N-ILC/ILC N-ILC/ILC 1 29/170 29/11 31/28 29/5 10/11 2 37/84 19/29 21/34 17/23 15/56 3 60/35 39/43 22/22 14/13 19/12 4 40/75 45/77 52/94 28/37 24/36 5 29/56 18/33 0/2 0/3 0/3 6 0/22 0/11 0/9 0/0 0/0 7 16/30 0/3 0/1 0/1 0/4 8 23/40 0/12 0/4 0/1 0/2 9 10/15 0/30 0/2 0/3 0/1 10 4/19 3/18 0/4 0/0 0/0 11 48/45 35/36 17/31 10/15 0/5 12 21/51 11/47 11/53 12/25 19/19 13 13/36 13/46 11/31 6/16 6/19 14 26/63 24/56 7/32 9/25 18/21 15 31/62 5/40 7/26 15/14 3/3 16 14/96 31/58 30/50 17/24 15/20 Mean 25Æ1/56Æ217Æ0/34Æ413Æ1/26Æ49Æ8/12Æ88Æ1/13Æ3 SD 16Æ0/38Æ315Æ3/20Æ215Æ0/24Æ79Æ8/11Æ58Æ9/15Æ3 Median 24Æ5/48Æ015Æ5/34Æ59Æ0/27Æ09Æ5/13Æ54Æ5/8Æ0 Min 0/15 0/3 0/1 0/0 0/0 Max 60/170 45/77 52/94 29/37 24/56
Table 2 Summary of lesion counts reduction (%) vs. baseline at week showing at least 90% reduction. The mean percentage reduc- 4, 8 and 12 of treatment and at follow-up tion of the lesion counts was still improving during follow-up, reaching 77Æ2% (P <0Æ001) for total, 78Æ4% (P <0Æ001) for Week 4 Week 8 Week 12 Week 16 total inflammatory and 63Æ2% (P <0Æ001) for total noninflam- Total count of all acne lesions matory lesion count. Mean 31Æ556Æ076Æ077Æ2 SD 35Æ335Æ319Æ222Æ4 Median 21Æ657Æ672Æ390Æ4 Overall acne severity Min )20 )27 43 41 The mean (range) investigator’s baseline score was 3Æ3 (3–4), Max 93 98 100 100 P-value 0Æ001 <0Æ001 <0Æ001 <0Æ001 reaching 1Æ3 (0–3) at end of treatment and 1Æ6 (0–3) at fol- Total count of inflammatory lesions low-up. The mean (range) patients’ baseline score was Mean 23Æ553Æ777Æ478Æ4 3Æ1 (2–4), reaching 1Æ3 (0–2) at end of treatment and 1Æ2 SD 49Æ136Æ218Æ321Æ0 (0–2) at follow-up. Both investigator’s and patients’ scores Median 27Æ758Æ676Æ287Æ8 improved significantly from week 4 onwards (P ¼ 0Æ002 and ) ) Min 100 25 51 33 P <0Æ001, respectively) reaching P £ 0Æ001 for subsequent Max 94 97 100 100 visits. P-value 0Æ018 <0Æ001 <0Æ001 <0Æ001 Total count of noninflammatory lesions Mean 29Æ945Æ858Æ363Æ2 Global assessment of effectiveness SD 55Æ660Æ239Æ838Æ7 Median 31Æ064Æ059Æ766Æ9 A statistically significant improvement in both the physician’s Min )121 )114 )21 )7 and the patients’ global efficacy score was obtained from week Max 100 100 100 100 4 onwards until the end of treatment and follow-up P-value 0Æ016 0Æ009 <0Æ001 <0Æ001 (P <0Æ001). The mean (range) global efficacy score at end of n ¼ 16 for each group, exact two-sided P-value from Wilcoxon treatment was 3Æ8 (1–5) when scored by the physician and signed rank test comparison with value at baseline. 3Æ9 (2–5) when scored by the patients. At follow-up, the scores were 3Æ6 (0–5) and 3Æ9 (0–5), respectively.
(week 16; P <0Æ001). At end of treatment, the mean percen- Safety and tolerability evaluation tage reduction was 77Æ4% (P <0Æ001) for total inflammatory and 58Æ3% (P <0Æ001) for total noninflammatory lesion count, R115866 was well tolerated, with 16 AEs recorded in nine of with respectively six of 16 (38%) and five of 16 (31%) patients 17 patients: 10 were mild, five moderate and one severe
2007 The Authors Journal Compilation 2007 Barrier Therapeutics NV • British Journal of Dermatology 2007 157, pp122–126 Oral R115866 in facial acne vulgaris, C.J. Verfaille et al. 125
(eczema); all were resolved at the end of the trial. Six AEs, (a) including the case of severe eczema, were interpreted as not drug related, and three as doubtfully, six as possibly and one (mild stomach discomfort) as very likely related to treatment. AEs reported in more than one patient were dry skin (n ¼ 3), contact dermatitis (n ¼ 2) and eczema (n ¼ 2). No deaths or serious AEs were reported. No clinically relevant ECG or laboratory abnormalities were reported. Elevation above normal values for triglycerides was observed in three patients during treatment but none was reported as an AE. At baseline, no cutaneous all-trans-RA-related side-effects (b) were reported. During treatment, the all-trans-RA-related side- effects were scored mild to moderate in the majority of the patients. Compared with baseline, a significant aggravation of xerosis (4 and 8 weeks, P £ 0Æ01; 12 weeks, P ¼ 0Æ008), cheilitis (4, 8 and 12 weeks, P £ 0Æ01) and pruritus (4 weeks, P ¼ 0Æ031) was observed. All all-trans-RA-related side-effects had disappeared after 4 weeks of treatment-free follow-up, except for xerosis (mild, n ¼ 1) and hair loss (mild, n ¼ 3).
Discussion Fig 2. Moderate papulopustular acne that almost completely cleared with minimal scarring after 12 weeks of treatment with R115866 Treatment with oral R115866 1 mg once daily for 12 weeks 1 mg daily: (a) before treatment and (b) after treatment. in patients with moderate to severe facial acne vulgaris resul- ted in a statistically significant improvement for all assessed efficacy variables (Figs 1, 2). Improvement was still pro- nounced 4 weeks after the last drug intake. This excellent treatment effect was confirmed by the acne severity scores and overall effectiveness scores by investigator and patients. (a) In a double-blind placebo-controlled study with oral tazaro- tene, a mean 56% reduction in noninflammatory lesion count was achieved after 12 weeks of treatment.7 These data are in line with our data showing a 58Æ3% mean reduction after 12 weeks of treatment, further improving to 63Æ2% at follow- up. The mean percentage reduction in total count of papules and pustules after 12 weeks of treatment with oral R115866 1 mg daily was 76% vs. 50% with oral tazarotene 3 mg and 52% with oral tazarotene 6 mg daily; the mean percentage reduction in the nodulocystic lesions was 81Æ3% for R115866 vs. 65% and 71% for oral tazarotene 3 mg and 6 mg, respect- ively.7 These preliminary results suggest that R115866 is equi- potent in improving the noninflammatory lesions but may (b) exhibit a greater efficacy in improving inflammatory lesions. In addition, the onset of action of R115866 was fast. In this patient group, R115866 was well tolerated with an acceptable safety profile. Elevated triglyceride levels, a side- effect often seen after oral isotretinoin treatment, were repor- ted in three of 17 patients but none was reported as an AE.8 Oral ingestion of synthetic retinoids results in a significant increase of the retinoid levels in blood and consequently in tissue, possibly resulting in undesirable effects (e.g. hepato- toxicity).9–11 R115866 selectively increases the intracellular levels of all-trans-RA in those cell systems where CYP enzymes are present that specifically metabolize endogenous all-trans- 1,12–14 Fig 1. Healing of severe papular facial acne with some residual RA, such as CYP 26. This targeted selective increase is discolouring in the submandibular area. (a) Before treatment and expected to result in a better safety profile compared with oral (b) after 12 weeks of treatment with oral R115866 1 mg daily. retinoids.15
2007 The Authors Journal Compilation 2007 Barrier Therapeutics NV • British Journal of Dermatology 2007 157, pp122–126 126 Oral R115866 in facial acne vulgaris, C.J. Verfaille et al.
The results in a limited number of patients indicate that 5 U.S. Food and Drug Administration. Center for Drug Evaluation R115866 is efficacious and well tolerated, with equally good and Research. Guidance for Industry. Acne Vulgaris: Developing Drugs for effects on inflammatory and noninflammatory lesions. Treatment. http://www.fda.gov/cder/guidance/6499dft.htm (last accessed 31 January 2007). R115866 deserves further clinical investigation for dose and 6 SAS version 8.2. Cary, NC: SAS Institute. duration in patients with moderate to severe facial acne 7 Tan JKL, Lew-Kaya D, Walker PS. Oral tazarotene reduces come- vulgaris. dones. J Am Acad Dermatol 2004; 50 (Suppl.):P20 (#78). 8 Accutane (Oral Isotretinoin Capsules) Prescribing Information. http://www.fda. gov/cder/foi/label/2005/018662s056lbl.pdf (last accessed 31 July Acknowledgments 2006). The authors thank Luc Wouters for the statistical analysis. 9 Penniston KL, Tanumihardjo SA. The acute and chronic toxic effects of vitamin A. Am J Clin Nutr 2006; 83:191–201. 10 Kurlandsky SB, Gamble MV, Ramakrishnan R et al. Plasma delivery References of retinoic acid to tissues in the rat. J Biol Chem 1995; 270:17850– 7. TM 1 Rambazole (R115866). Investigational Drug Brochure, version 4.0. Geel, 11 Allen JG, Bloxham DP. The pharmacology and pharmacokinetics of Belgium: Barrier Therapeutics (data on file), 2006. the retinoids. Pharmacol Ther 1989; 40:1–27. 2 Stoppie P, Borgers M, Borghgraef P et al. R115866 inhibits all-trans- 12 Napoli JL. Retinoic acid: its biosynthesis and metabolism. Prog retinoic acid metabolism and exerts retinoidal effects in rodents. Nucleic Acid Res Mol Biol 1999; 63:139–88. J Pharmacol Exp Ther 2000; 293:304–12. 13 Napoli JL. Interactions of retinoid binding proteins and enzymes TM 3 Stoppie P, van Wauwe J, Wouters L et al. Rambazole , a potent in retinoid metabolism. Biochim Biophys Acta 1999; 1440:139– inhibitor of all-trans-retinoic acid metabolism, applied topically to 62. mouse tail skin exerts retinoid mimetic activity. J Invest Dermatol 14 Petkovich PM. Retinoic acid metabolism. J Am Acad Dermatol 2001; 2004; 122:A70 (#420). 45:S136–42. 4 Cools M, Pavez-Lorie E, Borgers M et al. Effect of topical R115866 15 David M, Hodak E, Lowe NJ. Adverse effects of retinoids. Med TM (Rambazole ) gel on the expression of biomarkers in the skin of Toxicol 1988; 3:273–88. healthy volunteers. J Invest Dermatol 2006; 126:S40 (#240).
2007 The Authors Journal Compilation 2007 Barrier Therapeutics NV • British Journal of Dermatology 2007 157, pp122–126 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07947.x Mycophenolate mofetil for severe childhood atopic dermatitis: experience in 14 patients M. Heller,* H.T. Shin,* S.J. Orlow* and J.V. Schaffer* *Ronald O. Perelman Department of Dermatology and Department of Pediatrics, New York University School of Medicine, New York, NY 10016, U.S.A. Hackensack University Medical Center, Hackensack, NJ, U.S.A
Summary
Correspondence Background Reports of successful treatment of atopic dermatitis (AD) with myco- Julie V. Schaffer. phenolate mofetil (MMF) have thus far been limited to adults. Considering that E-mail: [email protected] the condition typically develops during childhood and is most active during this period, MMF would represent a valuable addition to the therapeutic armamentar- Accepted for publication 21 January 2007 ium for paediatric AD. Objectives To evaluate the safety and efficacy of MMF in the treatment of severe Key words childhood AD. atopic dermatitis, mycophenolate mofetil Methods A retrospective analysis was performed of all children treated with MMF as systemic monotherapy for severe, recalcitrant AD between August 2003 and Conflicts of interest August 2006 at New York University Medical Center. Fourteen patients meeting None declared. these criteria were identified. Results Four patients (29%) achieved complete clearance, four (29%) had > 90% improvement (almost complete), five (35%) had 60–90% improvement and one (7%) failed to respond. Initial responses occurred within 8 weeks (mean 4 weeks), and maximal effects were attained after 8–12 weeks (mean 9 weeks) ) ) at MMF doses of 40–50 mg kg 1 daily in younger children and 30–40 mg kg 1 daily in adolescents. The medication was well tolerated in all patients, with no infectious complications or development of leucopenia, anaemia, thrombocyto- penia or elevated aminotransferases. Conclusions This retrospective case series demonstrates that MMF can be a safe and effective treatment for severe, refractory AD in children. MMF represents a prom- ising therapeutic alternative to traditional systemic immunosuppressive agents with less favourable side-effect profiles, and prospective controlled studies are warranted, further to assess its benefits in paediatric AD.
Atopic dermatitis (AD) is a chronic inflammatory skin disease preclude the chronic use of systemic corticosteroids in such with a typical onset during infancy or early childhood. This instances, and phototherapy is often inconvenient and of lim- intensely pruritic dermatosis affects 10–20% of children and ited efficacy.5,6 Although ciclosporin is considered as a first- frequently leads to physical discomfort, skin infections, sleep line systemic agent for refractory AD, risks of nephrotoxicity disturbances, psychological distress, impaired social function- and hypertension restrict its long-term administration.4,5,7 ing and disrupted family dynamics.1,2 In most patients, AD More recently, azathioprine therapy with dosing guided by can be successfully managed with a therapeutic approach that the patient’s erythrocyte thiopurine methyltransferase level has employs use of topical corticosteroids for flares (supplemented been recognized as a relatively safe systemic option that can by systemic antibiotics and antihistamines as needed) and a be effective in children and adults with AD.8–10 Additional combination of emollients, avoidance of triggers and (in more systemic medications with an improved risk-benefit ratio severe cases) intermittent use of topical corticosteroids and ⁄or would represent important additions to the therapeutic arm- topical calcineurin inhibitors for maintenance.3,4 amentarium for paediatric AD. However, some children with severe, recalcitrant AD Mycophenolate mofetil (MMF) is a safer, more bioavail- require more aggressive treatment to achieve disease control. able, esterified form of mycophenolic acid (MPA). Following Unacceptable side-effects such as decreased linear growth and ingestion, MMF is hydrolysed to MPA (its active metabolite), osteoporosis as well as rebound exacerbations upon tapering which is rapidly inactivated by glucuronidation in the liver
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 127 128 Mycophenolate mofetil for childhood atopic dermatitis, M. Heller et al. and subsequently converted back to its active form by tial side-effects at each office visit and periodic evaluation of b-glucuronidase within the epidermis and gastrointestinal full blood count with differential, liver function tests and tract. The immunomodulatory effects of MMF result from the serum blood urea nitrogen and creatinine levels. In general, inhibition of inosine monophosphate dehydrogenase. This these laboratory studies were performed at baseline, every blocks the de novo pathway of purine synthesis, preferentially 2 weeks for the first month of MMF therapy, then monthly to inhibiting the proliferation of B and T lymphocytes (which every 2 months for the duration of treatment. Patients and lack a purine salvage pathway, and therefore depend upon their parents were informed that during treatment with MMF, de novo production). Because lymphocytes are the primary the use of live attenuated vaccines should be avoided and vac- target of MMF, unwanted effects on other cell types are mini- cinations may be less effective. mized. Moreover, lymphocytes represent the predominant inflammatory cells in lesions of AD. Results During the past decade, MMF has been utilized increasingly in immunosuppressive regimens for paediatric and adult solid During the study period, 14 children (nine boys and five organ transplant recipients.11 It has also become an important girls) with severe AD were treated with MMF. Patient charac- corticosteroid-sparing agent for the treatment of autoimmune teristics are summarized in Table 1. Their ages ranged from 2 and inflammatory disorders such as lupus erythematosus, to 16 years (mean 10 years) at the time of initiation of MMF, nephrotic syndrome, uveitis, vasculitides and autoimmune blis- and the onset of AD was at or before age 3 years in all cases. tering diseases in children as well as adults.12–20 Although sev- The duration of disease prior to MMF therapy was between 1 eral open-label pilot studies and small case series have shown and 15 years (mean 9 years). In all patients, aggressive topical that MMF is a beneficial, well-tolerated therapy for moderate regimens including very potent corticosteroids and calcineurin to severe AD in adults, the use of MMF for childhood AD has inhibitors had been inadequate in controlling the condition. not yet been reported.21–24 Herein we review the safety and Eleven of the patients had received intermittent oral or intra- efficacy of MMF in a series of 14 children with severe AD. muscular corticosteroids prior to initiation of MMF. Four patients had been treated previously with ciclosporin; two Patients and methods failed to respond and two developed side-effects (hypertension and pseudotumor cerebri) that necessitated discontinuation of A retrospective analysis was performed of all children (age the medication. All 14 patients had received multiple courses < 18 years) with severe, refractory AD (diagnosed according to of systemic antibiotics for staphylococcal superinfections. the criteria of Hanifin and Rajka25) who were treated with MMF The course and results of MMF therapy are presented in ) at New York University Medical Center between August 2003 Table 2. Initial doses of MMF ranged from 12 to 40 mg kg 1 and August 2006. Severe AD was defined as extensive in distri- daily divided twice daily and were titrated upwards until bution and magnitude and negatively impacting the affected in- patients either achieved disease clearance or reached a dose of ) dividual’s quality of life (e.g. sleep disturbance, psychological 75 mg kg 1 daily (with a 3 g daily maximum). Of the 14 distress or impaired social functioning). All patients had failed children treated, four (29%) cleared, four (29%) had an to respond to conventional topical treatment (including very po- ‘excellent’ response, five (35%) had a ‘good’ response and tent corticosteroids and calcineurin inhibitors) and ⁄or systemic one (7%) had an ‘inadequate’ response (as defined above). therapies such as oral corticosteroids and ciclosporin. They were Improvement was noted within 8 weeks in all responders each treated with MMF as systemic monotherapy for at least (mean 4 weeks), and the sole nonresponder failed to improve 2 months. No children meeting these criteria were excluded. adequately even with additions of ciclosporin and predniso- Patient information was obtained through chart review and lone to his regimen. Maximal effects were achieved after ) office visits. The response to MMF was evaluated based upon 8–12 weeks (mean 9 weeks) at doses of 25–48 mg kg 1 daily ) objective decreases in surface area involved, erythema, indura- (mean 38 mg kg 1 daily). Patients continued previously insti- tion ⁄papulation ⁄oedema, excoriations and lichenification docu- tuted topical therapies as required, and nine of 13 responders mented in records of the clinical examination as well as subjective were able to discontinue topical corticosteroids or decrease improvements in pruritus, sleep and general quality of life repor- their use to £ 2 days per week within 6 weeks. Patient 3 was ted by the patients and their parents. We categorized overall initially continued on oral prednisolone therapy that had been responses as complete clearance, excellent (> 90% improvement), initiated prior to the MMF, but this was tapered within a good (60–90% improvement) and inadequate (< 60% improve- month without any resultant disease exacerbation. ment).8 Although the retrospective nature of the study precluded The duration of MMF treatment ranged from 2 to the use of a validated scoring system, we believe that the informa- 24 months (mean 8 months). MMF was eventually discontin- tion routinely gathered in our medical records (including the ued in one responder, who has continued to have excellent aforementioned physician and parental observations) provided a disease control with intermittent topical therapy during a reasonable estimate of response to treatment based on a combin- 10-month follow-up period. ation of the appearance of the skin, symptoms and quality of life. MMF therapy was well tolerated in all patients. Two patients Safety data included standard inquiries regarding gastro- experienced transient, mild gastrointestinal upset during the intestinal symptoms, intercurrent infections and other poten- initial week of treatment. Patient 2 had a history of recurrent
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 Mycophenolate mofetil for childhood atopic dermatitis, M. Heller et al. 129
Table 1 Patient characteristics
Disease Previous disease-directed therapies Age duration Patient (years) Sex (years) Topical or phototherapy Systemic Previous adjunctive systemic therapies 1 2 M 1 Very potenta and potent CS Prednisolone Cetirizine,a doxepin, cefadroxil 2 4 M 3 Very potenta and potent CS; TCIa Prednisone, Cetirizine,a hydroxyzine, cyproheptadine, prednisolone cefalexin, aciclovir 3 6 F 3 Very potent and potent CS; TCIa Prednisolone,a Cetirizine,a hydroxyzine, doxepin,a cefadroxil ciclosporin 4 6 M 6 Very potent and potenta CS; TCI None Cetirizine,a hydroxyzine, cyproheptadine,a cefazolin 5 7 M 7 Very potenta and potent CS; TCI Prednisone, Cetirizine,a hydroxyzine, doxepin,a ciclosporin montelukast, cefalexin, cefadroxil,a clindamycin 6 8 M 7 Very potenta and potent CS; TCIa Prednisolone,a Cetirizine,a hydroxyzine, diphenhydramine, ciclosporina doxepin,a cyproheptadine, cefadroxil,a cefdinir 7 12 M 12 Very potent and potenta CS; TCI;a None Cetirizine, diphenhydramine, hydroxyzine,a UVB cyproheptadine,a montelukast, doxycycline, rifampin, co-trimoxazole, clindamycina 8 12 F 7 Very potenta and potent CS; TCI Prednisone Hydroxyzine,a diphenhydramine, cefalexin, cefazolin 9 13 F 13 Very potenta and potent CS; TCIa Prednisone Cetirizine,a fexofenadine, diphenhydramine, doxepin,a montelukast, co-trimoxazole, cefalexin 10 14 F 13 Very potent and potenta CS; TCI;a None Cetirizine, hydroxyzine,a clarithromycin, 7Æ5% coal tar in petrolatum erythromycin 11 15 F 12 Very potenta and potenta CS; TCI Prednisone Cetirizine, loratadine, fexofenadine,a hydroxyzine,a doxepin, cefadroxil, cefalexin 12 16 M 13 Very potenta and potent CS; TCI Prednisone Cetirizine, hydroxyzine,a cefalexin 13 16 M 15 Very potent and potenta CS; TCI; Prednisone Cetirizine,a hydroxyzine, diphenhydramine, PUVA; NB-UVB cefadroxil, levofloxacin, ciprofloxacin, valaciclovir 14 16 M 15 Very potenta and potent CS; TCI;a Prednisone, Cetirizine,a hydroxyzine, doxepin, cefalexin 10% coal tar in petrolatum ciclosporin
aContinued during treatment with mycophenolate mofetil (includes topical medications used > 2 days per week on a regular basis; see Table 2 for durations of concurrent disease-directed therapies). CS, corticosteroid; UV, ultraviolet; NB-UVB, narrowband UVB; PUVA, psora- len plus UVA; TCI, topical calcineurin inhibitor.
herpes simplex viral infections (including eczema herpeti- an initial open-label prospective study, Neuber et al.24 treated cum), which decreased in frequency from approximately 10 adults with severe AD with MMF at a dose of 1 g daily for monthly prior to initiation of MMF to only one mild, locali- 1 week, then 2 g daily for 11 additional weeks. After 12 weeks zed, self-limited episode during his 12-month course of MMF of therapy, the median SCORAD (severity SCORing for Atopic therapy. Only two patients, who had a history of numerous Dermatitis) index was reduced by 68% (P =0Æ007), with a episodes of extensive impetiginization and folliculitis (patient parallel decrease in the median serum IgE level (P =0Æ01). 5) and recurrent staphylococcal furunculosis (patient 7), con- No significant changes were noted in serial full blood counts tinued to develop superinfections requiring systemic antimi- and liver function tests, and the only adverse reactions were crobial therapy while undergoing treatment with MMF. No mild nausea (two patients) and a moderate increase in lactate substantial changes in laboratory values were observed in any dehydrogenase, mild thrombocytopenia, and insomnia (each in of the patients. In particular, leucopenia, anaemia, thrombocy- one patient). Benez and Fierlbeck21 described three adults with topenia and elevated aminotransferases did not develop. severe, recalcitrant AD that cleared within 1 month of initiating Patient 9 had b-thalassaemia trait and a baseline haematocrit MMF therapy at a dose of 2 g daily. Over the following 12– of 10, which was stable during MMF therapy. 29 months, these patients were successfully maintained at a dose of 1 g daily without adverse effects. In contrast, Hansen 26 Discussion et al. described five adults with AD refractory to multiple treatment modalities (including phototherapy and azathioprine Several studies have demonstrated that MMF is a safe and effect- in several individuals) that also failed to respond to MMF at ive therapeutic modality for moderate to severe AD in adults. In doses of 2–2Æ5 g daily for up to 12 weeks.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 130 Mycophenolate mofetil for childhood atopic dermatitis, M. Heller et al.
Table 2 Course and results of mycophenolate mofetil (MMF) therapy
Time to response, Initial MMF dose, Highest MMF dose, months Duration of ) ) mg kg 1 daily mg kg 1 daily Overall therapy, Concurrent treatmentsa Patient (total mg daily) (total mg daily) Initial Maximal response months (duration, months) 1 32 (400) 48 (600) 1 2 Complete 4 Very potent CS (1) clearance 2 33 (600) 44 (1000) 1 2 Complete 12b Very potent CS (1), TCI (11) clearance 3 32 (800) 48 (1200) 1 3 Complete 24b Prednisolone (1), TCI (2) clearance 4 40 (800) 40 (800) 0Æ5NAc Excellent 2c Potent CS (0Æ5) 5 26 (600) 35 (800) 1 3 Good 15b Very potent CS (3) 6 20 (400) 75 (1600) NA NA Inadequate 2 + 6 Prednisolone (2), ciclosporin (6), (two courses) very potent CS (2), TCI (6) 7 19 (800) 37 (1600) 1 2 Good 6b Potent CS (6), TCI (4) 8 22 (1000) 44 (2000) 0Æ75 2 Good 4b Very potent CS (1Æ5) 9 25 (1000) 37 (1500) 1Æ5 3 Excellent 4b Very potent CS (1Æ5), TCI (4) 10 25 (1500) 25 (1500) 0Æ5 2 Complete 9b Potent CS (1), TCI (9) clearance 11 17 (1000) 34 (2000) 1 2 Excellent 4b Very potent CS (2), potent CS (4) 12 18 (1000) 36 (2000) 0Æ5 2 Excellent 3b Very potent CS (1Æ5), potent CS (1), TCI (1) 13 16 (1000) 31 (2000) 2 3 Good 11 Potent CS (11) 14 12 (1000) 36 (3000) 1 NAc Good 3c Very potent CS (1Æ5), TCI (3)
aIncludes disease-directed systemic therapies and topical medications used > 2 days per week on a regular basis. bThe patient is still receiv- ing therapy at the minimal dose required for disease control. cDisease is continuing to improve (upon dosage increase from 2 to 3 g daily in patient 12). CS, corticosteroid; NA, not applicable; TCI, topical calcineurin inhibitor.
The potential for MMF to induce remission in AD was similar to or lower than that of multidrug regimens not con- investigated in another open-label prospective study. Grund- taining MMF).30,31 mann-Kollmann et al.22 treated 10 adults with moderate to Viral (especially herpes simplex) and bacterial infections severe AD with MMF at a dose of 2 g daily for 4 weeks fol- have been described in patients undergoing treatment with lowed by 1 g daily for 4 more weeks; they subsequently fol- MMF. For example, a 50-year-old woman with AD was repor- lowed the patients for an additional 12 weeks. The mean ted to develop staphylococcal septicaemia during a disease SCORAD index was reduced significantly by week 4, with flare (presumably accompanied by an impaired skin barrier in seven patients achieving complete clearance. Aside from the association with Staphylococcus aureus colonization) that occurred development of herpes retinitis in one individual, the MMF after 5 months of MMF therapy.32 However, our patient with was well tolerated. The seven patients who completed the a history of recurrent eczema herpeticum experienced a mark- study had a lasting remission, maintaining a 74% reduction edly decreased frequency and severity of these outbreaks while from the mean pretreatment SCORAD index at week 20 with receiving MMF, and only two of our responding patients the use of only topical corticosteroid therapy during the required antibiotic therapy during MMF treatment (although 12-week follow-up period. all had received frequent courses of systemic antibiotics previ- Nausea, vomiting, diarrhoea and abdominal discomfort rep- ously). These observations are likely to be related to the resent the most frequent side-effects of MMF, occurring in marked improvement of the patients’ primary skin condition. 10–30% of patients. These gastrointestinal symptoms are often The standard dose of MMF for solid organ transplant recipi- ameliorated by administering the same total amount of medi- ents receiving concomitant ciclosporin therapy is 2 g daily or ) cation in two or more doses daily.27 Significant bone marrow 30 mg kg 1 daily in adults. The biologically equivalent dose ) suppression due to MMF is uncommon. Mild increases in in children is 1200 mg m 2 daily, which corresponds to 30– ) serum levels of liver enzymes are occasionally observed, as 60 mg kg 1 daily and results in larger doses per body weight recently reported in two adults receiving MMF for AD.28,29 in smaller patients (due to their increased surface area- Malignancies (e.g. lymphoma) and post-transplant lymphopro- to-volume ratios).33 Because younger children have faster liferative disorder have not been reported in children receiving hepatic drug metabolism and therefore require higher doses MMF monotherapy, and only the latter has been described in per body weight, calculation of paediatric MMF doses based paediatric solid organ transplant recipients taking MMF on body surface area can help to provide adequate drug together with other immunosuppressive agents (with a risk levels.12,34 The younger patients in our series obtained a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 Mycophenolate mofetil for childhood atopic dermatitis, M. Heller et al. 131
) maximal benefit at MMF doses of 40–50 mg kg 1 daily, while eczema: a double-blind, randomised controlled trial. Lancet 2006; ) doses of 30–40 mg kg 1 daily were successful in older chil- 367:839–46. dren and adolescents; however, all of these doses represented 11 Ettenger R, Sarwal MM. Mycophenolate mofetil in pediatric ) renal transplantation. Transplantation 2005; 80(Suppl. 2):S201– approximately 1200 mg m 2 daily. Some authors have sugges- 10. ted that AD, like other dermatological conditions, may 12 Filler G, Hansen M, LeBlanc C et al. Pharmacokinetics of mycophe- occasionally require MMF doses as high as 3Æ5 g daily (corres- nolate mofetil for autoimmune disease in children. Pediatr Nephrol )1 ponding to 50–90 mg kg daily in children) to achieve a 2003; 18:445–9. therapeutic response.16,19,35 13 Buratti S, Szer IS, Spencer CH et al. Mycophenolate mofetil treat- As a retrospective case series, our study is limited by the ment of severe renal disease in pediatric onset systemic lupus lack of a standardized treatment protocol, use of validated erythematosus. J Rheumatol 2001; 28:2103–8. 14 Mendiza´bal S, Zamora I, Berbel O et al. Mycophenolate mofetil assessment tools, and a control group. However, it represents in steroid ⁄cyclosporine-dependent ⁄resistant nephrotic syndrome. the first report of MMF therapy for AD in the paediatric popu- Pediatr Nephrol 2005; 20:914–19. lation and provides information that will be valuable in 15 Thorne JE, Jabs DA, Qazi FA et al. Mycophenolate mofetil ther- designing future trials in children. Our series demonstrates apy for inflammatory eye disease. Ophthalmology 2005; 112:1472– ) ) that MMF at doses of 30–50 mg kg 1 daily (1200 mg m 2 7. daily) can be a safe, well-tolerated and effective treatment for 16 Mimouni D, Anhalt GJ, Cummins DL et al. Treatment of pemphigus severe childhood AD. More than half of the patients (eight of vulgaris and pemphigus foliaceus with mycophenolate mofetil. Arch Dermatol 2003; 139:739–42. 14) achieved complete clearance or > 90% improvement 17 Farley-Li J, Mancini AJ. Treatment of linear IgA bullous dermatosis within 3 months of initiating MMF. Of the 13 patients with of childhood with mycophenolate mofetil. Arch Dermatol 2003; > 60% improvement, one had a sustained remission upon dis- 139:1121–4. continuation of MMF and 12 were successfully maintained on 18 Tran MM, Anhalt GJ, Barrett T et al. Childhood IgA-mediated epi- MMF for up to 24 months without adverse events. dermolysis bullosa acquisita responding to mycophenolate mofetil MMF represents an appealing therapeutic alternative to as a corticosteroid-sparing agent. J Am Acad Dermatol 2006; 54:734– systemic immunosuppressive agents with less favourable 6. 19 Nousari HC, Sragovich A, Kimyai-Asadi A et al. Mycophenolate side-effect profiles that have traditionally been used to treat mofetil in autoimmune and inflammatory skin disorders. J Am Acad children with severe AD. The promising results of this series in Dermatol 1999; 40:265–8. children and previous studies in adults warrant the initiation of 20 Hartmann M, Erik A. Mycophenolate mofetil and skin diseases. controlled clinical trials, further to evaluate MMF therapy Lupus 2005; 14(Suppl. 1):S58–63. for AD. 21 Benez A, Fierlbeck G. Successful long-term treatment of severe atopic dermatitis with mycophenolate mofetil. Br J Dermatol 2001; 14:638–9. References 22 Grundmann-Kollmann M, Podda M, Ochsendorf F et al. Mycophe- nolate mofetil is effective in the treatment of atopic dermatitis. Arch 1 Simpson EL, Hanifin JM. Atopic dermatitis. Med Clin N Am 2006; Dermatol 2001; 137:870–3. 90:149–67. 23 Grundmann-Kollmann M, Korting HC, Behrens S et al. Successful 2 Carroll CL, Balkrishnam R, Feldman SR et al. The burden of atopic treatment of severe refractory atopic dermatitis with mycopheno- dermatitis: impact on the patient, family, and society. Pediatr Derma- late mofetil. Br J Dermatol 1999; 141:175–6. tol 2005; 22:192–9. 24 Neuber K, Schwartz I, Itschert G et al. Treatment of atopic 3 Akhavan A, Rudikoff D. The treatment of atopic dermatitis with eczema with oral mycophenolate mofetil. Br J Dermatol 2000; systemic immunosuppressive agents. Clin Dermatol 2003; 21:225–40. 143:385–91. 4 Ellis C, Luger T. International Consensus Conference on Atopic 25 Hanifin JM, Rajka G. Diagnostic features of atopic dermatitis. Dermatitis II (ICCAD II): clinical update and current treatment Acta Derm Venereol (Stockh) 1980; 92(Suppl.):S44–7. strategies. Br J Dermatol 2003; 148(Suppl. 63):3–10. 26 Hansen ER, Buus S, Deleuran M et al. Treatment of atopic der- 5 Hanifin JM, Cooper KD, Ho VC et al. Guidelines of care for atopic matitis with mycophenolate mofetil. Br J Dermatol 2000; dermatitis. J Am Acad Dermatol 2004; 50:391–404. 143:1324–6. 6 Eichenfield LF, Hanifin JM, Luger TA. Consensus conference on 27 Behrend M. Adverse gastrointestinal effects of mycophenolate pediatric atopic dermatitis. J Am Acad Dermatol 2003; 49:1088–9. mofetil: aetiology, incidence and management. Drug Saf 2001; 7 Dadlani C, Orlow SJ. Treatment of children and adolescents with 24:645–63. methotrexate, cyclosporine, and etanercept: review of the dermato- 28 Hantash B, Fiorentino D. Liver enzyme abnormalities in patients logic and rheumatologic literature. J Am Acad Dermatol 2005; with atopic dermatitis treated with mycophenolate mofetil. Arch 52:316–40. Dermatol 2006; 142:109–10. 8 Murphy LA, Atherton D. A retrospective evaluation of azathioprine 29 Balal M, Demir E, Paydas S et al. Uncommon side effect of MMF in in severe childhood atopic eczema, using thiopurine methyltrans- renal transplant recipients. Ren Fail 2005; 27:591–4. ferase levels to exclude patients at high risk of myelosuppression. 30 CellCept [package insert]. Nutley, NJ: Roche Laboratories Inc, Br J Dermatol 2002; 147:308–15. 2005. 9 Berth-Jones J, Takwale A, Tan E et al. Azathioprine in severe adult 31 Dharnidharka VR, Ho PL, Stablein DM et al. Mycophenolate, tacro- atopic dermatitis: a double-blind, placebo-controlled, crossover limus and post-transplant lymphoproliferative disorder: a report of trial. Br J Dermatol 2002; 147:324–30. the North American Pediatric Renal Transplant Cooperative Study. 10 Meggitt SJ, Gray JC, Reynolds NJ. Azathioprine dosed by thiopu- Pediatr Transplant 2002; 6:396–9. rine methyltransferase activity for moderate-to-severe atopic
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 132 Mycophenolate mofetil for childhood atopic dermatitis, M. Heller et al.
32 Satchell AC, Barnetson RS. Staphylococcal septicaemia complicating 34 Filler G, Foster J, Berard R et al. Age-dependency of mycophenolate treatment of atopic dermatitis with mycophenolate. Br J Dermatol mofetil dosing in combination with tacrolimus after pediatric renal 2000; 143:202–3. transplantation. Transplant Proc 2004; 36:1327–31. 33 Filler G, Lepage N. To what extent does the understanding of phar- 35 Grundmann-Kollmann M, Kaufmann R, Zollner TM. Treatment of macokinetics of mycophenolate mofetil influence its prescription. atopic dermatitis with mycophenolate mofetil. Br J Dermatol 2001; Pediatr Nephrol 2004; 19:962–5. 145:351–2.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp127–132 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07942.x Vehicle-controlled, randomized, double-blind study to assess safety and efficacy of imiquimod 5% cream applied once daily 3 days per week in one or two courses of treatment of actinic keratoses on the head A. Alomar, J. Bichel* and S. McRae Hospital de la Santa Creu i Sant Pau, Dermatology Service, Avda. San Antonio Marı´a Claret, 167, Barcelona, Spain *3M Medica, D-41460 Neuss, Germany 3M Pharmaceuticals, St Paul, MN 55144, U.S.A.
Summary
Correspondence Background Imiquimod has been investigated as a safe and effective therapeutic Augustı´n Alomar. option for the treatment of actinic keratosis (AK). E-mail: [email protected] Objectives To evaluate imiquimod vs. vehicle applied three times a week for 4 weeks in one or two courses of treatment for AK on the face or balding scalp. Accepted for publication 11 July 2006 Patients and methods Patients diagnosed with AK were enrolled in this multicentre, vehicle-controlled, double-blind study conducted in Europe. Twenty study cen- Key words tres enrolled a total of 259 patients in this study. Patients applied the study drug actinic keratoses, imiquimod, neoplasm, senile for 4 weeks, entered a 4-week rest period and if they did not have complete keratoses clearance, they then entered a second course of treatment. Conflicts of interest Results Patients in the imiquimod group had an overall complete clearance rate of This study was supported by 3M Pharmaceuticals, 55Æ0% (71 ⁄129) vs. a rate of 2Æ3% (3 ⁄130) for the vehicle group. There was a St Paul, MN, U.S.A. J.B. is an employee of 3M high rate of agreement between the clinical assessment and histological findings Medica and S.M. is a former employee of 3M with respect to AK lesion clearance. At both 8-week post-treatment visits, the Pharmaceuticals. A.A. has received fees for negative predictive value of the investigator assessment was 92Æ2% for clinical consulting with 3M Pharmaceuticals. assessments vs. histological results. Conclusions A 4-week course of treatment with three times weekly dosing of imi- quimod 5% cream, with a repeated course of treatment for those patients who fail to clear after the first course of treatment, is a safe and effective treatment for AK. The overall complete clearance rate (complete clearance after either course 1 or course 2) is comparable to the 16-week treatment regimen, while decreasing drug exposure to the patient and decreasing the overall treatment time.
Actinic keratoses (AKs), also known as solar keratoses or senile Imiquimod 5% cream is currently available as AldaraTM (3M keratoses, are intermediate epidermal neoplasms in the multi- Pharmaceuticals, St Paul, MN, U.S.A.) and is indicated for the stage process of carcinogenesis.1 They demonstrate histological treatment of external genital warts. It has been shown to and molecular genetic features of malignancy. According to stimulate the immune system by activating antigen-presenting new nomenclature they are discussed as in situ carcinoma or cells such as monocytes ⁄macrophages and dendritic cells to keratinocytic intraepidermal neoplasia (KIN).2 As most persons produce interferon and other cytokines and chemokines. These with AK present with multiple lesions, the condition is con- cytokines stimulate the nonspecific innate immune response sidered to be an aspect of field cancerization.3 This means that and are also important for directing the acquired immune there is a close regional coexistence of epithelial malignancies response.6,7 Because of this unique mechanism, imiquimod in different developmental stages, including preclinical and has been used for the treatment of AK. Previously reported clinical AK, potentially leading to invasive squamous cell carci- findings in five large, randomized, double-blind, parallel- noma (SCC).4,5 SCC of the skin has the potential to metasta- group, vehicle-controlled studies indicated that imiquimod 5% size and may account for up to 20% of deaths from skin cream applied two or three times weekly for 16 weeks is an cancer. effective and well-tolerated treatment for actinic keratosis.8–10
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 133 134 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al.
An investigator-initiated study lends some support to the incidence and severity of adverse events (AEs) and local skin idea that a shorter treatment duration (i.e. < 16 weeks) may reactions (LSRs), and reducing total drug exposure. be effective in clearing AK lesions. Twenty-five patients with Enrolment for this study began in December 2003, and all AK lesions on the scalp, forehead or cheeks were treated with study procedures were completed by December 2004. Patients imiquimod 5% cream once daily 3 days per week for 4 weeks in this study attended a pre-study visit. Eligible patients were followed by a 4-week rest period. If any AK lesions remained randomized to either imiquimod 5% cream or vehicle cream in the treatment area at the end of the first 8-week cycle, in a 1 : 1 ratio. patients continued with another 8-week cycle of treatment. Prior to entering the study, the investigator or a designated Including both courses of treatment, 18 of 39 (46%) patients assistant explained to each patient or legally acceptable repre- were completely clear of their lesions; however, 16 of 39 sentative the nature of the study, its purpose, the procedures (41%) patients were completely clear of their lesions by the and expected duration of the study, and the benefits and risks end of the course 1 (C1) treatment period. of study participation. All patients were informed of their The present vehicle-controlled, randomized study was con- rights. Before any study procedures were performed, each ducted to corroborate previous trial results and determine if subject voluntarily signed and dated an informed consent imiquimod is effective in clearing AK lesions when adminis- document. tered over a 4-week treatment period followed by a 4-week Patients applied study cream to the treatment area once rest period (up to two courses of treatment). One pre- daily on three alternate days per week for 4 weeks of treat- treatment confirmatory biopsy of a target AK lesion and one ment (C1) followed by a 4-week post-treatment period. post-treatment biopsy for evaluation of efficacy in a predetermined Patients selected a dosing schedule of alternate days (e.g. Mon- AK lesion were taken to prove clinical response by histology. day, Wednesday, Friday), and they were encouraged to use that schedule during the entire dosing period. Patients Patients and methods returned to the study centre for evaluations at week 1, week 2, C1 end-of-treatment (EOT) (week 4), and the C1 4-week This was a multicentre, vehicle-controlled, double-blind study post-treatment visit (week 8). Patients who were completely conducted in Europe. Twenty study centres (four in France, clear at the C1 4-week post-treatment visit returned to the six in Germany, three in Italy, two in the Netherlands, three study centre 4 weeks later for the C1 8-week post-treatment in Spain and two in the United Kingdom) enrolled a total of visit (week 12) for a biopsy of the target lesion site and a 259 patients. Investigators at the study centres were medical clinical count of the remainder of the treatment area. Patients doctors who specialize in dermatology and therefore experi- who did not completely clear all of their AK lesions in the enced in the practice of diagnosing and treating AK. treatment area at the C1 4-week post-treatment visit (week 8) participated in a second 4-week course of treatment (C2) and returned for visits at week 10, C2 EOT (week 12), C2 4-week Patients post-treatment visit (week 16), and C2 8-week post-treatment Patients were eligible for the study if they were at least visit (week 20). 18 years of age and had a pre-study clinical diagnosis of five to nine AK lesions within a contiguous 25-cm2 treatment area Biopsies on the head (balding scalp or face, but not both). One repre- sentative lesion within the treatment area had to be confirmed For each patient in the study, a 3–4-mm punch biopsy was by a pretreatment biopsy as histologically positive for AK. taken from a representative lesion at the pre-study visit for Patients were excluded from study participation if they had histological confirmation of the clinical diagnosis. Patients had malignant tumours of the skin within the treatment area or a maximum of two lesions (marked as no. 1 and no. 1A) if they had any dermatological disease or condition in the selected for pretreatment biopsy. A biopsy of another lesion treatment or surrounding area that could impede local skin (no. 2) within the treatment area, also identified at the initi- assessments. ation visit, was taken 8 weeks’ post-treatment to histologically Patients were also excluded from study participation if they assess the lesion site for the presence of AK. All biopsy speci- had clinically significant unstable cardiovascular, immunosup- mens were examined for histological evaluation at a cen- pressive, haematological, hepatic, neurological, renal, endo- tral dermatopathology laboratory. Dermatopathologists who crine, collagen–vascular or gastrointestinal abnormalities as reviewed the specimens were blind to treatment assignment. assessed by the investigator. Study cream application Interventions The study drug was supplied in the form of a topical cream in This double-blind study was designed to examine a treatment single-use sachets (packets). Patients were instructed to gently regimen that could improve the benefit–risk profile of imiqui- clean and dry the treatment area before applying the study mod 5% cream for the treatment of AK lesions by maintain- drug, which was applied at the same time of day just prior to ing or improving the efficacy response rates while decreasing their normal sleeping hours. Patients rubbed the cream into
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al. 135 the treatment area until it was no longer visible and allowed it Statistical analysis to remain on the skin for at least 8 h. Investigators could prescribe rest periods from treatment if The primary dataset analysed for efficacy and safety was the a patient was unable to apply doses because of LSRs or treat- intention-to-treat (ITT) dataset, which included all random- ment site AEs. The treatment period was not extended beyond ized patients. Sample size determinations were made to be 4 weeks even if patients missed doses or took rest periods. able to have at least 90% power to detect clinically meaningful differences in complete clearance rates after C1 of 0Æ11 for the vehicle group vs. 0Æ32 for the imiquimod group, at a level of Histology significance of 0Æ025 and allowing for a 20% drop-out rate. A biopsy of lesion no. 2, identified and photographed at the All statistical analyses were performed using SAS version treatment initiation visit, was taken 8 weeks post-treatment to 8Æ2 (SAS Institute Inc., Cary, NC, U.S.A.). A Cochran–Mantel– histologically assess the lesion site for the presence of AK. All Haenszel (CMH) test, which adjusts for the effects of multiple biopsy specimens were sent to a central dermatopathology study centres, was the primary statistical test used to compare laboratory for histological evaluation. Two independent derm- treatment groups with respect to the primary variables, the atopathologists, blinded to treatment assignments, reviewed complete clearance rate and the complete clearance rate after the slide specimens. Both dermatopathologists conferred with C1. Each CMH test was carried out at the a =0Æ05 ⁄2=0Æ025 each other and had to be in agreement regarding the diagnosis level of significance using the ITT data set. Testing each pri- of a specimen. mary variable at the 0Æ025 level preserved the overall level of significance at 0Æ05. The complete clearance rate and the com- plete clearance rate after C1 [with 95% confidence interval Efficacy and safety measurements (CI)] were calculated for each treatment group. For the sec- ondary variables, partial clearance rate and partial clearance Efficacy rate after C1, a CMH test was used to compare treatment There were two efficacy variables. One variable was the com- groups using the ITT data set. plete clearance rate after C1, defined as those patients who For a given patient, if the post-treatment biopsy was posi- had no histological evidence of AK in the biopsy of the post- tive for AK or if the post-treatment biopsy was missing, then treatment target lesion and no (zero) AK lesions in the this lesion was considered present in the post-treatment count remainder of the treatment area. The other variable was the of AK lesions. If the post-treatment biopsy exhibited evidence complete clearance rate after C1 or C2. Secondary efficacy var- of other diagnosis (diagnostic of another skin disease) then iables included the partial clearance rate, defined as the pro- the lesion was not considered present in the post-treatment portion of patients at their last study visit with at least 75% count of AK lesions, and it also was considered not present in reduction in the number of AK lesions counted at baseline; the baseline count of AK lesions. In this case the baseline the partial clearance rate after C1; and the clearance of individ- count was reduced by 1. ual AK lesions (reduction in all AK lesions from all patients Fisher’s exact test was used to compare treatment groups totalled together). with respect to the incidence of AEs and Wilcoxon rank-sum test was used to compare treatment groups with respect to the most intense LSRs experienced by the patient. For each LSR Safety category, the relationship between complete clearance and the Lesions within the treatment area were counted and mapped most intense reaction experienced by a patient was evaluated (with the use of a clear plastic template) at the initiation and by means of the Cochran–Armitage Test for Trend in Pro- at each study centre visit beginning with week 2; photographs portions (two-sided); this relationship was also assessed for were obtained beginning with the pre-study visit until the complete clearance after C1 and the most intense reaction end-of-study visit. The presence and intensity of specific LSRs experienced by a patient in C1. For laboratory parameters, (erythema, oedema, erosion ⁄ulceration, scabbing ⁄crusting, Wilcoxon’s signed-rank test was used to assess the significance weeping ⁄exudate, vesicles and flaking ⁄scaling ⁄dryness) were of within-treatment shifts and changes from baseline. Wilcox- assessed by the investigator at each visit and rated on a four- on’s rank-sum test was used to compare treatment groups point scale: none (0), mild (1), moderate (2) and severe (3). with respect to shifts and change from baseline. For patient Any suspected skin infections were swabbed and cultured. demographics, Fisher’s exact test was used to compare treat- Pre-study and post-study safety evaluations included phys- ment groups with respect to sex, race and skin type; analysis ical examinations, clinical laboratory tests (haematology, blood of variance was used to compare treatment groups with chemistry and urinalysis), and urine pregnancy tests for respect to age. women of childbearing potential. A physical examination, including vital sign measurements, was conducted at the Results pre-study, week 4 and week 12 visits. AEs and the use of concomitant medications were monitored at each study visit There were 259 patients enrolled in the study; 129 (14 beginning with treatment initiation. female) patients were randomized to imiquimod and 130
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 136 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al.
Patients screened n = 293
Patients randomized Patients ineligible n = 259 n = 34
Imiquimod Vehicle n = 129 n = 130
Completed C1 Withdrew in C1 Completed C1 Withdrew in C1 treatment period treatment period treatment period treatment period n = 126 n = 3 n = 130 n = 0
Clear Not clear Clear Not clear Clear Not clear Clear Not clear n = 47 n = 79 n = 1 n = 2 n = 1 n = 129 n = 0 n = 0 Stop! Stop! Stop! Stop! Stop! Stop!
Completed C2 Withdrew in C2 Did not go to C2 Completed C2 Withdrew in C2 Did not go to C2 treatment period treatment period treatment period treatment period treatment period treatment period a n = 74 n = 1 n = 5 n = 126 n = 3 n = 0
Did not Did not Did not Completed C2 Completed C2 Completed C2 Completed C2 complete C2 complete C2 complete C2 Did not complete C2 post-treatment post-treatment post-treatment post-treatment post-treatment post-treatment post-treatment post-treatment assessment assessment assessment assessment assessment assessment assessment n = 3 n = 74 n = 1 n = 121 n = 0 n = 0 n = 0 n = 5
Fig 1. Patient disposition. aOne patient withdrew in the C1 treatment period but returned for the C1 4-week post-treatment visit; the patient continued on to C2 treatment and post-treatment visits. C1, course 1; C2, course 2.
(17 female) patients were randomized to vehicle at 20 study the baseline AK lesions. For the per protocol data set, the centres in Europe. Violations of inclusion or exclusion cri- median number of baseline AK lesions in the treatment area teria were the most common reasons for study ineligibility. was six lesions for the imiquimod treatment group and seven A total of four patients randomized to imiquimod and three lesions for the vehicle treatment group (P =0Æ63). patients randomized to vehicle discontinued the study (Fig. 1). Efficacy results For the ITT data set, no statistically significant differences between treatment groups were found with respect to any of Imiquimod 5% cream was statistically significantly better the demographic characteristics: sex, age, race and skin type than vehicle with respect to the complete clearance rate (Table 1). All of the patients in the study were white. The (P <0Æ0001). In the imiquimod-treated group, 71 of the median age for patients assigned to imiquimod was 70 years 129 (55Æ0%) patients achieved overall complete clearance vs. vs. 72 years for patients in the vehicle group. Patients ranged three of the 130 (2Æ3%) patients randomized to vehicle. At in age from 44 to 94 years. the end of C1, 48 of the 129 (37Æ2%) patients in the In the imiquimod group, 105 of the 129 (81Æ4%) patients imiquimod-treated group and one of 130 (0Æ8%; P < had previously received at least one treatment for AK vs. 91 of 0Æ0001) of the patients in the vehicle group had complete the 130 (70Æ0%) vehicle patients. The median number of clearance (Fig. 2). baseline AK lesions in the treatment area for the ITT data set The difference in complete clearance rates (imiquimod was six lesions for the imiquimod treatment group and seven minus vehicle) was 52Æ7% (55Æ0% minus 2Æ3%) and the lesions for the vehicle treatment group (P =0Æ97). There was corresponding 95% CI for the difference ranged from 43Æ8% no statistically significant treatment difference with respect to to 61Æ7%. The Mantel–Haenszel estimate of the common odds
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al. 137
Table 1 Patient demographics
Imiquimod 3 · week Vehicle 3 · week Total Overall Variable description Statistics ⁄level (n = 129) (n = 130) (n = 259) P-value Sex Female 14 (10Æ9%) 17 (13Æ1%) 31 (12Æ0%) 0Æ702 Male 115 (89Æ1%) 113 (86Æ9%) 228 (88Æ0%) Age (years) n 129 130 259 0Æ114 Mean 70Æ371Æ85 71Æ08 Standard deviation 8Æ45 7Æ254 7Æ895 Median 70 72 71 Minimum 44 53 44 Maximum 90 94 94 Race White 129 (100Æ0%) 130 (100Æ0%) 259 (100Æ0%) Fitzpatrick skin type I: always burns easily; never tans 22 (17Æ1%) 19 (14Æ6%) 41 (15Æ8%) 0Æ794 II: always burns easily; tans minimally 64 (49Æ6%) 65 (50Æ0%) 129 (49Æ8%) III: burns moderately; tans gradually 36 (27Æ9%) 39 (30Æ0%) 75 (29Æ0%) IV: burns minimally; always tans well 7 (5Æ4%) 7 (5Æ4%) 14 (5Æ4%)
individual lesions cleared in the imiquimod and vehicle groups, respectively. 100 When analysing for subgroups, the overall and C1 complete 90 clearance rates were analysed by lesion location, age, sex, race, 80 skin type, sunscreen use, dosing compliance and previous AK 70 treatment for the ITT data set. For the imiquimod group, the overall complete clearance rates were similar for patients 60 55·0% 56·8% < 65 years (48Æ4%) and ‡ 65 years (57Æ1%) as well as for 50 female (50Æ0%) and male patients (55Æ7%). Imiquimod com- 40 plete clearance rates were higher for treatment areas on 30 the face (64Æ6%) than for scalp (49Æ4%). For the C1 com-
20 plete clearance rates, the imiquimod group showed a higher
Overall complete clearance rate complete clearance Overall clearance rate in treatment areas on the face (50Æ0%), 10 2·3% 1·7% patients ‡ 65 years (39Æ8%), and male patients (39Æ1%) com- 0 Æ Æ Imiq Veh Imiq Veh pared with scalp (29 6%), patients < 65 years (29 0%), and ITT PP female patients (21Æ4%). A shorter treatment period for those patients who cleared after 4 weeks of treatment (12 doses) was a benefit of the Fig 2. Overall complete clearance rates (imiquimod vs. vehicle). ITT, intention-to-treat; PP, per protocol. dosing regimen because it minimized drug exposure in com- parison with the 16-week dosing regimen (48 doses). Only ratio (across study centres) for the complete clearance rate patients who were not clear after an initial 4-week treatment (imiquimod to vehicle) was 43Æ7 with a CI of 13Æ56–140Æ9. period received an additional 4-week course of treatment. This means that the odds that imiquimod-treated patients This increased the complete clearance in more patients would completely clear were more than 43 times the odds (37Æ2–55%). At the same time, the LSR profile for patients that vehicle patients would completely clear. treated in C2 was lower compared with C1. This flexible Overall, 65Æ9% (85 ⁄129) of the subjects in the imiquimod approach to the dosing duration makes allowance for individual treatment group and 3Æ8% (5 ⁄130) of the subjects in the vehi- variability in the complex immune response that is necessary cle treatment group achieved partial clearance (P <0Æ0001). to clear the AK lesions, and this approach offers the best For the individual lesion clearance after C1, an analysis of balance of efficacy and safety. imiquimod-treated patients and vehicle-treated patients with This study demonstrated a high rate of agreement between lesion counts at both baseline and C1 4 weeks post-treatment the clinical assessment and histological findings with respect showed that 522 of 855 (61Æ1%) and 96 of 852 (11Æ3%) of to AK lesion clearance. At both 8-week post-treatment visits, individual AK lesions cleared in the imiquimod and vehicle the negative predictive value of the investigator assessment groups, respectively. For the individual lesion clearance during was 92Æ2% for clinical assessments vs. histological results. the entire study, counts for 129 imiquimod and 130 vehicle These histological results confirm that imiquimod cream can patients with lesion counts at both baseline and end-of-study eliminate basal neoplastic keratinocytes, and that the entire showed that 647 of 855 (75Æ7%) and 161 of 852 (18Æ9%) of depth of the epidermis can be cleared of AK.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 138 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al.
groups (P <0Æ001) with respect to the distribution of the Safety results most intense values for erythema, flaking ⁄scaling ⁄dryness and scabbing ⁄crusting. Adverse events Additionally, statistically significantly different distributions All patients were queried at each study visit regarding any cur- of the most intense LSRs were also noted between the imiqui- rent AEs or any AEs that had occurred since the last visit. AEs mod and vehicle groups for oedema (P <0Æ001), vesicles occurred in both treatment groups, with 69 of 129 (53Æ5%) (P <0Æ001), erosion ⁄ulceration (P <0Æ001) and weeping ⁄ imiquimod-treated patients and 40 of 130 (30Æ8%) vehicle exudate (P <0Æ001) (Table 3). patients reporting AEs (P <0Æ001). Two imiquimod-treated No unexpected safety issues were observed in this study patients and no vehicle-treated patients discontinued from the population with a median age of 71 years. treatment period due to AEs (C1); two imiquimod-treated patients and no vehicle-treated patients discontinued from the Discussion treatment period due to LSRs. Five randomized patients reported 10 serious AEs (SAEs) The results from this study demonstrate that imiquimod 5% (one imiquimod, four vehicle). The patient randomized to cream dosed once daily three times weekly for 4 or 8 weeks imiquimod reported falling, and as a result suffered abrasions is an effective and well-tolerated treatment for AK lesions on of the arm that required hospitalization and therapy. There the face or scalp. The flexible dosing regimen with one or was one death during the study. This patient, randomized to two courses of treatment allows a shorter treatment period for vehicle, was diagnosed with pancreatic cancer and died as a those patients who clear after 4 weeks of treatment (12 result before completing the study. One patient was ineligible doses), minimizing exposure to imiquimod and further sup- for the study, but had signed a consent form at the time of porting an improved benefit–risk profile. Even for those the reported SAE; therefore the event was recorded in the data patients who are not clear after an initial 4-week treatment base. None of the SAEs were considered to be related to the period, an additional 4 weeks of treatment (total of 24 doses) study drug. still represents a marked reduction in the exposure to imiqui- mod over 16 weeks of treatment (48 doses). The benefits of treatment with topical imiquimod 5% cream Local skin reactions included complete clearance or partial clearance (‡ 75% LSRs occurred in both treatment groups, and they were identi- reduction in the number of baseline lesions) of AK lesions for fied by type as erythema, oedema, vesicles, erosion ⁄ulceration, a majority of patients. Complete clearance was attained by 71 weeping ⁄exudate, flaking ⁄scaling ⁄dryness and scabbing ⁄crust- of 129 (55Æ0%) imiquimod patients vs. three of 130 (2Æ3%) ing. Over the course of the study, the three most common vehicle patients after one or two courses of treatment. Partial investigator-assessed treatment area LSRs were erythema, fla- clearance was attained by 85 of 129 (65Æ9%) imiquimod king ⁄scaling ⁄dryness and scabbing ⁄crusting (Table 2). There patients vs. five of 130 (3Æ8%) vehicle patients. Partial clear- were significant differences between imiquimod and vehicle ance is beneficial for those patients who have extensive AK
Table 2 Most common investigator-assessed Type of reaction and Imiquimod Vehicle local skin reactions summary Intensity n = 129 (%) n = 130 (%) Erythema Most intense during None 4 (3Æ1) 38 (29Æ2) the study Mild 21 (16Æ3) 65 (50Æ0) Moderate 64 (49Æ6) 27 (20Æ8) Severe 40 (31Æ0) 0 (0Æ0) Any during study Mild, moderate, or severe 125 (96Æ9) 92 (70Æ8) Flaking ⁄scaling ⁄dryness Most intense during None 8 (6Æ2) 34 (26Æ2) the study Mild 55 (42Æ6) 63 (48Æ5) Moderate 51 (39Æ5) 32 (24Æ6) Severe 15 (11Æ6) 1 (0Æ8) Any during study Mild, moderate, or severe 121 (93Æ8) 96 (73Æ8) Scabbing ⁄crusting Most intense during None 15 (11Æ6) 65 (50Æ0) the study Mild 31 (24Æ0) 52 (40Æ0) Moderate 52 (40Æ3) 11 (8Æ5) Severe 31 (24Æ0) 2 (1Æ5) Any during study Mild, moderate, or severe 114 (88Æ4) 65 (50Æ0)
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al. 139
Table 3 Statistically significant, most intense local skin reactions (investigator assessment) Imiquimod once daily Vehicle once daily 3 days per week 3 days per week Type of reaction Intensity n = 129 (%) n = 130 (%) P-value Oedema None ⁄not present 39 (30Æ2) 120 (92Æ3) Mild 51 (39Æ5) 6 (4Æ6) Moderate 30 (23Æ3) 4 (3Æ1) Severe 9 (7Æ0) 0 (0Æ0) <0Æ001 Vesicles None ⁄not present 81 (62Æ8) 129 (99Æ2) Mild 32 (24Æ8) 1 (0Æ8) Moderate 14 (10Æ9) 0 (0Æ0) Severe 2 (1Æ6) 0 (0Æ0) <0Æ001 Erosion ⁄ulceration None ⁄not present 34 (26Æ4) 108 (83Æ1) Mild 40 (31Æ0) 17 (13Æ1) Moderate 41 (31Æ8) 4 (3Æ1) Severe 14 (10Æ9) 1 (0Æ8) <0Æ001 Weeping ⁄exudate None ⁄not present 47 (36Æ4) 124 (95Æ4) Mild 38 (29Æ5) 4 (3Æ1) Moderate 38 (29Æ5) 1 (0Æ8) Severe 6 (4Æ7) 1 (0Æ8) <0Æ001
lesions, as single lesions remaining after therapy can be treated The histological results suggest that imiquimod treats the more practically with a destructive therapy, potentially lower- entire thickness of the AK lesion rather than the superficial ing the number of repeat treatment visits as well as pain and portion of the lesion. This finding may be due to imiquimod’s discomfort for the patient. induction of the immune response, which facilitates the The primary variable was the complete clearance rate, movement of activated immune cells from the bloodstream defined as the proportion of patients with 100% clearance of into the cutaneous tissue. These immune cells migrate from their AK lesions; however, this measurement tends to underesti- the dermal blood vessel plexus underlying the epidermis mate the true effectiveness of a treatment as it discounts the towards the lesion. The complete histological clearance of AK clinical effect of the therapy in patients who did not experience lesions observed in this study confirms that treatment with 100% clearance. A measurement of the effectiveness for patients imiquimod cream eliminates all abnormal cells, not just the who did not experience complete clearance but had a clinical visible portion of the AK lesion. response is the total individual lesion clearance. An analysis of Although most of the patients who received imiquimod had imiquimod-treated patients who had lesion assessments at both LSRs and about one-half of imiquimod-treated patients reported baseline and 8 weeks post-treatment revealed that 75Æ7% of the application site reactions, the treatment was very well tolerated. total number of 855 individual lesions counted at baseline had The side-effects were consistent with local stimulation of an cleared at the 8-week post-treatment visit. This individual lesion immune response. Unlike patients with genital warts, where clearance rate is similar to the rates reported for other treat- erythema was not observed in many patients who cleared their ments, including cryotherapy or fluorouracil 5% cream.9,11–13 warts,14 the LSRs could be considered a pharmacodynamic Because of their characteristic appearance, AKs are usually marker of the effects of imiquimod in AK patients. However, diagnosed clinically, without confirmation by biopsy. How- despite the high incidence of erythema, surprisingly, most ever, biopsies were performed to demonstrate the validity patients denied accompanying discomfort. In fact, only 2Æ3% of the clinical assessments in this study. The negative pre- and 5Æ4% of the patients in the imiquimod treatment group dictive value of the investigator clinical assessment of the reported pain and burning at the treatment site, respectively. post-treatment biopsy site was 92Æ2% (71 ⁄77). This meant Although there were statistically significant changes in some that in 71 out of 77 cases where the investigator said there haematological markers in the imiquimod treatment group, was no AK at the site, the histological result confirmed the these changes were not clinically meaningful because the absence of AK. The positive predictive value was 64Æ6% counts remained well within the normal ranges. The standard (31 ⁄48), which meant that in 31 of 48 cases where the errors of these significance tests were small because of the investigator said AK was present, the histological result con- large sample sizes, thus making it possible to detect small dif- firmed the presence of AK. The investigators correctly ferences as statistically significant. These changes have been assessed AK for 31 of the 37 (83Æ8%) patients who truly observed previously and may reflect a random effect due to (histologically) had AK. In addition, investigators correctly the advanced age of the study population. Cytokines have been assessed no AK for 71 of 88 patients (80Æ7%) who were reported to have a suppressive effect on bone marrow;15 small histologically negative for AK. amounts of locally produced cytokines released into the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 140 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al. systemic circulation may explain the small decrease in platelets fu¨r Dermatologie und Venerologie) for performing der- and haemoglobin. matopathology interpretations; Jens Bichel MD (3M Medica Although collected as a safety parameter, the findings from medical monitor), Terrance Fox, Ping Liu and Scott E. McRae the skin quality assessments suggest that imiquimod treatment (3M biostatistician); Henning Reuter (study manager, 3M may improve cosmetic outcome. This observation is consistent Medica); Stephanie Stock (3M Medica); and Lynda Shindley with studies that showed that aspects of the immune response, (manuscript preparation). especially cytokines, can affect tissue remodelling and wound healing.16–18 References Imiquimod has been shown to stimulate the immune 1 Callen JP, Bickers DR, Moy RL. Actinic keratosis. J Am Acad Dermatol system by activating antigen-presenting cells such as mono- 1997; 36:650–3. cytes ⁄macrophages and dendritic cells to produce interferon 2 Yantsos VA, Conrad N, Zabawski E et al. Incipient intraepidermal and other cytokines and chemokines, which leads to the cutaneous squamous cell carcinoma: a proposal for reclassifying stimulation of the adaptive immune response. Based on this and grading solar (actinic) keratoses. Semin Cutan Med Surg 1999; mechanism, imiquimod has the potential not only to clear 18:3–14. AK lesions but also control field cancerization.4 This could 3 Kocsard E. Solar keratoses and their relationship to non-melanoma lead to low recurrence rates within the treatment area. skin cancers. Australas J Dermatol 1997; 38 (Suppl. 1):S30. 4 Uhoda I, Quatresooz P, Pierard-Franchimont C, Pierard GE. Furthermore, the cutaneous immune response has also been Nudging epidermal field cancerogenesis by imiquimod. Dermatology shown to play an important role in the transformation of AK 2003; 206:357–60. 19 into invasive SCC. Therefore, it is plausible that the 5 Carlson JA, Scott D, Wharton J et al. Incidental histopathologic stimulation of cutaneous immunity by imiquimod could also pattern: possible evidence of field cancerisation surrounding skin decrease the transformation rate of AK into invasive tumors. Am J Dermpathol 2001; 23:494–7. SCCs; however, further studies are needed to confirm this 6 Testerman TL, Gerster JF, Imbertson LM et al. Cytokine induction hypothesis. by the immunomodulators imiquimod and S-27609. J Leukoc Biol 1995; 58:365–72. In conclusion, imiquimod 5% cream has been shown to be 7 Imbertson LM, Beaurline JM, Couture AM et al. Cytokine induction a safe and effective treatment for AK. The efficacy results of in hairless mouse and rat skin after topical application of the this study evaluating the 4 or 8 weeks of dosing were similar immune response modifiers imiquimod and S-28463. J Invest Dermatol to earlier studies evaluating longer dosing durations. With the 1998; 110:734–9. recommended dosage, imiquimod was well tolerated and 8 Lebwohl M, Dinehart S, Whiting D et al. Imiquimod 5% cream for demonstrated high rates of complete and partial clearance of the treatment of actinic keratosis: results from two phase III, ran- multiple AK lesions, confirmed by clinical and histological domized, double-blind, parallel group, vehicle-controlled trials. J Am Acad Dermatol 2004; 50:714–21. assessment. A flexible approach to the dosing duration makes 9 Szeimies RM, Gerritsen MJ, Gupta G et al. Imiquimod 5% cream allowance for individual variability in the complex immune for the treatment of actinic keratosis: results from a phase III, response that is needed to clear the AK lesions, and the results randomized, double-blind, vehicle-controlled, clinical trial with indicate that flexibility in dosing offers the best balance of histology. J Am Acad Dermatol 2004; 51:547–55. efficacy and safety. 10 Korman N, Moy R, Ling M et al. Dosing with 5% imiquimod cream 3 times per week for the treatment of actinic keratosis: results of two phase 3, randomized, double-blind, parallel- Acknowledgments group, vehicle-controlled trials. Arch Dermatol 2005; 141:467– 73. This study was funded by 3M Pharmaceuticals, St Paul, MN, 11 Thai K-E, Fergin P, Freeman M et al. A prospective study of the use U.S.A. The following investigators recruited patients: Professor of cryosurgery for the treatment of actinic keratoses. Int J Dermatol Pascal Joly, Rouen, France; Jean-Paul Ortonne MD, Nice, 2004; 43:687–92. France; Mireille Ruer-Mulard MD, Martiques, France; Patrick 12 Loven K, Stein L, Furst K, Levy S. Evaluation of the efficacy and Brun MD, Cannes, France; Axel Hauschild MD, Kiel, Germany; tolerability of 0.5% fluorouracil cream and 5% fluorouracil cream Julia Reifenberger MD, Du¨sseldorf, Germany; Harald Grenz applied to each side of the face in patients with actinic keratosis. Clin Ther 2002; 24:990–1000. MD, Friedberg, Germany; Karl-Gustav Meyer MD, Berlin, Ger- 13 Gupta AK, Davey V, McPhail H. Evaluation of the effectiveness of many; Georg Popp MD, Augsburg, Germany; Karin Weis MD, imiquimod and 5-fluorouracil for the treatment of actinic kerato- Reichenbach, Germany; Professor Claudio Varotti MD, sis: critical review and meta-analysis of efficacy studies. J Cutan Bologna, Italy; Professor Andrea Peserico MD, Padova, Italy; Med Surg 2005; 9:209–14. Incorporating Medical and Surgical Dermat- Professor Giuseppe Micali, Catania, Italy; Professor Petrus ology [serial online]. 2 Mar 2006. Available from http://www. Cornelis Maria van de Kerkhof MD, Nijmegen, the Nether- springerLink.com (accessed 3 April 2006). lands; Lourens Jan de Groot MD, Drachten, the Netherlands; 14 Edwards L, Ferenczy A, Eron L et al. Self-administered topical 5% imiquimod cream for external anogenital warts. Arch Dermatol 1998; Agustı´n Alomar MD (coordinating investigator), Barcelona, 134:25–30. Spain; Carlos Guille´n MD, Valencia, Spain; Jose´ M. Arrazola 15 Reichard O, Gunnar N, Fryden A et al. Randomised, double- MD, Madrid, Spain; Professor Andrew Y. Finlay MD, Cardiff, blind, placebo-controlled trial of interferon [alpha]-2b with and U.K.; Girish Gupta MD, Lanarkshire, U.K. We gratefully without ribavirin for chronic hepatitis C. Lancet 1998; 351:83– acknowledge Professor Helmut Kerl MD (Universita¨tsklinik 7.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 Imiquimod 5% cream – safety and efficacy study, A. Alomar et al. 141
16 Duncan M, Berman B. Gamma interferon is the lymphokine 18 Gillitzer R, Goebeler M. Chemokines in cutaneous wound healing. and beta interferon the monokine responsible for inhibition of J Leukoc Biol 2001; 69:513–21. fibroblast collagen production and late but not early fibroblast pro- 19 Tucci M, Offidani A, Lucarini G et al. Advances in the under- liferation. J Exp Med 1985; 162:516–27. standing of malignant transformation of keratinocytes: an 17 Granstein RD, Flotte TJ, Amento EP. Interferons and collagen pro- immunohistochemical study. J Eur Acad Dermatol Venereol 1998; duction. J Invest Dermatol 1990; 95 (6 Suppl.):75S–80S. 10:118–24.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp133–141 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07950.x Corticosteroid-induced clinical adverse events: frequency, risk factors and patient’s opinion L. Fardet,* A. Flahault, A. Kettaneh,* K.P. Tiev,* T. Ge´ne´reau, C. Tole´dano,* C. Lebbe´§ and J. Cabane* *Department of Internal Medicine, Hoˆpital Saint-Antoine, 184 rue du Faubourg Saint-Antoine, 75571 Paris Cedex 12, France Department of Public Health, Hoˆpital Tenon; INSERM UMR-S707; University Pierre et Marie Curie; 4 rue de la Chine, 75020 Paris, France Department of Internal Medicine, Nouvelles Cliniques Nantaises, 4 rue Eric Tabarly 44277 Nantes Cedex 2, France §Department of Dermatology, Hoˆpital Saint-Louis, 1 Av Claude Vellefaux, 75010 Paris, France
Summary
Correspondence Background More than 50 years after the introduction of corticosteroids, few stud- Laurence Fardet. ies have focused on corticosteroid-induced adverse events after long-term system- E-mail: [email protected] ic therapy. Objectives To assess the frequency, risk factors and patient’s opinion regarding Accepted for publication 28 December 2006 clinical adverse events occurring early during prednisone therapy. Patients and methods We conducted a cohort study in two French centres. All con- ) Key words secutive patients starting long-term (‡ 3 months), high dosage (‡ 20 mg day 1) adverse events, corticosteroids prednisone therapy were enrolled. The main clinical adverse events attributable to corticosteroids were assessed after 3 months of therapy, by comparison with Conflicts of interest baseline status. The patient’s opinion regarding the disability induced by these None declared. adverse events was recorded. Risk factors of frequently observed adverse effects were identified by using logistic regression. Results Eighty-eight patients were enrolled and 80 were monitored for at least 3 months (women 76%; mean age 59Æ1±18Æ7 years; giant cell arteritis 39%; ) mean baseline prednisone dosage 54 ± 17 mg day 1). Lipodystrophy was the most frequent adverse event [63Æ0% (51Æ0–73Æ1)], was considered the most dis- tressing by the patients and was most frequent in women and young patients. Neuropsychiatric disorders occurred in 42 patients [52Æ5% (41Æ0–63Æ8)], necessi- tating hospitalization in five cases. Skin disorders were noted by 37 patients [46Æ2% (35Æ0–57Æ7)] and were more frequent in women. Muscle cramp and proximal muscle weakness were reported by 32Æ5% (22Æ5–43Æ9) and 15% (8Æ0– 24Æ7) of patients, respectively. Newly developed hypertension occurred in 8Æ7% (2Æ9–20Æ3) of patients. Lastly, 39% (19Æ7–61Æ4) of the premenopausal women reported menstrual disorders. Conclusions Lipodystrophy and neuropsychiatric disorders are common adverse events of long-term prednisone therapy and are particularly distressing for the patients concerned. The impact of these adverse events on adherence to cortico- steroid therapy is not known.
Since their introduction more than 50 years ago, corticosteroids follow-up study of patients starting long-term systemic high- have assumed a major role in the management of a wide variety dose prednisone therapy, focusing on the rate and risk factors of of immunological, neoplastic and allergic diseases. However, major clinical adverse events. their use is limited by frequent adverse events, which may sometimes outweigh their beneficial effects. The literature on Patients and methods corticosteroid adverse events is surprisingly sparse. Apart from adverse osseous and ocular effects,1–5 corticosteroid adverse Patient population events have rarely been the focus of prospective studies. More- over, comparison of the results of available data is hindered Subjects were recruited in two French tertiary hospitals by differences in the study populations, treatment periods and between June 2003 and May 2005. All consecutive patients the definitions used.6–11 We therefore conducted a prospective who were prescribed long-term systemic prednisone were
2007 The Authors 142 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 Corticosteroids and adverse events, L. Fardet et al. 143 invited to participate. The following inclusion criteria were We did not systematically assess corticosteroid adverse applied: age over 18 years, need to start long-term (planned events on bones and eyes, already well documented in the to last more than 3 months) systemic therapy with predni- medical literature. All patients treated concomitantly with ) sone, initial dosage over 20 mg day 1, enrolment within other potassium-lowering drugs were given potassium chlor- ) 15 days of treatment outset, and written informed con- ide (1200 mg day 1). All the patients received calcium and sent. The following exclusion criteria were applied: human vitamin D supplementation, with or without biphosphonate. immunodeficiency virus (HIV) infection, inflammatory bowel The study complied with the ethical standards of the 2000 disease (because of the difficulty to interpret digestive com- revision of the Helsinki Declaration. All the subjects gave their plaints), pregnancy and previous systemic corticosteroid ther- written inform consent. apy ending < 6 months before enrolment in this study. Patients who said they had lost more than 5% of their initial Data analysis body weight owing to their underlying health disorder were not eligible for the study because of the difficulty of interpret- At the end of the study, three senior physicians independently ing the morphological changes on the photographs. and simultaneously examined the M0 and M3 photographs of As there is no consensus definition of ‘long-term’ cortico- each patient and classified each patient as lipodystrophic steroid therapy, we adopted a period of at least 3 months, like yes ⁄no ⁄unclassifiable. They were not blinded to the sequence other authors.12 Patients with < 3 months of follow-up were in which the photographs were taken. To help their decision excluded from the analysis. making, examiners were advised to take into account four cri- teria (moon face, double chin, wrinkles filling, nasogenian fold filling). However, final judgement was empirical without Study design and data collection a formalized scoring system. Corticosteroid-induced hyperten- After enrolment, all the patients were seen every month for sion was defined as a newly developed systolic blood pressure 3 months. At the baseline visit (M0), the following demogra- > 140 mmHg and ⁄or diastolic blood pressure > 90 mmHg at phic and clinical data were recorded on a standard question- two consecutive on-treatment visits. Neuropsychiatric dis- naire: sex, age, the reason for prednisone therapy, the initial orders were defined as a newly developed anxiety, irritability, steroid dosage (and the number of pulses, if any), use of other euphoria, hyperactivity, depression or manic episode. To medications prescribed before the initiation of prednisone, cur- report cutaneous disorders, we assessed the presence of hirsu- rent weight, weight before onset of the underlying disease, tism, spontaneous bruising or disturbed wound healing. hirsutism, spontaneous bruising, altered wound healing, Rates of adverse events were calculated only for disorders sodium and water retention (swollen ankles) and hand tremor. reported on the M3 questionnaire but not on the M0 ques- Arterial blood pressure was measured after 10 min of rest. tionnaire. Using this method, we sought to minimize the risk Digital photographs were taken of the face and dorsocervical of reporting symptoms not related to the treatment [e.g. region. We also asked patients if they suffered from muscle symptoms related to the underlying disease or to the patient cramp, proximal muscle weakness, insomnia, neuropsychiatric him (or her) self]. and emotional disorders, hyperphagia, epigastric pain and We used logistic regression models to identify potential risk menstrual disorders during the previous 3 months (i.e. the factors for adverse events occurring in at least 30% of patients. 3 months preceding the prednisone therapy). Because of the possibility that our models were overfitted, we At subsequent visits (M1, M2, M3), arterial blood pressure included only four items in our models: age (< vs. ‡ mean was assessed after 10 min of rest. At the M3 visit, the patients value), sex, previous systemic therapy lasting more than were examined for hirsutism, spontaneous bruising, altered 3 months (yes vs. no) and the cumulative M0–M3 steroid wound healing, sodium and water retention, and hand tre- dose divided by the M0 body weight (‡ vs. < mean value). mor. Infectious events were recorded. A second standard self- Adjusted odds ratios (OR) and 95% confidence intervals (CI) questionnaire was administered, covering subjective symptoms were calculated. The Clopper–Pearson exact method was used known to be associated with corticosteroids (see above) dur- to calculate the 95% CI. Differences were considered statisti- ing the previous 3 months. Patients were free to add any other cally significant when P <0Æ05 (two-tailed). Analyses were symptoms they attributed to prednisone therapy. The patients performed using SAS software (version 8Æ2, SAS Institute, were asked which adverse event they found most distressing. Cary, NC, U.S.A.). Values are reported as mean ± SD, rates We chose to question patients on the clinical adverse events (95% CI) and ORs (95% CI). related to corticosteroids only at the M3 visit (and not at the M1 and M2 visits) to have two periods of comparable times, Results i.e. 3 months before therapy and 3 months after therapy. For the calculation of the frequency of adverse events, we took Patients into account only adverse events reported on the second ques- tionnaire (i.e. those induced by corticosteroids) but not on During the 2-year inclusion period, 88 patients were enrolled. the first one (already present before the initiation of cortico- Five of these patients were excluded from the analysis because steroids). of follow-up shorter than 3 months (n = 3), steroid therapy
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 144 Corticosteroids and adverse events, L. Fardet et al.
Table 1 Baseline characteristics of the patients lasting < 3 months (n = 1) or poor adherence to treatment (n = 1). Three other patients rapidly and spontaneously Patients (n = 80) stopped taking their treatment because of the fear of adverse Sex events (a woman with Sweet disease who wanted to avoid Female, n (%) 61 (76) weight gain and two diabetic patients with suspected giant Male, n (%) 19 (24) cell arteritis who wanted to avoid a worsening of their Mean age, years ± SD 59Æ1±18Æ7 diabetes). Underlying disease The characteristics of the 80 participants are shown in Giant cell arteritis, n (%) 31 (39) Table 1. All received prednisone at a mean baseline dosage Connective tissue disease, n (%) 21 (26) )1 Vasculitis, n (%) 10 (13) of 54 ± 17 mg day . The mean M3 dosage was 31 ± )1 Neutrophilic dermatitis, n (%) 4 (5) 15 mg day prednisone. The mean M0–M3 cumulative )1 Systemic sarcoidosis, n (%) 3 (4) dosage was 4111 ± 1963 mg or 64 ± 30 mg kg of weight. Others, n (%) 11 (13) No patient died during the study period. Pulsed steroids, n (%) 9 (11) ) Mean initial steroid dosage (range) 54 mg day 1 (25–90) Other diseases Objective clinical adverse events Diabetes ⁄glucose intolerance 5 Hypertension 23 Table 2 shows the proportion of patients with adverse events 9,13–22 Use of other drugs during the first 3 months of prednisone therapy. By Gastroprotective drugs 23 using photographic evaluation, corticosteroid-induced lipo- Antihypertensive drugs 22 dystrophy was present after 3 months of therapy in two-thirds Other immunosuppressive therapy 0 of the patients. Lipodystrophy was not always associated with weight gain. Mean weight gain during the first 3 months of
Table 2 Proportion of patients with adverse events during the first 3 months of prednisone therapy
Number of Proportion of patients Published data in % Symptoms patients in % (95% CI) (therapy £ 3 months) Lipodystrophy (photographic evaluation) 50 63Æ0 (51Æ0–73Æ1) 15–409,20,21 Newly developed hypertension: 53 patients 58Æ7(2Æ9–20Æ3) 0–379,22 without past history of hypertension Cutaneous disorders 37 46Æ2 (35Æ0–57Æ7) 2120 Hirsutism 26Æ0 (15Æ8–39Æ1) Spontaneous bruising 13Æ8(7Æ1–23Æ3) Altered wound healing 17Æ5(9Æ9–27Æ6) Swollen ankles 8 10Æ0(4Æ4–18Æ8) 0–520,21 Hand tremors 20 25Æ0 (16Æ0–35Æ9) NDA Muscle cramps 26 32Æ5 (22Æ5–43Æ9) NDA Hands 26Æ3 (17Æ1–37Æ3) Feet 10Æ0(4Æ4–18Æ8) Legs 15Æ0(8Æ0–24Æ7) Proximal muscle weakness 12 15Æ0(8Æ0–24Æ7) 5313 Insomnia 43 53Æ8 (42Æ2–65Æ0) 24–5814–17 Neuropsychiatric disorders 42 52Æ5 (41Æ0–63Æ8) 0–189,14–16,18 Irritability 25Æ0 (16Æ0–35Æ9) Anxiety ⁄depression 11Æ3(5Æ3–20Æ3) Euphoria ⁄hyperactivity 12Æ5(6Æ2–21Æ8) Manic episode 3Æ8(0Æ8–10Æ6) Epigastric pain 10 12Æ5(6Æ2–21Æ8) 3–159,14,19 With gastroprotective drug (n = 23) 11Æ9(3Æ6–29Æ8) Without gastroprotective drug (n = 57) 13Æ1(5Æ1–26Æ8) Hyperphagia 38 47Æ5 (36Æ2–59Æ0) NDA Menstrual disorders: 23 premenopausal women 9 39Æ1 (19Æ7–61Æ4) 1216 Other Ichthyosis 32Æ5 (22Æ5–43Æ9) NDA Acneiform eruption 2Æ5(0Æ3–8Æ7) NDA Flush 3Æ8(0Æ8–10Æ6) NDA Dysphonia 3Æ8(0Æ8–10Æ6) NDA
NDA, no data available.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 Corticosteroids and adverse events, L. Fardet et al. 145 treatment was 1Æ4±3Æ1 kg but at M3, weight was stable or 28 66 75 79 Æ Æ Æ lower than at baseline in 33% of the lipodystrophic patients Æ (vs. 54% for the nonlipodystrophic patients, P <0Æ05). Cuta- P 5) 0 1) 0 5) 0 0) 0 Æ Æ Æ neous disorders were observed in half of the patients. Hirsut- Æ 6–4 5–3 3–2 ism was diagnosed in 16 of 61 (26%) women and in no 3–4 Æ Æ Æ Æ men. Mean blood pressure (BP) was stable during the study 7(0 2(0 8(0 2(0 Æ Æ Æ Æ OR (95% CI) period. Corticosteroid-induced hypertension occurred in five Model 6 Hyperphagia (8Æ7%, 2Æ9–20Æ3) out of 57 patients who had strictly normal 24 1 76 1 08 0 01 1 Æ Æ Æ BP at baseline. None of the five patients with corticosteroid- Æ induced hypertension had clinical signs of sodium ⁄water P 3) 0 3) 0 0) 0 retention. Six patients developed an acute infectious episode 7) < 0 Æ Æ Æ Æ during the study period (acute pyelonephritis or pneumonia 6–5 3–2 9–9 1–0 Æ Æ Æ n = 2, symptomatic strongyloidiasis n = 1, transient local Æ 9(0 9(0 8(0 2(0 Æ Æ Æ Æ Model 5 Neuropsychiatric disorders herpes zoster n = 3). OR (95% CI) 66 0 17 0 46 2 04 0 Æ Æ Æ Æ Patients’ complaints P 2) 0 4) 0 1) 0 0) 0 Æ Æ Æ The mean number of complaints per patient were 2Æ06 ± Æ 3–2 2–1 5–5 1–1 Æ Æ Æ 1Æ52 among men and 2Æ33 ± 1Æ19 among women (P =0Æ4) Æ 8(0 5(0 6(0 (Table 2). Nine out of 23 premenopausal women reported 3(0 Æ Æ Æ Æ Model 4 Insomnia menstrual disorders with irregular (n = 5) or shorter (n =2) OR (95% CI) menses and ⁄or less abundant menstrual bleeding (n = 3). 01 0 33 0 06 1 59 0 Æ Æ Æ Æ Muscle cramp was reported by 26 patients and was usually P distal. Hyperphagia was frequent and eight patients reported a 1) 0 5) 0 Æ Æ 2) 0 4) 0 Æ need to eat during the night. Insomnia and neuropsychiatric Æ 0–15 3–12 6–5 4–7 Æ Æ Æ disorders were reported by 43 and 42 patients, respectively. Æ 6(1 1(1 7(0 Six patients had severe mental disorders during the study per- 4(0 Æ Æ Æ Æ Model 3 Muscle cramps iod (manic episode n = 3, severe depression with suicidal OR (95% CI) 01 3 08 4 62 1 thoughts n = 2, aggressiveness n = 1), and five were hospital- 61 1 Æ Æ Æ Æ ized because of them. Only one of these patients, all of whom P were women, had a relevant history of minor depression. 3) < 0 Æ 1) 0 2) 0 6) 0 Æ Æ Æ 2–29 2–1 3–2 4–4 Æ Æ Æ Risk factors for frequently observed adverse events Æ 0(1 5(0 8(0 4(0 Æ Æ Æ Æ
Six adverse events were observed in more than 30% of Model 2 Cutaneous disorders OR (95% CI) patients. Their risk factors are reported in Table 3. The risk 02 6 01 0 02 0 75 1 Æ Æ Æ of lipodystrophy was higher in women and in younger Æ subjects and increased with the cumulative dosage of P prednisone. Cutaneous disorders were observed more 7) 0 0) < 0 5) 0 Æ Æ Æ 4) 0 frequently in women and tended to increase with age. The Æ 2–14 9–24 2–12 3–5 Æ Æ Æ risk of muscle cramps decreased with age and tended to be Æ 2(1 8(1 9(1 more frequently reported by women than by men. Insom- 3(0 Æ Æ Æ Æ Model 1 Lipodystrophy OR (95% CI) nia and neuropsychiatric disorders were less frequently 3 1
reported by patients with a past history of long-term corti- ) costeroid therapy.
Subjective impact of adverse events
During the first 3 months of treatment, 57 patients (71%) 64 vs. < 64 mg kg reported at least one adverse event and 53 (66%) reported at ‡ least one distressing adverse event. The adverse event cited as 60 years 6 ‡ the most distressing was morphological changes (primarily the moon face) cited by 39% of patients, followed by neuro- psychiatric disorders (17%), muscle cramps (11%), hand tre- Factors associated with adverse events affecting more than 30% of patients (multivariate analysis) mor (9%) and insomnia (7%); 17% of patients described Age: < 60 vs. Sex: women vs. men 4 Previous systemic therapy: yes vs. no 1 Cumulative dosage:
another adverse event as the most distressing. Table 3
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 146 Corticosteroids and adverse events, L. Fardet et al.
36,37 Discussion are uncertain. Available data suggest no link between cor- ticosteroid-induced hypertension and sodium retention.37,38 This cohort study focused on the clinical adverse events of As corticosteroid-induced adverse events on eyes and bones prolonged prednisone therapy and associated risk factors. The are well described in the literature, we did not focus on it. study population consisted only of patients who received However, these complications must be kept in mind. An high-dose prednisone for at least 3 months, and who had not increased risk of posterior subcapsular cataract and open-angle received such treatment in the recent past. Adverse events glaucoma has been demonstrated.5,39 The risk increases with were examined prospectively and systematically and rates were the daily cumulative dosage and the duration of treatment.5,39 calculated only for disorders reported under treatment and not Corticosteroid-treated patients also develop an increased risk of present during the 3 months having preceded the initiation of osteoporosis and bone fracture.40,41 The bone loss is major prednisone therapy. We found that 71% of patients reported during the first 6 months of therapy (about 10%) and decreas- adverse events during the first 3 months of steroid therapy. es thereafter (about 2–5% per year).42,43 A measurement of Lipodystrophy and neuropsychiatric disorders were the most the bone mass density is recommended before initiation of frequent events and despite usually being considered ‘minor’ systemic corticosteroid therapy at a baseline daily dosage ) by physicians, were the most distressing adverse events from ‡ 7Æ5 mg day 1 and planned to last more than 1 month.44 the patients’ standpoint. Nevertheless, while most physicians are aware of the risk of The incidence of corticosteroid-induced lipodystrophy (CIL) osteoporosis, 50–70% of patients on significant doses of corti- has rarely been assessed. Shubin9 found that 15% of 47 patients costeroids are not investigated for this adverse event45,46 and ) had CIL after 8 weeks of therapy with 10–30 mg day 1 of pred- 40–70% do not receive adequate prevention.46,47 nisone equivalent, while others20,21 reported incidence rates This prospective cohort study has several limitations. Firstly, ranging from 33% to 40% after 8–12 weeks of treatment (mean as 39% of patients had giant cell arteritis, our population was ) prednisone dosage: 23 mg day 1). The higher incidence in our older than the general population receiving prolonged predni- study may be explained by the higher mean prednisone dosage sone therapy. This could have biased our rates, some of the ) (42 mg day 1 during the first 3 months). The risk factors, adverse events such as hypertension and hand tremor being associated metabolic disorders and pathophysiological mech- more frequent in older subjects. However, we analysed only anisms of CIL are unknown, contrary to HIV-associated clinical disorders that occurred during therapy, thereby min- lipodystrophy.23,24 imizing this potential source of over-estimation. Secondly, Neuropsychiatric disorders were reported by more than except for some adverse events, this study was based mainly 50% of patients. Even if most neuropsychiatric complaints on the patients’ declarations. This could have led to an overes- were mild (irritability or anxiety), 8% of patients developed timation of some subjective adverse events, such as those major psychiatric disorders usually soon after beginning ster- noted on the packet insert. Thirdly, the patients considered oid therapy. This is in keeping with results from Hall et al.,25 that lipodystrophy was the most distressing adverse event after who found that the risk of psychiatric reactions was highest 3 months of prednisone therapy. Nevertheless, some other shortly after prednisone introduction. In our study, the risk of adverse events (e.g. neuropsychiatric complaints, muscle neuropsychiatric disorders was not associated with the cumu- cramps) were maximal during the first month of therapy and lative steroid dosage. In contrast, the Boston Collaborative tended to wane thereafter, whereas lipodystrophy showed the Drug Study18 showed that the frequency of corticosteroid- opposite trend. This point may have biased the impression of induced neuropsychiatric disorders increased with the dosage patients. Lastly, we attributed all patient-reported health dis- ) (1Æ3% of patients treated with < 30 mg day 1 prednisone, orders to prednisone, even if other causes may have contribu- ) 18Æ4% in those treated with > 80 mg day 1). However, all ted to these symptoms (e.g. the underlying disease and our patients received high daily dosages of prednisone, reduc- neuropsychiatric disorders). ing the chances that we would identify such a link. Such biases are difficult to avoid; however, placebo- A two- to three-fold increase in the frequency of peptic dis- controlled studies of systemic corticosteroid therapy are rarely orders has been observed in patients receiving corticoster- justified and a control group for comparison is thus difficult oids,26,27 but two meta-analyses28,29 showed that fewer than to provide. In that way, this study more accurately represents 2% of corticosteroid-treated patients developed peptic ulcer. In the rate of meaningful adverse events than an incidence that is our study, epigastric pain was frequent in patients who were somehow adjusted to subtract the placebo effect. However, already receiving proton pump inhibitors at enrolment. How- we think our methodology is pertinent because patients’ com- ever, there are no available data that prove the efficacy of plaints—even if attributable to placebo effect—are particularly proton pump inhibitors in the prevention of corticosteroid- relevant in that the patient’s perception often alters compliance induced epigastric pain, which may not be related to gastric to therapy and prescribing practices. disorders. The lack of reliable data on steroid adverse events is regret- Corticosteroid-induced arterial hypertension was infrequent table for several reasons. Firstly, about 0Æ5% of the 1Æ2 billion in our study, as also reported elsewhere,8,9,30–34 and we con- people in industrialized countries48 receive prolonged systemic firm that it tends to occur during the first days (or weeks) of corticosteroid therapy in a given time,12 yet they cannot be treatment.22,35 The underlying pathophysiological mechanisms properly informed of some frequent and debilitating adverse
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 Corticosteroids and adverse events, L. Fardet et al. 147 events such neuropsychiatric disorders. Secondly, adverse rheumatoid arthritis: clinical efficacy, disease-modifying properties, events that physicians generally consider ‘minor’ may none- and side effects: a randomized, double-blind, placebo-controlled theless alter patients’ quality of life and affect their willingness clinical trial. Ann Intern Med 2002; 136:1–12. 20 Bar-Meir S, Chowers Y, Lavy A et al. Budesonide versus prednisone to adhere to therapy.49 Thirdly, some of these complica- in the treatment of active Crohn’s disease. The Israeli Budesonide tions may be overestimated or, on the contrary, underesti- Study Group. Gastroenterology 1998; 115:835–40. 50,51 mated with potential financial and epidemiological 21 Campieri M, Ferguson A, Doe W et al. Oral budesonide is as effect- consequences. ive as oral prednisolone in active Crohn’s disease. The Global Budesonide Study Group. Gut 1997; 41:209–14. 22 Sato A, Funder JW, Okubo M et al. Glucocorticoid-induced hyper- References tension in the elderly. Relation to serum calcium and family his- 1 Kanis JA, Johansson H, Oden A et al. A meta-analysis of prior corti- tory of essential hypertension. Am J Hypertens 1995; 8:823–28. costeroid use and fracture risk. J Bone Miner Res 2004; 19:893–9. 23 Grinspoon S, Carr A. Cardiovascular risk and body-fat abnormal- 2 van Staa TP, Leufkens HG, Cooper C. The epidemiology of cortico- ities in HIV-infected adults. N Engl J Med 2005; 352:48–62. steroid-induced osteoporosis: a meta-analysis. Osteoporos Int 2002; 24 Carr A. HIV lipodystrophy: risk factors, pathogenesis, diagnosis 13:777–87. and management. AIDS 2003; 17 (Suppl. 1):S141–8. 3 Homik JE, Cranney A, Shea B et al. A metaanalysis on the use of 25 Hall RC, Popkin MK, Stickney SK et al. Presentation of the steroid bisphosphonates in corticosteroid induced osteoporosis. J Rheumatol psychoses. J Nerv Ment Dis 1979; 167:229–36. 1999; 26:1148–57. 26 Hernandez-Diaz S, Rodriguez LA. Steroids and risk of upper gastro- 4 Delcourt C, Cristol JP, Tessier F et al. Risk factors for cortical, intestinal complications. Am J Epidemiol 2001; 153:1089–93. nuclear, and posterior subcapsular cataracts: the POLA study. 27 Wolfe F, Hawley DJ. The comparative risk and predictors of Pathologies Oculaires Liees a l’Age. Am J Epidemiol 2000; 151:497– adverse gastrointestinal events in rheumatoid arthritis and osteo- 504. arthritis: a prospective 13 year study of 2131 patients. J Rheumatol 5 Garbe E, LeLorier J, Boivin JF, Suissa S. Risk of ocular hypertension 2000; 27:1668–73. or open-angle glaucoma in elderly patients on oral glucocorticoids. 28 Messer J, Reitman D, Sacks HS et al. Association of adrenocortico- Lancet 1997; 350:979–82. steroid therapy and peptic-ulcer disease. N Engl J Med 1983; 309:21– 6 Smyllie HC, Connolly CK. Incidence of serious complications of 4. corticosteroid therapy in respiratory disease. A retrospective survey 29 Conn HO, Poynard T. Corticosteroids and peptic ulcer: meta-analy- of patients in the Brompton hospital. Thorax 1968; 23:571–81. sis of adverse events during steroid therapy. J Intern Med 1994; 7 Lieberman P, Patterson R, Kunske R. Complications of long- 236:619–32. term steroid therapy for asthma. J Allergy Clin Immunol 1972; 30 Gabriel SE, Sunku J, Salvarani C et al. Adverse outcomes of antiin- 49:329–36. flammatory therapy among patients with polymyalgia rheumatica. 8 Thomas TP. The complications of systemic corticosteroid therapy Arthritis Rheum 1997; 40:1873–78. in the elderly. A retrospective study. Gerontology 1984; 30:60–5. 31 Caporali R, Cimmino MA, Ferraccioli G et al. Prednisone plus 9 Shubin H. Long term (five or more years) administration of corti- methotrexate for polymyalgia rheumatica: a randomized, double- costeroids in pulmonary diseases. Dis Chest 1965; 48:287–90. blind, placebo-controlled trial. Ann Intern Med 2004; 141:493–500. 10 Akerkar GA, Peppercorn MA, Hamel MB et al. Corticosteroid-associ- 32 Walsh LJ, Wong CA, Oborne J et al. Adverse effects of oral ated complications in elderly Crohn’s disease patients. Am J Gastro- corticosteroids in relation to dose in patients with lung disease. enterol 1997; 92:461–64. Thorax 2001; 56:279–84. 11 Saag KG, Koehnke R, Caldwell JR et al. Low dose long-term cortico- 33 Jackson SH, Beevers DG, Myers K. Does long-term low-dose steroid therapy in rheumatoid arthritis: an analysis of serious corticosteroid therapy cause hypertension? Clin Sci (Lond) 1981; adverse events. Am J Med 1994; 96:115–23. 61 (Suppl. 7):381s–3s. 12 Walsh LJ, Wong CA, Pringle M et al. Use of oral corticosteroids in 34 Rizzato G, Riboldi A, Imbimbo B et al. The long-term efficacy the community and the prevention of secondary osteoporosis: a and safety of two different corticosteroids in chronic sarcoidosis. cross sectional study. BMJ 1996; 313:344–6. Respir Med 1997; 91:449–60. 13 Batchelor TT, Taylor LP, Thaler HT et al. Steroid myopathy in can- 35 Kelly JJ, Mangos G, Williamson PM et al. Cortisol and hypertension. cer patients. Neurology 1997; 48:1234–38. Clin Exp Pharmacol Physiol Suppl 1998; 25:S51–6. 14 Kirwan JR, Hallgren R, Mielants H et al. A randomised placebo con- 36 Whitworth JA, Schyvens CG, Zhang Y et al. Glucocorticoid-induced trolled 12 week trial of budesonide and prednisolone in rheuma- hypertension: from mouse to man. Clin Exp Pharmacol Physiol 2001; toid arthritis. Ann Rheum Dis 2004; 63:688–95. 28:993–96. 15 Chibane S, Feldman-Billard S, Rossignol I et al. [Short-term toler- 37 Schacke H, Docke WD, Asadullah K. Mechanisms involved in the ance of three days pulse methylprednisolone therapy: a prospective side effects of glucocorticoids. Pharmacol Ther 2002; 96:23–43. 38 Whitworth JA, Kelly JJ. Evidence that high dose cortisol-induced study in 146 patients]. Rev Med Interne 2005; 26:20–6. + 16 Radakovic-Fijan S, Furnsinn-Friedl AM, Honigsmann H et al. Oral Na retention in man is not mediated by the mineralocorticoid dexamethasone pulse treatment for vitiligo. J Am Acad Dermatol receptor. J Endocrinol Invest 1995; 18:586–91. 2001; 44:814–17. 39 Urban RC Jr, Cotlier E. Corticosteroid-induced cataracts. Surv 17 Mignogna MD, Lo Muzio L, Ruoppo E et al. High-dose intravenous Ophthalmol 1986; 31:102–10. ‘pulse’ methylprednisone in the treatment of severe oropharyngeal 40 Cooper C, Coupland C, Mitchell M. Rheumatoid arthritis, cortico- pemphigus: a pilot study. J Oral Pathol Med 2002; 31:339–44. steroid therapy and hip fracture. Ann Rheum Dis 1995; 54:49–52. 18 Boston Collaborative Drug Study. Acute adverse reactions to pred- 41 McEvoy CE, Ensrud KE, Bender E et al. Association between cortico- nisone in relation to dosage. Clin Pharmacol Ther 1972; 13:694–98. steroid use and vertebral fractures in older men with chronic 19 van Everdingen AA, Jacobs JW, Siewertsz Van Reesema DR et al. obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 157 Low-dose prednisone therapy for patients with early active (3 Pt 1):704–9.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 148 Corticosteroids and adverse events, L. Fardet et al.
42 LoCascio V, Bonucci E, Imbimbo B et al. Bone loss in response to 47 Yood RA, Harrold LR, Fish L et al. Prevention of glucocorticoid- long-term glucocorticoid therapy. Bone Miner 1990; 8:39–51. induced osteoporosis: experience in a managed care setting. Arch 43 Laan RF, van Riel PL, van de Putte LB et al. Low-dose prednisone Intern Med 2001; 161:1322–27. induces rapid reversible axial bone loss in patients with rheuma- 48 Population Reference Bureau. Available at: www.prb.org (accessed toid arthritis. A randomized, controlled study. Ann Intern Med 1993; on 31 January 2007). 119:963–68. 49 Morrison E, Crosbie D, Capell HA. Attitude of rheumatoid arthritis 44 Goldstein MF, Fallon JJ Jr, Harning R. Chronic glucocorticoid ther- patients to treatment with oral corticosteroids. Rheumatology (Oxford) apy-induced osteoporosis in patients with obstructive lung disease. 2003; 42:1247–50. Chest 1999; 116:1733–49. 50 Bae SC, Corzillius M, Kuntz KM et al. Cost-effectiveness of low dose 45 Bell R, Carr A, Thompson P. Managing corticosteroid induced corticosteroids versus non-steroidal anti-inflammatory drugs and osteoporosis in medical outpatients. J R Coll Physicians Lond 1997; COX-2 specific inhibitors in the long-term treatment of rheuma- 31:158–61. toid arthritis. Rheumatology (Oxford) 2003; 42:46–53. 46 Hougardy DM, Peterson GM, Bleasel MD, Randall CT. Is enough 51 Pisu M, James N, Sampsel S et al. The cost of glucocorticoid-associ- attention being given to the adverse effects of corticosteroid therapy? ated adverse events in rheumatoid arthritis. Rheumatology (Oxford) J Clin Pharm Ther 2000; 25:227–34. 2005; 44:781–88.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp142–148 THERAPEUTICS DOI 10.1111/j.1365-2133.2007.07974.x A multicentre, randomized, controlled study of the efficacy, safety and cost-effectiveness of a combination therapy with amorolfine nail lacquer and oral terbinafine compared with oral terbinafine alone for the treatment of onychomycosis with matrix involvement R. Baran, B. Sigurgeirsson,* D. de Berker, R. Kaufmann, M. Lecha,§ J. Faergemann,– N. Kerrouche** and F. Sidou** Nail Disease Centre, 42 rue des Serbes, 06400 Cannes, France *Dermatology Centre, Smaratorg 1, 200 Kopavogur, Iceland Department of Dermatology, Bristol Royal Infirmary, Bristol BS2 8HW, U.K. Universita¨tsklinikum Frankfurt am Main, Dermatologie, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany §Hospital Clinic, Dermatologia, C/Villarroel 170, 08036 Barcelona, Spain –Sahlgrenska Universitetssjukhuset, Hudkliniken, Gro¨na Stra˚ket 16, S-413 45 Go¨teborg, Sweden **Galderma Research & Development, 2400 route des Colles, 06902 Sophia Antipolis, France
Summary
Correspondence Background Onychomycosis is common, accounting for up to 50% of all nail dis- R. Baran. orders. Toenail onychomycosis can cause nail deformity, embarrassment, pain E-mail: [email protected] and walking difficulties. Some populations, such as individuals with diabetes, are at higher risk for developing secondary complications such as infections. Treat- Accepted for publication 11 November 2006 ment takes many months and therapeutic choices can increase clinical effective- ness, lower toxicity and minimize healthcare costs. Key words Objectives Based on the results of a previous pilot study, the objective of the pre- amorolfine, combination therapy, cost-effectiveness, sent study was to show, in a larger population, the enhanced efficacy of a com- onychomycosis, randomized controlled trial, bination of amorolfine nail lacquer and oral terbinafine in the treatment of terbinafine onychomycosis with matrix involvement. In addition, a cost-effectiveness analysis Conflicts of interest was performed. The investigating authors received honoraria for Methods In this multicentre, randomized, open-label, parallel group study, patients conducting the studies, and work as consultants for were randomized to receive either a combination of amorolfine hydrochloride Galderma Laboratories; Farzaneh Sidou and Nabil 5% nail lacquer once weekly for 12 months plus terbinafine 250 mg once daily Kerrouche are employees of Galderma R&D. The for 3 months (AT group) or terbinafine alone once daily for 3 months authors have no other financial interests to disclose. (T group). The study duration was 18 months including a 6-month treatment- free phase following the 12-month active treatment phase for the AT group and a 15-month treatment-free phase following the 3-month active treatment phase for the T group. The primary efficacy criterion was overall response, dichotomized into success or failure, success being the combination of clinical cure and negative mycology at month 18. This criterion was used as the effectiveness measure in the pharmacoeconomic analysis, conducted from a payer perspective. Results In total, 249 patients were included into the study: 120 in the AT group and 129 in the T group. A significantly higher success rate was observed for patients in the AT group relative to those in the T group at 18 months (59Æ2% vs. 45Æ0%; P =0Æ03). Both treatment regimens were safe and well tolerated. Treatment cost per cured patient was lower for the combination than for terbina- fine alone in all countries.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 149 150 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al.
Conclusions Study results confirmed that, in the treatment of dermatophytic toenail onychomycosis with matrix involvement, amorolfine nail lacquer in combination with oral terbinafine enhances clinical efficacy and is more cost-effective than terbinafine alone.
Onychomycosis is a common fungal infection of the nail unit, involvement. The study also included two exploratory varia- accounting for up to 50% of all nail disorders.1–4 Approxi- bles, the presence of streaks and/or onycholysis on the target mately 2–10% of the adult population are affected by the con- toenail at baseline, which were analysed for a possible predic- dition and the prevalence rises with age, with approximately tive relationship with the overall response rate of combination 14–28% of those above 60 years of age affected.5–7 Onyc- therapy. Finally, the cost-effectiveness of the combination of homycosis is mainly caused by dermatophyte infections, with amorolfine nail lacquer with terbinafine was compared with greater than 80% involving Trichophyton rubrum.8 Without that of terbinafine alone. proper treatment, onychomycosis can cause psychosocial and 9,10 physical problems, significantly impacting quality of life. Materials and methods Various topical and oral treatment options are available for the management of onychomycosis.5,11 Treatment guidelines The study was conducted in accordance with the principles of suggest that topical treatments are appropriate for early and current international ethical standards originating from the mild forms of the disease, such as when the distal part of the Declaration of Helsinki and in compliance with local regula- nail is affected. In patients who did not respond to a 6-month tory requirements and was reviewed and approved by Inde- topical treatment or in more severe cases, such as those pendent Ethics Committees. All patients provided their written involving the nail matrix, or with lateral edge involvement, informed consent prior to entering the study. systemic treatment is indicated. However, with relapses and re-infections being common Study design and standard systemic treatments resulting in less than 50% of patients achieving disease-free nails,12 onychomycosis remains The efficacy and safety of the combination of amorolfine nail difficult to treat. In the past, combination therapy has been lacquer plus terbinafine were compared with those of terbina- demonstrated to be an effective and safe alternative treatment fine alone in an 18-month, multicentre, randomized, open- approach in patients who do not respond to monotherapy or label, and parallel group study in patients with dermatophytic as first-line therapy for patients who may benefit from a more onychomycosis and matrix involvement. The study was con- aggressive therapy.5 ducted at 20 centres in nine European countries between Feb- Management of onychomycosis is costly. The direct cost of ruary 2002 and September 2004. A 1:1 ratio randomization treating onychomycosis was estimated as $43 million per year list was specifically generated for this study. Suitable patients by U.S. Medicare.13 Hence, the therapeutic approach should received in chronological order, corresponding to their inclu- be based on a long-term strategy, taking into account the basic sion into the study at baseline, either a combination therapy concepts of pharmacoeconomics: weighing costs of treatment of amorolfine hydrochloride (Loceryl 5% nail lacquer; Gal- (input) against efficacy (outcome). derma, Lausanne, Switzerland) once weekly for 12 months Amorolfine, a morpholine derivative with potent antifungal and terbinafine (Lamisil 250 mg tablets; Novartis, Basel, properties,14 is available in a 5% nail lacquer in many coun- Switzerland) once daily for 3 months (AT group) or terbina- tries worldwide. Harman et al. demonstrated a synergistic fine alone once daily for 3 months (T group). The study activity in vitro for some combinations of topical amorolfine duration was 18 months; a 6-month treatment-free phase fol- and oral antifungal therapies, suggesting that this synergy may lowed the 12-month active treatment phase for the AT group improve clinical cure rates.15 Further research in humans dem- and a 15-month treatment-free phase followed the 3-month onstrated that enhanced efficacy exists when treatment with active treatment phase for the T group. Patients were evaluated once-weekly amorolfine nail lacquer is combined with stand- at baseline and at months 1, 3, 6, 9, 12, 15 and 18. A preg- ard systemic therapies for onychomycosis.5,14–17 nancy test was required at screening and after completion of In a previous pilot study, the combination of topical amor- terbinafine therapy (3-month visit) for all women of child- olfine and oral terbinafine, the most widely prescribed oral bearing potential. antifungal agent, resulted in markedly improved mycological At screening visit, investigators selected a pathological target and clinical outcomes as well as a better cost per cure ratio nail to be sampled for mycological follow-up. All samples for combination therapy.16 The objective of the present study were sent to a single reference laboratory (Professor E.G.V. was to confirm these results and to determine if an enhanced Evans, Cardiff, U.K.) for direct microscopic examination and efficacy exists when amorolfine nail lacquer is used in mycological culture. For direct microscopy, nail specimens combination with oral terbinafine in a larger cohort of pat- were placed in 20% potassium hydroxide solution. Cultures ients with dermatophytic toenail onychomycosis and matrix were maintained at room temperature for up to 4 weeks using
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. 151
Sabouraud dextrose agar with 0Æ005% chloramphenicol and France, Germany and Italy, where the questionnaire was cul- 0Æ05% actidione.16 turally validated.9 In addition, two exploratory criteria were assessed during the study: the presence of streaks and of ony- cholysis on the target nail was assessed at baseline, followed Patients by a subgroup analysis of the primary efficacy criterion accord- Men and women aged 18–70 years with dermatophytic ony- ing to the presence or absence of these characteristics. A yellow chomycosis affecting at least one great toenail (target nail) streak is a variant of dermatophyma. It presents as a distal to and matrix involvement were screened for enrolment. Eligible proximal yellow-white spike. It may be single or multiple. patients were required to have dermatophytes from both Safety was assessed through adverse events and laboratory direct microscopic and mycological culture examinations. monitoring. Throughout the course of the study, adverse Patients should have had a washout period of at least events were recorded by the investigators with a clinical deter- 3 months for antimycotic nail lacquers and of 6 months for mination of severity and relationship to study drug. Liver and oral antifungals. Exclusion criteria prohibited enrolment of kidney function were checked at screening and after comple- patients with predisposing conditions such as diabetes or other tion of terbinafine therapy (month 3 visit). peripheral circulatory disorders, psoriasis, lichen planus or human immunodeficiency virus; patients with known sensitiv- Cost-effectiveness evaluation ities to study treatments, patients requiring interfering treat- ment; and patients with impaired liver and/or kidney The primary efficacy criterion, overall response at the last visit, function. Women were excluded if they were pregnant, was chosen as efficacy measure. The pharmacoeconomic breast-feeding, or planning a pregnancy. evaluation was conducted from the payer’s perspective in every country. Only direct costs related to the drug acquisition were included. Other costs were not considered as both Efficacy and safety assessments groups followed the same evaluations (diagnostic tests, phys- The primary efficacy variable was the overall response at end- ician visits, laboratory tests etc.) and because only few adverse point [month 18, intent-to-treat (ITT) population, last obser- events occurred. The mean quantity of amorolfine nail lacquer vation carried forward (LOCF)] using a dichotomous scale of used per patient was based on a specific study evaluating the success or failure. Success was defined as the combination quantity of nail lacquer required to treat an onychomycosis- of clinical cure (i.e. disappearance of all lesions on each nail infected nail.18 For terbinafine, the quantity used was based or residual disease of no more than 10% of the original total on the protocol indication. Local public prices of amorolfine diseased surface) and negative mycology comprising both 5% nail lacquer and terbinafine 250 mg were considered. negative direct microscopy and negative culture (Table 1). When several pack sizes were available, drug costs were calcu- Secondary efficacy assessments included: clinical response lated to be minimized. (failure, improvement or cure), mycological examination of the target nail (direct microscopy and mycological culture), Statistical analyses and total percentage diseased surface. A 19-question ony- chomycosis quality-of-life questionnaire was completed at A sample size of 125 patients per group was deemed neces- baseline and at endpoint by a subset of study patients in sary to detect a 20% clinically relevant difference between the
Table 1 Definitions for efficacy assessments
Primary efficacy criterion Overall response: dichotomous scale, at month 18 (ITT, LOCF) Failure Failure Clinical failure or positive direct microscopy or positive culture Improvement Clinical improvement and negative direct microscopy and negative culture Success Cure Clinical cure and negative direct microscopy and negative culture Secondary efficacy criteria Clinical response: full scale (ITT, LOCF) Failure Reduction of total diseased surface by < 20% or worsening of condition from baseline Improvement At least 20% reduction in the total diseased surface from baseline Cure Disappearance of all lesions on each nail or residual disease of no more than 10% of the original total diseased surface (the remaining diseased surface on all affected nails should involve only the distal third) Mycological examination: dichotomous scale (ITT, LOCF) Negative Both negative direct microscopy and culture results Positive Positive direct microscopy and/or positive culture results
ITT, intent-to-treat; LOCF, last observation carried forward.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 152 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. two treatment groups based on the results of a previous age of patients who discontinued due to lack of efficacy (5Æ4% study16 and assuming a dropout rate of 20%. in the T group vs. 0Æ8% in the AT group) and to patient The primary statistical analysis was performed on the ITT request (5Æ4% in the T group vs. 1Æ7% in the AT group). Base- population, i.e. in all patients enrolled and randomized. The line characteristics of the ITT population are summarized in LOCF methodology was used to account for missing data for Table 2. These characteristics were comparable between the the ITT population analysis. The safety population was defined two treatment groups. Figure 1 presents the patient flow as all patients randomized and treated at least once. throughout the study. The analysis of all efficacy criteria was performed using the Major protocol deviations were reported for 67 patients Cochran–Mantel–Haenszel test stratified by ‘pseudocentre’ (27%), with a higher incidence in the T group (44 patients; (cluster of study centres set up by regional area) and ridit 34%) compared with the AT group (23 patients; 19%). The transformation. All tests were two-sided and used the 0Æ05 most common protocol deviation in both treatment groups level to declare significance. No adjustment for multiplicity was associated with the 18-month visit either not being was made. completed (16%) or being outside of the protocol-specified Adverse events were based on the safety population. Results timeframe (8%). were summarized by their relationship to the study drugs and by their severity (i.e. mild, moderate or severe). Efficacy evaluation
Results Success rates at each study visit are presented in Figure 2. At endpoint (month 18; ITT, LOCF), treatment with the amorol- fine-terbinafine combination resulted in a significantly higher Patient disposition and baseline characteristics success rate (59Æ2%) compared with treatment with terbina- In total, 249 patients were randomized and included into the fine alone (45Æ0%; P =0Æ03). During the course of the study, ITT population: 120 received amorolfine plus terbinafine the success rate in the AT group increased at each evaluation (AT group) and 129 received terbinafine alone (T group). following baseline, with an evident numerical difference Overall, 83Æ5% of all patients completed the study. Discontinu- between the treatment groups beginning at month 12. In ation rates were higher in the T group (20Æ2%) than in the terms of clinical response, patients in the AT group were sig- AT group (12Æ5%), mainly due to differences in the percent- nificantly more likely to be clinically cured at month 18 than
Table 2 Baseline demographics and disease AT group T group Total characteristics (intent-to-treat population) (n = 120) (n = 129) (n = 249) Sex, n (%) Male 82 (68Æ3) 87 (67Æ4) 169 (67Æ9) Female 38 (31Æ7) 42 (32Æ6) 80 (32Æ1) Age, years Mean ± SD 46Æ8±13Æ347Æ8±12Æ647Æ3±12Æ9 Median 46Æ849Æ748Æ6 Range 20Æ8–70Æ820Æ4–72Æ520Æ4–72Æ5 Race, n (%) White 115 (95Æ8) 126 (97Æ7) 241 (96Æ8) Black 2 (1Æ7) 2 (1Æ6) 4 (1Æ6) Asian 2 (1Æ7) 1 (0Æ8) 3 (1Æ2) Other 1 (0Æ8) – 1 (0Æ4) Mycological examination positivea, n (%) 120 (100Æ0) 129 (100Æ0) 249 (100Æ0) % total diseased surfaceb Mean ± SD 253 ± 197 234 ± 171 244 ± 184 Range 30–880 25–720 25–880 Causative organisms, n (%) Acremonium species 0 1 (0Æ8) 1 (0Æ4) Scopulariopsis brevicaulis 02(1Æ6) 2 (0Æ8) Trichophyton interdigitale 1(0Æ8) 0 1 (0Æ4) Trichophyton mentagrophytes 5(4Æ2) 6 (4Æ7) 11 (4Æ4) Trichophyton rubrum 112 (93Æ3) 120 (93Æ0) 232 (93Æ2) Missing 2 (1Æ7) 0 2 (0Æ8)
The AT group received amorolfine + terbinafine, while the T group received terbinafine only. aCombination of direct microscopy and culture results. bTotal diseased surface was determined by summation of the diseased surfaces of all affected nails.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. 153
Fig 3. Clinical response at month 18 (intent-to-treat population, last observation carried forward). *P <0Æ04.
Fig 1. Patient flow chart.
Fig 2. Success rates (intent-to-treat population, last observation carried forward) at each timepoint throughout the study. **P =0Æ03. those treated with terbinafine (66Æ7% vs. 53Æ5%, respectively; P <0Æ04; Fig. 3). Both treatment groups showed important reductions in the percentage of the total diseased surface on the target nail, with mean reductions of 85Æ1% for the AT group and 78Æ5% for the T group at month 18. Similarly, the mycological examin- ation showed a higher number of mycologically cured patients (defined as negative results on both direct microscopy and culture) in the AT group than in the T group from month 6 onwards (Fig. 4a). The overall mycological examination results mirrored the direct microscopy results (Fig. 4b), as mycotic structures visible microscopically may still be present long after all of the dermatophytes have been eliminated. Results from the mycological culture showed larger differences between the treatment groups compared with direct micro- Fig 4. Incidence of negative mycological examination. (a) P <0Æ05, scopy. Significant differences in the percentage of patients (b) P <0Æ01, (c) *P <0Æ001 and P <0Æ01.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 154 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al.
Fig 5. Effect of combination therapy with amorolfine plus terbinafine on onychomycosis with matrix involvement. (a) Baseline; (b) month 3; (c) month 12; (d) month 18.
with negative mycological cultures were observed as early as Table 3 Results of exploratory analyses (presence of streaks and/or month 3 (94Æ2% in the AT group vs. 59Æ7% in the T group; onycholysis at baseline and success rate) at month 18 (intent-to-treat P <0Æ001), and at months 6, 9, 12 and 15 (Fig. 4c). Figure 5 population, last observation carried forward) illustrates the effect of the amorolfine-terbinafine combination on dermatophytic toenail onychomycosis affecting the matrix Success rate, n (%) region during the course of the study. AT group T group P-value Two exploratory variables, the presence of streaks and/or of onycholysis at baseline, were analysed for a possible predictive Presence of streaks at baseline No 57 (58Æ2%) 46 (45Æ1%) relationship with the overall response rate of the combination 0Æ79 Yes 14 (63Æ6%) 12 (44Æ4%) therapy at month 18 (Table 3). Patients in the AT group with Presence of onycholysis at baseline the presence of streaks at baseline achieved higher success No 48 (59Æ3%) 40 (46Æ0%) 0Æ79 rates (63Æ6%) than patients in the AT group without streaks Yes 23 (59Æ0%) 18 (42Æ9%) (58Æ2%), while success rates in the AT group were the same The AT group received amorolfine + terbinafine, while the with (59Æ0%) or without (59Æ3%) the presence of onycholysis. T group received terbinafine only. In the T group, success rates were slightly lower in the pres- ence of streaks (44Æ4% with streaks vs. 45Æ1% without streaks) and of onycholysis (42Æ9% with onycholysis vs. 46Æ0% without onycholysis). However, the study was insufficiently The quality-of-life assessment was conducted in 39 patients powered for a proper statistical evaluation of these subgroup (15Æ7%) in Italy, Germany and France. Scores decreased simi- analyses. larly in both groups from baseline to endpoint. There was no
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. 155 difference in quality of life appreciation between the combin- Discussion ation and the monotherapy. Despite the availability of safe and efficacious therapies for onychomycosis, treatment failures with monotherapy are rou- Safety evaluation tinely reported and relapses are common, particularly in more Both amorolfine and terbinafine were well tolerated during extensive, severe cases. Recent in vitro and clinical studies this study. Most adverse events were mild or moderate in showed that combination therapies with oral and topical anti- severity. Treatment-related adverse events occurred in 15 fungal agents can be an attractive option to improve thera- (11Æ6%) patients in the T group and 19 (15Æ9%) patients in peutic efficacy in appropriate patients.15–17 The aim of this the AT group. Only two (1Æ7%) adverse events were deemed study was to assess the efficacy, safety and cost-effectiveness to be related to amorolfine (in-grown toenail and constipa- of a combination of amorolfine nail lacquer with oral terbina- tion). In total, 40 adverse events related to terbinafine were fine vs. terbinafine monotherapy in a large cohort of pati- experienced by 32 (12Æ9%) patients; the most frequently ents with dermatophytic toenail onychomycosis and matrix reported terbinafine-related adverse events were gastrointest- involvement. inal disorders. Efficacy results confirmed those previously published:19,20 a Seven (2Æ8%) patients discontinued due to adverse events combination of amorolfine nail lacquer with oral terbinafine [four (3Æ3%) in the AT group and three (2Æ3%) in the enhances treatment efficacy compared with terbinafine alone. T group], all of them deemed to be possibly related to The primary efficacy criterion, the success rate (clinical cure terbinafine. Nine patients (3Æ6%) experienced 11 serious and negative direct microscopic examination and negative cul- adverse events during the study, all of which were deemed to ture) after 18 months was shown to be significantly higher in be unrelated to study treatments. Laboratory monitoring the combination group (59Æ2%) compared with the terbina- showed no abnormalities in kidney or liver function. fine group (45Æ0%; P =0Æ03). Secondary efficacy assessments showed similar improvements in favour of the combination group. The study further demonstrated that the impact of the Cost-effectiveness analysis combination in the cure process was visible from month 3 In this study, a mean of 3Æ20 nails per patient were infected. onwards with significantly (P <0Æ001) more patients in the One 5-mL bottle (or two 2Æ5-mL bottles) allowed for 251 AT group presenting with negative mycological cultures com- applications,18 deemed sufficient for a 12-month treatment pared with patients in the T group, except for month 18. period. Three patients (2Æ4%) in the terbinafine group were inclu- Total costs in all countries involved were shown to be ded with nondermatophyte moulds at study entry. In order to higher for a treatment with the amorolfine-terbinafine com- assess the robustness of the ITT analysis, an additional analysis bination than with terbinafine alone. However, and because was run on all ITT population except the three patients with of the significantly higher efficacy of the combination ther- nondermatophyte moulds. It showed that the success rate was apy over the same time, costs per cured patient can be con- still higher with the combination treatment than with the sidered lower with the amorolfine-terbinafine combination monotherapy (59Æ2% vs. 46Æ0%) and this difference was still therapy than with terbinafine alone in all studied countries statistically significant (P =0Æ04). (Fig. 6). It was hypothesized that the addition of a topical antifungal treatment directly to the nail plate would improve the response rate over oral therapy alone in patients with import- ant nail invasion by fungi or in those with nail plates separ- ated from the nail bed. Therefore, two exploratory variables, the presence of streaks and/or of onycholysis on the target toenail at baseline, were analysed with a view to a possible predictive relationship with the overall response rate obtained with the combination therapy. No statistically significant improvements in overall response rate were observed for com- bination therapy patients with streaks and/or onycholysis, although the study was not sufficiently powered for a proper statistical evaluation of this subgroup analysis due to the rela- tively small sample size of the patients with streaks and ony- cholysis at baseline. However, the numerical increase in success rate for AT patients with streaks (AT group: 63Æ6% with streaks vs. 58Æ2% without streaks) warrants further study of this hypothesis in a larger cohort of patients. Fig 6. Cost per cured patient (€). Source for exchange rates: European The quality-of-life questionnaire was completed in only a Central Bank, 6 January 2006. limited number of countries, limiting meaningfulness of the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 156 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. statistical analysis. However, the analysis showed that scores There are three noteworthy differences between this study decreased in both groups similarly from baseline to endpoint, and the previous pilot study: firstly, treatment duration with confirming that quality of life is heavily impacted by diseases amorolfine was longer (15 months) in the pilot study than such as onychomycosis.9 in the presently reported study (12 months); secondly, the Both treatment regimens were safe and well tolerated, with primary endpoint of the current study was chosen to set a a similar incidence of treatment-related adverse events occur- high threshold for achieving success by combining the meas- ring in both groups (15Æ9% of patients in the AT group and ures of cured nail morphology (clinical success) with myco- 11Æ6% in the T group). All of the adverse events deemed rela- logical examination into an 18-month overall response ted to the study treatments were secondary to terbinafine, assessment, while the previous study evaluated mycological with the exception of two adverse events. Patient tolerance examination at 3 months as the primary efficacy variable; may be a concern with combination therapy; however, the finally, the duration of the current study was 18 months, addition of amorolfine nail lacquer to terbinafine therapy did 3 months longer than that of the pilot study, reflecting the not present an increase in safety risk relative to terbinafine currently preferred study length to allow sufficient time for alone in this study, consistent with safety findings in previous toenail regeneration.5 studies.16,17 The enhanced overall response with the amorolfine-terbina- Given the increasing prevalence of fungal nail infections and fine combination results partly from the complementary nat- the associated costs and burden to the healthcare systems, it is ure of topical and systemic treatments.21,22 Amorolfine and important to select the most cost-effective treatments,13 bearing terbinafine both inhibit the production of ergosterol, but they in mind that a successful long-term strategy should be sought in target different enzymes along the ergosterol biosynthesis the treatment of onychomycosis. Indeed, a combination of treat- pathway. The combination of an oral and a topical agent also ments that are effective in the long term is more cost-effective provides the benefit of two different routes of administration, than a treatment used for a shorter period but with lower long- targeting the infection at the nail bed via the bloodstream as term efficacy with possible complications related to the relapse well as through the nail plate. or chronicity of the disease, adding to the patient’s overall In summary, amorolfine nail lacquer, when used in com- healthcare costs. The pharmacoeconomic investigation carried bination with terbinafine in the treatment of dermatophytic out in the present study, accounting for the cure rate as effect- toenail onychomycosis with matrix involvement, provides iveness measure and based on the public prices, showed that in enhanced efficacy relative to terbinafine alone, hence confirm- all participating countries, combination therapy with amorolfine ing results of a previous pilot study. The addition of amorol- nail lacquer is more cost-effective than oral terbinafine mono- fine nail lacquer to terbinafine therapy did not present an therapy. The cost per cured patient with the combination increased safety risk relative to terbinafine alone and is more therapy was lower, depending on the country, by up to cost-effective than a treatment with terbinafine alone. 15Æ47% compared with oral terbinafine monotherapy. Only the direct costs of the drugs were considered. Adverse event- Acknowledgments related costs were not included as treatments were relatively safe without severe side-effects. Furthermore, we did not con- The authors thank the following investigators: Dr Serge Dahan, sider the cost related to treatment failure. This is a conserva- Toulouse, France; Dr Martine Feuilhade, Paris, France; tive approach as an additional 15% of patients (22 patients) in Dr Miche`le Havet, Bordeaux Cauderan, France; Dr Patrick the terbinafine group compared with the combination group Combemale and Dr Sandra Ronger, Lyon, France; Dr Georges were not cured and thus might have received a new treatment Reuter, Strasbourg, France; Prof. Dr Med. Thomas Luger, and might have undergone a series of new mycological tests. Mu¨nster, Germany; Prof. Dr Percy Lehmann, Wuppertal, In a previous study,16 the cost-effectiveness analysis, com- Germany; Prof. Ruggero Caputo, Milan, Italy; Prof. Antonio paring the amorolfine-terbinafine combination with terbina- Garcovich, Rome, Italy; Dr Christa De Cuyper, Bruges, fine alone, was conducted only in France. Current study Belgium; Dr Bertrand Richert, Lie`ge, Belgium; Dr Josette results confirm the findings and extend the analysis to eight Andre, Brussels, Belgium; Prof. Andrew Finlay, Cardiff, U.K.; other European countries. These results are also in line with Dr Mark Goodfield, Leeds, U.K.; Prof. Ame´rico Figueiredo, recent publications showing that amorolfine in association Coimbra, Portugal; Dr Jose´ L. Sanchez, Valencia, Spain; with terbinafine or itraconazole is more cost-effective than Dr Carlos Ferrandiz, Badalona, Barcelona, Spain; Dr Ola each of these treatments alone.13 The synergy between the Rollman, Uppsala, Sweden; Dr Hannele Heikkila, Helsinki, two drugs increases the success rate and might improve Finland; Dr Jon H. Olafsson and Dr Jon T. Steinsson, Kopavogur, patient compliance. This study adds to the body of knowledge Iceland; and Zeina Saab, Medical Writer, Galderma, for editorial supporting the finding that combination therapy is the most assistance. cost-effective treatment. The present study results support the existing published Dedication data showing the efficacy of combination therapy for the treatment of onychomycosis and confirm the results observed This publication is dedicated to the memory of our friend and in an earlier pilot study of amorolfine and terbinafine.5,16 colleague, Professor Glyn Evans.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 Amorolfine + terbinafine for onychomycosis with matrix involvement, R. Baran et al. 157
References 13 Arenas-Guzman R, Tosti A, Hay R et al. National Institute for Clin- ical Excellence. Pharmacoeconomics – an aid to better decision- 1 Baran R, Hay R, Haneke E et al. Onychomycosis: the Current Approach to making. J Eur Acad Dermatol Venereol 2005; 19 (Suppl. 1):34–9. Diagnosis and Therapy, 2nd edn. Oxford: Taylor and Francis, 2006. 14 Haria M, Bryson HM. Amorolfine: a review of its pharmacological 2 Andre J, Achten G. Onychomycosis. Int J Dermatol 1987; 26: 481– properties and therapeutic potential in the treatment of onychomy- 90. cosis and other superficial fungal infections. Drugs 1995; 49:103–20. 3 Faergemann J, Baran R. Epidemiology, clinical presentation and 15 Harman S, Ashbee HR, Evans EGV. Testing of antifungal combin- diagnosis of onychomycosis. Br J Dermatol 2003; 149 (Suppl. ations against yeast and dermatophytes. J Dermatol Treat 2004; 65):1–4. 15:104–7. 4 Hay R. Literature review. Onychomycosis. J Eur Acad Dermatol Venereol 16 Baran R, Feuilhade M, Datry A et al. A randomized trial of amorol- 2005; 19 (Suppl. 1):1–7. fine 5% solution nail lacquer associated with oral terbinafine com- 5 Baran R, Kaoukhov A. Topical antifungal drugs for the treatment pared with terbinafine alone in the treatment of dermatophytic of onychomycosis: an overview of current strategies for mono- toenail onychomycosis affecting the matrix region. Br J Dermatol therapy and combination therapy. J Eur Acad Dermatol Venereol 2005; 2000; 142:1177–83. 19:21–9. 17 Lecha M, Alsina M, Torres-Rodriguez JM et al. An open-label, multi- 6 Roberts DT. Onychomycosis: Current treatment and future chal- center study of the combination of amorolfine nail lacquer and oral lenges. Br J Dermatol 1999; 141 (Suppl. 56):1–4. itraconazole compared with oral itraconazole alone in the treatment 7 Gupta AK, Jain HC, Lynde CW et al. Prevalence and epidemiology of severe toenail onychomycosis. Curr Ther Res 2002; 63:366–79. of unsuspected onychomycosis in patients visiting dermatologists’ 18 Marty JP, Lambert J, Jackel A et al. Treatment costs of three nail lac- offices in Ontario, Canada – a multicenter survey of 2001 patients. quers used in onychomycosis. J Dermatolog Treat 2005; 16:299–307. Int J Dermatol 1997; 36:783–7. 19 Sigurgeirsson B, Billstein S, Rantanen T et al. L.I.O.N. study: efficacy 8 Scher RK. Onychomycosis: therapeutic update. J Am Acad Dermatol and tolerability of continous terbinafine (Lamisil ) compared to 1999; 40:S21–6. intermittent itraconazole in the treatment of toenail onychomyco- 9 Drake LA, Patrick DL, Fleckman P et al. The impact of onychomy- sis. Br J Dermatol 1999; 141 (Suppl. 56):5–14. cosis on quality of life: development of an international onycho- 20 Sigurgeirsson B, Olafsson JH, Steinsson JB et al. Long-term effect- mycosis-specific questionnaire to measure patient quality of life. JAm iveness of treatment with terbinafine vs itraconazole in onycho- Acad Dermatol 1999; 41:189–96. mycosis: a 5-year blinded prospective follow-up study. Arch 10 Turner RR, Testa MA. Measuring the impact of onychomycosis on Dermatol 2002; 138:353–7. patient quality of life. Qual Life Res 2000; 9:39–53. 21 Gupta AK, Einarson TR, Summerbell RC et al. An overview of 11 Roberts DT, Taylor WD, Boyle J. Guidelines for treatment of onyc- topical antifungal therapy in dermatomycoses: a North American homycosis. Br J Dermatol 2003; 148:402–10. perspective. Drugs 1998; 55:645–74. 12 Epstein E. How often does oral treatment of toenail onychomycosis 22 Evans E. The rationale for combination therapy. Br J Dermatol 2001; produce a disease-free nail? Arch Dermatol 1998; 134:1551–4. 145 (Suppl. 60):9–13.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp149–157 CONCISE COMMUNICATION DOI 10.1111/j.1365-2133.2007.07922.x Endothelial cells in infantile haemangiomas originate from the child and not from the mother (a fluorescence in situ hybridization-based study) S. Re´gnier, N. Dupin, C. Le Danff,* M. Wassef, O. Enjolras § and S. Aractingi– Department of Dermatology, Hoˆpital Tarnier-Cochin (APHP), Paris F-75006, France *CEA, Hoˆpital Saint-Louis, Paris, France Department of Pathology, Hoˆpital Lariboisie`re (APHP), Paris F-75010, France Department of Maxillofacial and Plastic Surgery, Hoˆpital d’Enfants Armand Trousseau (AHHP), Paris F-75012, France §Universite´ Pierre et Marie Curie-Paris 6, UFR de Me´decine Pierre et Marie Curie, Paris F-75005, France –Department of Dermatology, Hoˆpital Tenon (APHP), Paris F-75020, France
Summary
Correspondence Background Infantile haemangiomas are benign vascular tumours of infancy of Stephanie Re´gnier. unknown origin. Their aetiological relationship to maternal cells has been ques- E-mail: [email protected] tioned given that they develop during the neonatal period. Objectives As endothelial cells in the placenta may be of maternal or fetal origin, Accepted for publication 12 January 2007 we questioned whether vascular haemangioma cells originated from fetal or maternal tissue. Key words Methods We aimed to detect, by using fluorescence in situ hybridization, maternal endothelial cells, infantile haemangioma, XX cells in the male XY tissue in four specimens of infantile haemangiomas microchimerism obtained from boys. A sample of a female infantile haemangioma was used as Conflicts of interest a positive control and a male specimen of melanocytic naevus as a negative None declared. control. Results In one case of infantile haemangioma, a single XX female—probably maternal—cell was detected in the infantile haemangioma. All the other cells from this male as well as the three other informative specimens were uniformly negative for XX cell detection. Conclusion Our results support the hypothesis that endothelial cells of infantile haemangiomas appear to derive from the child itself, in accordance with other studies.
Infantile haemangiomas are benign vascular tumours of placental trauma, a condition known to increase the occur- infancy that appear within the first weeks of life. After a per- rence of infantile haemangiomas. As endothelial cells in the iod of rapid growth, the proliferating phase, a prolonged placenta may be of maternal or fetal origin, our objective was involuting phase occurs that lasts over a period of 2–10 years. to determine whether haemangioma endothelial cells origin- Several studies have shown that in infantile haemangiomas, ated from fetal or maternal tissue. the endothelial cells always express glucose transporter 1 (GLUT1), Fcc RII, Lewis Y antigen and merosin, a vascular Patients and methods immunophenotype similar to endothelial cells in the placenta.1 Furthermore, it was recently demonstrated that endothelial Patients cells in infantile haemangiomas are clonal and proliferate abnormally.2 Together, these results suggest that placental In order to study the maternal origin, we searched for routine endothelial cells reaching the fetal skin, and experiencing clo- biopsies of infantile haemangiomas in boys. Fluorescence nal expansion, could lead to the development of infantile in situ hybridization (FISH) allows detection of maternal XX haemangiomas.1 Within the placenta, there is a direct close cells in the background of male XY tissue. Skin biopsy is not contact of maternal and fetal cells, responsible for a bidirec- routinely carried out anymore in infantile haemangiomas, tional trafficking between the child and the mother. In add- but 16 specimens of infantile haemangiomas were obtained ition, admixture of maternal and fetal cells is enhanced by from boys and studied by FISH. Unfortunately, only four
2007 The Authors 158 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp158–160 Origin of endothelial cells in infantile haemangiomas, S. Re´gnier et al. 159 formalin-fixed and paraffin-embedded biopsies were informa- male melanocytic naevus, used as a control, no female cells tive. The other 12 specimens were fixed in Bouin solution, were noted. In contrast, a sample of female infantile haeman- which alters DNA and does not allow FISH realization. A sam- gioma showed nicely stained double XX cells. ple of a female infantile haemangioma was used as a positive control and a male specimen of melanocytic naevus as a Discussion negative control. The presence of a single maternal cell found in only one case strongly suggests that neither the endothelial cells nor other Fluorescent in situ hybridization infiltrating cells of the infantile haemangiomas were derived FISH was carried out on paraffin-embedded sections as from the mother but were of fetal origin. Indeed, although in described.3 Briefly, 5-lm skin sections were deparaffinized the control from a male child, we did not find any maternal and then rehydrated in ethanol series. DNA was denatured by cells, other studies have demonstrated the presence of maternal ) incubation in 0Æ2 mol L 1 HCl for 15 min. After incubation cells in normal skin in children.4 Similarly, Maloney et al.5 with proteinase K, the slides were treated with 4% formalde- showed the presence of maternal cells in the peripheral blood hyde and dehydrated. The XY cocktail probes (Vysis ; Abbot of healthy normal subjects decades after birth. Therefore, the Molecular Diagnostics, Rungis, France) were applied and after single maternal cell found in one case is most probably inde- closure with rubber cement, the probes and the DNA were pendent of haemangioma development. Of note, the four denatured and the samples incubated overnight at 42 C. The males that we studied had surgically resected infantile haeman- ) next day, counterstaining was carried out with 0Æ03 lgmL 1 giomas. This indicates a bias as most infantile haemangiomas 4¢,6-diamidino-2-phenylindole. FISH readings were carried are not usually surgically removed. However, all had typical out using confocal microscopy. and pathological infantile haemangiomas which were removed because of their location. In addition, the main pathogenic Results events, particularly the cellular origin, are probably the same in all infantile haemangiomas. We therefore believe that the The infantile haemangiomas analysed consisted of tumour or results found here apply to infantile haemangiomas in general. infiltrated plaques in three cases and a discrete lesion in one Although the pathogenesis of infantile haemangiomas case. They were located on the cheek, the eyelid, the upper remains unknown, North et al.1 showed that the endothelial lip and the forehead. Two of four were still in a proliferating cells of infantile haemangiomas highly express Lewis Y anti- stage while two were in the first part of the involuting phase. gen, FccRII, merosin and GLUT1. The endothelial cells of var- In three specimens of infantile haemangioma, FISH did not ious other cutaneous vascular tumours, including congenital disclose any female maternal cells. In one case, a single XX haemangiomas, tufted angioma, kaposiform haemangioendo- female cell was detected in the haemangioma (Fig. 1). In the thelioma, and a number of infantile vascular tumours now clearly individualized from infantile haemangiomas, did not express these antigens. The presence of all these markers is similar to that on the vessels of placental chorionic villi.1 The authors proposed therefore two possible mechanisms for development of the haemangiomas: colonization of the child’s mesenchyme by angioblasts which later aberrantly switch to a placental endothelial ‘immature’ phenotype, or placental endothelial cells embolizing from chorionic villi within the fetal tissue. The latter hypothesis is supported by the increased incidence of haemangiomas following chorionic villus samp- 6 XY cell ling, as placental trauma is known to enhance embolization. Importantly, Boye et al.2 analysed a series of infantile haeman- giomas and showed that in all, endothelial cells were clonal and differed from normal endothelial cells in rate of migration and proliferation in vitro. Circulating AC133/CD34 double- XX cell positive cells as well as CD34/VEGF/KDR double-positive cells—considered to be specific for endothelial progenitor cells—were increased 15-fold in patients with infantile haem- angiomas.7 In accordance, Yu et al.8 showed the presence of endothelial progenitor cells in infantile haemangiomas during Fig 1. Fluorescence in situ hybridization showing a single XX cell in the proliferating phase. Furthermore, an interesting study tissue from an infantile haemangioma of a male patient. XX cell, comparing the transcriptomes of human placenta and infantile nuclei containing two X chromosomes (green); XY cell, nuclei haemangiomas showed sufficient similarity to suggest a containing an X chromosome (green) and a Y chromosome (red). placental origin for infantile haemangiomas.9 Together, the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp158–160 160 Origin of endothelial cells in infantile haemangiomas, S. Re´gnier et al. above results strongly suggest that endothelial cells of infantile References haemangiomas derive from the migration of a unique placen- 1 North PE, Waner M, Mizeracki A et al. A unique microvascular tal endothelial cell. phenotype shared by juvenile hemangiomas and human placenta. As stated before, the placenta is a complex interactive struc- Arch Dermatol 2001; 137:559–70. ture mixing maternal and fetal cells. Endothelial cells in the 2 Boye E, Yu Y, Paranya G et al. Clonality and altered behavior placenta may be fetally or maternally derived and both may of endothelial cells from hemangiomas. J Clin Invest 2001; 10:745– express the GLUT1 antigen. Fetal endothelial progenitor cells 52. regularly enter the maternal circulation and, conversely, 3 Johnson KL, Zhen DK, Bianchi DW. The use of fluorescence in situ maternal endothelial progenitor or differentiated cells may hybridization (FISH) on paraffin-embedded tissue sections for the study of microchimerism. Biotechniques 2000; 29:1220–4. enter the fetal circulation. Therefore, if infantile haemangio- 4 Khosrotehrani K, Guegan S, Fraitag S et al. Presence of chimeric mas derive from placental endothelial cells, the question arises maternally derived keratinocytes in cutaneous inflammatory dis- of whether they are of maternal or fetal origin. Our results eases of children: the example of pityriasis lichenoides. J Invest suggest that these were fetal. This is in accordance with the Dermatol 2006; 126:345–8. study on clonality by Boye et al.2 using the X-inactivation tech- 5 Maloney S, Smith A, Furst DE et al. Microchimerism of maternal nique; this method relies on the fact that there are indeed two origin persists into adult life. J Clin Invest 1999; 104:41–7. copies of the X chromosome, and as it was informative, this 6 Burton BK, Schulz CJ, Angle B et al. An increased incidence of hemangiomas in infants born following chorionic villus sampling implies that female tissues were XX, and therefore also of fetal (CVS). Prenat Diagn 1995; 15:209–14. 10 origin. In addition, Pittman et al. recently showed the 7 Kleinman ME, Tepper OM, Capla JM et al. Increased circulating absence of maternal cells in seven infantile haemangiomas AC133+ CD34+ endothelial progenitor cells in children with using FISH, as here, but also using microsatellite markers on hemangioma. Lymphat Res Biol 2003; 1:301–7. sorted endothelial cells, as well as an elegant analysis of 8 Yu Y, Flint AF, Mulliken JB et al. Endothelial progenitor cells in GLUT1 polymorphism in the haemangiomas. infantile hemangioma. Blood 2004; 103:1373–5. Maternal cells have been detected by polymerase chain reac- 9 Barnes CM, Huang S, Kaipainen A et al. Evidence by molecular pro- filing for a placental origin of infantile hemangioma. Proc Natl Acad tion (PCR) analysis in fetal liver, lung, heart, thymus, spleen, Sci U S A 2005; 102:19097–102. adrenal gland, kidney and placenta in a second-trimester fetus 10 Pittman KM, Losken HW, Kleinman ME et al. No evidence for 11 with malformations. Using FISH and PCR, maternal micro- maternal–fetal microchimerism in infantile hemangioma: a mole- chimerism was identified in the liver of infants with biliary cular genetic investigation. J Invest Dermatol 2006; 126:2533–8. atresia.12 Maternal myocytes have been found in altered heart 11 Gotherstrom C, Johnsson AM, Mattsson J et al. Identification of of small infants with neonatal lupus syndrome.13 Also, mater- maternal hematopoietic cells in a 2nd-trimester fetus. Fetal Diagn nal keratinocytes were identified in several types of epidermal Ther 2005; 20:355–8. 12 Suskind DL, Rosenthal P, Heyman MB et al. Maternal microchimer- damage in children.4 Other studies identified maternal micro- ism in the livers of patients with biliary atresia. BMC Gastroenterol chimerism in peripheral blood cells of the skin and muscle of 2004; 4:14. 14,15 children with juvenile dermatomyositis. In one of these, 13 Stevens AM, Hermes HM, Rutledge JC et al. Myocardial-tissue- the authors were able to detect maternal T lymphocytes allore- specific phenotype of maternal microchimerism in neonatal lupus active against the children’s cells. These original and interest- congenital heart block. Lancet 2003; 365:1617–23. ing data demonstrate that cells transferred from mothers to 14 Artlett CM, Ramos R, Jiminez SA et al. Chimeric cells of maternal sons during pregnancy may either be stem cells helping tissue origin in juvenile idiopathic inflammatory myopathies. Child- hood Myositis Heterogeneity Collaborative Group. Lancet 2000; repair or sometimes be deleterious.14,15 Nevertheless, although 356:2155–6. there is significant and diverse trafficking of maternal cells 15 Reed AM, Picornell YJ, Harwood A et al. Chimerism in children towards fetuses, endothelial cells of infantile haemangiomas with juvenile dermatomyositis. Lancet 2000; 356:2156–7. appear to derive from the child itself.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp158–160 CONCISE COMMUNICATION DOI 10.1111/j.1365-2133.2007.07925.x Transforming growth factor-b receptor II is preferentially expressed in the companion layer of the human anagen hair follicle H.M. Sowden, R.O.S. Karoo and D.J. Tobin Medical Biosciences Research, School of Life Sciences, University of Bradford, Bradford BD7 1DP, U.K.
Summary
Correspondence Background Transforming growth factor (TGF)-b is a multifunctional growth factor Desmond J. Tobin. with multiple roles in skin including hair follicle development and cycling, E-mail: [email protected] where it regulates cell proliferation, differentiation and apoptosis, as well as in wound healing. While TGF-b receptor I (TGF-b RI) and receptor II (TGF-b RII) Accepted for publication 3 February 2007 expression helps define early human hair follicle morphogenesis, expression in the adult human hair follicle remains to be established. Key words Objectives To assess TGF-b receptor expression in human scalp anagen hair companion layer, hair follicle, transforming growth follicles. factor-b receptors Methods Immunohistochemical and double immunofluorescence analysis of TGF-b Conflicts of interest RI and RII was conducted on frozen sections of haired human scalp obtained None declared. from 10 healthy individuals. Results TGF-b RI expression was detected in the outer root sheath of anagen hair follicles while TGF-b RII was expressed almost exclusively in the companion layer of inner root sheath and less so in premedulla keratinocytes. Both receptors were colocalized in the companion layer of the proximal and mid follicle. Conclusions The well-described role of TGF-b in keratinocyte apoptosis during catagen is likely to involve anagen-specific hair follicle components including the companion layer, as this layer provides the slippage plane supporting the inner root sheath and hair shaft as they ascend to the skin surface. Results of this study suggest that the colocalization of TGF-b RI/RII complexes at the companion layer would facilitate TGF-b signalling at this site to regulate apoptosis of the compan- ion layer keratinocytes, facilitating shrinkage/contraction of this cell layer during hair follicle regression/catagen.
The complexity of the regulatory controls underlying hair (ORS) and IRS Henle’s layer. However, unlike ORS cells, com- growth and cycling continues to present numerous challenges panion layer keratinocytes migrate upwards together with the to the development of clinical interventions. The adult hair rest of the IRS to ensheath the growing hair shaft within.3 follicle (HF) continually cycles through phases of growth Both IRS and companion layer cells are produced only during (anagen), apoptosis-driven regression (catagen) and relative anagen and are switched off at the end of anagen and start of rest or quiescence (telogen).1 However, the identity of the catagen. While many growth factors have been implicated in molecular stimuli responsible for initiating these changes has the regulation of the hair growth cycle,1 factors associated yet to be fully elucidated. with catagen induction remain elusive. The anagen hair bulb is composed of highly proliferative This is in part due to the somewhat paradoxical observation epithelial cells that give rise to multiple keratinocyte lineages that while transforming growth factor (TGF)-b is associated including trichocytes of hair shaft (cuticle, cortex and with tissue development during HF morphogenesis this cyto- medulla) and the multicomponent inner root sheath (IRS) kine can also induce tissue involution in this same mini-organ including the cuticle, Huxley’s and Henle’s layer as well as the when it cycles in the adult.4–6 Thus, this cytokine has wide- recently characterized ‘companion layer’.2 The latter is a dis- ranging and opposing effects on many cellular processes tinct layer consisting of a single sheath of relatively flattened including differentiation, proliferation, cellular activation, keratinocytes, located at the interface of the outer root sheath immune and inflammatory responses. TGF-b regulates cellular
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp161–164 161 162 TGF-b RII expression in the hair follicle, H.M. Sowden et al.
MED
Fig 1. Transforming growth factor (TGF)-b receptor II (RII) expression in the companion layer of the human anagen hair follicle. TGF-b RII was detected almost exclusively in the companion layer of the anagen hair follicle (a, b). When companion layer cells were viewed obliquely the TGF-b RII expression was cytoplasmic and granular (c). Weak expression was also seen in presumptive medullary keratinocytes near the upper pole of the follicular papilla (d). TGF-b RII was not expressed in the outer root sheath (ORS), except for very rare, singly scattered cells (e). In cells of the companion layer, stronger TGF-b RII expression was found on the regions of the cells adjoining the ORS (e). Original magnifications: (a) · 20, (b, d) · 40, (c, e) · 100. CL, companion layer; He, Henle’s layer; IRS, inner root sheath; HS, hair shaft; MED, medulla; DP, follicular dermal papilla. processes via signalling through three distinct and functionally RII cannot generate responses independently of RI, and so sig- different TGF-b receptors,7 although these can be coexpressed nalling is only activated when RI and RII form a heteromeric in many cell types. Both TGF-b type 1 receptor (RI) complex.8,9 (65–70 kDa) and type 2 receptor (RII) (85–110 kDa) are Both TGF-b1 and TGF-b2 are implicated in regulating transmembrane serine/threonine kinases.7 TGF-b initially murine and human catagen via inhibition of keratinocyte differ- binds to RII, which has constitutive kinase activity, and only entiation10 and induction of keratinocyte apoptosis.6 Data thereafter binds RI, resulting in TGF-b RI phosphorylation by from knockout mouse models have indicated that TGF-b1 and TGF-b RII. Thus, TGF-b RI cannot bind free TGF-b molecules, but TGF-b2 are important for regulating HF number, whereas rather to one that is already bound to RII. Thus, it is suggested that TGF-b3-null mice show no apparent HF defects.10 No data are
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp161–164 TGF-b RII expression in the hair follicle, H.M. Sowden et al. 163 yet available from TGF-b receptor-null mice regarding HF conjugated secondary antibody diluted in 1% normal donkey abnormalities. Further support for the role of TGF-b in catagen serum for 1 h. For detection of the second primary antibody induction in the human HF may be provided by the demon- (TGF-b RII), the sections were further blocked with 10% nor- stration of the relevant TGF-b receptors in anagen-specific mal donkey serum for 90 min, before incubation with the components of the HF, as these components need to be tar- second primary antibodies for 90 min. The sections were then geted/remodelled to facilitate HF regression. To this end, we incubated with tetramethylrhodamine isothiocyanate (TRITC)- conducted an immunohistochemical and double immunofluo- conjugated secondary antibody for 1 h, washed, and mounted rescence analysis of TGF-b RI and RII expression in human in Vectashield mounting medium with 4¢,6-diamidino-2- anagen HF. phenylindole (Vector Laboratories, Peterborough, U.K.). Non- immune serum (1% normal donkey serum) in place of the Materials and methods primary antibody served as a negative control. The sections were viewed with a Nikon Eclipse 80i fluorescence micro- Normal human scalp tissue (postauricular) was obtained from scope, and photodocumented using a Nikon DS digital camera 10 healthy individuals after elective plastic surgery with Local and the ACT-2U graphics program. The images produced with Ethics Committee approval. Cryosections of 7 lm were cut the two fluorochromes, TRITC (red) and FITC (green), were TM and collected on to poly-L-lysine-coated slides, air-dried at merged using the Paint Shop Pro 7 graphics program (Jasc room temperature for 1 h, and fixed in ice-cold acetone. The Software, Minneapolis, MN, U.S.A.). Colocalization of TGF-b sections were blocked with 10% normal goat serum for RI and TGF-b RII was indicated by the production of a 90 min, and incubated for 90 min with primary antibodies to yellow/orange colour. TGF-b RI and TGF-b RII (Santa Cruz Biotechnology, Santa Cruz, CA, U.S.A.) at predetermined optimal dilutions. The sec- Results tions were washed and incubated with a biotinylated anti- rabbit/antimouse conjugated secondary link antibody (Dako, Transforming growth factor-b receptor II is expressed Carpinteria, CA, U.S.A.) before incubation with peroxidase- in the inner root sheath companion layer of anagen hair conjugated streptavidin (Dako) each for 25 min. After wash- follicles ing, aminoethylcarbazole chromagen (Dako) and substrate buffer (H2O2) were added to sections for 10 min before The expression of TGF-b RII was strongly and specifically washing in distilled H2O and mounting. The primary antibody detected in the companion layer of the IRS in anagen HFs in was omitted in negative control sections, instead being all 10 individuals examined (Fig. 1a–e). Companion layer ker- replaced with nonimmune 1% goat serum. Slides were exam- atinocytes appeared to exhibit a polarity of TGF-b RII expres- ined by brightfield microscopy and photodocumented using a sion, whereby greatest peptide expression was detected on the Nikon DS digital camera and the ACT-2U graphics program surface adjacent to the ORS (Fig. 1b). Peptide expression was (Nikon, Badhoevedorp, the Netherlands). broadly cytoplasmic, where it assumed a granular pattern For double immunofluorescence the tissue was treated as (Fig. 1c,e). Weak TGF-b RII immunoreactivity was also detec- above and blocked with 10% normal donkey serum in phos- ted in presumptive medullary keratinocytes located above the phate-buffered saline for 90 min. The first primary antibody upper pole of the follicular dermal papilla (Fig. 1d). TGF-b (TGF-b RI) was diluted in 1% normal donkey serum and RII expression was detected early in the differentiation of incubated with the sections for 18 h at 4 C. Sections were keratinocytes that later become keratinocytes of the companion washed and incubated with fluorescein isothiocyanate (FITC)- layer. These TGF-b RII-positive keratinocytes were located just
Fig 2. Transforming growth factor (TGF)-b receptor I (RI) and receptor II (RII) are colocalized in the companion layer of the anagen follicle. Colocalization of TGF-b RI and RII was seen in the companion layer (CL) of the proximal (a) and distal anagen hair follicle (b). Original magnification: (a, b) · 40. PC, precortex; ORS, outer root sheath.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp161–164 164 TGF-b RII expression in the hair follicle, H.M. Sowden et al. above the level of the Auber’s critical level or level of maximal integrity that occur during catagen regression. Our observation proliferation (Fig. 1d). Interestingly, very rarely singly scat- that both TGF-b RI and TGF-b RII are coexpressed only in the tered keratinocytes located in the mid-ORS also expressed this companion layer of the IRS provides for apoptosis-related receptor (Fig. 1e). TGF-b RI immunoreactivity was detected signal transduction activity via receptor complex formation at throughout the ORS and precortex of the anagen HF. this site.4–6
Coexpression of transforming growth factor-b References receptor I and II 1 Stenn KS, Paus R. Controls of hair follicle cycling. Physiol Rev 2001; Colocalization of TGF-b RI and RII was seen in the companion 81:449–94. layer of the proximal anagen HF (Fig. 2a). Both receptors 2 Rothnagel JA, Roop DR. Hair follicle companion layer: reacquaint- ing an old friend. J Invest Dermatol 1995; 104:S42–3. were also colocalized, although with reduced expression 3 Orwin DF. Cell differentiation in the lower outer sheath of the levels, in the mid-shaft portions of the companion layer Romney wool follicle: a companion cell layer. Aust J Biol Sci 1971; (Fig. 2b). 24:989–99. 4 Foitzik K, Paus R, Doetschman T et al. The TGF-beta2 isoform is both a required and sufficient inducer of murine hair follicle mor- Discussion phogenesis. Dev Biol 1999; 212:278–89. TGF-bs are potent growth inhibitors for most cell types, but 5 Soma T, Ogo M, Suzuki J et al. Analysis of apoptotic cell death in human hair follicles in vivo and in vitro. J Invest Dermatol 1998; also regulate cell differentiation, migration, extracellular 8,9 111:948–54. matrix production and modulate immune function. Hair 6 Foitzik K, Lindner G, Mueller-Roever S et al. Control of murine hair growth in mice is associated with the sequential expression of follicle regression (catagen) by TGF-beta1 in vivo. FASEB J 2000; TGF-b and its receptors during the different phases of the 14:752–60. cycle.4–6 We report here the preferential expression of TGF-b 7 Okunieff P, Cornelison T, Mester M et al. Mechanism and modifica- RII in the companion layer in the human anagen scalp HF. tion of gastrointestinal soft tissue response to radiation: role of This adds to a short, although expanding, list of markers growth factors. Int J Radiat Oncol Biol Phys 2005; 62:273–8. 8 Massague J. TGF-beta signal transduction. Annu Rev Biochem 1998; demarcating this highly specialized cell layer in the growing 67:753–91. HF, which now includes microtubule-associated protein type 9 Wrana JL, Attisano L, Wieser R et al. Mechanism of activation of 11 12 13 2, calretinin, plasminogen activator inhibitor type 2, the TGF-beta receptor. Nature 1994; 370:341–7. type 2 epithelial keratin K6hf14 and desmogleins 1/3.15 The 10 Paus R, Foitzik K, Welker P et al. Transforming growth factor-beta precise localization of both TGF-b RI and RII at this location receptor type I and type II expression during murine hair follicle is important as TGF-b has been implicated as a growth factor development and cycling. J Invest Dermatol 1997; 109:518–26. involved in the precipitation of the anagen follicle into apop- 11 Hallman JR, Fang D, Setaluri V, White WL. Microtubule associated protein (MAP-2) expression defines the companion layer of the tosis-driven regression (catagen).4,6 While in these studies anagen hair follicle and an analogous zone in the nail unit. J Cutan TGF-b RII expression was reported in the ORS of murine HFs, Pathol 2002; 29:549–56. the low magnification used in these studies precludes com- 12 Poblet E, Jimenez F, de Cabo C et al. The calcium-binding protein ment on selective additional expression in the companion calretinin is a marker of the companion cell layer of the human layer. Moreover, these murine studies may reflect species-spe- hair follicle. Br J Dermatol 2005; 152:1316–20. cific expression profiles. TGF-b1 is observed in the murine 13 Jensen PJ, Yang T, Yu DW et al. Serpins in the human hair follicle. hair cuticle and connective tissue sheath, while TGF-b2is J Invest Dermatol 2000; 114:917–22. 14 Wang Z, Wong P, Langbein L et al. Type II epithelial keratin 6hf expressed in the outer layer of the ORS. TGF-b3 is found in (K6hf) is expressed in the companion layer, matrix, and medulla precortical hair matrix keratinocytes. Soma et al. have shown in anagen-stage hair follicles. J Invest Dermatol 2003; 121:1276–82. that TGF-b2 can induce apoptosis in keratinocytes of the com- 15 Hanakawa Y, Li H, Lin C et al. Desmogleins 1 and 3 in the com- panion layer in the ex vivo cultured human HF.5 Furthermore, panion layer anchor mouse anagen hair to the follicle. J Invest Peters et al. have recently demonstrated a tyrosine kinase Dermatol 2004; 123:817–22. B-mediated upregulation of TGF-b2 during the anagen to 16 Peters EMJ, Hansen MG, Overall RW et al. Control of human hair catagen switch.16 Thus, it is likely that induction of apoptosis growth by neurotrophins: brain-derived neurotrophic factor inhib- its hair shaft elongation, induces catagen, and stimulates follicular in a component of the growing HF that interfaces the non- transforming growth factor b2 expression. J Invest Dermatol 2005; terminally differentiated ORS with the terminally differentiated 124:675–85. IRS could contribute to the significant alterations of cellular
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp161–164 CONCISE COMMUNICATION DOI 10.1111/j.1365-2133.2007.07993.x Associations of promoter region polymorphisms in the tumour necrosis factor-a gene and early-onset psoriasis vulgaris in a northern Polish population B. Nedoszytko, A. Szczerkowska-Dobosz, M. Zabłotna, J. Glen´,K.Re˛bała* and J. Roszkiewicz* Department of Dermatology, Venereology and Allergology and *Institute of Forensic Medicine, Medical University of Gdansk, De˛binki str. 7, 80-211 Gdan´sk, Poland
Summary
Correspondence Background Tumour necrosis factor (TNF)-a is considered to be an important Bogusław Nedoszytko. mediator in the pathogenesis of psoriasis. Increased levels and activity of this E-mail: [email protected] cytokine have been observed in blood and skin of patients with psoriasis. As certain allelic variants of the TNF-a gene are associated with increased or Accepted for publication 2 March 2007 decreased production of TNF-a, the disturbed cytokine balance may be under genetic control. Key words Objectives To investigate the potential association of TNF-a promoter alleles amplification refractory mutation system– within subtypes of psoriasis compared with healthy controls in a northern Polish polymerase chain reaction, early-onset psoriasis, population. promoter gene polymorphism, tumour necrosis Methods We analysed 166 patients with psoriasis vulgaris (134 with type I factor-a and 32 with type II) and 65 healthy controls. The polymorphisms –238G ⁄A Conflicts of interest and –308G ⁄A in the promoter region of the TNF-a gene were typed None declared. using the amplification refractory mutation system–polymerase chain reaction method. Results We found that the TNF-a –308A allele frequency was significantly decreased among patients with early-onset psoriasis in comparison with control subjects (7Æ5% vs. 15Æ4%, P =0Æ022), whereas in the same patients the fre- quency of the TNF-a –238A allele was significantly increased as compared with the controls (16Æ8% vs. 3Æ1%, P =0Æ000017, odds ratio 8Æ79, 95% confidence interval 2Æ606–29Æ678). Patients with early-onset psoriasis with –238 genotype GA or AA were found more often among those with age at onset < 25 years in comparison with those with genotype GG (31Æ7% vs. 9Æ1%, P =0Æ0312). We also found that the mean ± SD age at onset among –238A carriers was signifi- cantly lower in comparison with that associated with the –238GG genotype (13Æ5±7Æ4 vs. 19Æ2±9Æ9 years, P =0Æ0132). Conclusions Our study confirming the association between –238 G ⁄A TNF-a pro- moter polymorphism and early-onset psoriasis vulgaris in the northern Polish population suggests that the –238A variant may contribute not only to a pre- disposition to psoriasis vulgaris but also to the disease phenotype.
Exaggerated production of tumour necrosis factor (TNF)-a Caucasians consist of G to A transitions in the promoter region may play an important role in the pathogenesis of psoriasis.1,2 at positions –238 and –308, although there is considerable Increased levels of this proinflammatory cytokine have been diversity in the distribution among different populations with detected in blood and psoriatic lesions of patients with psoria- psoriasis.1 Moreover, variants of TNF-a polymorphisms may sis vulgaris, whereas anti-TNF-a therapy has produced dra- be associated with specific psoriasis subgroups defined by matic improvements.2 As certain allelic variants of the TNF-a early and late onset of the disease.1 In our study we compared gene are associated with increased or decreased production of the frequency of TNF-a –238 and –308 promoter polymor- TNF-a, the disturbed cytokine balance may be under genetic phisms in patients with psoriasis vulgaris and in healthy con- control. Commonly described variants of polymorphisms in trols in a northern Polish population.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp165–167 165 166 TNF-a promoter polymorphisms and psoriasis vulgaris, Nedoszytko et al.
Materials and methods Table 1 Frequency of tumour necrosis factor (TNF)-a –238G ⁄A and –308G ⁄A genotypes in patients with psoriasis and controls Patients and controls Early-onset Late-onset Psoriasis One hundred and sixty-six unrelated patients with psoriasis psoriasis psoriasis vulgaris, (84 men and 82 women) of mean age 37 years (range 22– Controls vulgaris vulgaris total 67) were included in this study. Each patient was evaluated (n =65), (n = 134), (n = 32), (n = 166), according to the standard protocol including a complete his- n (%) n (%) n (%) n (%) tory and physical examination. All patients had the classical TNF-a –238 genotype pattern of skin lesions (chronic plaque lesions, psoriasis vul- GG 62 (95Æ4) 94 (70Æ1) 31 (96Æ9) 125 (75Æ3) garis), confirmed by a dermatologist. They were divided into GA 2 (3Æ1) 35 (26Æ1) 1 (3Æ1) 36 (21Æ7) AA 1 (1Æ5) 5 (3Æ7) 0 5 (3Æ0) two subgroups: those with age at onset of psoriasis before the Allele age of 40 years (type I, n = 134) and those with the age at G 126 (96Æ9) 223 (83Æ2) 63 (98Æ4) 286 (86Æ1) onset above 40 years (type II, n = 32). Patients with psoriatic A4(3Æ1) 45 (16Æ8)** 1 (1Æ6) 46 (13Æ9) arthritis were excluded from the study. TNF-a –308 genotype Controls (n = 65) were unrelated healthy individuals, GG 46 (70Æ8) 114 (85Æ1) 27 (84Æ4) 141 (84Æ9) including mostly university healthcare personnel. GA 18 (27Æ7) 20 (14Æ9) 4 (12Æ5) 24 (14Æ4) AA 1 (1Æ5) 0 1 (3Æ1) 1 (0Æ7) Allele Tumour necrosis factor-a genotyping by amplification G 110 (84Æ6) 248 (92Æ5) 58 (90Æ6) 306 (92Æ2) refractory mutation system–polymerase chain reaction A 20 (15Æ4) 20 (7Æ5)* 6 (9Æ4) 26 (7Æ8)
From each patient genomic DNA was extracted from periph- Fisher’s exact test: *P =0Æ02222; **P =0Æ000017. eral blood leucocytes using Blood DNA Prep Plus (A&A Biotechnology, Gdynia, Poland). To assess TNF-a gene poly- morphism, we used amplification refractory mutation system– polymerase chain reaction according to the method described P =0Æ022), whereas in the same patients the frequency of the by Perrey et al.3 TNF-a –238A allele was significantly increased as compared The study was approved by the Local Research Ethics Com- with the controls (16Æ8% vs. 3Æ1%, P =0Æ000017, OR 8Æ79, mittee of the Medical University of Gdansk. 95% CI 2Æ606–29Æ678). Restriction of the analyses to patients with early-onset psor- iasis vulgaris showed that patients with –238 genotype GA or Statistical analysis AA were found more often among those with age at onset Observed and expected genotype frequency were compared by < 25 years in comparison with those with genotype GG a Monte Carlo goodness-of-fit test in patients with psoriasis (31Æ7% vs. 9Æ1%, P =0Æ0312). We also found that the vulgaris and in the control population, and significance was mean ± SD age at onset among –238A carriers (GA + AA evaluated by the v2 test with Yates’ correction or a two-tailed genotype) was significantly lower in comparison with that Fisher’s exact test. Odds ratios (ORs) and exact 95% confi- associated with the –238GG genotype (13Æ5±7Æ4 vs. dence intervals (CIs) were calculated to compare genotype fre- 19Æ2±9Æ9 years, P =0Æ0132). quencies, and the Mann–Whitney test was used to test for significance of differences in data distribution. Discussion
Results Although a psoriasis susceptibility gene(s) has not been yet identified, several candidate genes have been studied, with TNF-a genotype frequencies in the investigated groups are evidence for a major locus being located within the major shown in Table 1. Analysis of TNF-a –308 promoter poly- histocompatibility complex (PSORS1). The Cw6 allele is the morphism revealed that the GG genotype was the most com- most extensively investigated candidate gene. However, the mon in the control population (70Æ8%), while genotype GA increased frequency of the Cw6 allele in patients with psoria- was observed in 27Æ7% and genotype AA in 1Æ5% of controls. sis may be due to linkage with other genes of PSORS1. Various Regarding TNF-a –238 promoter polymorphism, the genotype studies in Caucasian populations with psoriasis have shown GG was found in 95Æ4% of controls, whereas genotypes GA that several polymorphisms in the TNF-a gene may predispose and AA were detected in 4Æ6% of subjects. The genotypes to the development of psoriasis. were in Hardy–Weinberg equilibrium, and did not differ from Our observation that carriage of the polymorphism variant those in other Caucasian populations.4–6 TNFa –308A is significantly less common among patients The TNF-a –308A allele frequency was significantly with early-onset psoriasis than among healthy controls is in decreased among patients with early-onset psoriasis vulgaris agreement with the reports of Mo˘ssner et al.6 and Craven in comparison with control subjects (7Æ5% vs. 15Æ4%, et al.,7 and differs from other studies that have failed to find
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp165–167 TNF-a promoter polymorphisms and psoriasis vulgaris, Nedoszytko et al. 167 a specific association of psoriasis vulgaris with TNF-a 2 Bonifati C, Ameglio F. Cytokines in psoriasis. Int J Dermatol 1999; –308G ⁄A.5,8,9 38:241–51. We also found a strong association between psoriasis with 3 Perrey C, Pravica V, Sinnott PJ et al. Genotyping for polymorphisms in interferon-gamma, interleukin-10, transforming growth factor- early onset and polymorphism at the –238 locus, confirming beta 1 and tumour necrosis factor-alpha genes: a technical report. previous reports of the –238A TNF-a gene polymorphism Transplant Immunol 1998; 6:193–7. 4,5,8,10 conferring a higher risk of early-onset psoriasis. This 4 Hohler T, Kruger A, Schneider PM et al. A TNF-alpha promoter 11,12 association was not detected in either Japanese or Korean polymorphism is associated with juvenile onset psoriasis and psori- populations,9 probably reflecting population-dependent diver- atic arthritis. J Invest Dermatol 1997; 109:562–5. sity in TNF-a gene polymorphisms. 5 Reich K, Westphal G, Schulz T et al. Combined analysis of poly- Most importantly, we found that patients with type I psor- morphisms of the tumor necrosis factor-a and interleukin-10 pro- motor regions and polymorphic xenobiotic metabolizing enzymes iasis with the G ⁄A and A ⁄A genotypes of the –238 TNF-a in psoriasis. J Invest Dermatol 1999; 113:214–20. promoter have a significantly lower age at onset of the disease 6Mo˘ssner R, Kingo K, Kleensang A et al. Association of TNF –238 compared with patients with the G ⁄G genotype, suggesting and –308 promoter polymorphisms with psoriasis vulgaris and that genetic screening might predict disease susceptibility. This psoriatic arthitis but not with pustulosis palmoplantaris. J Invest novel observation has not been previously described. Dermatol 2005; 124:282–4. In conclusion, our study confirms the association between 7 Craven N, Jackson C, Kirby B et al. Cytokine gene polymorphism in –238G ⁄A TNF-a promoter polymorphism and early-onset psoriasis. Br J Dermatol 2001; 144:849–53. 8 Long F, Sun C, Deng D et al. TNF–238A is associated with juvenile psoriasis vulgaris in the northern Polish population and sug- onset psoriasis in patients of Han population in southwest China. gests that the –238A variant may contribute not only to a J Dermatol Sci 2004; 36:109–11. predisposition to psoriasis vulgaris but also to the disease 9 Kim TG, Pyo CW, Hur SS et al. Polymorphisms of tumor necrosis phenotype. factor (TNF) alpha and beta genes in Korean patients with psoria- sis. Arch Dermatol Res 2003; 295:8–13. 10 Arias AI, Giles B, Eiermann TH et al. Tumor necrosis factor- Acknowledgments alpha gene polymorphism in psoriasis. Exp Clin Immunogenet 1997; 14:118–22. This work was supported by grant 3PO5A 12022 of the State 11 Nishibu A, Oyama N, Nakamura K et al. Lack of association of Committee for Scientific Research of Poland. TNF–238A and –308A in Japanese patients with psoriasis vulgaris, psoriatic arthritis and generalized pustular psoriasis. J Dermatol Sci References 2002; 29:181–4. 12 Tsunemi Y, Nishibu A, Saeki H et al. Lack of association between 1 Reich K, Mo˘ssner R, Ko˘nig IR et al. Promoter polymorphisms of the promoter polymorphisms at positions –308 and –238 of the the genes encoding tumor necrosis factor-alpha and interleukin-1 tumor necrosis factor alpha gene and psoriasis vulgaris in Japanese beta are associated with different subtypes of psoriasis character- patients. Dermatology 2003; 207:371–4. ized by early and late onset. J Invest Dermatol 2002; 118:155–63.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp165–167 CASE REPORT DOI 10.1111/j.1365-2133.2007.07920.x Acquired palmoplantar keratoderma and immunobullous disease associated with antibodies to desmocollin 3 M.C. Bolling, J.R. Mekkes,* W.F.M. Goldschmidt,* C.J.M. van Noesel, M.F. Jonkman and H.H. Pas Centre for Blistering Diseases, Department of Dermatology, University Medical Centre Groningen, University of Groningen, PO. Box 30.001, 9700 RB Groningen, the Netherlands Departments of *Dermatology and Pathology, Academic Medical Centre Amsterdam, the Netherlands
Summary
Correspondence We present a case of immunobullous disease with an impressive acquired palmo- H.H. Pas. plantar keratoderma (PPK) and unique antigenicity. The palms of the patient E-mail: [email protected] showed hyperkeratotic ridges with a tripe pattern that decreased with the amelior- ation of the immunobullous condition. The histopathology of perilesional skin Accepted for publication 20 November 2006 (blister) demonstrated eosinophilic spongiosis and suprabasal blistering as in pemphigus vulgaris. In palmar skin, acantholysis, intraepidermal pustules, papillo- Key words matosis and marked hyperkeratosis were observed. Direct and indirect immuno- desmocollin, hyperkeratosis, keratoderma, fluorescence displayed intraepidermal intercellular IgG staining as well as linear palmoplantar, pemphigus IgG staining along the epidermal basement membrane zone. Immunochemical Conflicts of interest assays revealed IgG antibodies to the desmosomal protein desmocollin 3 and to None declared. the hemidesmosomal proteins BP230 and LAD-1. Affinity-purified antidesmocollin 3 serum IgG bound to monkey oesophagus in the typical pemphigus pattern. Desmocollins are transmembrane proteins of the desmosome. Desmosome dis- eases may cause hereditary PPK. In our patient with acquired PPK, we hypothesize that the antibodies to desmocollin 3 were, apart from their role in eliciting the pemphigus-like blistering disease, also implicated in the pathogenesis of the PPK.
Palmoplantar keratoderma (PPK), also known as palmoplantar epidermal differentiation and morphogenesis.12–16 The desmo- hyperkeratosis or hyperkeratosis palmaris et plantaris, is a dis- collins exist in three different isoforms, Dsc1, 2 and 3, that order characterized by marked thickening of the skin of the demonstrate differentiation-specific expression through the palms and soles. A clinical division is made between different epidermal layers. Dsc1 is found mainly in the upper spinous PPKs by the pattern of the hyperkeratosis: diffuse, focal and and granular layers, Dsc2 in the basal layer, and Dsc3 punctate. PPK can occur as an inherited or as an acquired dis- throughout the whole epidermis, but mainly in the basal and order. The causes of acquired PPK are diverse. It can be drug- spinous layers.17–23 related, infective, reactive, inflammatory, paraneoplastic, Here we describe a patient with an autoimmune bullous climacteric and associated with systemic disease. Inherited PPK disease, who presented with an exacerbation of blistering and is associated with mutations in a selective group of genes, all subsequently developed an impressive diffuse PPK. Besides the affecting the scaffolding network of the keratinocyte.1–3 Muta- hemidesmosomal antigens, the desmosomal cadherin Dsc3 tions in several keratin genes lead to a variety of skin diseases was also the target of IgG autoantibodies. We hypothesize that characterized by fragility and/or overgrowth (hyperkeratosis) antibodies to Dsc3 induced the PPK in this patient. of epithelial tissues, including palmoplantar skin. Examples are epidermolysis bullosa simplex type Dowling–Meara, epider- Case report molytic hyperkeratosis of palms and soles types 1–3, and pach- yonychia congenita. Further genes in which mutations may In 2004 a 63-year-old man presented with severe PPK and cause PPK code for proteins of the desmosomes, plakophilin cutaneous blisters on the trunk and extremities. His medical 1,4,5 plakoglobin,6 desmoplakin7,8 and desmoglein 1.9–11 history included urolithiasis, hypertension for which he took Direct adhesion between neighbouring keratinocytes is medi- metoprolol, and a Billroth II gastrectomy for gastric ulcers. He ated through the transmembrane desmogleins and desmocol- did not smoke and used no alcohol. His family history lins of the desmosome, that together form the desmosomal revealed neither hereditary skin diseases nor malignancies. cadherin family. Cadherins are also thought to be involved in In 2002 the patient had a blistering episode. Tense bullae
2007 The Authors 168 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 Acquired PPK and antibodies to desmocollin 3, M.C. Bolling et al. 169 together with annular erythematous plaques spread all over inflammatory hyperpigmentation on the arms and legs. Both the body and a burning, itching feeling was present in the flaccid and more tense bullae on erythematous skin were pre- palms and soles. No mucosal involvement was observed. The sent on the left lower leg and feet (Fig. 1a). Nikolsky’s sign clinical diagnosis was bullous pemphigoid (BP). However, on was positive. The most remarkable clinical sign observed histological examination of the perilesional skin of the left though was a transgrediens distribution of diffuse PPK of lower arm, acantholysis as well as what seemed subepidermal feet (Fig. 1b) and hands (Fig. 1c) with a ridged surface pat- blistering was observed. Direct immunofluorescence (DIF) tern resembling tripe. No palmar or plantar erythema was pre- showed linear IgG and C3 deposition along the epidermal sent. The nails showed subungual hyperkeratosis. Mucous basement membrane as well as IgG in a honeycomb pattern membranes of the eyes, mouth and genitals had a normal along the keratinocyte surface in the lower part of the epider- appearance as well as the groins and axillae. The nose and mis. This combined pattern can be seen in paraneoplastic helices of the ears showed no hyperkeratosis or scaling as is pemphigus (PNP).24 However, the patient did not have sto- often seen in acrokeratosis paraneoplastica (Bazex syndrome). matitis, and no histological features were present that could Histological examination of a perilesional skin biopsy of the sustain the diagnosis of PNP. For that reason the working left lower leg revealed eosinophilic spongiosis and cleavage diagnosis remained BP. Topical steroids and oral prednisolone near the subepidermal layer (Fig. 2a). Pan-keratin staining were started and this resulted in complete remission of the (clone: AE1/AE3, DAKO, Glostrup, Denmark) demonstrated skin lesions. After a symptom-free year, in 2004 blisters devel- that the blister level was in fact suprabasal as keratin was pre- oped all over the patient’s body again, some tense and others sent in both the roof and the floor of the blister (data not flaccid, on erythematous skin. Subsequently an extreme PPK shown). Histopathology of palmar skin showed features of with pain and itching developed. Physical examination pemphigus vegetans with acantholysis in the lower layers of revealed multiple dried up bullae, crusts, erosions and post- the epidermis, intraepidermal pustules filled with eosinophilic
(a) (b)
(c) (d)
Fig 1. (a) Intact and ruptured flaccid blisters (arrows) on erythematous skin of the ankle. Also partially visible is the plantar hyperkeratosis, (b) thick plantar hyperkeratosis with desquamation, (c) palmar hyperkeratosis with (d) tripe pattern.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 170 Acquired PPK and antibodies to desmocollin 3, M.C. Bolling et al.
(a) Fig 2. Histology. (a) Peribullous skin showing eosinophilic spongiosis and suprabasal clefting with acantholysis (*), (b) palmar skin showing papillomatosis, eosinophilic spongiosis with slight acantholysis (arrows), an intraepidermal pustule (#), and hyperkeratosis (x), (c) detail of an eosinophilic acantholytic pustule. cells, papillomatosis and marked hyperkeratosis (Fig. 2b and 2c). Only a few necrotic keratinocytes were seen. Again, as in 2002, no histological signs of PNP were present.
Laboratory diagnosis
DIF of perilesional skin showed intercellular substance (ICS) deposition of IgG in the more basal layers of the epidermis and linear C3c along the epidermal basement membrane zone (BMZ). Indirect immunofluorescence (IIF) on monkey oeso- phagus showed linear binding to IgG along the epithelial BMZ and in a pemphigus-like ICS pattern (Fig. 3a). On salt-split skin IgG (3+) and IgA (+/2+) bound the epidermal side of (b) the split. IgG and IgA enzyme-linked immunoabsorbent assays (ELISA) to desmogleins 1 and 3 were negative. Immunoblot with normal human keratinocyte extract and conditioned medium extract demonstrated IgG binding to BP230 (Fig. 4a) and LAD-1 (Fig. 4b). Binding to BP230 was confirmed by BP control sera (data not shown). No IgA binding antigen was found. Furthermore IgG-ELISA for the NC16A domain of BP180 was positive (index value 118). An immunoblot did not demonstrate binding to the PNP-related plakin proteins periplakin and envoplakin. Instead four additional bands in the 100–110-kDa region were visible (Fig. 4a). Binding to these four bands, of which two are major, could be repro- duced with a monoclonal antibody (clone: Dsc3-U114) speci- fic to the anchor domain of human desmocollin 3 (Dsc3) (Fig. 4a). Considering the negative desmoglein ELISAs, the anti-Dsc3 antibodies were the most likely cause of the ICS staining observed in IF. To verify this we eluted the patient IgG that bound at the Dsc3 position from the blot.25 When these affinity-purified antibodies were brought on to a section (c) of monkey oesophagus they indeed stained in an ICS pattern, identical to the pattern obtained with the serum but without the linear BMZ IgG (Fig. 3b).
Further course
Initially we considered a paraneoplastic syndrome in this patient from the polymorphic clinical appearance and the find- ing of antibodies to both hemidesmosomal and desmosomal antigens as found in paraneoplastic pemphigus. Furthermore acquired PPK is often connected with an underlying neo- plasm.26–28 However no malignancy was found after extensive screening; physical examination, blood tests, computed tomography of the abdomen and chest X-ray, ultrasound scan of pelvis and abdomen, and oesophagogastroduodenoscopy including biopsies did not reveal any neoplasm; no occult blood in stool was detected and head and neck evaluation was normal. Topical steroids, oral tetracycline 500 mg three times
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 Acquired PPK and antibodies to desmocollin 3, M.C. Bolling et al. 171
(a) (a) (b)
(b)
Fig 4. Identification of IgG recognized antigens by Western blotting. (a) Immunoblotting using human keratinocyte cell extract substrate. Except to BP230, the patient’s serum bound to four additional protein bands with molecular weights around 100–110 kDa, of which two were major (lane 1). Exactly the same four-band pattern was observed with the Dsc3 specific monoclonal antibody Dsc3-U114 (Progen Biotechnik, Heidelberg, Germany) (lane 2). Biotinylated molecular marker proteins (Biotinylated SDS-PAGE Standards, BioRad Laboratories, Hercules, CA, U.S.A.) (lane 3). (b) Western blotting using spent keratinocyte culture medium extract demonstrated binding of patient IgG to the 120-kDa LAD-1 molecule (lane 4). LAD-1 specific bullous pemphigoid control serum (lane 5).
Discussion
The patient we describe here had pemphigus-like skin lesions and antibodies to Dsc3 with a concomitant PPK as the most striking feature. Biopsies demonstrated spongiosis, acantho- lysis, intraepidermal vesicles and suprabasal cleavage, which Fig 3. (a) Indirect immunofluorescence on monkey oesophagus are all characteristics of pemphigus. Since antibodies to showed patient serum IgG binding to the epithelial basement membrane zone (oblique arrow) and intraepidermally along the desmoglein 1 and/or 3 were absent, it is likely that the anti- keratinocyte surfaces (horizontal arrow) (· 20), (b) anti-Dsc3 affinity- bodies to Dsc3 initiated these pemphigus-like features. The purified IgG antibodies from patient serum bound intraepidermal sites distribution of Dsc3 being more abundant in the basal layer only along the keratinocyte surfaces (horizontal arrow) and not along and diminishing higher in the epidermis is consistent with the the basement membrane zone (· 20). Bars are 50 lm. location of the acantholysis and blistering observed in IIF and the histopathology of the patient’s lesional skin.13,15,16,20,23 It daily, and nicotinamide 500 mg three times daily were started is unlikely that the antibodies to BP230 and LAD-1 contribut- together with topical 10% salicylic acid ointment for the PPK. ed to the acantholysis and the suprabasal cleavage, as such This improved the skin well, no new blisters formed and old features have never been observed in pemphigoid. In addition, lesions healed. The PPK diminished. Medication was then the palmar skin showed papillomatosis and hyperkeratosis. gradually decreased and eventually stopped. Blistering did not The almost simultaneous appearance of the PPK with an return and the PPK was only slightly present. After being dis- exacerbation of cutaneous blistering and its marked dimi- charged the patient stayed under close follow-up because the nution on immunosuppressive therapy are features that sup- suspicion of a malignancy still existed. At the last follow-up, port an immunopathological basis for this PPK in our patient. 16 months after discharge, still no signs for malignancy were The exact pathomechanism is unclear. Clues may be found in present and the blistering and PPK had not returned. the genetic PPKs that point to disturbed desmosome signalling.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 172 Acquired PPK and antibodies to desmocollin 3, M.C. Bolling et al.
Mutations in the desmosomal proteins plakophilin 1,4,5 plako- remissions of the pemphigoid. Although in our case the PPK globin,6 desmoplakin7,8 and desmoglein 19–11 may cause PPK. did not really manifest itself in the first disease period in These proteins represent the three major families that consti- 2002, the patient at that time complained about itching and tute the desmosome: the cadherins, the armadillo proteins burning palms. Considering the acantholysis and intercellular and the plakins. Compelling evidence suggests that these pro- intraepidermal IgG antibodies it seems most likely that some teins, especially the cadherins, function not only as adhesion anti-Dsc3 IgG was already then present in addition to the molecules but also in cell signalling29. Overexpression of pemphigoid antibodies. This unique combination of anti- Dsc3 in the upper layers of the skin led to disturbed intra- bodies may have arisen through epitope spreading.40,41 In cellular signalling, involving b-catenin, and morphogenetic autoimmunity epitope spreading refers to the development of defects including thickened skin (hyperkeratosis).14 It is feas- immune responses against secondary epitopes caused by the ible that binding of antibodies to Dsc3 could interfere with autoimmune response to the primary epitope.42 On the other its cell signalling function and thereby induce disturbed hand it is equally possible that both antigen specificities differentiation. evolved simultaneously during the primary process and that The hyperkeratosis found in palmoplantar skin is thought to the clinical pattern over time merely reflects the individual be caused by the distinctive differentiation of palmar skin.30 titres of the different antibodies. However, definite proof for Unlike Dsc1, Dsc3 is also found in glandular ducts, the outer either of these statements must still be delivered. root sheath of the hair follicles and in other stratified nonkera- To conclude, we present a case of immunobullous disease tinizing epithelia, such as the vagina, tongue, oesophagus, cer- with concomitant acquired PPK and unique antigen specificity. vix and buccal mucosa.31,32 Nevertheless, our patient did not We hypothesize a pivotal role for anti-Dsc3 antibodies in have mucosal lesions. In antidesmoglein pemphigus the occur- inducing the acantholysis and intraepidermal blistering and in rence of lesions on either the mucosa or the skin is thought to the pathogenesis of the acquired PPK. depend on the compensating effect of the nonaffected desmoglein that is differently distributed in epidermal and References mucosal epithelium.33 In the patient here described it seems that in mucosae, unlike skin, the other cadherins are able to 1 Kimyai-Asadi A, Kotcher LB, Jih MH. The molecular basis of heredi- maintain appropriate keratinocyte adhesion in spite of tary palmoplantar keratodermas. J Am Acad Dermatol 2002; 47:327–43. antibodies to Dsc3. The exact properties and functions of the 2 Itin PH, Fistarol SK. Palmoplantar keratodermas. Clin Dermatol 2005; 23:15–22. different desmosomal cadherins in the different tissues remain 3 McGrath JA. Hereditary diseases of desmosomes. J Dermatol Sci to be elucidated. 1999; 20:85–91. Autoantibodies to Dsc3 are rarely reported. The few cases 4 McGrath JA, McMillan JR, Shemanko CS et al. Mutations in the described were always in combination with antidesmoglein plakophilin 1 gene result in ectodermal dysplasia/skin fragility antibodies.34–37 All these showed the features of atypical pem- syndrome. Nat Genet 1997; 17:240–4. phigus with diverse clinical pictures, but none had clinical 5 McGrath JA. A novel genodermatosis caused by mutations in plako- palmoplantar involvement. In all these cases the concomitant philin 1, a structural component of desmosomes. J Dermatol 1999; 26:764–9. presence of anti-Dsc antibodies was detected by immunoblot, 6 McKoy G, Protonotarios N, Crosby A et al. Identification of a dele- but their actual titre could not be measured by IIF because of tion in plakoglobin in arrhythmogenic right ventricular cardiomyo- the simultaneous presence of antidesmoglein antibodies. It may pathy with palmoplantar keratoderma and woolly hair (Naxos be that higher titres, as in our patient, are needed to initiate disease). Lancet 2000; 355:2119–24. the PPK. Another possibility is that the Dsc3 molecule contains 7 Armstrong DK, McKenna KE, Purkis PE et al. Haploinsufficiency of pathogenic and nonpathogenic epitopes, as recently demon- desmoplakin causes a striate subtype of palmoplantar keratoderma. strated for desmoglein38 and that the PPK-inducing effect is Hum Mol Genet 1999; 8:143–8. 8 Whittock NV, Ashton GH, Dopping-Hepenstal PJ et al. Striate connected with antibody binding to a particular epitope. palmoplantar keratoderma resulting from desmoplakin haplo- We had two reasons to suspect a paraneoplastic syndrome. insufficiency. J Invest Dermatol 1999; 113:940–6. Firstly, acquired diffuse PPK is often associated with an under- 9 Rickman L, Simrak D, Stevens HP et al. N-terminal deletion in a lying neoplasm26–28 (90% of cases27). Secondly, the combined desmosomal cadherin causes the autosomal dominant skin disease BMZ/ICS staining in immunofluorescence is one of the char- striate palmoplantar keratoderma. Hum Mol Genet 1999; 8:971–6. acteristics of PNP. As we did not find any malignancy and, 10 Keren H, Bergman R, Mizrachi M et al. Diffuse nonepidermolytic most importantly, the patient responded well to the medica- palmoplantar keratoderma caused by a recurrent nonsense muta- tion in DSG1. Arch Dermatol 2005; 141:625–8. tion in 2002 as well as in 2004 we feel that a paraneoplastic 11 Whittock NV, Bower C. Targetting of desmoglein 1 in inherited syndrome in this patient is very unlikely. and acquired skin diseases. Clin Exp Dermatol 2003; 28:410–15. ‘Tripe palms’ associated with immunobullous disease have 12 Chidgey M, Brakebusch C, Gustafsson E et al. Mice lacking desmo- been reported before by Razack et al. in 1987 in a case of collin 1 show epidermal fragility accompanied by barrier defects acquired PPK associated with pemphigoid.39 However, in this and abnormal differentiation. J Cell Biol 2001; 155:821–32. older report no immunofluorescence or immunoblot was per- 13 Chidgey MA, Yue KK, Gould S et al. Changing pattern of desmo- formed to investigate the characteristics of the autoantibodies. collin 3 expression accompanies epidermal organisation during skin development. Dev Dyn 1997; 210:315–27. The PPK, as in our case, coincided with the exacerbations and
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 Acquired PPK and antibodies to desmocollin 3, M.C. Bolling et al. 173
14 Hardman MJ, Liu K, Avilion AA et al. Desmosomal cadherin mis- 30 Wan H, Dopping-Hepenstal PJ, Gratian MJ et al. Desmosomes exhi- expression alters beta-catenin stability and epidermal differenti- bit site-specific features in human palm skin. Exp Dermatol 2003; ation. Mol Cell Biol 2005; 25:969–78. 12:378–88. 15 North AJ, Chidgey MA, Clarke JP et al. Distinct desmocollin iso- 31 Nuber UA, Schafer S, Stehr S et al. Patterns of desmocollin synthesis forms occur in the same desmosomes and show reciprocally grad- in human epithelia: immunolocalization of desmocollins 1 and 3 in ed distributions in bovine nasal epidermis. Proc Natl Acad Sci U S A special epithelia and in cultured cells. Eur J Cell Biol 1996; 71:1–13. 1996; 93:7701–5. 32 King IA, Sullivan KH, Bennett R Jr et al. The desmocollins of 16 Runswick SK, O’Hare MJ, Jones L et al. Desmosomal adhesion regu- human foreskin epidermis: identification and chromosomal assign- lates epithelial morphogenesis and cell positioning. Nat Cell Biol ment of a third gene and expression patterns of the three isoforms. 2001; 3:823–30. J Invest Dermatol 1995; 105:314–21. 17 Buxton RS, Magee AI. Structure and interactions of desmosomal 33 Amagai M, Tsunoda K, Zillikens D et al. The clinical phenotype of and other cadherins. Semin Cell Biol 1992; 3:157–67. pemphigus is defined by the anti-desmoglein autoantibody profile. 18 Buxton RS, Cowin P, Franke WW et al. Nomenclature of the des- J Am Acad Dermatol 1999; 40:167–70. mosomal cadherins. J Cell Biol 1993; 121:481–3. 34 Gooptu C, Mendelsohn S, Amagai M et al. Unique immunobullous 19 Collins JE, Legan PK, Kenny TP et al. Cloning and sequence analysis disease in a child with a predominantly IgA response to three of desmosomal glycoproteins 2 and 3 (desmocollins): cadherin- desmosomal proteins. Br J Dermatol 1999; 141:882–6. like desmosomal adhesion molecules with heterogeneous cytoplas- 35 Hisamatsu Y, Amagai M, Garrod DR et al. The detection of IgG and mic domains. J Cell Biol 1991; 113:381–91. IgA autoantibodies to desmocollins 1–3 by enzyme-linked immuno- 20 King IA, Tabiowo A, Purkis P et al. Expression of distinct desmocollin sorbent assays using baculovirus-expressed proteins, in atypical isoforms in human epidermis. J Invest Dermatol 1993; 100:373–9. pemphigus but not in typical pemphigus. Br J Dermatol 2004; 21 Koch PJ, Franke WW. Desmosomal cadherins: another growing 151:73–83. multigene family of adhesion molecules. Curr Opin Cell Biol 1994; 36 Kozlowska A, Hashimoto T, Jarzabek-Chorzelska M et al. Pemphigus 6:682–7. herpetiformis with IgA and IgG antibodies to desmoglein 1 and IgG 22 Parker AE, Wheeler GN, Arnemann J et al. Desmosomal glycopro- antibodies to desmocollin 3. J Am Acad Dermatol 2003; 48:117–22. teins II and III. Cadherin-like junctional molecules generated by 37 Preisz K, Horvath A, Sardy M et al. Exacerbation of paraneoplastic alternative splicing. J Biol Chem 1991; 266:10438–45. pemphigus by cyclophosphamide treatment: detection of novel 23 Yue KK, Holton JL, Clarke JP et al. Characterisation of a desmocollin autoantigens and bronchial autoantibodies. Br J Dermatol 2004; isoform (bovine DSC3) exclusively expressed in lower layers of 150:1018–24. stratified epithelia. J Cell Sci 1995; 108:2163–73. 38 Ishii K, Harada R, Matsuo I et al. In vitro keratinocyte dissociation 24 Camisa C, Helm TN. Paraneoplastic pemphigus is a distinct neoplasia- assay for evaluation of the pathogenicity of anti-desmoglein 3 IgG induced autoimmune disease. Arch Dermatol 1993; 129:883–6. autoantibodies in pemphigus vulgaris. J Invest Dermatol 2005; 25 Pas HH, Kloosterhuis GJ, Heeres K et al. Bullous pemphigoid and 124:939–46. linear IgA dermatosis sera recognize a similar 120-kDa keratinocyte 39 Razack EM, Premalatha S, Rao NR, Zahra A. Acanthosis palmaris in a collagenous glycoprotein with antigenic cross-reactivity to BP180. patient with bullous pemphigoid. J Am Acad Dermatol 1987; 16:217– J Invest Dermatol 1997; 108:423–9. 19. 26 Breathnach SM, Wells GC. Acanthosis palmaris: tripe palms. A dis- 40 Chan LS, Vanderlugt CJ, Hashimoto T et al. Epitope spreading: tinctive pattern of palmar keratoderma frequently associated with lessons from autoimmune skin diseases. J Invest Dermatol 1998; internal malignancy. Clin Exp Dermatol 1980; 5:181–9. 110:103–9. 27 Cohen PR, Grossman ME, Silvers DN et al. Tripe palms and cancer. 41 Salato VK, Hacker-Foegen MK, Lazarova Z et al. Role of intramolec- Clin Dermatol 1993; 11:165–73. ular epitope spreading in pemphigus vulgaris. Clin Immunol 2005; 28 Mullans EA, Cohen PR. Tripe palms: a cutaneous paraneoplastic 116:54–64. syndrome. South Med J 1996; 89:626–7. 42 Powell AM, Black MM. Epitope spreading: protection from patho- 29 Getsios S, Huen AC, Green KJ. Working out the strength and flexi- gens, but propagation of autoimmunity? Clin Exp Dermatol 2001; bility of desmosomes. Nat Rev Mol Cell Biol 2004; 5:271–81. 26:427–33.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp168–173 CASE REPORT DOI 10.1111/j.1365-2133.2007.07968.x Bartonella-related pseudomembranous angiomatous papillomatosis of the oral cavity associated with allogeneic bone marrow transplantation and oral graft-versus-host disease C. Vassallo, M. Ardigo`, V. Brazzelli, M. Zecca,* F. Locatelli,* P.E. Alessandrino, M. Lazzarino, S. Corona, P. Lanzerini, M. Benazzo,§ M. Fabbi,– and G. Borroni Department of Dermatology, *Oncoematologia Pediatrica, Centro Trapianti di Midollo Osseo, Istituto di Ematologia, Laboratory of Electron Microscopy, Department of Infectious Diseases, and §Department of Otolaryngology, Head and Neck Surgery, University of Pavia, Fondazione IRCCS Policlinico San Matteo, Piazzale Golgi 2, 27100 Pavia, Italy –IZSLER Sezione Diagnostica, Pavia, Italy
Summary
Correspondence Patients undergoing allogeneic stem cell transplantation are at high risk for infec- Camilla Vassallo. tion with a variety of pathogens during different phases of the procedure. E-mail: [email protected] Human infections due to Bartonella spp. are viewed as emerging diseases typical in, although not exclusive to, immunosuppressed patients, in particular those Accepted for publication 24 January 2007 with AIDS, organ transplants and haematological malignancies. We describe four patients, three children and one adult, who developed vegetating papillomatous Key words lesions exclusively on the oral mucosae. They shared a history of haematological Bartonella henselae, chronic graft-versus-host malignancy and allogeneic bone marrow ⁄stem cell transplantation, and later disease, haematological malignancies, oral mucosa, developed chronic graft-versus-host disease, also involving the oral mucosae. papillomatosis Histopathologically, the vegetating lesions were characterized by a diffuse neo- Conflicts of interest angiogenesis, granulation-like tissue, and a mixed cell infiltrate predominantly None declared. composed of neutrophils. Gram-negative bacteria were found in the endothelial cells of the vessels in the deeper portion of the corium by electron microscopy. In three cases, DNA of B. henselae was detected by polymerase chain reaction (PCR), and confirmed by sequencing of the PCR products. All the lesions healed after systemic antibiotic therapy, although some recurred after months, and regressed again after systemic antibiotic treatment associated with conservative surgical excision.
Bartonella is a small, 0Æ3–0Æ5 · 0Æ7–1Æ0 lm aerobic, Gram- as demonstrated in bacillary angiomatosis and veruga peru- negative, pleomorphic bacillus.1 The genus Bartonella currently ana.14 In 1990, Garcia et al., in a study performed on B. hense- comprises 14 species; B. henselae, B. quintana and B. bacilliformis lae, were the first to detect an angiogenic factor from are the most common serotypes known to be causative agents B. bacilliformis.15 We describe four bone marrow-transplanted of human infections, especially in immunocompromised patients who developed multiple, recurrent, yellowish-pink, patients.2–9 Bartonella causes a variety of disorders in humans pseudomembranous, painless vegetations on their tongue, including bacillary angiomatosis, peliosis hepatis, meningitis, palate and cheek mucosae. pneumonia, neuroretinitis, and culture-negative endocarditis predominantly in the immunocompromised host. The diagno- Case reports sis of Bartonella can be difficult and the organism has been found to be able to escape immune defence.10–12 Patient 1 Cats are known as the natural reservoir for Bartonella. Human infection is likely to be the consequence of a cat scratch, bite A 17-year-old Russian girl, with a history of allogeneic bone or lick, but other cases are possibly seen as the consequence marrow transplantation for myelodysplastic syndrome, was of arthropod bite (i.e. sandfly), with cat-to-human transmis- referred to the Department of Dermatology for multiple, yel- sion of the bacillus.13 Among bacterial pathogens, a feature lowish-pink, pseudomembranous, painless pedunculated vege- typical of Bartonella is its capability to induce neoangiogenesis, tations, 1–2 cm in length, on the lateral aspects and ventral
2007 The Authors 174 Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp174–178 Angiomatous papillomatosis and oral chronic GvHD, C. Vassallo et al. 175
Fig 1. Patient 1. Yellow-pink multiple vegetations, present for Fig 2. After bone marrow transplantation, patient 2 presented with 1 month on the lateral aspects and ventral surface of the tongue and multiple vegetating, yellow-pink, fibrin-covered lesions involving the oral mucosa. oral mucosa, predominantly on the tongue. surface of her tongue and oral mucosa that had appeared covered by a yellow membrane of fibrin, on the left lateral 1 month previously (Fig. 1). She had a history of graft-versus- margin of his tongue (Fig. 3). Surgical excision was per- host disease (GvHD) of the oral mucosa and was being treated formed under local anaesthesia. The patient was subsequently with ciclosporin 125 mg daily. A punch biopsy was taken treated with ceftazidime, following a sudden onset of fever. from a vegetating lesion under local anaesthesia, and few days Serology for B. henselae and HIV was negative. later a shave excision of two of the several lesions was also After 14 months of complete remission, fever and a new performed. Serology for human immunodeficiency virus vegetation recurred at the site of the previous lesions (Fig. 4). (HIV), Epstein–Barr virus (EBV) and B. henselae was negative, A new surgical excision was performed under local anaesthe- and no circulating EBV DNA was detected. Culture on blood sia, followed by treatment with erythromycin 1 g three times agar plates from a biopsy failed to demonstrate the presence daily for 12 days. Histopathology was characterized by the of Bartonella. Treatment with clarithromycin 500 mg daily was same features as the first excisional biopsy. Polymerase chain started, with complete resolution of the residual lesions after reaction (PCR) of peripheral blood for B. henselae, however, 1 month of treatment. However, recurrence of the vegetating was negative. lesions was reported after 2 months. Patient 4 Patient 2 A 3-year-old boy with juvenile myelomonocytic leukaemia An 8-year-old boy with acute lymphocytic leukaemia who received allogeneic bone marrow transplantation from a non- was being treated with ciclosporin 80 mg daily, following consanguineous donor. Five months later, a relapse of leukae- allogeneic bone marrow transplantation, was referred to the mia was observed and a second infusion of stem cells from Department of Dermatology with multiple vegetating, pedun- the same donor was made. After 1 year the patient presented culated, yellowish-pink lesions involving the oral mucosa, pre- a severe, diffuse, mucocutaneous chronic GvHD. Moreover, he dominantly the tongue (Fig. 2). A history of chronic GvHD presented with multiple vegetating, pedunculated, yellowish- with mucous membrane involvement was given. Serology was pink lesions involving the oral mucosa, predominantly the negative for B. henselae, Chlamydia spp., HIV and EBV. Histo- tongue. Serology was negative for B. henselae, Chlamydia spp., pathology of two excised lesions presented identical findings HIV and EBV. The histopathology of two lesions excised pre- as in patient 1; treatment with clarithromycin 150 mg daily sented identical findings to those of patient 2; treatment with was given for 20 days, with partial remission of the lesions. cefaclor 450 mg daily and topical erythromycin was started, The treatment was continued with cefaclor 450 mg daily and with complete remission of the lesions in 4 weeks. topical erythromycin for 25 days, with complete remission of the lesions in 5 weeks. Materials and methods
Patient 3 Molecular biology and sequencing technique
A 47-year-old man presented with a history of allogeneic Total DNA was extracted from about 25 mg of tissue follow- bone marrow transplantation for chronic myeloid leukaemia, ing the tissue protocol described in the QIAamp Mini Kit diagnosed 22 months previously. He was being treated with instructions (Qiagen, Milan, Italy). Five microlitres of the ciclosporin 200 mg daily and prednisone 7Æ5 mg daily. extracts were analysed by a PCR assay specific for B. henselae Twenty days previously, he had developed a single vegetation, subtyping as described by Bergmans et al.16 The assay is based
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp174–178 176 Angiomatous papillomatosis and oral chronic GvHD, C. Vassallo et al.
stained with 1% borated methylene blue and examined by light microscopy. Thin sections (about 60 nm thick) were double stained with uranyl acetate in 50% acetone and Reynolds’ lead citrate solution. All the samples were examined and photo- graphed with a Philips CM12 TEM electron microscope.
Results
Histopathology of all the vegetating lesions biopsied or excised gave an overall impression of a granulation-like tissue (Fig. 5); the newly formed capillaries were embedded in an oedematous stroma; a mixed cell infiltrate was composed Fig 3. Patient 3 received allogeneic bone marrow transplantation for mostly of neutrophils, scattered throughout the granulation- chronic myeloid leukaemia and later developed a single yellow-pink like tissue, with different degrees of leucocytoclasis (ranging vegetation on the lateral aspects of his tongue; the presence of chronic from scant to diffuse) and without vasculitis, i.e. fibrin depos- graft-versus-host disease of his tongue is evident as variegated patches its around or in the vessel walls, or within the lumina associ- of atrophic mucosae alternating with whitish areas. ated with neutrophil infiltration and nuclear dust (Fig. 6). Special stains (periodic acid–Schiff, Gram and Warthin–Starry)
Fig 4. After 14 months of remission patient 3 developed fever and a new vegetation, recurring at the site of the previous lesion. Fig 5. Patient 1. Histopathological findings of a vegetation were characterized by granulation-like tissue, with prominent newly formed capillaries embedded in an oedematous stroma (haematoxylin and upon the variation in 16S gene between subtypes I and II of eosin; original magnification · 10). B. henselae and is species specific. Previous laboratory observa- tions (data not shown) had shown this assay to be more sen- sitive than PCR assays targeting other genomic regions of Bartonella. To avoid contamination, dedicated areas and equip- ment for pre- and post-PCR were adopted. The PCR products obtained from the biopsies were sequenced. The PCR fragment of the expected length was purified from agarose gel by the QIAquick gel extraction kit (Qiagen) and the resulting PCR- purified products were sequenced with an ABI Prism device with BigDye terminators (PE Applied Biosystems, Monza, Italy). The nucleotide sequences were analysed by BLAST (www.ncbi.nlm.nih.gov/BLAST).
Transmission electron microscopy
Specimens were fixed for 2 h with 4% glutaraldehyde solution Fig 6. Patient 2. Histopathology of the fibrinoid material covering the )1 in 0Æ1 mol L cacodylate buffer (pH 7Æ2) at 4 C, postfixed sessile lesion shows the presence of fibrin mixed with newly formed for 1Æ5 h with 1% OsO4 in the same buffer (pH 7Æ2) at 4 C, capillaries. Beneath, a frank granulation-like tissue with diffuse dehydrated in graded ethanol solutions and embedded in Epon leucocytoclasis and no vasculitis is observed (haematoxylin and eosin; 812. Preliminarily, semithin sections (0Æ5 lm thick) were original magnification · 25).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp174–178 Angiomatous papillomatosis and oral chronic GvHD, C. Vassallo et al. 177
a monomorphous pattern of neoangiogenesis. The first general histological impression was that of a granulation-like tissue, suggestive of a reactive process, possibly secondary to GvHD Case 1 Case 2 Controls ulceration. However, unlike true granulation tissue, newly formed capillaries were not associated with oedematous stroma, plumped interstitial fibroblasts, and a polymorphous inflammatory infiltrate. In our cases, despite the absence of a true leucocytoclastic vasculitis, leucocytoclasis was evident throughout the corium, and was particularly evident in one case around vessels (patient 3). On clinical grounds, the pseudo- membranous aspect ruled out an angiomatous nature. Among the angiomatous reactive processes, pyogenic granuloma was Fig 7. Patients 1 and 2. Polymerase chain reaction for 16S gene also considered, but both histologically and clinically this specific for Bartonella henselae I and II. diagnosis was not consistent. Histologically, in fact, pyogenic granuloma is regarded as a variant of capillary haemangioma, made early by a distinct proliferation of capillaries embedded in an oedematous stroma, and later by a more evident septum of collagen.17 Pyogenic granuloma, when ulcerated, may be characterized by neutrophils, while leucocytoclasis is usually absent. Furthermore, clinically, the rarely described oral pyogenic granuloma is reported as a red-purpuric exophytic mass that bleeds easily.18 Neoangiogenesis and a diffuse neutrophilic infiltrate associ- ated with leucocytoclasis prompted us to consider a bacterial nature for the lesions, confirmed by electron microscopy, which revealed Gram-negative cocci in endothelial cells in the deepest part of all the lesions of our patients. With reference to neoangiogenesis, Bartonella spp. are known for their ability to induce capillary proliferation, as demonstrated in bacillary 19–23 Fig 8. Gram-negative pleomorphic bacilli were found in the deeper angiomatosis and veruga peruana. portion of the corium, in particular in endothelial cells, but they were Molecular biology confirmed the presence of Bartonella in not found in the upper portion of the corium. Original magnification three of our four patients, and therefore a causative role of · 6800; inset · 10 000. this bacteria can be suggested. We were not able to identify B. henselae in vitro by tissue culture as the genus Bartonella grows for bacteria were not conclusive, or negative. In all four cases, slowly (doubling time: 10 h) and probably because of its molecular analysis and PCR for microorganisms was done complex nutritional requirements.22 Moreover, changes in directly on tissues (Fig. 7). Analyses proved negative for EBV bacterial flora of the oral cavity, due to immunosuppression, and Chlamydia spp., while positive results on tissue were could interfere with isolation of Bartonella in vitro.23 observed for B. henselae in three cases out of four. These find- As already reported in most HIV-infected patients, an anti- ings were also confirmed by sequencing of the PCR products. body response to Bartonella infection could not be detected in On electron microscopy, in all the biopsies taken from all the our group of patients and, even in immunocompetent individ- patients, the presence of Gram-negative bacteria was demon- uals, specific antibodies may not be found.22 Furthermore, strated in endothelial cells of vessels of the deepest portions of ciclosporin inhibits a proper immunological response, with the corium (Fig. 8). impairment of serological response. In contrast to our patients, the rarely described cases of Discussion bacillary angiomatosis with oral involvement present with a concomitant widespread cutaneous involvement.24 Those oral Both the clinical and the histopathological features of all four lesions are described as red angiomatoid papules or nodules, patients, who presented in a short time frame, were peculiar frequently bleeding, rarely multiple and more commonly and constant. The clinical features were characterized by the localized at the gingiva, palate or floor of the oral cavity and rapid onset of single or multiple, vegetating, nonbleeding, only rarely involving the tongue.25 The constant clinical painless, yellowish-pink formations, covered by a fibrinous mucosal involvement, the common history, and the demon- pseudomembrane, localized on the oral mucosae and ⁄or stration of the causative agent in three patients, without tongue (with a prevalent involvement of the tongue). Even evidence of further cutaneous involvement, together suggest though clinically the lesions were not suggestive of having an that these cases are a specific B. henselae infection, that may be angiomatous nature, histologically they were characterized by seen in immunosuppressed, haematological patients with a
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp174–178 178 Angiomatous papillomatosis and oral chronic GvHD, C. Vassallo et al.
previous history of allogeneic bone marrow transplantation 12 Schroder G, Dehio C. Virulence-associated type IV secretion and oral chronic GvHD. systems of Bartonella. Trends Microbiol 2005; 13:336–42. A review of the literature indicates at least six cases, which 13 Plettenberg A, Lorenzen T, Burtsche BT et al. Bacillary angiomatosis in HIV-infected patients – an epidemiological and clinical study. have been described as ‘oral pyogenic granuloma’, that share 26–30 Dermatology 2000; 201:326–31. many histological and clinical features with our cases. 14 Ciceroni L, Fabbi M, Ciarrocchi S et al. Characterization of the first None of these articles gives evidence of the causative role of Bartonella henselae strain isolated from a cat in Italy. Comp Immun Micro- B. henselae infection. biol Infect Dis 2002; 25:217–28. The specificity of these cases suggests the term of pseudo- 15 Garcia FU, Wojta J, Broadley KN et al. Bartonella bacilliformis stimulates membranous (because of fibrin-covered flattish lesions) endothelial cells in vitro and is angiogenic in vivo. Am J Pathol 1990; angiomatous (because of their histopathological features char- 136:1125–35. 16 Bergmans AMC, Schellekens JFP, van Embden JDA, Schouls LM. acterized by neoangiogenesis) papillomatosis (because of their Predominance of two Bartonella henselae variants among cat-scratch pedunculated, vegetating aspect) of the oral cavity in bone disease patients in the Netherlands. J Clin Microbiol 1996; 34:254– marrow-transplanted patients. 60. 17 Grosshans E. Pyogenic granuloma: who are you? J Eur Acad Dermatol Venereol 2001; 15:106–7. Acknowledgments 18 Angelopoulos AP. Pyogenic granuloma of the oral cavity: statistical We thank Luciana De Giuli for her indispensable assistance in analysis of its clinical features. J Oral Surg 1971; 29:840–7. 19 LeBoit PE. Bacillary angiomatosis. Mod Pathol 1995; 8:218–22. the first part of the study for the molecular biology and 20 Kempf VAJ, Volkmann B, Schaller M et al. Evidence of a lead- sequencing technique (PCR for 16S gene). ing role for VEGF in Bartonella henselae-induced endothelial cell proliferation. Cell Microbiol 2001; 9:623–32. References 21 Dehio C. Bartonella interactions with endothelial cells and erythro- cytes. Trends Microbiol 2001; 9:279–85. 1 Maurin M, Raoult D. Bartonella (Rochalimaea) quintana infections. Clin 22 Schabereiter-Gurtner C, Lubitz W, Ro¨lleke S. Application of broad- Microbiol Rev 1996; 9:273–92. range 16S rRNA PCR amplification and DGGE fingerprinting for 2 Liang Z, La Scola B, Lepidi H, Raoult D. Production of Bartonella genus- detection of tick-infecting bacteria. J Microbiol Methods 2003; specific monoclonal antibody. Clin Diag Lab Immun 2001; 8:847–9. 52:251–60. 3 Anderson BE, Neuman MA. Bartonella spp. as emerging human patho- 23 Myers WF, Osterman JV, Wisseman CL Jr. Nutritional studies of gens. Clin Microbiol Rev 1997; 10:203–19. Rickettsia quintana: nature of the hematin requirement. J Bacteriol 4 Paul MA, Fleischer AB, Wieselthier JS, White WL. Bacillary angio- 1972; 109:89–95. matosis in an immunocompetent child: the first reported case. 24 Jacomo V, Raoult D. Human infections caused by Bartonella spp. Pediatr Dermatol 1994; 11:338–41. Part 2. Clin Microbiol Newsl 2000; 22:9–13. 5 Cockerel CJ, Bergstresser PR, Myrie-Williams C, Tierno PM. Bacil- 25 Glick M, Cleveland DB. Oral mucosal bacillary epithelioid angioma- lary epithelioid angiomatosis occurring in an immunocompetent tosis in a patient with AIDS associated with rapid alveolar bone individual. Arch Dermatol 1990; 126:787–90. loss: case report. J Oral Pathol Med 1993; 22:235–9. 6 Tappero JW, Koehler JE, Berger TG et al. Bacillary angiomatosis and 26 Hofman P, Raspaldo H, Michiels JF et al. Angiomatose bacillare de bacillary splenitis in immunosuppressed adults. Ann Intern Med 1993; la cavita` buccale au cours du SIDA. Rev Stomatol Chir Maxillofac 1993; 118:363–5. 94:375–8. 7 Cline MS, Cummings OW, Goldman M et al. Bacillary angiomatosis 27 Akyol MU, Yalc¸iner EG, Dog˘an AI. Pyogenic granuloma (lobular in a renal transplant recipient. Transplantation 1999; 67:296–8. capillary hemangioma) of the tongue. Int J Pediatr Otorhinolaryngol 8 Torok L, Viragh SV, Borka I, Tapai M. Bacillary angiomatosis in a 2001; 58:239–41. patient with lymphocytic leukaemia. Br J Dermatol 1994; 130:665–8. 28 Itin PH, Lautenschlager S, Fluckiger R, Rufli T. Oral manifestations 9 Koehler JE, Cederberg L. Intra-abdominal mass associated with in HIV-infected patients: diagnosis and management. J Am Acad gastrointestinal hemorrhage: a new manifestation of bacillary Dermatol 1993; 29:749–60. angiomatosis. Gastroenterology 1995; 109:2011–14. 29 Itin PH, Fluckiger R, Zbinden R, Frei R. Recurrent pyogenic granulo- 10 Lefkowitz M, Wear DJ. Cat-scratch disease masquerading as a solit- ma with satellitosis – a localized variant of bacillary angiomatosis? ary tumor of the breast. Arch Pathol Lab Med 1989; 113:473–5. Dermatology 1994; 189:409–12. 11 Resto-Ruiz S, Burgess A, Anderson BE. The role of the host 30 Bachmeyer C, Devergie A, Mansouri S et al. Pyogenic granuloma of immune response in pathogenesis of Bartonella henselae. DNA Cell Biol the tongue in chronic graft versus host disease. Ann Dermatol Venereol 2003; 22:431–40. 1996; 123:552–4.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp174–178 Gene Corner
Novel COL7A1 mutations in a Japanese family with transient bullous dermolysis of the newborn associated with pseudosyndactyly
DOI: 10.1111/j.1365-2133.2007.07955.x also apparent. An X-ray confirmed the fused toes as pseudo- syndactyly and revealed shortened distal phalanges of the first Transient bullous dermolysis of the newborn (TBDN) is a and second toes (Fig. 2, lower panel). neonatal blistering disease evident at birth or shortly there- For genetic analysis, genomic DNA was extracted from peri- after.1 A key feature of TBDN is that the blistering decreases pheral blood lymphocytes of the proband, his parents and his with increased age, and most lesions heal within several half-brother using a standard protocol. Mutation screening of months to leave minimal scarring and pigmentation. Although the COL7A1 gene was performed by denaturing high-perform- the blistering can be widespread and occasionally affects the ance liquid chromatography (dHPLC; Transgenomic WAVE ; mucous membranes, TBDN rarely involves skeletal abnormal- Transgenomic, Omaha, NE, U.S.A.) and direct sequencing in ities such as pseudosyndactyly, so this subtype of epidermo- an ABI 310 Genetic Analyzer (Applied Biosystems, Foster City, lysis bullosa (EB) is thought to be relatively benign. Retention CA, U.S.A.). Specific primers and polymerase chain reaction of collagen VII within the epidermis is observed histologically, (PCR) conditions have been described in detail elsewhere.2 and dilated rough endoplasmic reticulum and inclusion bod- Allele-specific PCR was performed using the following primers ies, known as stellate bodies, are visible by electron micro- for the mutant allele: forward, 5¢-CAGGGAAAGCCAGGC- scopy in basal ⁄suprabasal keratinocytes of the patient’s skin. GAGG*T-3¢ (*, the location of the deleted wild-type A5504); We herein report novel COL7A1 gene mutations in a Japanese and reverse, 5¢-TGTAGGTGTGCTGGCGTTTC-3¢. The PCR con- family with TBDN associated with pseudosyndactyly. ditions were: 4 min pre-incubation at 94 C, followed by 40 cycles of 45 s at 94 C, 45 s at 63 C and 45 s at 72 C. Case and methods Results and discussion A 7-day-old male infant was referred to our department for erosions and blisters on the feet, legs and trunk, which pre- In the proband, dHPLC analysis of the PCR products spanning sented just after delivery. There was no family history of blis- exon 64 and the flanking introns showed a pattern shift, and tering disease. Physical examination revealed well-demarcated direct sequencing revealed a heterozygous deletion of the erosions on the feet and legs and, most seriously, the first toes adenine residue at nucleotide position 5504 (c.5504delA). were completely denuded. Small bullae were visible on the This deletion produces a premature termination codon (PTC) back and on the left thigh where adhesive tape had been at nucleotide position 5518. Allele-specific PCR demonstrated attached, and the oral mucosa was also eroded. A biopsy was that the mutation was carried by the proband’s mother, but performed at 7 days of age to confirm the clinical diagnosis of not by the father or the half-brother (by a different father) congenital EB. Electron microscopy revealed separation of the (Fig. 1d). Mutation analysis disclosed two additional point dermoepidermal junction below the lamina densa of the base- mutations, a heterozygous G to A substitution at position ment membrane, which was associated with decreased and 2969 of exon 22 and a heterozygous G to A substitution at impaired anchoring fibrils (Fig. 1a). In addition, dilated rough position 6023 of exon 73 (Fig. 1e, f); the former converts an endoplasmic reticula were noted in the basal and suprabasal arginine to a glutamine (p.R990Q), and the latter changes an keratinocytes (Fig. 1b). The erosion of the oral mucosa arginine to a histidine (p.R2008H). Direct sequencing and resolved within 2 weeks and blister formation ceased, leaving restriction fragment length polymorphism analysis demonstra- atrophic scars, pigmentation and toenail dystrophy that were ted that the c.2969 GfiA mutation was carried by the mother, evident when the proband was examined 2 months later. The whereas the c.6023 GfiA was carried by the father (Fig. 1e, f). patient was diagnosed as having TBDN, and follow-up exami- These two missense mutations were not identified in 102 nations revealed no blister formation by the time he was healthy controls. 29 months old. However, the mother noticed that the patient TBDN is currently recognized as a subtype of dystrophic EB seemed to be in pain while walking, and physical examination on the basis of the underlying COL7A1 gene defect and several revealed shortened webbing between the first and second toes reports on the phenotype.3 The inheritance of TBDN has been of the right foot (Fig. 2, upper panel). Toenail dystrophy was shown to be sporadic in most cases, although autosomal
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp179–182 179 180 COL7A1 mutations in TBDN associated with pseudosyndactyly, H. Nakano et al.
(a)
(c)
(b) (d)
(e) (f)
Fig 1. Transmission electron microscopy of the proband’s skin. (a) Dermoepidermal separation below the lamina densa. Decreased and impaired anchoring fibrils are also noted. (b) Dilated rough endoplasmic reticula containing filamentous material are visible in basal and suprabasal keratinocytes (arrows). Asterisk indicates nucleus. Original magnification: (a), (b) · 40 000. (c) Pedigree of the Japanese family with transient bullous dermolysis of the newborn. The arrow in the pedigree indicates the proband. (d) Allele-specific polymerase chain reaction (PCR) amplification using primers specific for the mutant sequence. DNA from the proband and his mother yielded an amplified band of 105 bp. C, a normal control. (e) Direct sequencing revealed a heterozygous c.2969 GfiA mutation in samples from the proband and his mother (upper panel). In the normal allele, BbvI cleaved the 500-bp PCR product into fragments of 463 and 37 bp, while cleavage of the mutated allele resulted in fragments of 276, 187 and 37 bp (lower panel). (f) Sequence analysis disclosed a heterozygous c.6023 GfiA mutation in the proband’s and his father’s COL7A1 genes. The c.6023 GfiA mutation generated a NcoI site that led to cleavage of the 280-bp PCR product into fragments of 201 and 79 bp (lower panel).
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp179–182 COL7A1 mutations in TBDN associated with pseudosyndactyly, H. Nakano et al. 181
Fig 2. (Upper panel) Pseudosyndactyly of the right foot associated with toenail dystrophy at 29 months of age. The first and second toes are shortened. (Lower panel) Growth retardation of the first and second distal phalanges is apparent on X-ray (arrows).
dominant and recessive families have been reported. Thus far, in the COL7A1 gene. The pathogenic role of the maternal mis- four mutations in the COL7A1 gene have been identified in sense mutation, however, remains to be elucidated. Transcripts three families with TBDN: a heterozygous c.4120-1 GfiC pre- from the mutant maternal allele harbouring the c.5504delA dicted to result in in-frame skipping of exon 36,4 compound mutation are likely to undergo nonsense-mediated mRNA heterozygous mutations causing the amino acid substitutions decay, and thus to result in loss of expression of the additional p.G1519D and p.G2251E,5 and a heterozygous c.4565 GfiA mutation carried by the same allele. mutation generating p.G1522E.6 In the present study, we have Pseudosyndactyly occurs commonly in recessive dystrophic identified two novel missense mutations, c.2969 GfiA and EB (50–60%), while patients with dominant dystrophic EB c.6023 GfiA, leading to p.R990Q and p.R2008H, respective- and other subtypes of EB develop this complication much less ly, and a deletion mutation, c.5504delA, which results in a frequently.10 This skeletal abnormality is progressive and PTC. The c.5504delA mutation was previously demonstrated resistant to most therapeutic modalities, including surgical in combination with a recurrent donor splice site mutation intervention. In 29 previously reported cases of TDBN, no c.6573+1 GfiC, which also results in a PTC, in a Japanese association with skeletal abnormality was documented, while family with non-Hallopeau–Siemens type recessive dystrophic nail loss and dystrophy, characteristic features of dystrophic EB.7 The arginine residue at amino acid position 2008 was EB, were common. The pseudosyndactyly in the present case reported to be mutated in a glycine or a cysteine residue in is presumably attributable to the severe denudation of the toes families with recessive dystrophic EB.8,9 These findings, at birth, and the associated shortening of the phalanges although demonstrated in only a limited number of cases, might have occurred secondary to the scarring of the toes. In suggest that there is no phenotype-specific mutation for TBDN almost all TBDN cases except that reported by Gedde-Dahl,11
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp179–182 182 COL7A1 mutations in TBDN associated with pseudosyndactyly, H. Nakano et al. blistering was of ‘transient’ nature, leading us to hypothesize 5 Hammami-Hauasli N, Raghunath M, Kuster W, Bruckner-Tuder- that the pseudosyndactyly in our case may be treatable by sur- man L. Transient bullous dermolysis of the newborn associated gical intervention. Although foot involvement does not gener- with compound heterozygosity for recessive and dominant COL7A1 mutations. J Invest Dermatol 1998; 111:1214–19. ally require intervention because toe webbing contractions do 6 Fassihi H, Dida VC, Wessagowit V et al. Transient bullous dermo- not usually affect ambulation, surgical therapy may be appro- lysis of the newborn in three generations. Br J Dermatol 2005; priate for the proband if pain upon ambulation persists. 153:1058–63. 7 Ishiko A, Masunaga T, Ota T, Nishikawa T. Does the position of Departments of Dermatology and *Orthopedics, H. NAKANO the premature termination codon in COL7A1 correlate with the Hirosaki University School of Medicine, Y. TOYOMAKI clinical severity in recessive dystrophic epidermolysis bullosa? Exp Dermatol 2004; 13:229–33. 5 Zaifu-cho, Hirosaki 036-8562, Japan S. OHASHI 8 Hovnanian A, Rochat A, Bodemer C et al. Characterization of 18 Department of Dermatology, Aomori City A. NAKANO new mutations in COL7A1 in recessive dystrophic epidermolysis Hospital, Aomori, Japan H. JIN* bullosa provides evidence for distinct molecular mechanisms àDivision of Gene Therapy Science, Graduate T. MUNAKATA underlying defective anchoring fibril formation. Am J Hum Genet School of Medicine, Osaka University, N. AKITA 1997; 61:599–610. Osaka, Japan K. TAMAIà 9 Kon A, Pulkkinen L, Ishida-Yamamoto A et al. Novel COL7A1 muta- §Department of Dermatology, Yamagata Y. MITSUHASHI§ tions in dystrophic forms of epidermolysis bullosa. J Invest Dermatol University School of Medicine, Yamagata, Japan 1998; 111:534–7. 10 Fine J-D, Johnson LB, Moshell A, Suchindran C. The risk of selec- E-mail: [email protected] ted major extracutaneous outcomes in inherited epidermolysis bullosa. In: Epidermolysis Bullosa (Fine J-D, Bauer EA, McGuire J, References Moshell A, eds). Baltimore: The Johns Hopkins University Press, 1999; 193–205. 1 Hashimoto K, Matsumoto M, Iacobelli D. Transient bullous dermo- 11 Gedde-Dahl T Jr. The childhood course of recessive epidermolysis lysis of the newborn. Arch Dermatol 1985; 121:1429–38. bullosa dystrophica inversa. In: Epidermolysis Bullosa: A Comprehensive 2 Christiano AM, Hoffman GG, Zhang X et al. Strategy for identifica- Review of Classification, Management and Laboratory Studies (Priestley GC, tion of sequence variants in COL7A1, and a novel 2 bp deletion Tidman MJ, Weiss JB, Eady RAJ, eds). Crowthorne, Berkshire: mutation in recessive dystrophic epidermolysis bullosa. Hum Mutat DEBRA UK, 1990; 84–6. 1997; 10:408–14. 3 Fine J-D, Eady RAJ, Bauer EA et al. Revised classification system for Accepted for publication: 4 February 2007 inherited epidermolysis bullosa: report of the second international Key words: COL7A1, epidermolysis bullosa, mutation, transient bullous dermolysis consensus meeting on diagnosis and classification of epidermolysis of the newborn bullosa. J Am Acad Dermatol 2000; 42:1051–66. 4 Christiano AM, Fine J-D, Uitto J. Genetic basis of dominantly Conflicts of interest: none declared. inherited transient bullous dermolysis of the newborn: a splice site mutation in the type VII collagen gene. J Invest Dermatol 1997; 109:811–14.
2007 The Authors Journal compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp179–182 Correspondence
Cutaneous lesions in neurofibromatosis 1: existing concordance in the clinical classification of certain confused terminology cutaneous lesions, was carried out as described below. Three dermatologists, all members of the NFN, selected a DOI: 10.1111/j.1365-2133.2007.07903.x series of 113 clinical pictures of lesions commonly found in NF1 patients and ascribed to each one of six terms used in 1 1,2,7–9 SIR, Neurofibromatosis (NF1) is one of the most common the literature and text books: cafe´-au-lait spots, freckles autosomal dominant disorders, with a prevalence of 1 in or lentigines, dermal neurofibromas, nodular neurofibromas, 3500.1,2 Diagnosis is based on seven National Institutes of plexiform neurofibromas and juvenile xanthogranulomas. Health (NIH) consensus criteria,3 the following three of Twenty-four members of the NFN who were highly experi- which relate to cutaneous lesions: presence of six or more enced in dealing with NF1 (they were all involved in neuro- cafe´-au-lait spots; presence of axillary freckling; and presence fibroma clinics and had seen at least three NF1 cases per week of at least two neurofibromas or one plexiform neurofibroma. for more than 10 years) were sent copies of the pictures and Some lesions are thought to have prognostic value. For asked to classify them according to the six diagnoses. example, plexiform and subcutaneous neurofibromas have The concordance level (CL) or agreement ratio was calculated been linked to malignancy and internal neurofibromas.4–6 for each picture as the ratio between the number of experts giv- Recognition and classification of dermatological lesions there- ing the same diagnosis and the total number of assessments. fore have important implications for clinical practice. Unfortu- Concordance was classified as perfect (CL ¼ 100%), high nately, however, there is no consensus on terminology, (CL ¼ 75–99%), low (CL ¼ 50–74%) or poor (< 50%). particularly with regard to neurofibromas. Sixteen of 24 experts sent the questionnaire responded (14 In the literature, including textbooks, particular lesions may dermatologists, one paediatrician, one geneticist). Concor- be described in different ways. Cutaneous neurofibroma, for dance was perfect for only 16 of 113 pictures (14%), high instance, may be referred to as localized neurofibroma, super- for 46 of 113 (41%), low for 34 of 113 (30%) and poor in ficial neurofibroma, dermal neurofibroma or nodular neuro- 17 of 113 (15%). Table 1 summarizes the levels of agreement fibroma. Variations in terminology used by clinicians and between experts with regard to the proposed diagnoses. CL pathologists only add to the confusion. The term ‘plexiform’ was mainly perfect or high for the diagnosis of cafe´-au-lait has a specific meaning in pathology, and clinicians should not spots, freckles, juvenile xanthogranuloma and plexiform neuro- use it to classify a neurofibroma unless they have histological fibroma, but mainly low or poor in cases of dermal neuro- evidence for doing so.1,2,7–9 fibroma and nodular neurofibroma. The French Neurofibromatoses Network (NFN) is develop- This first step in our project to establish a common termin- ing a common clinical terminology which experts from var- ology showed that concordance was good for pigmented lesions ious relevant disciplines can use to classify NF1 skin lesions, (cafe´-au-lait spots and freckles), large plexiform neurofibromas thereby improving concordance. The first step, to evaluate and juvenile xanthogranulomas, but classification of neurofibromas
Table 1 Type of diagnosis and concordance level between 16 experts concerning the Number Concordance levela between experts (%) identification of 113 pictures of cutaneous of pictures lesions in neurofibromatosis Type of diagnosis identified Perfectb Highc Mildd Poore
Cafe´-au-lait spots 13 3 (23) 7 (54) 3 (23) 0 Freckles 5 3 (60) 2 (40) 0 0 Dermal neurofibroma 35 0 11 (31) 18 (51) 6 (17) Nodular neurofibroma 22 0 10 (45) 6 (27) 6 (27) Plexiform neurofibroma 36 10 (28) 14 (39) 7 (19) 5 (14) Juvenile xanthogranuloma 2 0 2 (100) 0 0
Total 113 16 (14) 46 (41) 34 (30) 17 (15)
aConcordance level (CL): ratio between the number of experts giving the same diagnosis and the total number of experts. bCL ¼ 100%; cCL ¼ 75–99%; dCL ¼ 50–74%; eCL < 50%.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 183 184 Correspondence
MD, Patrick Combemale MD, Jean-Franc¸ois Cuny MD, Chris- tian Derancourt MD, He´le`ne Dollfus MD, Salah Ferkhal MD, Claude Guillot MD, Christine Labre`ze MD, Dan Lipsker MD, Alain Taieb MD, Pierre Vabres MD, Jacques Zeller MD. This study was supported in part by the Association Neurofibroma- toses et Recklinghausen, Blagnac, France.
Departments of Dermatology, S. BARBAROT *Clinical Research, Methodological Unit and C. NICOL Photology, Nantes University Hospital, C. VOLTEAU* CHU Hoˆtel Dieu, place Alexis Ricordeau, D. LE F ORESTIER F-44093 Nantes, France J.M. N’GUYEN* Department of Dermatology and Grand Paris E. MANSAT Neurofibromatoses Center, Hoˆpital Henri-Mondor, P. WOLKENSTEIN AP-HP, Paris XII University, Cre´teil, France J.F. STALDER E-mail: [email protected]
References
1 Huson SM, Hughes RAC. The Neurofibromatoses: A Pathogenetic and Clinical Overview. London: Chapman and Hall, 1994. 2 Friedman JM, Gutmann DH, MacCollin M, Riccardi VM. Neurofibroma- tosis: Phenotype, Natural History and Pathogenesis. Baltimore, MD: Johns Hopkins University Press, 1999. 3 Neurofibromatosis. Conference statement. National Institutes of Health consensus development conference. Arch Neurol 1988; 45: 575–8. 4 Khosrotehrani K, Bastuji-Garin S, Riccardi VM et al. Subcutaneous neurofibromas are associated with mortality in neurofibromatosis 1: a cohort study of 703 patients. Am J Med Genet 2005; 132:49–53. 5 Tucker T, Wolkenstein P, Revuz J et al. Association between benign and malignant peripheral nerve sheath tumors in NF1. Neurology 2005; 65:205–11. 6 Leroy-Viard K, Dumas V, Voisin MC et al. Malignant peripheral nerve sheath tumors associated with neurofibromatosis type 1: a clinical, pathologic and molecular study of 17 patients. Arch Dermatol 2001; 137:908–13. 7 Friedman JM, Birch PH. Type 1 neurofibromatosis – a descriptive analysis of the disorder in 1728 patients. Am J Med Genet 1997; 70:138–43. 8 Enzinger FM, Weiss SW. Benign tumors of peripheral nerves. In: Soft Tissue Tumors, 3rd edn. (Enzinger FM, Weiss SW eds). St Louis, MO: CV Mosby, 1995; 851–62. Fig 1. Dermal and nodular neurofibromas with poor diagnostic 9 Megahed M. Histopathological variants of neurofibroma. A study of concordance level (< 50%) between the 16 experts. 114 lesions. Am J Dermatopathol 1994; 16:486–95. was confused, particularly that of recent and discrete dermal and nodular neurofibromas (Fig. 1). During a debriefing work- Conflicts of interest: none declared. shop, the panel of experts offered two main explanations for these findings: (i) pictures do not provide information on pal- pation; and (ii) there is absolutely no consensus on termin- ology. In order to take our next step and reach that consensus, we will recruit experts from other medical disciplines and pro- vide more clinical information about each type of lesion (palpa- Lentigo maligna involving the tumour nests tion, pain, localization and skin surface modifications). and stroma of a nodular basal cell carcinoma
Acknowledgments DOI: 10.1111/j.1365-2133.2007.07904.x
The authors are grateful to their fellow members of the French SIR, Lentigo maligna and basal cell carcinoma (BCC) are two Neurofibromatosis Network: Henri Adamski MD, Franck Bora- distinct tumours that classically develop in elderly people levi MD, Pierre Castelneau MD, Jacqueline Chevrant-Breton on sun-exposed skin including the face. The coexistence of
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 185 lentigo maligna and BCC is, none the less, a rare occurrence, The BCC appeared completely excised (closest radial margin with only five previously reported cases, to our knowledge.1–5 2Æ5 mm, closest deep margin 1 mm) but lentigo maligna In such cases assessing the overall prognosis may be further extended focally beyond the lateral margins of the carcinoma, complicated by colonization of malignant melanocytic cells up to the radial margin of excision. Notably, atypical melano- within the carcinoma tumour nests,3–5 raising the question as cytes, both as single cells and small clusters, appeared to pop- to whether this represents true dermal ‘invasion’. ulate not only the basaloid islands (Fig. 2c) but also the We report a 78-year-old man who presented with a stroma of the BCC, highlighted by Melan-A (Fig. 2d,e) and 2-month history of an asymptomatic pigmented lesion on the S-100. Although Melan-A confirmed the presence of lentigo nasal tip. He had moderate previous sun exposure, including maligna extending beyond the lateral margin of the BCC, there trips to Australia on business, and sun burning on beaches was no evidence of invasion of atypical melanocytes into the in his youth. Clinical examination showed an indistinct ‘normal’ dermis beyond the peripheries of the tumour 7 · 5 mm papular lesion bearing eccentric and irregular (Fig. 2f). Additional stains performed included Ber-EP4, epi- pigmentation (Fig. 1). The clinical differential diagnoses inclu- thelial membrane antigen, cytokeratin-5, Bcl-2, Ki-67, CD10, ded pigmented BCC, pigmented actinic keratosis and lentigo p53, cyclin D1 and b-catenin. The most striking observations maligna melanoma. On general skin inspection, a 10 · 7mm were strong staining of the basaloid carcinoma cells with irregular pigmented macular lesion was incidentally observed Ber-EP4 and cytokeratin-5, indirectly highlighting, by absence on the right side of his neck, which was also deemed clinical- of staining, the atypical melanocytic population within the ly suspicious. The lesions were excised with clinical margins tumour nests. There was moderately strong staining of cyclin of 4 mm for the nasal lesion and 5 mm for the neck lesion. D1 in atypical melanocytes within tumour nests, in addition to Histology of the nasal lesion showed a nodular BCC, depth a proportion of adjacent lentigo maligna, in contrast to virtu- 2 mm, evident at low power, in addition to solar elastosis of ally absent staining in basaloid cells of the BCC. The second the adjacent dermis (Fig. 2a). Higher power confirmed the lesion, on his neck, proved to be melanoma in situ (closest typical morphology of BCC, with islands of basaloid cells radial margin 2Æ5 mm). showing peripheral palisading and numerous apoptotic bodies The association of malignant melanoma and other skin can- (Fig. 2b). However, an atypical intraepidermal lentiginous cers is well known, and is partly attributable to previous solar melanocytic proliferation, showing severe cytological atypia ultraviolet irradiation and field effect. In a previous case– with pleomorphism and prominent nucleoli, could also be control study, an overall fourfold increase in risk of melanoma appreciated overlying the BCC, diagnostic for lentigo maligna. was shown when BCC or squamous cell carcinoma was present on the face.6 However, ‘collision tumours’ are uncommon. In a retrospective audit of 78 000 excisions of primary cutaneous cancers, 11 cases of contiguous malignant melanoma and BCC were identified.7 These were all located at nonfacial sites, the melanoma was not of lentigo maligna melanoma subtype, and melanoma and BCC were sharply demarcated from each other in all cases. With specific regard to lentigo maligna and BCC ‘collision tumours’, we have iden- tified only five previous cases in the literature.1–5 Two reports refer to an incidental finding of lentigo maligna in juxtaposi- tion to, but distinct from BCC in the same excision speci- men,1,2 whereas in three reports lentigo maligna was found to intermingle directly with the BCC.3–5 BCCs are thought to arise from the follicular outer root sheath epithelium. Ber-EP4 consistently stains normal outer root sheath in addition to BCC, in keeping with this hypoth- esis.8 The presence of benign melanocytes populating BCC has been established,9 and is perhaps unsurprising given that mel- anocytes usually reside in the basal epidermal and follicular epithelium. As a BCC proliferates, melanocytes may become entrapped within tumour nests. It has previously been demon- strated in vitro that melanocyte growth and dendricity are directly promoted by keratinocyte-derived factors,10 and it is likely that melanocytes within BCCs may be influenced by tumour-derived growth factors in an analogous manner. Lentigo maligna is frequently noted to involve follicular epi- thelium and malignant melanocytes may become incorporated Fig 1. Papule on tip of nose bearing irregular pigmentation. into BCC by similar mechanisms.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 186 Correspondence
(a) (d)
(b) (e)
(c) (f)
Fig 2. (a) Nodular basal cell carcinoma (BCC) and adjacent solar elastosis evident at low power. (b) Severely atypical junctional melanocytic proliferation overlying BCC. (c) BCC tumour islands populated by atypical melanocytes. (d) Melanocytic population of tumour islands, in addition to junctional lentiginous proliferation, evident at low power. (e) Atypical melanocytes populating tumour islands and stroma of BCC, evident at high power. (f) Absence of dermal invasion of atypical melanocytes beyond margins of BCC. (a–c) Haematoxylin and eosin; (d–f) Melan-A staining; original magnification: (a,d) · 50; (b,f) · 200; (c,e) · 400.
In one study of 10 typical BCCs using immunostaining with periphery of tumour islands in five cases, or evenly scattered Melan-A, HMB-45, S-100 and CD1a, melanocytes were dem- throughout the nests in the remaining five cases, indicating onstrated within BCC tumour nests in all cases, with additional that neoplastic keratinocytes in BCC are still able to regulate Langerhans cells in nine of the 10 tumours.9 The melanocytes the growth and morphology of benign melanocytes that pop- appeared distributed as single dendritic units, either at the ulate them. This contrasted with the example of BCC and
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 187 melanoma in situ ‘collision tumour’ also studied by the same of BCC would suggest otherwise, and that an extensive re-exci- group, in which tumour nests contained a higher density of sion margin, e.g. 1–2 cm involving the tip of the nose, could malignant, less obviously dendritic, melanocytes, disposed not be justified. The further management of this patient was both singly and, unlike the 10 straightforward BCCs, in discussed at our skin cancer multidisciplinary team meeting. In clusters of two or more cells, suggesting a proliferative dys- view of the focal extension of atypical melanocytic hyperplasia regulation. and the significant photoageing in surrounding ‘normal’ skin The close relationship between the populations of melano- coupled with the centrofacial position of the lesion we decided cytes and basaloid cells in such cases (including our own) may to offer our patient regular clinical follow-up rather than any indicate that the tumour combination reflects more than a specific further intervention at this stage. However, we chance event. For example, there are previous reports of acknowledge that predicting the biology of this melanocytic tumours showing dermal nests comprising an intimate admix- component, given the paucity of similar cases, is difficult. ture of epithelial and melanocytic malignant cells, referred to 11,12 as malignant ‘basomelanocytic’ tumour. This entity refers Acknowledgments to a genuinely biphasic tumour, with cells believed to arise from a common ectodermal precursor showing biphasic We thank Dr Karen Blessing, Glasgow, for her helpful immunophenotypic differentiation, as suggested by a propor- comments. tion of cells showing immunopositivity for both melanocytic and epithelial (cytokeratin) markers and also, significantly, as Departments of Dermatology and S.M. TAIBJEE 11 in the case reported by Erickson et al., by the absence of a Histopathology, Warwick Hospital, B.C. GEE junctional malignant melanocytic component. Although our Warwick CV34 5BW, U.K. D.S.A. SANDERS case has some similarities, it is unlikely to represent a genu- Correspondence: Richard A. Carr. A. SMITH inely basomelanocytic tumour given the presence of a junc- E-mail: [email protected] R.A. CARR tional lentiginous melanocytic proliferation extending beyond the margins of the BCC, in keeping with lentigo maligna, and References the distinct immunophenotype of the atypical melanocytes within BCC nests (S-100, Melan-A, cyclin D1 positive) com- 1 Sina B, Samorodin C. Basal cell carcinoma surrounded by lentigo pared with BCC tumour cells (Ber-EP4, cytokeratin-5 positive). maligna. Cutis 1989; 44:81–2. Our case appears to be unique, differing from previous 2 Hirakawa E, Miki H, Kobayashi S et al. Collision tumor of cutane- ous malignant melanoma and basal cell carcinoma. Pathol Res Pract reports by the presence of malignant melanocytic cells within 1998; 194:649–53. the stroma in addition to tumour islands of the BCC. Notably, 3 Burkhalter A, White WL. Malignant melanoma in situ colonizing there are previous reports of tumours featuring distinctive basal cell carcinoma. A simulator of invasive melanoma. Am J stromal components, including trichoepithelioma and, indeed, Dermatopathol 1997; 19:303–7. BCC, found in combination with melanocytic naevi.13,14 It has 4 Wang H, Benda PM, Piepkorn MW. Parasitism of basal cell carci- been hypothesized that these neoplasms are the direct result of noma by lentigo maligna melanoma. J Am Acad Dermatol 2003; 48 epithelial and stromal induction caused by the pre-existing (Suppl. 5):S92–4. 5 Belisle A, Gautier MS, Ghozali F et al. A collision tumor involving melanocytic naevus. Potentially, malignant melanocytic prolif- basal cell carcinoma and lentigo maligna melanoma. Am J Dermato- erations including lentigo maligna may similarly predispose to pathol 2005; 27:319–21. 3 secondary tumours including BCC. 6 Green AC, O’Rourke MGE. Cutaneous malignant melanoma in asso- The extremely unusual situation of lentigo maligna colon- ciation with other skin cancers. J Natl Cancer Inst 1985; 74:977–80. izing BCC poses a particular prognostic and therapeutic 7 Pierard GE, Fazaa B, Henry F et al. Collision of primary malignant dilemma. The prognosis of malignant melanoma usually corre- neoplasms on the skin: the connection between malignant mela- lates with the Breslow thickness, and if this were interpreted as noma and basal cell carcinoma. Dermatology 1997; 194:378–9. 8 Jimenez FJ, Burchette JL Jr, Grichnik JM, Hitchcock MG. Ber-EP4 the melanocytic component involving the BCC islands it would immunoreactivity in normal skin and cutaneous neoplasms. measure 2 mm. The consensus in the previous literature is that Mod Pathol 1995; 8:854–8. melanoma in situ colonizing the epidermal, although neoplastic, 9 Florell SR, Zone JJ, Gerwels JW. Basal cell carcinomas are popu- environment of BCC is analogous to appendageal extension of lated by melanocytes and Langerhans cells. Am J Dermatopathol 2001; lentigo maligna, and should not be regarded as ‘invasive’. 23:24–8. However, Belisle et al.5 specifically mention that malignant 10 Gordon PR, Mansur CP, Gilchrest BA. Regulation of human mel- melanocytic cells were confined to BCC tumour nests in their anocyte growth, dendricity, and melanization by keratinocyte derived factors. J Invest Dermatol 1989; 92:565–72. case, and classify this as in situ melanoma on the basis that after 11 Erickson LA, Myers JL, Mihm MC et al. Malignant basomelanocytic careful scrutiny such cells were undetectable between dermal tumor manifesting as metastatic melanoma. Am J Surg Pathol 2004; collagen bundles. (Interestingly, in the subsequent re-excision 28:1393–6. in their patient, true dermal melanoma invasion was identified 12 Rodriguez J, Nonaka D, Kuhn E et al. Combined high-grade basal beyond the limits of the BCC). By their definition, our particu- cell carcinoma and malignant melanoma of the skin (‘malignant lar case would constitute truly ‘invasive’ melanoma. We feel basomelanocytic tumor’): report of two cases and review of the that the absence of a dermal component beyond the margins literature. Am J Dermatopathol 2005; 27:314–18.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 188 Correspondence
13 Brownstein MH, Starink TM. Desmoplastic trichoepithelioma and (a) intradermal nevus: a combined malformation. J Am Acad Dermatol 1987; 17:489–92. 14 Rosenblum GA. Large basal cell carcinoma in a congenital nevus. J Dermatol Surg Oncol 1986; 12:166–8.
Conflicts of interest: none declared.
Eruptive vellus hair cysts presenting as bluish-grey facial discoloration masquering as naevus of Ota
DOI: 10.1111/j.1365-2133.2007.07926.x
SIR, Eruptive vellus hair cysts (EVHCs) typically present as multiple skin-coloured papules affecting the chest. We describe an unusual case of EVHCs in a 19-year-old Chinese woman who had dyspigmentation over her face and proximal upper limbs since childhood. Examination showed remarkable bluish-grey facial discoloration mimicking bilateral naevus of Ota. To our knowledge, this is the second case of EVHCs reported in a Chinese individual. (b) A 19-year-old Chinese woman was referred to the derma- tology clinic with a greater than 10-year history of a persistent asymptomatic bluish discoloration over her face, bilateral shoulders and extensor surfaces of both arms. The discolor- ation had intensified over the previous few years. Her health had previously been good, with no history of facial dermatitis, excessive sun exposure or cosmetic application. Her father had similar lesions on his face. Examination showed bilateral bluish discoloration on her forehead, temples and cheeks (Fig. 1a) mimicking bilateral naevus of Ota. Closer examination revealed myriad tiny non- follicular dermal papules on these areas. Similar lesions were also found on the shoulders and upper arms. The papules var- ied from 1 to 2 mm in size and were nontender. The over- Fig 1. (a) The bluish-grey discoloration was symmetrical on both lying epidermis appeared unaffected. Inflammatory acne sides, mimicking bilateral naevus of Ota. (b) Photomicrograph papules were not a feature. Both sclera were normal. A provi- showing typical vellus hair cysts without significant melanosis sional diagnosis of dermal melanosis associated with milia, (haematoxylin and eosin; original magnification · 20). calcinosis cutis, closed comedones or juvenile colloid milium was considered. Skin biopsy of a lesion on the right temple revealed typical vellus hair cysts without significant melanosis locations involved are the chest and upper extremities.1 The (Fig. 1b). Our patient was put on a list for carbon dioxide neck or face may also be affected.4 They usually present at 1 (CO2) laser treatment. birth, childhood and in the first and second decades. EVHCs were first described by Esterly et al. in 1977.1 They Histopathology of EVHCs shows a cystic structure in the reported two children presenting with symmetrically distri- mid-dermis which is lined by squamous epithelium that is buted papules over the chest and flexor extremities. Although two to five cells thick. They arise from the infundibulum of a there is clinical variability in the appearance of these cysts, hair follicle and contain multiple vellus hairs. The differential they usually present as smooth 1–2-mm skin-coloured or vari- diagnosis of pilosebaceous cysts includes EVHCs, steato- ably pigmented papules.2 They may also present as multiple cystoma multiplex and infundibular cysts. All of them usually cysts. In one Korean case report, more than 400 cysts were present as asymptomatic papules and involve the anterior seen in a patient with generalized EVHCs.3 The most common chest. The cysts of steatocystoma multiplex have a crenulated,
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 189 eosinophilic, hyaline lining on the cyst wall. They arise in the 4 Kumakiri M, Takashima I, Iju M et al. Eruptive vellus hair sebaceous duct and may contain sebaceous glands arising cysts—a facial variant. J Am Acad Dermatol 1982; 7:461–7. within the cyst. Infundibular cysts are obstructed or occluded 5 Requena L, Sa´nchez Yus E. Follicular hybrid cysts. An expanded spectrum. Am J Dermatopathol 1991; 13:228–33. follicles which do not contain sebaceous glands or vellus hairs. 6 Mayron R, Grimwood RE. Familial occurrence of eruptive vellus Several reports revealed cases containing elements of both hair cysts. Pediatr Dermatol 1988; 5:94–6. 5 EVHCs and steatocystoma multiplex. ‘Hybrid cysts’ describe 7 Bovenmyer DA. Eruptive vellus hair cysts. Arch Dermatol 1979; cysts containing combinations of EVHCs, steatocystoma multi- 115:338–9. plex and epidermoid cysts. 8 Fisher DA. Retinoic acid in the treatment of eruptive vellus hair In most reported cases, EVHCs are isolated findings. Some cysts. J Am Acad Dermatol 1981; 5:221–2. cases were associated with other skin disorders such as pachy- 9 Huerter CJ, Wheeland RG. Multiple eruptive vellus hair cysts trea- ted with carbon dioxide laser vaporization. J Dermatol Surg Oncol onychia congenita, or they may be inherited as an autosomal 6 1987; 123:299–301. dominant trait. As the father of the proband is also affected, 10 Kageyama N, Tope WD. Treatment of multiple eruptive vellus hair it is likely that our case is inherited. Around 25% of EVHCs cysts with the erbium: YAG laser. Dermatol Surg 1999; 25:819–22. resolve spontaneously.7 This may occur via transepithelial 7 elimination of cyst products. Most of the lesions will persist. Conflicts of interest: none declared. Reports suggested treatment success using different modalities including topical retinoic acid8 and topical 12% lactic acid.6 Incision and drainage did not result in significant improve- ment. Huerter and Wheeland reported successful cyst removal 9 on the face using CO2 laser. They used 5 W of power, irradi- ) ance of 160 W cm 2, a 2-mm spot size, and a pulse duration Angiomyofibroblastoma of the vulva with of 0.2 s for the facial lesions and 3 W of power and a penile appearance ) 100 W cm 2 irradiance for the eyelid lesions. Only slight hyperpigmentation occurred at some of the treated sites. There DOI: 10.1111/j.1365-2133.2007.07929.x was no hypertrophic scarring or regeneration of cysts noted. Pulsed erbium : yttrium–aluminium–garnet laser (2940 nm SIR, We describe a case of angiomyofibroblastoma with a wavelength) using a 2-mm spot size, 250-ls pulse duration at peculiar appearance. An 82-year-old woman presented with a ) 60.5–63.7 J cm 2, and 1.9–2 J pulse energy has also been 1-year history of an asymptomatic tumour on her right vulva. reported effective in treating EVHCs on the trunk.10 Three to The tumour had slowly enlarged. Physical examination five pulses were delivered to the lesions. The cysts were then revealed a polypoid, well-circumscribed, mobile and elastic expressed with digital pressure and extracted with forceps. soft tumour on the vulva, 35 · 20 mm in size. An ulcerative The base of the cyst wall was further ablated with three pulses change was recognized on top of the lesion. The tumour and the areas were allowed to heal by secondary intention. showed an unusual penis-like appearance (Fig. 1). There was There was also no hypertrophic scarring, hyperpigmentation no inguinal lymphadenopathy and the tumour was completely or recurrence of cysts reported at the telephone follow-up. excised under general anaesthesia. In summary, we present a young Chinese woman who had Histopathologically, the lesion was relatively well-circum- EVHCs initially thought to be melanotic or melanocytic dys- scribed. The tumour was characterized by both hypercellular pigmentation mimicking bilateral naevus of Ota. We propose and hypocellular areas with numerous thin-walled small blood that EVHCs should be added to the differential diagnosis of vessels (Fig. 2a,b). Most of the neoplastic cells were oval and facial dyspigmentation.
Department of Dermatology, K.H.N. CHAN Centre of Health Protection, W.Y.M. TANG Pamela Youde Nethersole Eastern Hospital, W.Y. LAM* Chai Wan, Hong Kong, China K.K. LO *Department of Pathology, Tuen Mun Hospital, Hong Kong, China E-mail: [email protected]; [email protected]
References
1 Esterly NB, Fretzin DF, Pinkus H. Eruptive vellus hair cysts. Arch Dermatol 1977; 113:500–3. 2 Watson A. Eruptive vellus hair cysts. Int J Dermatol 1982; 21:273–4. 3 Kwon KS. A case of generalized eruptive vellus hair cysts. J Dermatol Fig 1. The tumour showing a penis-like appearance on the right 1997; 24:556–7. vulva.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 190 Correspondence
(a) shown). During the 6-month follow-up period after surgery, there was no evidence of recurrence. Angiomyofibroblastoma is a benign soft tissue tumour ori- ginally described by Fletcher et al. in 1992.1 This rare tumour occurs mainly in the vulval region of middle-aged premeno- pausal women, and less frequently in the spermatic cord and scrotum of men. The clinical features of angiomyofibroblasto- ma are characterized by a well-circumscribed, firm nodule mostly located in the subcutaneous tissue. Clinically, it is often misdiagnosed as Bartholin’s cyst. However, the main differen- tial diagnosis is aggressive angiomyxoma,2 which is a locally infiltrative but nonmetastasizing tumour of pelvic soft tissue that has a high rate of recurrence. In contrast, angiomyofibro- blastoma is an apparently nonrecurring neoplasm that can be (b) adequately treated by simple excision, but is comparable with aggressive angiomyxoma. It is important for clinicians to dif- ferentiate carefully between these two tumours because of the differences in their clinical course and treatment. Histopatho- logically, angiomyofibroblastoma is better demarcated than aggressive angiomyxoma, and contains numerous small thin- walled vessels. The stromal cells are plump and oval in shape, whereas those in aggressive angiomyxoma are spindle-shaped or stellate. In addition, angiomyofibroblastoma also demon- strates a more varied cellularity.3 The immunohistochemical profile of angiomyofibroblastoma includes the almost uniform expression of vimentin and desmin, and a variable expression of a-smooth muscle actin and both oestrogen and prog- esterone receptors.4–6 Although the clinical and histological (c) features of both neoplasms sometimes overlap, aggressive angiomyxoma tends to be larger than angiomyofibroblastoma and it is also more infiltrative.1 The histogenesis of angiomyofibroblastoma remains unclear. It has been suggested that neoplastic cells are derived from mesenchymal cells in the subepithelial myxoid stromal zone which extends from the endocervix to the vulva.4,5 The case reported here is clinically peculiar because it occurred in an elderly postmenopausal woman and showed a very obvious penis-like appearance.
Division of Dermatology, T. WATANABE Department of Medicine of Sensory and Y. YOSHIDA Motor Organs, Faculty of Medicine, O. YAMAMOTO Fig 2. (a) Hypercellular area of the tumour with numerous thin- Tottori University, 86 Nishi-cho, walled small blood vessels and (b) hypocellular area (haematoxylin Yonago 683-8503, Japan and eosin; original magnification · 100). (c) Oval and spindle-shaped E-mail: [email protected] neoplastic cells with bipolar eosinophilic cytoplasmic processes (haematoxylin and eosin; original magnification · 400). References spindle-shaped with bipolar eosinophilic cytoplasmic processes 1 Fletcher CDM, Tsang WYW, Fisher C et al. Angiomyofibroblastoma (Fig. 2c). The oedematous stroma contained wavy collagen of the vulva. Am J Surg Pathol 1992; 16:373–82. fibres and variable numbers of inflammatory cells, mainly 2 Steeper TA, Rosai J. Aggressive angiomyxoma of the female pelvis lymphocytes. and perineum. Am J Surg Pathol 1983; 7:463–75. 3 van der Griend MD, Burda P, Ferrier AJ. Angiomyofibroblastoma of Immunohistochemically, the tumour cells stained positively the vulva. Gynecol Oncol 1994; 54:389–92. for vimentin, desmin, and a-smooth muscle actin (not 4 Granter SR, Nucci MR, Fletcher CDM. Aggressive angiomyxoma: shown). However, there was no positive reaction for S100 reappraisal of its relationship to angiomyofibroblastoma in a series protein, CD34, or oestrogen and progesterone receptors (not of 16 cases. Histopathology 1997; 30:3–10.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 191
5 Nielsen GP, Rosenberg AE, Young RH et al. Angiomyofibroblastoma (a) of the vulva and vagina. Mod Pathol 1996; 9:284–91. 6 Laskin WB, Fetsch JF, Tavassoli FA. Angiomyofibroblastoma of the female genital tract. Hum Pathol 1997; 28:1046–55.
Conflicts of interest: none declared.
Successful treatment of severe psoriatic arthritis with infliximab in an 11-year-old child suffering from linear psoriasis along lines of Blaschko
DOI: 10.1111/j.1365-2133.2007.07928.x
SIR, The case of an 11-year-old boy who suffered from severe psoriatic arthritis and concomitant linear psoriasis following the lines of Blaschko is presented. At 6 months of age skin lesions appeared for the first time and a clinical diagnosis of psoriasis was made, confirmed by a skin biopsy taken at the age of 6 years. Lesions appeared in a linear fashion following the lines of Blaschko. Mainly the left part of his body was affected; pits were present in most nails, as well as onycho- dystrophy at the left thumb and hallux. At the age of 3 years the patient was additionally diagnosed with psoriatic arthritis. The lesions persisted at the same sites with varying inten- (b) sity, showing only partial and temporary response to treat- ment. Psoriatic arthritis developed into a highly inflammatory type with painful swollen joints requiring methotrexate ) 5–10 mg m 2 body surface weekly and later ciclosporin ) 3–4 mg kg 1 as higher doses of methotrexate could not be tolerated by the patient. Skin lesions were treated with topical calcipotriol and corticosteroids, but complete remission was never achieved. When he was first seen at our clinic in November 2004 the patient was experiencing a severe polyarticular course of his psoriatic arthritis (left ankle and distal interphalangeal joint of the third toe tender and swollen, dactylitis of both halluces) ) despite systemic treatment with methotrexate 10 mg m 2 ) weekly together with ciclosporin 4 mg kg 1 daily. Treatment Fig 1. Clinical pictures of the patient before infliximab treatment was with etanercept 2 · 25 mg weekly was therefore initiated in started. ) combination with a weekly dose of methotrexate 10 mg m 2 body surface given subcutaneously, but the clinical picture remained unchanged until May 2005 and then aggravated sub- infusions of infliximab were given after 3 and 8 weeks, before stantially in June. In August 2005 the patient could not walk therapy was continued regularly every 8 weeks. In May 2006 without crutches because of swollen and tender joints, and an the patient suffered from a throat infection and developed a acute exacerbation of psoriasis of the skin required short-term flare of guttate psoriasis which ceased after the next infusion topical therapy with corticosteroids under plastic film occlu- and short-term systemic antibiotic treatment. To date the boy ) sion. Treatment with infliximab 5 mg kg 1 was started has received a total of 10 infusions over a period of ) together with methotrexate 10 mg m 2 body surface subcuta- 14 months, which were all well tolerated. neously weekly. Within 2 h of the first infusion, the arthralgia Routine laboratory tests and determination of antinuclear almost completely disappeared. The skin lesions also improved and antidouble-stranded DNA antibodies did not reveal any at first but were less responsive to treatment (Fig. 1). Further abnormalities at baseline and during therapy. C-reactive
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 192 Correspondence
) protein decreased from 41Æ5mgL 1 before infliximab treat- 4 Yu HJ, Ko JY, Kwon HM et al. Linear psoriasis with porokeratotic ) ment to < 3Æ0mgL 1 after the first infusion. eccrine ostial and dermal duct nevus. J Am Acad Dermatol 2004; Linear psoriasis is a rare phenomenon and its existence has 50:S81–83. 5 St Clair EW. Infliximab treatment for rheumatic disease: clinical been discussed for some time.1–4 Our diagnosis was based on and radiological efficacy. Ann Rheum Dis 2002; 61(Suppl. 2):ii67–9. the classic clinical picture of skin lesions with concomitant 6 Breedveld FC, Emery P, Keystone E et al. Infliximab in active early psoriatic arthritis and nail dystrophy. Skin biopsy was consist- rheumatoid arthritis. Ann Rheum Dis 2004; 63:149–55. ent with psoriasis and no typical signs of inflammatory linear 7 Gottlieb AB, Evans R, Li S et al. Infliximab induction therapy for verrucous epidermal naevus were present. The onset of guttate patients with severe plaque-type psoriasis: a randomized, double- psoriasis after a throat infection also confirms the diagnosis. blind, placebo-controlled trial. J Am Acad Dermatol 2004; 51:534–42. The coincidence of linear psoriasis and psoriatic arthritis has 8 Lahdenne P, Vahasalo P, Honkanen V. Infliximab or etanercept in the treatment of children with refractory juvenile idiopathic arth- not been reported in the literature to date. Clinical response of ritis: an open label study. Ann Rheum Dis 2003; 62:245–47. our patient to infliximab was dual. The arthritis symptoms 9 Weinberg JM, Saini R. Biologic therapy for psoriasis: the tumor completely disappeared under therapy and have not re-occurred necrosis factor inhibitors infliximab and etanercept. Cutis 2003; but improvement of the skin lesions remains delayed. 71:25–9. Tumour necrosis factor (TNF)-a antagonists are a relatively 10 Chew AL, Bennett A, Smith CH et al. Successful treatment of severe new therapeutic option for the treatment of severe psoriatic psoriasis and psoriatic arthritis with adalimumab. Br J Dermatol arthritis and chronic plaque-type psoriasis. Treatment with 2004; 151:492–96. these agents is well established for rheumatoid arthritis and Crohn’s disease and in recent years TNF-a inhibitors have Conflicts of interest: none declared also been efficacious in the treatment of juvenile idiopathic arthritis.5–10 To date, no clinical studies on the efficacy and safety of inf- liximab in children with psoriatic arthritis or chronic plaque- type psoriasis have been published. Previous data about the efficacy of TNF-a antagonists in the treatment of children with Two siblings with neonatal pemphigus vulgaris refractory juvenile idiopathic arthritis showed a significant associated with mild maternal disease improvement of articular signs and symptoms and indicate that these drugs might offer an important clinical advance in DOI: 10.1111/j.1365-2133.2007.07927.x the treatment of paediatric rheumatoid diseases in the future.8 It has also been shown that TNF-a inhibitors in combination SIR, Pemphigus vulgaris (PV) is a chronic immune-mediated, with methotrexate can arrest the progression of structural blistering disease which affects the skin and mucous mem- 5,6 damage and even improve radiographic joint damage. branes. Neonatal PV is caused by transplacental transfer of IgG Therefore patients with severe forms of arthritis and rapid dis- antibodies from a mother with PV to the fetus. As the onset of ease progression should benefit from early treatment with PV occurs predominantly in the fourth to sixth decades of life, it these drugs, which may lead to significant improvements in is rare to find patients with PV who become pregnant or have an 6 functional disability and quality of life. Long-term safety and abortion.1 In one-third of those who become pregnant, the neo- efficacy, however, remain unknown at present and more clin- nates have PV lesions.2 We report two siblings with neonatal ical studies are needed to prove the currently documented low PV. We also discuss the differences between mothers and neo- degree of side-effects over a longer period of time. nates in clinical phenotype and expression of PV antibodies. The mother was a 30-year-old Japanese woman who had Psoriasis Center at the Department of Dermatology, S. ROTT not been pregnant previously. Two years 4 months before giv- University of Kiel, Schittenhelmstr. 7, R.M. KU¨ STER* ing birth to her first child, she developed oral erosions and 24105 Kiel, Germany U. MROWIETZ was subsequently referred to our institute. A diagnosis of PV *Department of Paediatric Rheumatology, was made on the basis of an oral biopsy. One year 3 months Rheumaklinik Bad Bramstedt, Bad Bramstedt, after the diagnosis, the eruptions had spread over about 30% Germany of her entire skin surface. As a result, she was treated with E-mail: [email protected] prednisolone 60 mg daily after steroid pulse therapy. The ster- oid dosage was then gradually tapered, and thereafter she pre- References sented with only a few oral erosions which were controlled with prednisolone 12Æ5 mg daily. In Februtary 2001, she 1 Raza N, Iqbal P, Anwer J. Unilateral psoriasis along Blaschko lines. became pregnant with her first child. During her pregnancy, J Ayub Med Coll Abbottabad 2005; 17:87–8. the oral erosions worsened slightly. 2 Ginarte M, Fernandez-Redondo V, Toribio J. Unilateral psoriasis: a After the birth, her 2-day-old baby boy (baby 1) presented case individualized by means of involucrin. Cutis 2000; 65:167– 70. with skin blisters and erosions on the neck, chest and hands, 3 Ghorpade A. Linear naevoid psoriasis along lines of Blaschko. J Eur but there was no mucosal involvement (Fig. 1a). No other Acad Dermatol Venereol 2004; 18:726–27. abnormalities were identified on physical examination. The
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 193
(a) (b)
(c)(d) (e)
Fig 1. (a) Bullae on neck of baby 1 at day 2. Skin erosions and bullae on the back (b), left wrist (c) and right lower extremity (d) of baby 2 at day 1. (e) An indirect immunofluorescence examination of the second neonate’s serum diluted 1 : 20. baby had Apgar scores of 9 at 1 min and 10 at 5 min. A skin examination. Direct IMF of her skin showed bright intercellular biopsy was not performed, but indirect immunofluorescence staining with IgG. The pemphigus antibody titres shown by (IMF) of the infant’s serum showed circulating IgG antibodies indirect IMF of her serum were positive at 1 : 20 (Fig. 1e). (at a titre of 1 : 40) which were observed to bind to the Anti-Dsg3 and Dsg1 IgG ELISAs showed index values of 121 and intercellular spaces. An enzyme-linked immunosorbent assay 9 for Dsg3 and Dsg1, respectively. A diagnosis of neonatal PV (ELISA) for antidesmoglein (Dsg) 3 IgG and anti-Dsg1 IgG was established for both babies. The lesions observed in both showed index values of 143 and 10 for Dsg3 and Dsg1, children healed completely within 1 week, with decreases in respectively (Table 1). indirect IMF titres and Dsg3 ELISA index value. In Feburary 2002, she became pregnant with a baby girl Neonatal PV is caused by the transplacental transfer of IgG (baby 2). Baby 2 had Apgar scores of 8 at 1 min and 9 at 5 min. antibodies from a mother with PV to her fetus. In 1983, She had walnut-sized cutaneous bullae and erosions on the Kaplan and Callen reported that about 35% of neonates born neck, back and extremities (Fig. 1b–d). She also had palm-sized to mothers with PV had PV lesions,2 referred to as neonatal erosions over an area from her groin to her anal region, and one PV. However, it is very rare that patients with PV become oral ulceration. No other abnormalities were found on physical pregnant and have successful deliveries, as the onset of PV
Table 1 Indirect antibody titres and enzyme- linked immunosorbent assay index values for Mother antidesmoglein (Dsg) 1 and Dsg3 IgG in the (at delivery) Baby 1 Baby 2 serum of the mother and neonates Baby 1 Baby 2 Day 2 Day 35 Day 2 Day 7 Indirect antibody titre 1 : 40 1 : 160 1 : 40 1 : 5 1 : 20 Not done Anti-Dsg1 IgG 45 116 10 Not done 9 5 Anti-Dsg3 IgG >200 >200 143 Not done 121 53
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 194 Correspondence occurs predominantly in the fourth to sixth decades of life.1 3 Chowdhury MMV, Natarajan S. Neonatal pemphigus vulgaris asso- Such a case is therefore considered to be highly unusual. A ciated with mild oral pemphigus vulgaris in the mother during number of reports have described no relationship between the pregnancy. Br J Dermatol 1998; 139:500–3. 4 Campo-Voegeli A, Muniz F, Mascaro JM Jr et al. Neonatal pemphigus clinical presentation in the mothers and the neonates.3,4 They vulgaris with extensive mucocutaneous lesions from a mother with also report that antibody titres and clinical presentation in the oral pemphigus vulgaris. Br J Dermatol 2002; 147:801–5. 3,4 mothers with PV cannot predict the severity of neonatal PV. 5 Middelkamp Hup J, Bruinsma RA, Boersma ER et al. Neonatal pem- In our case, while the mother only had a few oral erosions phigus vulgaris: transplacental transmission of antibodies. Pediatr during her two pregnancies, the neonates presented with skin Dermatol 1986; 3:468–72. bullae and erosions at birth. In addition, when comparing the 6 Parlowsky T, Welzel J, Amagai M et al. Neonatal pemphigus vul- two neonates, the disease condition of the second baby was garis: IgG4 autoantibodies to desmoglein 3 induce skin blisters in newborns. J Am Acad Dermatol 2003; 48:623–5. much more severe than that of the first. Therefore, we con- 7 Mercelo HG, Zamora E, Avinoach I et al. Pemphigus neonatorum. sider that our case supports the findings of the reports Pediatr Dermatol 1993; 10:169–70. mentioned above. Regarding the ELISA index values of the 8 Tope WD, Kamino H, Briggaman RA et al. Neonatal pemphigus neonates, although baby 2 had a lower level of Dsg3, her vulgaris in a child born to a mother in remission. J Am Acad Dermatol symptoms were more severe than those of her brother. This 1993; 29:480–5. suggests that a lower Dsg3 antibody titre in neonates is not 9 Ruach M, Ohel G, Rahav D et al. Pemphigus vulgaris and preg- necessarily associated with a lower disease severity. Since nancy. Obstet Gynaecol Surv 1995; 50:755–60. 10 Wu H, Wang ZH, Yan A et al. Protection against pemphigus folia- 1966 there have been 45 reports worldwide of the birth of ceus by desmoglein 3 in neonates. N Engl J Med 2000; 343:31–5. children from patients with PV, including our two cases. In 18 of these cases, the babies were born with PV lesions. Conflicts of interest: none declared. Moreover, in nine cases,3–6 the mothers had only mucous membrane lesions, while the babies had some skin lesions. Three other cases have been reported of neonates with PV who had severe skin lesions but whose mothers had no active disease.7–9 There are no other reports of siblings with neonatal PV. Our case is therefore considered to be very rare but 4 An infantile case of pityriasis lichenoides similar to the cases reported by Campo-Voegeli et al. and et varioliformis acuta Parlowsky et al.6 In these reports, anti-Dsg3 and Dsg1 IgG ELISA index values were also studied. In both cases, while DOI: 10.1111/j.1365-2133.2007.07931.x the mothers had high index values for anti-Dsg3 IgG and anti-Dsg1 IgG, the neonates had a high index value for only SIR, Pityriasis lichenoides et varioliformis acuta (PLEVA) is anti-Dsg3 IgG, as in our cases. These observations support the uncommon among children, and infantile cases are remarkably findings of previous studies,8,9 and suggest that the distribu- rare. We here present an infantile case of PLEVA. tion of Dsg3 in the neonatal epidermis is different from that An 11-month-old Japanese boy was referred to our hospital in the adult epidermis and is more similar to its distribution with eruptions on his upper arms. These lesions had appeared in the adult mucous membrane. Subcorneal keratinocytes in 2 weeks previously and had spread over his whole body. newborn infants contain both Dsg1 and Dsg3, whereas the There was no apparent episode of TORCH (Toxoplasma gondii, suprabasal layers contain much lower Dsg1 levels, with Dsg3 syphilis, rubella virus, cytomegalovirus or herpes simplex constituting most of the strength of the intercellular desmo- virus) infection associated with congenital abnormalities from somal bridges.10 This may be why antibody to Dsg3 induced maternal infection. Until 11 months of age, he had been gen- severe blisters in our cases.10 erally healthy and had shown no symptoms suggesting immu- nodeficiency. He had no medical history of herpes simplex or Department of Dermatology, Graduate School, T. UGAJIN varicella-zoster virus infection and had had no apparent epi- Tokyo Medical and Dental University, H. YAHARA sode of human herpesvirus type 6 infection. He had been vac- 1-5-45 Yushima, Bunkyo-ku, Y. MORIYAMA cinated against tuberculosis, diphtheria, pertussis, tetanus, Tokyo 113-8519, Japan T. SATO polio, measles, rubella and influenza, and no infection by the Correspondence: Hiroo Yokozeki. K. NISHIOKA pathogens or side-effects of vaccination were seen. Upon E-mail: [email protected] H. YOKOZEKI physical examination, there were numerous, dark-reddish areas of erythema on his trunk, arms and legs (Fig. 1a). Each References patch of erythema ranged from 3 to 10 mm in diameter and some of them were erosive and partially covered with a black- 1 Ouahes N, Tabarak AQ, Razzaque A. Infertility in women with ish crust. He had slight pruritus. We performed a skin biopsy pemphigus vulgaris and other autoimmune diseases. J Am Acad Dermatol 1997; 36:383–7. from a fresh palpable area of erythema on his buttock. Histo- 2 Kaplan RP, Callen JP. Pemphigus associated diseases and induced pathological observations demonstrated focal parakeratosis in pemphigus. Clin Dermatol 1983; 1:42–71. the stratum corneum, and spongiosis and satellite cell necrosis
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 195
(a) (b)
Fig 1. (a) Numerous reddish macules and papules partially covered with a blackish crust were found on the back. (b) The eruptions had almost disappeared after 2 months, leaving areas of pigmentation. in the basal and spinous cell layers of the epidermis. The Both CD4+ cells and CD8+ cells were seen, although CD8+ dermoepidermal junction was obscured due to lymphocytic cells were predominant (Fig. 2c,d). Only a small number of infiltration, and oedema of the dermal papillae was found CD20+ cells and CD30+ cells was found. Direct immunofluo- (Fig. 2a,b). Infiltrating lymphocytes were small without rescence staining revealed IgM deposition on the vessel walls nuclear atypia. Immunohistological investigation on the infil- in the upper dermis. These findings led to the diagnosis of trating cells demonstrated a large number of CD45RO (LCA)- PLEVA. Other papulonecrotic erythematous disorders, such as positive cells at the dermoepidermal junction and perivascular varicella, papulonecrotic tuberculid and lymphomatoid papu- regions. CD3+ cells were observed mainly in the superficial losis, were raised as differential diagnoses and the diagnosis dermis close to the dermoepidermal junction, and CD68+ of PLEVA was established by the clinical course and histo- cells were seen in perivascular regions in the upper dermis. pathological features. We started treatment with topical
(a) (b)
(c) (d) Fig 2. Histopathological features of a skin lesion. (a) Focal parakeratosis within the cornified layer, and spongiosis. The dermoepidermal junction was obscured due to lymphomatoid cell exocytosis. (b) Satellite cell necrosis (arrowhead) and oedema of dermal papillae (haematoxylin and eosin; original magnification: a, · 40; b, ·100). (c, d) Immunohistological staining revealed that CD8+ T cells (c) predominated over CD4+ T cells (d) among the infiltrating cells at the dermoepidermal junction (original magnification: c, d, · 100).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 196 Correspondence corticosteroid ointment (betamethasone valerate) and oral References antihistamine, and his eruptions slowly improved. After 1 Weinberg JM, Kristal L, Chooback L et al. The clonal nature of pity- 2 months of the treatment, his erythema had completely riasis lichenoides. Arch Dermatol 2002; 138:1063–7. cleared leaving only postinflammatory pigmentation (Fig. 1b). 2 Gelmitti C, Rigoni C, Alessi E et al. Pityriasis lichenoides in children: Pityriasis lichenoides is an erythematous, papulosquamous, a long-term follow-up of eighty-nine cases. J Am Acad Dermatol 1990; T cell-mediated dermatosis.1 Two subtypes are known: the 23:473–8. acute type, i.e. classical PLEVA or Mucha–Habermann disease, 3 Rogers M. Pityriasis lichenoides and lymphomatoid papulosis. Semin and a chronic subtype, i.e. pityriasis lichenoides chronica Dermatol 1992; 11:73–9. (PLC). The disease duration in PLEVA is not always shorter 4 Dupont C. Pityriasis lichenoides in a family. Br J Dermatol 1995; 133:338–9. than that in PLC.2 The underlying cause of PLEVA is unknown, 5 Klein PA, Jones EC, Nelson JL et al. Infectious causes of pityriasis but various infectious agents have been implicated based on a lichenoides: a case of fulminant infectious mononucleosis. J Am Acad cluster of cases, familial outbreaks and elevated serum titres to Dermatol 2003; 49:S151–3. inciting antigens, such as Epstein–Barr virus, T. gondii and 6 Weiss LM, Wood GS, Ellisen LW et al. Clonal T-cell populations human immunodeficiency virus.3–5 Several reports have dem- in pityriasis lichenoides et varioliformis acuta (Mucha–Habermann onstrated clonal T-cell receptor rearrangements from PLEVA disease). Am J Pathol 1987; 126:417–21. patient specimens, suggesting that PLEVA is a lymphoprolifera- 7 Buckley RH. The immunologic system and disorders. In: Nelson Textbook of Pediatrics (Behrman RE, Kliegman RM, Jenson HB, eds), tive disorder involving T-cell subtypes, despite its clinically 17th edn. Philadelphia: Elsevier, 2004; 688. benign course and the absence of morphologically atypical 1,6 cells in the skin lesions. Conflicts of interest: none declared. Although PLEVA can be found in patients at any age, it occurs predominantly in the second and third decades of life.5 PLEVA is an uncommon skin disorder in childhood, and infantile cases are extremely rare. As far as we know, only one infantile case of PLEVA has been reported in the English language literature.2 Gelmitti et al.2 reported an 8-month-old boy Gomez–Lopez–Hernandez syndrome: another with PLEVA, whose lesions continued for more than 8 years. consideration in focal congenital alopecia Our case is one of the first reports of this rare disease in infancy. As a common characteristic in postnatal lymphopoiesis, DOI: 10.1111/j.1365-2133.2007.07952.x virtually all T cells in cord blood are derived from the CD45RA (naive) isoform, and a dominance of CD45RA over CD45RO T SIR, Gomez–Lopez–Hernandez syndrome (GLHS), also known cells persists during the first 2–3 years of life. Then, the num- as cerebellotrigeminal-dermal dysplasia (OMIM 601 853), is bers of cells bearing these two isoforms gradually equalize.7 a rare neurocutaneous disorder. Features include cerebellar Cord blood T cells have the capacity to develop an antigen- malformation, trigeminal anaesthesia, focal scalp alopecia and specific T-cell response from birth, as shown by the vigorous craniosynostosis. We report a case of GLHS, initially referred tuberculin reactivity obtained at a few weeks following bacille as possible temporal triangular alopecia (TTA). Calmette–Gue´rin vaccination even on day 1 of life. These facts The patient is the second son of healthy unrelated parents. indicate that, in general, the T cell-mediated immunity of the Antenatal scans detected hydrocephalus, and delivery was by infant is immature, but sufficient to be effective. emergency caesarean section at term for fetal distress. Bitem- In the present case, CD45RO (LCA)-positive cells were pre- poral alopecia was present from birth. During the first year dominant among infiltrating cells in the skin lesion. This result of life he had hypotonia and plagiocephaly. Chromosomes, may indicate that T cell-mediated immunity in this patient is metabolic screening and head ultrasound scan were normal. comparatively mature. In addition, a large number of CD8+ His gross motor milestones were delayed and he received cells was seen in the skin lesion, similar to the immunohisto- developmental physiotherapy. He had strabismus requiring logical features shown in PLEVA in adults.1 These findings patching and recurrent middle ear effusions requiring grom- may suggest that, in this infantile case, skin lesions of PLEVA mets, but his hearing was normal. Between the ages of 2 were not a result of an immature immune system or immuno- and 7 years the patient had several generalized febrile con- logical abnormalities such as a skewed distribution of periph- vulsions each lasting less than 2 min. He stopped physiother- eral T cells, but were caused by pathomechanisms similar to apy at 4Æ5 years as he was functioning normally for his age, those in adult cases. although with poor balance. He attended mainstream school where he was academically above average, and participated Department of Dermatology, D. HOSHINA in rugby and cricket. At 9 years he sustained a mild head Hokkaido University Graduate School of Medicine, M. AKIYAMA injury with subsequent double vision and headaches, which North 15 West 7, Kita-ku, Sapporo 060-8638, Japan K. HAMASAKA* gradually resolved. *Hamasaka Skin Clinic, Sapporo, Japan H. SHIMIZU Examination revealed a slim 10-year-old boy with height E-mail: [email protected] 129 cm (2nd centile), weight 22Æ6kg(0Æ4th centile) and
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 197
The patient’s brother also had a small oval patch of tem- poral alopecia on the right, present since birth. He had no brachycephaly or neurological symptoms. MRI of the head was normal. GLHS was first described by Gomez in 1979, in a girl with brachycephaly, interstitial keratitis, analgesia of the face and cornea, focal congenital alopecia of the occiput and posterior scalp, hypotonia, ataxia and moderate mental retar- dation.1 Subsequently a further 11 patients have been repor- ted with GLHS.2–5 Cerebellar abnormalities are reported in all cases, with 11 of the 12 cases having rhombencephalo- synapsis. Supratentorial abnormalities are often also present, including ventricular dilatation, cortical thinning, dysgenesis of the corpus callosum, absent septum pellucidum, and arachnoid cysts. Mental ability is usually impaired, although a case has been reported where the patient had significant gross motor delay in early childhood followed by good academic performance at school, similar to our patient.2 Self- injurious behaviour is reported, but this may be attributable to trigeminal anaesthesia. One patient has been described with growth hormone deficiency, hyperactivity and bipolar disorder.3 Rhombencephalosynapsis consists of absence of the cere- bellar vermis, fusion of the dentate nuclei and fusion of the cerebellar hemispheres in the midline. Deficiency of the sep- tum pellucidum, dysgenesis of the corpus callosum, fused thalami and fused fornices may also be present.6 It is rare, Fig 1. Lateral views of the patient illustrating bitemporal alopecia and but increasingly detected since the advent of MRI. It has brachycephaly. The alopecia is nonscarring, with a central tuft. also been detected on antenatal ultrasound.4 Clinically, rhombencephalosynapsis results in variable degrees of motor head circumference 50Æ5 cm (2nd centile) with brachycephaly. and cognitive deficit. The degree of mental impairment He had patches of temporal alopecia bilaterally. Each had reflects the extent of associated supratentorial abnormalities. absent hair follicles and a central tuft of darker hair, but no Rhombencephalosynapsis has been found incidentally in a scarring (Fig. 1). Neurological examination, including facial 55-year-old man with normal intelligence and neuropsychi- sensation and corneal reflexes, was normal. atric testing,7 suggesting that it may remain undetected into Magnetic resonance imaging (MRI) of the head showed an adult life. absent cerebellar vermis and fusion of the cerebellar hemi- Scalp alopecia in GLHS is usually symmetrical and may spheres consistent with rhombencephalosynapsis, and two involve the frontal, temporal, parietal or occipital regions in arachnoid cysts (Fig. 2). a band-like pattern.4 TTA is characterized by unilateral or
(a) (b)
Fig 2. Horizontal magnetic resonance images of the brain. (a) A normal brain; (b) the patient’s scan showing rhombencephalo- synapsis with absence of the cerebellar vermis and fusion of the cerebellar hemispheres (arrow).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 198 Correspondence bilateral lancet- or oval-shaped patches of alopecia on the frontotemporal region of the scalp. The lesions are usually A case of phosphaturic mesenchymal tumour recognized within the first few years of life, and are non- (mixed connective tissue variant) that scarring, permanent, and not associated with neurological developed in the subcutaneous tissue of a deficit.8,9 There is one report of TTA in a girl with epi- patient with oncogenic osteomalacia and lepsy, mental retardation and a Dandy-Walker malforma- produced fibroblast growth factor 23 tion.10 Both TTA and GLHS are usually sporadic disorders, although some familial cases of TTA have been reported. DOI: 10.1111/j.1365-2133.2007.07940.x Interestingly, our patient’s brother had a lesion consistent with TTA but no other signs of GLHS. Other causes of SIR, Oncogenic osteomalacia (OOM) is an extremely rare bitemporal alopecia include Setleis syndrome and focal facial osteomalacia associated with tumours. It is characterized by dermal dysplasia; however, both result in scar-like defects. the presence of a tumour, hypophosphataemia caused by renal Setleis syndrome is also associated with leonine facies, low phosphate wasting, and low serum concentration of 1,25- hairline and abnormalities of the eyelashes, eyebrows and dihydroxyvitamin D [1,25(OH) D ] with clinical and histo- eyelids. 3 2 3 logical evidence of osteomalacia.1 As removal of the tumour GLHS should be considered among the differential diagno- results in rapid resolution of the symptoms and signs of ses of patients with focal congenital alopecia. As some patients OOM, it is believed that a humoral phosphaturic factor with underlying structural abnormalities of the cerebellum are derived from the tumour is responsible for these abnormal- relatively asymptomatic it is possible that alopecia may be the ities.1 In 2001, fibroblast growth factor 23 (FGF23) was iden- initial presenting feature. Neuroimaging should be considered tified as a causative factor of OOM.2 A high level of FGF23 for those with abnormal neurological findings or abnormal is detected in either the tumour3,4 or serum from patients skull shape. with OOM.5 It is considered that overproduction of FGF23 by tumours induces hypophosphataemia by inhibiting the Department of Dermatology, D.J. PURVIS re-absorption of phosphate in the kidney, although the exact Great Ormond Street Hospital for Children, A. RAMIREZ molecular mechanism is not clear. It is hypothesized that London WC1N 3JH, U.K. N. ROBERTS FGF23 is also secreted by one or more normal tissues as a E-mail: [email protected] J.I. HARPER phosphate-regulating hormone, as a low level of FGF23 is detected in healthy people.6 References Most OOM-associated tumours have been diagnosed as 1 Gomez MR. Cerebellotrigeminal and focal dermal dysplasia: a newly benign mesenchymal tumours such as haemangiopericytoma, 7,8 recognised neurocutaneous syndrome. Brain Dev 1979; 1:253–6. haemangioma, giant cell tumour or osteoblastoma. Several 2 Munoz MV, Santos AC, Graziadio C, Pina-Neto JM. Cerebello- investigators, however, have described that OOM-associated trigeminal-dermal dysplasia (Gomez–Lopez–Hernandez syndrome): mesenchymal tumours were in most cases histologically description of three new cases and review. Am J Med Genet 1997; distinctive and unlike any other known mesenchymal tumours, 72:34–9. and have suggested that most OOM-associated tumours are of a 3 Brocks D, Irons M, Sadeghi-Najad A et al. Gomez–Lopez–Hernandez single histopathological type called phosphaturic mesenchymal syndrome: expansion of the phenotype. Am J Med Genet 2000; 94:405–8. tumour, mixed connective tissue variant (PMTMCT). 4 Tan TY, McGillivray G, Goergen SK, White SM. Prenatal magnetic We report a case of PMTMCT that developed as a subcuta- resonance imaging in Gomez–Lopez–Hernandez syndrome and neous nodule on the back of a patient with OOM. The patient review of the literature. Am J Med Genet 2005; 138A:369–73. was a 47-year-old man who presented with typical clinical 5 Bowdin S, Phelan E, Watson R et al. Rhombencephalosynapsis present- and biochemical features of osteomalacia. Laboratory tests ing antenatally with ventriculomegaly ⁄hydrocephalus in a likely case (Table 1) showed hypophosphataemia, hyperparathyroidism of Gomez–Lopez–Hernandez syndrome. Clin Dysmorphol 2007; 16:21–5. and low 1,25(OH) D . An increase in alkaline phosphatase 6 Utsunomiya H, Takano K, Ogasawara T et al. Rhombencephalosyna- 2 3 psis: cerebellar embryogenesis. Am J Neuroradiol 1998; 19:547–9. was also noted. An elastic hard subcutaneous tumour of about 7 Bell BD, Stanko HA, Levine RL. Normal IQ in a 55-year-old with 3 cm in diameter was noted in the right side of his lower newly diagnosed rhombencephalosynapsis. Arch Clin Neuropsychol back. The surface skin was erythematous and slightly dome 2005; 20:613–21. shaped. On the basis of these findings the patient was diag- 8 Tan E, Ng M, Giam YC. Temporal triangular alopecia: report of nosed as having OOM. The tumour was surgically removed five cases in Asian children. Pediatr Dermatol 2002; 19:127–8. en bloc. After surgery, the patient’s clinical abnormalities gradu- 9 Elmer KB, George RM. Congenital triangular alopecia: a case report ally improved and at 10 weeks after surgery the pain was and review. Cutis 2002; 69:255–6. 10 Ruggieri M, Rizzo R, Pavone P et al. Temporal triangular alopecia almost eradicated. The clinical response after surgery is shown in association with mental retardation and epilepsy in a mother in Table 1. The levels of serum phosphate and 1,25(OH)2D3 and daughter. Arch Dermatol 2000; 136:426–7. became normal immediately after surgery. FGF23 in serum samples taken before and after surgery was assayed at the Conflicts of interest: none declared. same time, as described previously9 (Table 1). FGF23 was
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 199
Table 1 Laboratory findings before and after resection
After resection
Before resection Day 1 Day 2 Day 5 Day 7 Day 14 Normal range ) Phosphate (mg dL 1)1Æ21Æ61Æ92Æ72Æ93Æ81Æ4–4Æ5 ) Calcium (mg dL 1)8Æ88Æ49Æ09Æ48Æ98Æ68Æ4–9Æ9 ) Parathyroid hormone (pg mL 1) 98 ND ND ND 88 ND 10–65 )1 1,25-dihydroxyvitamin D3 (pg mL ) 8 ND ND ND 129 ND 20–60 )1 25-hydroxyvitamin D3 (ng mL )11Æ5NDND12Æ9NDND9Æ0–33Æ9 ) Alkaline phosphatase (IU L 1) 832 646 656 712 679 679 109–321 ) Fibroblast growth factor 23 (pg mL 1) 220 30 21 36 35 37 10–50
ND, not determined.
elevated preoperatively, while on the day following surgery (Fig. 1a). They were round in shape with ovoid nuclei, scanty the FGF23 concentration fell to within the normal range. cytoplasm, and indistinct cellular boundaries. Neither cyto- The tumour was embedded in subcutaneous fat tissue. It logical atypia nor atypical mitotic figures were seen. Slight was a well-circumscribed and encapsulated mass that meas- calcification was observed in the matrix and intercellular ured 2Æ8 · 1Æ0 · 1Æ0 cm. The cut surface was grey-white with spaces (Fig. 1a, arrow). Many foci of eosinophilic chondroid partial areas of prominent haemorrhage. Histopathologically, matrix were observed in the periphery of the tumour a nodular lesion with a clear boundary was present in the (Fig. 1b), in which some cells with vacuolated cytoplasm subcutaneous area. The tumour cells were embedded at low were scattered (Fig. 1c, white arrows). A small number of density within an eosinophilic and distinctive smudgy matrix osteoclast-like giant cells was also present in the matrix
(a) (b)
Fig 1. Histopathological findings (a–d) and (c) (d) immunohistochemical analysis of fibroblast growth factor 23 (FGF23)-producing cells in tumour tissues (e, f). (a) Typical tumour cells (haematoxylin and eosin; original magnification · 400). (b) Many foci of eosinophilic chondroid matrix in the periphery of the tumour (haematoxylin and eosin; original magnification · 100). (c) Magnified image of (b) (haematoxylin and eosin; original magnification · 400). (d) Prominent haemorrhage and vascular (e) (f) structures (haematoxylin and eosin; original magnification · 40). (e) Deparaffinized sections of isolated tumour were examined immunohistochemically using antihuman FGF23 antibody as described previously.10 FGF23 immunoreactivity was present in tumour cells (original magnification · 100). (f) Sections treated without primary antibody were used as negative controls (original magnification · 100).
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 200 Correspondence
(Fig. 1c, black arrows). There were numerous blood vessels in overexpressed by tumors that cause phosphate wasting. J Clin the peripheral region of the tumour (Fig. 1d), especially Endocrinol Metab 2001; 86:497–500. where matrix deposition was absent. Many of these blood ves- 4 Nelson AE, Bligh RC, Mirams M et al. Fibroblast growth factor 23: a new clinical marker for oncogenic osteomalacia. J Clin Endocrinol sels were dilated and showed a large sinusoidal configuration. Metab 2003; 88:4088–94. The major histological features of PMTMCT described by Folpe 5 Takeuchi Y, Suzuki H, Ogura S et al. Venous sampling for fibroblast 8 et al. are: (i) spindle- to stellate- or round-shaped tumour growth factor-23 confirms preoperative diagnosis of tumor- cells with normochromatic, small nuclei and indistinct nuc- induced osteomalacia. J Clin Endocrinol Metab 2004; 89:3979–82. leoli, (ii) myxoid to myxochondroid matrix, (iii) calcification 6 Jonsson KB, Zahradnik R, Larsson T et al. Fibroblast growth factor with osteoclast-like giant cells, and (iv) prominent blood ves- 23 in oncogenic osteomalacia and X-linked hypophosphatemia. sel formation with a pattern similar to that of a haemangio- N Engl J Med 2003; 348:1656–63. 7 Weidner N, Santa Cruz D. Phosphaturic mesenchymal tumors. pericytoma. Our patient showed all of these features and thus Cancer 1987; 59:1442–54. we diagnosed him as having PMTMCT associated with OOM. 8 Folpe AL, Fanburg-Smith JC, Billings SD et al. Most osteomalacia- Immunohistochemically, the tumour cells showed reactivity associated mesenchymal tumors are a single histopathologic entity. for vimentin (data not shown). The tumour cells were not Am J Surg Pathol 2004; 28:1–30. reactive with S100, CD34 or epithelial membrane antigen 9 Yamazaki Y, Okazaki R, Shibata M et al. Increased circulatory level (data not shown). The tumour matrix contained variable of biologically active full-length FGF-23 in patients with hypo- amounts of mucus, which stained with alcian blue (data not phosphatemic rickets ⁄osteomalacia. J Clin Endocrinol Metab 2002; 87:4957–60. shown). Blood vessels were not reactive with a-smooth mus- 10 Kobayashi K, Imanishi Y, Koshiyama H et al. Expression of FGF23 cle actin. To determine whether the tumour expressed FGF23 is correlated with serum phosphate level in isolated fibrous protein, immunohistochemical analysis using antihuman dysplasia. Life Sci 2006; 78:2295–301. FGF23 antibody was performed as described previously.10 The majority of tumour cells stained with FGF23 (Fig. 1e). No Conflicts of interest: none declared. FGF23 expression was observed in the smudgy matrix. Our case supports the involvement of FGF23 produced from PMTMCT in the pathogenesis of OOM. Identification of the origin of the tumour cells and elucidation of the mechanism of high expression of FGF23 in the tumour cells are necessary for understanding the pathophysiology of PMTMCT. Unilateral periorbital oedema due to sarcoid infiltration of the eyelid: an unusual presentation of sarcoidosis with facial nerve Acknowledgments palsy and parotid gland enlargement We greatly thank Dr Keisuke Kobayashi (Osaka City University Graduate School of Medicine) for his technical assistance in DOI: 10.1111/j.1365-2133.2007.07939.x measuring FGF23 and immunohistochemical staining. SIR, Causes of unilateral eyelid swelling include cellulitis, angio- Divisions of Dermatology and M. OKA oedema, granulomatous blepharitis and malignant lymphoma. *Endocrinology ⁄Metabolism, T. KAMO However, cases of sarcoidosis that present with unilateral eyelid Neurology and Hematology ⁄Oncology, E. SASAKI swelling are rare. We report a patient with sarcoidosis who Department of Clinical Molecular Medicine, H. KAJI* had a remarkable, unilateral eyelid swelling. Interestingly, the Kobe University Graduate School of Medicine, H. NISHIZAWA* patient also had facial nerve palsy in the first branches of Kobe 650-0017, Japan Y. IMANISHI the right trigeminal nerve resulting facial weakness, in just the Department of Metabolism, Endocrinology C. NISHIGORI same area as the skin involvement. and Molecular Medicine, Osaka City A 54-year-old man was initially seen with a 5-month his- University Graduate School of Medicine, tory of persistent and asymptomatic eruptions on his right Osaka 545-8585, Japan eyelid with low-grade fever and malaise. Physical examination E-mail: [email protected] revealed remarkable swelling on the right eyelid with a brown-coloured erythema but without heat, tenderness or References pruritus (Fig. 1a). Wrinkling around the right side of the fore- head was indistinct. In addition, he had a right parotid gland 1 Nelson AE, Robinson BG, Mason RS. Oncogenic osteomalacia: is swelling. The initial clinical diagnosis was granulomatous ble- there a new phosphate regulating hormone? Clin Endocrinol 1997; pharitis. Histopathological examination of his upper right eye- 47:635–42. lid showed numerous noncaseating granulomas in the dermis, 2 Shimada T, Mizutani S, Muto T et al. Cloning and characterization composed of epithelioid cells with a few scattered lympho- of FGF23 as a causative factor of tumor-induced osteomalacia. Proc Natl Acad Sci USA 2001; 98:6500–5. cytes and histiocytes surrounding the granulomas (Fig. 2a, b). 3 White KE, Jonsson KB, Carn G et al. The autosomal dominant Laboratory data showed elevated levels of angiotensin- ) hypophosphatemic rickets (ADHR) gene is a secreted polypeptide converting enzyme (ACE) at 29Æ7UL1 (normal 8Æ3–21Æ5),
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 201
ab
Fig 1. (a) Upon admission, the right upper and lower eyelid and right cheek were swollen. There was slight contraction of the right frontalis muscle on attempting to wrinkle the forehead. (b) After systemic prednisolone therapy. The eyelid and cheek swelling had disappeared, enabling the right eye to open.
upper mediastinum, both sides of the aorta, and bilateral (a) inguinal lymph nodes, whereas no bilateral hilar lymphadeno- pathy was detected. Based on the clinical and histological findings, a diagnosis of sarcoidosis was made. In addition, the patient’s symptoms, including fever, facial nerve palsy and swelling of the right parotid gland, were compatible with an incomplete type of Heerfordt syndrome. Because of arrhythmia, facial nerve palsy and remarkable eyelid swelling, we started the patient on (b) oral prednisolone 40 mg daily. Within several days, rapid subsidence of the periorbital swelling was obtained (Fig. 1b). Furthermore, ACE levels normalized and the extrasystole improved slightly. Sarcoidosis is a multisystemic granulomatous disorder of unknown aetiology, most frequently manifest in the lung, skin, lymph node and eye. Cases of sarcoidosis-related eyelid swell- ing have previously been reported,1–8 but in only one case were sarcoidosis lesions confirmed by skin biopsy.7 In the other cases, eyelid swelling might have occurred as a result of an in- flammatory reaction occurring in cutaneous sarcoidosis. In add- ition, eyelids were involved bilaterally in these reports, whereas our patient showed only unilateral eyelid involvement. Furthermore, our patient was also diagnosed as having Heerfordt syndrome, which is a rare association of sarcoidosis Fig 2. (a) Histopathological features of the right upper eyelid features characterized by fever, uveitis, swelling of the parotid showed multiple nodules throughout the dermis. (b) Multiple gland and facial nerve palsy.9 This syndrome can be further noncaseating granulomas composed of epithelioid cells were present divided into complete and incomplete types. Our case is classi- in the dermis. Haematoxylin and eosin; (a) low power magnification; fied as an incomplete type, as uveitis was absent. The facial (b) high power magnification. nerve palsy in this syndrome might occur as a result of com- pression of the facial nerve by the parotid gland swelling or and an erythrocyte sedimentation rate of 20 mm in the first direct infiltration of sarcoidal granulomas to the nerve. In our hour. Intradermal injection of purified protein derivative for patient, the eyelid swelling, swelling of the parotid gland and Mycobacterium tuberculosis gave negative results. Ophthalmological facial nerve palsy were observed on the same (right) side of examination was unremarkable. An electrocardiogram revealed the face, with the first branches of the right trigeminal nerve. a ventricular extrasystole of 2000 beats per day. Computed To our knowledge, this is the first case of sarcoidosis with tomography and magnetic resonance imaging showed swelling Heerfordt syndrome that was simultaneously limited to just of his right parotid gland, and lymph node enlargement at the one side of the face.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 202 Correspondence
As our patient had shown arrhythmia, facial nerve palsy and marked swelling of the eyelid, we decided to treat him with a I systemic corticosteroid at 40 mg daily. Rapid subsidence of the eyelid swelling and facial nerve palsy was obtained within sev- eral days, and no recurrence was seen at 7 months. Here we have reported an interesting case of sarcoidosis with remarkable unilateral eyelid swelling associated with II Heerfordt syndrome. In conclusion, we believe that eyelid swelling should be added to the list of clinical signs associated 12 with sarcoidosis.
Department of Dermatology, Hokkaido University M. YAOSAKA Graduate School of Medicine, N 15 W 7, R. ABE III Kita-ku, Sapporo 060-8638, Japan H. UJIIE E-mail: [email protected] Y. ABE 12 H. SHIMIZU
Fig 1. Family pedigree showing hereditary pityriasis rubra pilaris. References
1 Brownstein S, Liszauer AD, Carey WD et al. Sarcoidosis of the eyelid Patient 1 (generation II, offspring 1). The mother, now aged skin. Can J Ophthalmol 1990; 25:256–9. 33 years, developed PRP when she was 12 years old. There 2 Jabs DA, Johns CJ. Ocular involvement in sarcoidosis. Am J Ophthalmol was no prior family history of PRP, psoriasis or atopy. She has 1986; 102:297–301. 3 Hall JG, Cohen KL. Sarcoidosis of the eyelid skin. Am J Ophthalmol the classical type of adult-onset PRP typified by a cephalocau- 1995; 119:100–1. dal eruption of follicular hyperkeratotic papules progressing to 4 Khatri KA, Chotzen VA, Burrall BA. Lupus pernio: successful treat- generalized erythroderma with islands of sparing (Fig. 2a,b). ment with a potent topical corticosteroid. Arch Dermatol 1995; She also has a scaly scalp, yellow thickening of the palms and 131:617–18. soles and subungual hyperkeratosis. The clinical diagnosis of 5 Biswas J, Krishnakumar S, Raghavendran R et al. Lid swelling and PRP was supported by histological findings of hyperkeratosis, diplopia as presenting features of orbital sarcoid. Indian J Ophthalmol patchy parakeratosis varying in both horizontal and vertical 2000; 48:231–3. 6 Bienfait MF, Hoogsteden HC, Baarsma GS et al. Diagnostic value of planes and follicular plugging. bronchoalveolar lavage in ocular sarcoidosis. Acta Ophthalmol 1987; She has been treated with potent topical steroids, psoralen 65:745–8. and ultraviolet A phototherapy, narrowband ultraviolet B 7 Diestelmeier MR, Sausker WF, Pierson DL et al. Sarcoidosis manifest- phototherapy, isotretinoin, methotrexate and ciclosporin with ing as eyelid swelling. Arch Dermatol 1982; 118:356–7. variable response. She used emollients only during pregnancy 8 Taga M, Abe T, Nishimura S et al. A case of Sjo¨gren’s syndrome when she noticed some improvement followed by a postnatal associated with chronic hepatitis C and sarcoidosis. Jpn J Clin Immunol relapse. 2003; 26:336–40. 9 Greenberg G, Anderson R, Sharpstone P et al. Enlargement of parotid gland due to sarcoidosis. Br Med J 1964; ii:861–2. Patient 2 (generation III, offspring 1). The first daughter was born with thick scales on the scalp (Fig. 2c) and areas of superficial Conflicts of interest: none declared. peeling on the face (Fig. 2d), genital region, palms and soles. Within weeks she developed follicular erythematous papules on the face, trunk and limbs (Fig. 2e,f) which enlarged into pink plaques with a scaly edge. Some of the plaques coalesced on the trunk. There was no palmoplantar thickening. A skin biopsy showed alternating orthokeratosis and parakeratosis in Pityriasis rubra pilaris in a mother and both the vertical and horizontal directions with lipping of two daughters the follicular ostia and associated follicular plugging. These appearances were consistent with PRP. She is now 3 years old DOI: 10.1111/j.1365-2133.2007.07938.x and her skin is mostly clear, requiring emollients and inter- mittent use of mild topical steroids for redness affecting the SIR, Pityriasis rubra pilaris (PRP) is a disorder of keratinization creases. of unknown cause. Most cases of PRP are sporadic but familial clustering has been reported.1–4 We report three members of Patient 3 (generation III, offspring 2). The second daughter was one family with clinical and histological features of PRP born with similar findings to her sister, although less severely (Fig. 1). To our knowledge, this is the first report of congen- affected. She had greasy scales on the front of her scalp and ital PRP affecting two daughters. forehead extending to the eyelids (Fig. 2g). She also had
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 203
a d g
e
h
b
c f
Fig 2. (a,b) Mother with classical adult-onset pityriasis rubra pilaris showing confluent erythematous scaly skin with islands of sparing. (c) The first daughter was born with thick scales on the scalp and (d) areas of superficial peeling around and behind the ear which evolved into (e) thick scales on the head and neck and (f) erythematous papules on the trunk by age 4 weeks. (g) The second daughter had greasy scales on the scalp and forehead and (h) discrete scaly patches on the cheeks and sides of the face at birth. erythematous scaly patches with discrete margins on the Familial PRP is the rarest form of PRP. Reports of a posi- cheeks and sides of the face (Fig. 2h). The rest of her skin tive family history in patients with PRP range from zero to was clear with no sign of palmoplantar thickening. A biopsy 6Æ5%.5–7 Most published series of familial PRP show auto- was deemed unnecessary given her mild disease and the somal dominant inheritance4,8 with variable expression8 benign course observed in her sister. She is now 9 months old although one family with five affected members showed and her skin disease is well controlled with emollients and an autosomal recessive pattern.1 Our pedigree is consistent mild topical steroids. with autosomal dominant or X-linked inheritance and the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 204 Correspondence differing extent of disease in the daughters suggests variable expression. Cutaneous Mycobacterium neoaurum Griffiths classified PRP into five types based on clinical infection causing scarring alopecia in an features, age at onset and prognosis: classical adult (type I), immunocompetent host atypical adult (type II), classical juvenile (type III), circumscribed juvenile (type IV) and atypical juvenile (type DOI: 10.1111/j.1365-2133.2007.07953.x V). Most familial cases belong to type V and present at birth or in the first few years of life with widespread scaly erythema SIR, Differentiation of cutaneous granulomatous disorders, and follicular papules. Palmoplantar keratoderma ranges from including sarcoidosis and mycobacteria, may be difficult. The mild thickening to pronounced keratoderma resulting in a relationship between sarcoidosis and putative infectious causes contracted sclerodermoid appearance. The daughters in our remains inconclusive.1 Sensitive molecular diagnostic tech- case have some features of type V PRP but lack palmoplantar niques are resulting in increased detection of atypical patho- keratoderma. Type V PRP also tends to run a chronic course gens; however, criteria for pathogenicity in these cases are but the disorder has mostly cleared in the older daughter and controversial.2 We report here the first known case of cutane- we anticipate that the second daughter’s PRP will follow the ous infection with Mycobacterium neoaurum. same benign course. Despite these discrepancies, the pattern A 53-year-old woman presented with a 12-month history of disease in these daughters still fits best with type V as com- of progressive hair loss on the right parietal scalp. She was of pared with type III or IV. We have considered other diagnoses Fijian Indian descent and had been living in Australia for for the daughters, the most plausible being very extensive 20 years although she had travelled extensively. She had a seborrhoeic dermatitis. However, the congenital onset and background of mitral valve replacement for mitral regurgita- the rare occurrence of clinical features compatible with PRP tion and was on antihypertensive medication, but was other- affecting two neonates in one family suggests that the family wise very well. She had received the bacille Calmette–Gue´rin history must be relevant. The histological findings also support vaccination as a child. a diagnosis of PRP. On examination, there was a well-demarcated 6 · 3cm Congenital PRP has been reported only twice previously2,7 plaque of scarring alopecia on the right parietal scalp. The and may be aetiologically significant. We speculate that margin was raised and firm and the centre was smooth and maternal transmission of circulating factors may cause a PRP atrophic with telangiectases (Fig. 1a). There were no pustules phenotype which improves as maternal factors decline. This and she had no lymphadenopathy or scar sarcoid. would explain the congenital presentation and time course of A punch biopsy showed extensive sarcoidal granulomas in the daughters’ condition in our case. Theories based on mater- the dermis extending into the subcutaneous tissue (Fig. 1b). nal transmission as opposed to genetics would also explain how There was surrounding lymphocytic inflammation with a mother with type I PRP had two daughters with type V PRP. isolated plasma cells and focally prominent multinucleated histio- cytes in the dermis. There was no necrobiosis or suppuration. Birmingham Children’s Hospital, Steelhouse Lane, M.A. THOMSON Hair follicles were reduced in number and replaced by sar- Birmingham B4 6NH, U.K. C. MOSS coidal granulomas. Putt-modified Ziehl–Neelsen stains for E-mail: [email protected] acid-fast bacilli were negative and polarization failed to reveal any foreign material. A lip biopsy was taken to look for References evidence of sarcoidosis elsewhere, but was unremarkable with 1 Griffiths WAD. Pityriasis rubra pilaris. Clin Exp Dermatol 1980; 5:105– no granulomas. Chest X-ray was unremarkable, and full blood 12. count, serum chemistry, angiotensin-converting enzyme, 2 Piamphongsant T, Akaraphant R. Pityriasis rubra pilaris: a new serum calcium and inflammatory markers were normal. proposed classification. Clin Exp Dermatol 1994; 19:134–8. Treatment had been initiated elsewhere for cutaneous sar- 3 Sanchez-Regana M, Creus L, Umbert P. Pityriasis rubra pilaris. coidosis with methotrexate 7Æ5 mg weekly, hydroxychloroqu- A long-term study of 25 cases. Eur J Dermatol 1994; 4:593–7. ine 200 mg daily and intralesional steroids. However, the 4 Vanderhooft SL, Francis JS, Holbrook KA et al. Familial pityriasis lesion continued to progress despite treatment and after rubra pilaris. Arch Dermatol 1995; 131:448–53. 5 Gelmetti C, Shiuma AA, Cerri D, Gianotti F. Pityriasis rubra pilaris 6 months, the patient was referred to our institution for fur- in childhood: a long term study of 29 cases. Pediatr Dermatol 1986; ther evaluation. The edge of the plaque was rebiopsied includ- 3:446–51. ing a biopsy for mycobacterial polymerase chain reaction 6 Clayton BD, Jorizzo JL, Hitchcock MG et al. Adult pityriasis rubra (PCR) and culture. The repeat biopsy showed similar histo- pilaris: a 10-year case series. J Am Acad Dermatol 1997; 36:959–64. logical features. Mycobacterial culture was negative; however, 7 Allison DS, el-Azhary RA, Calobrisi SD, Dicken CH. Pityriasis rubra mycobacterial PCR detected M. neoaurum. Computed tomography pilaris in children. J Am Acad Dermatol 2002; 47:386–9. of the head demonstrated no evidence of bony involvement. 8 Kint A, De Bie S, Geerts ML, T’Kint R. Pityriasis rubra pilaris, a familial condition. Arch Belges Dermatol Syphilol 1972; 158:371–6. Mycobacterium neoaurum has exhibited sensitivity to a wide range of antimicrobial agents on in vitro susceptibility testing.3 Conflicts of interest: none declared. Treatment was initiated with combination moxifloxacin and
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 205
(a) A mycobacterial origin for sarcoidosis has been postulated. Several studies, summarized by Du Bois et al., have demonstra- ted presence of mycobacterial nucleic acid in sarcoidal tissues.1 The rates of detection in these studies varied from 4Æ2% to 80%, although this variability can be explained to some extent by the methodology used.8 In addition, a small number of passage experiments in animal models,9 and the isolation of acid-fast cell wall-deficient forms,10 have lent weight to the mycobacterial argument. Others have postulated that myco- bacteria may represent a trigger for granuloma formation in sarcoidosis, rather than a pathogenic infection. Overall the evidence is inconclusive, and while mycobacteria appear to be implicated in some cases, the mycobacterial hypothesis does 1 (b) not explain all cases of sarcoid. Nontuberculous mycobacteria are ubiquitous in the envi- ronment and differentiation between colonization, pathogenic infection and laboratory contamination can be difficult. Cuta- neous nontuberculous mycobacterial infections are often pau- cibacillary, and failure to culture is not unusual. With more sensitive molecular diagnostic techniques available, the criteria required to establish an infectious aetiology need to be recon- sidered.2 Attributing causality on the basis of PCR alone requires caution; however, nucleic acid amplification technol- ogy remains a powerful tool for identification of previously undetectable pathogens. We believe that the progression with immunosuppressants and response to antimycobacterial ther- Fig 1. (a) Plaque of scarring alopecia on the right parietal scalp. The apy support a pathogenic role for M. neoaurum in this case. purple circles indicate the location of biopsies from the raised margin. (b) Prominent sarcoidal granulomas in the dermis with surrounding *Departments of Dermatology and Infectious Diseases, L.K. MARTIN* lymphocytic inflammation and Langerhans giant cells. No hair St George Hospital, Kogarah, Sydney, NSW 2217, R. LAWRENCE follicles are seen in this section (haematoxylin and eosin; original Australia S. KOSSARD § magnification · 100). University of New South Wales, Sydney, Australia D.F. MURRELL* §Skin and Cancer Foundation, Sydney, Australia roxithromycin for 4 months. This combination was successful Correspondence: De´de´e Murrell. in flattening the plaque; however, due to destruction of hair E-mail: [email protected] follicles hair regrowth did not occur. Mycobacterium neoaurum is a rapidly growing pigmented species belonging to the M. parafortuitum complex which is commonly References found in soil and water.4 Mycobacterium neoaurum was first des- 1 Du Bois RM, Goh N, McGrath G, Cullinan P. Is there a role for 5 cribed as a human pathogen in 1988. There have been nine microorganisms in the pathogenesis of sarcoidosis? J Intern Med 6 case reports of infection in humans. Most of these involved 2003; 254:4–17. catheter-associated bacteraemia; other cases included perito- 2 Fredricks DN, Relman DA. Sequence-based identification of micro- nitis, endocarditis, urinary tract infection and fatal meningo- bial pathogens: a reconsideration of Koch’s postulates. Clin Microbiol encephalitis. Most were diagnosed on culture, although one Rev 1996; 9:18–33. case was diagnosed with PCR alone.6 This is the first reported 3 George SL, Schlesinger LS. Mycobacterium neoaurum – an unusual cause of infection of vascular catheters: case report and review. Clin Infect case of cutaneous infection with M. neoaurum. The majority of Dis 1999; 28:682–3. previous cases occurred in immunocompromised individuals, 4 Tsukamura M. A new species of rapidly growing, scotochromogenic whereas this patient has no known predisposing factors. mycobacteria. Mycobacterium neoaurum. Med Biol 1972; 85:229–33. Histopathological differentiation of cutaneous mycobacterial 5 Davison MB, McCormack JG, Blacklock ZM et al. Bacteremia caused infection from other granulomatous processes relies predom- by Mycobacterium neoaurum. J Clin Microbiol 1988; 26:762–4. inantly on the granuloma type and inflammatory cell infiltrate, 6 Heckman GA, Hawkins C, Morris A et al. Rapidly progressive although these features have low specificity. There is evidence dementia due to Mycobacterium neoaurum meningoencephalitis. Emerg Infect Dis 2004; 10:924–7. that the histological appearance of cutaneous mycobacterial 7 Bartralot R, Pujol RM, Garcia-Patos V et al. Cutaneous infections due infection is related to the immune status of the host, with sar- to nontuberculous mycobacteria: histopathological review of 28 coidal granulomas being more common than tuberculoid or cases. Comparative study between lesions observed in immunosup- suppurative granulomas in immunocompetent hosts.7 pressed patients and normal hosts. J Cutan Pathol 2000; 27:124–9.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 206 Correspondence
8 Mangiapan G, Hance AJ. Mycobacteria and sarcoidosis: an overview (a) and summary of recent molecular biological data. Sarcoidosis 1995; 12:20–37. 9 Mitchell DN, Rees RJW, Rees JR. A transmissible agent from sarcoid tissue. Lancet 1969; ii:81–4. 10 Almenoff PL, Johnson A, Lesser M, Mattman LH. Growth of acid fast L forms from the blood of patients with sarcoidosis. Thorax 1996; 51:530–3.
Conflicts of interest: none declared.
(b)
The use of intravenous immunoglobulin in cutaneous and recurrent perforating intestinal Degos disease (malignant atrophic papulosis)
DOI: 10.1111/j.1365-2133.2007.07951.x
SIR, Malignant atrophic papulosis, or Degos disease, is a very rare disorder with pathognomonic lesions, occlusive vasculopathy and resultant ischaemia. When associated with the gastrointesti- nal tract, patients have a very poor prognosis and a high mor- tality. Treatment of this disease is difficult. We report a patient with Degos disease who had recurrent abdominal pain and intes- Fig 1. Clinical findings. (a) Well-demarcated papule with a tinal perforations, as well as characteristic skin lesions. The porcelain-white atrophic centre and a slightly raised reddish rim. patient was not responsive to the usual treatments including anti- (b) A perforation on the wall of the ileum (arrow). coagulants and antiplatelet therapy, but intravenous immuno- globulin (IVIG) infusion produced a satisfactory outcome. After the operation, antibiotics, anticoagulants (warfarin) A 38-year-old Chinese woman presented to the emergency and antiplatelet therapy [dipyridamole and lipoprostaglandin department in December 2005 with severe acute abdominal E1 (Tide Pharmaceutica, Beijing, China)] were given. Her gen- pain and fever of 38Æ6 C. Physical examination indicated eral status improved, although the platelet count remained rigidity of abdominal muscles, tenderness and rebound tender- elevated (three repeated platelet counts ranged from ) ) ness of the whole abdomen, shifting dullness and hypoactive 348 · 109 L 1 to 629 · 109 L 1). Ten days later, a severe bowel sounds. About 40–50 scattered well-demarcated papules abdominal pain developed again with nausea, vomiting and a 3–7 mm in diameter were found all over the body. Each fever of 38Æ2 C. At the same time, new lesions arose on her papule had a porcelain-white atrophic centre with a slightly neck and shoulders. Small intestine perforation was considered raised reddish rim (Fig. 1a). Laboratory studies revealed eleva- after an abdominal computed tomographic scan demonstrated ) ted platelets at 428 · 109 L 1. Erythrocyte sedimentation rate free fluid. The patient underwent small bowel resection due to was 56 mm in the first hour. Biochemistry, autoimmune pro- multiple perforation ulcers in the ileum (Fig. 1b). Histological file, anticardiolipin antibody test, coagulation time test and examination of both intestine and skin (old lesions and newly platelet aggregation study findings were unremarkable. Plain developed lesions) was performed. Biopsy of a fully developed abdominal X-ray showed bowel distention, and intestinal papule showed an area of epidermal atrophy, with focal obstruction was suggested. Ultrasonography demonstrated a necrosis overlying a wedge-shaped zone of sclerosis in the distended fluid-filled bowel and intestinal obstruction. Three dermis and a large acellular area with a moderate amount of months previously, the patient had had similar abdominal amorphous material. Interface change was observed as focal symptoms and was operated on at a local hospital with a diag- vacuolization of the dermal–epidermal junction and oedema nosis of mesenteric ischaemia. Asymptomatic skin lesions had of the papillary and reticular dermis. There was a sparse developed over the course of 1 year. The patient underwent superficial perivascular infiltrate of mononuclear inflammatory an exploratory laparotomy. Multiple variably sized yellowish cells (Fig. 2a). Biopsy of a newly developed papule showed lesions were observed with extensive oedema and hyperaemia thrombosed vessels in the deep dermis. Affected vessels on the wall of small intestines. No perforation was found. A showed endothelial swelling and proliferation leading to nar- diagnosis of primary acute extensive peritonitis was made. rowing of the lumina. Platelet–fibrin thrombi were found par- Degos disease was considered by the dermatologist and a skin tially or completely occluding the lumen (Fig. 2b). A section biopsy was planned. of the intestine tissue showed severe necrotizing enteritis with
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 207
(a) disease with cutaneous and recurrent perforating intestinal lesions. We tried to treat her with warfarin, dipyridamole and
prostaglandin E1 which has been reported to be effective in a Japanese patient with perforating intestinal Degos disease who achieved long-term survival.5 However, all these therapies proved to be unsuccessful. As our patient needed effective treatment to control the progression of her disease we felt that IVIG was a good option. The reported successful response to IVIG in patients with Kawasaki disease, which is a generalized vasculitis with possible thrombus of the involved vessels,6 supported our decision to use one course of high-dose IVIG ) therapy (0Æ4gkg 1 daily for 5 days). A week later a dramatic improvement of the lesions and general condition was noticed. (b) During the following 11 months of follow-up, the patient has had no new skin lesions or gastrointestinal complaints. To our knowledge, IVIG has not previously been reported for use in patients with Degos disease. The beneficial con- sequences in our present patient with recurrent perforation of the intestines is encouraging, although the potential mechan- ism of action remains unknown.
Department of Dermatology, Sir Run Run Shaw Hospital, K.J. ZHU Zhejiang University School of Medicine, Q. ZHOU No. 3 Qingchun Road East, Hangzhou, Zhejiang, China A.H. LIN Correspondence: Hao Cheng. Z.M. LU E-mail: [email protected] H. CHENG Fig 2. Histopathological examination of skin lesions. (a) Epidermal atrophy and a large acellular area in the dermis in a well-developed References lesion. (b) An altered vessel below the dermal necrobiotic zone with the lumen occluded by a thrombus in a newly developed lesion. 1 Loewe R, Palatin R, Petzelbauer P. Degos disease with an incon- Haematoxylin and eosin; original magnification: (a) · 100; (b) · 400. spicuous clinical course. J Eur Acad Dermatol Venereol 2005; 19:477–80. 2 Snow JL, Muller SA. Degos syndrome: malignant atrophic papulosis. Semin Dermatol 1995; 14:99–105. interstitial haemorrhage and stenotic endovascular proliferation 3 Bogenrieder T, Kuske M, Landthaler M, Stolz W. Benign Degos’ with thrombosis in the presence of inflammation. disease developing during pregnancy and followed for 10 years. Seven days later, the patient once again had a recurrence of Acta Derm Venereol (Stockh) 2002; 82:284–7. the acute abdominal pain and fever of 38 C. As all the above 4 Ojeda Cuchillero RM, Sanchez Regana M, Umbert Millet P. Benign treatments had failed to control the disease, IVIG (human cutaneous Degos’ disease. Clin Exp Dermatol 2005; 28:145–7. immunoglobulin, PH4; Boya Biopharmaceutical Co., China) 5 Shimazu S, Imai H, Kokubu S et al. Long-term survival in malignant was administered for five consecutive days at a dosage of atrophic papulosis: a case report and review of the Japanese ) literature. Nippon Geka Gakkai Zassh 1988; 89:1748–51. 0Æ4gkg 1 daily. Her condition soon improved, with neither 6 Morgan GJ, Macleod C, Jenkins J et al. IVIG, aspirin, and Kawasaki new lesions nor abdominal pain developing. She was disease. J Pediatr 2003; 143:280–1. discharged in January 2006. Up to now, at 11 months of follow-up, her general health has remained good. Conflicts of interest: none declared. Since the first case report in 1942, approximately 200 cases have been published of Degos disease.1 It is a disorder charac- terized by multiple infarcts in the skin and internal organs, related to a thrombotic vasculopathy of unknown origin.2 Clinically, in addition to the characteristic atrophic and porcelain white centred papules, involvement of the gastrointestinal tract A novel deletion mutation in the EDAR gene in is observed in 50% of cases, with intestinal perforation being a Pakistani family with autosomal recessive the most common cause of death.3 These patients usually die hypohidrotic ectodermal dysplasia in several months. Unfortunately, until now there has been no effective treatment for Degos disease. Most reported agents DOI: 10.1111/j.1365-2133.2007.07949.x used are anticoagulants, antiplatelet agents, stimulators of fibrinolysis, immunosuppressants and corticosteroids, but not SIR, Hypohidrotic (anhidrotic) ectodermal dysplasia (HED) is all cases respond to the therapy.4 Our patient had typical Degos the most common form of ectodermal dysplasias (EDs) and is
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 208 Correspondence characterized by absence or deficient function of the deriva- (a) tives of the ectoderm: teeth, hair and sweat glands. The mode of inheritance of HED may be X-linked or autosomal (domin- ant or recessive). X-linked HED (XLHED; MIM 305100) is caused by mutations in the ectodysplasin gene (EDA1), located on chromosome Xq12–q13.1.1 A minority of patients with the HED (MIM 224900) phenotype display an autosomal (recessive or dominant) inheritance pattern that is due to mutations in the ectodysplasin A1 isoform (EDA-A1) receptor (EDAR; MIM 604095), located on chromosome 2q11–q13.2,3 Furthermore, autosomal recessive inheritance with HED has been eventually found due to a mutation in the EDAR-associ- ated death domain (EDARADD; MIM 606603) gene, located on chromosome 1q42.2–q43.4 EDAR is a NF-jB-activating member of the tumour necrosis factor (TNF) receptor family. EDAR protein contains an extra- cellular ligand binding N-terminal domain, single transmem- brane region and intracellular region containing death domain (b) (DD).3 EDA-A1 is a ligand of EDAR, activating it by physically interacting through its TNF-ligand motif in the C-terminal region. EDAR interacts with its adapter EDARADD and activates the NF-jB pathway using TAB2, TRAf6 and TAK1.5 This linear pathway explains the genetic heterogeneity among HED patients. In the present study we ascertained a Pakistani family dem- onstrating the autosomal recessive form of HED (ARHED). The study was approved by the Quaid-i-Azam University Insti- Fig 1. (a) Pedigree of a Pakistani family with autosomal recessive tutional Review Board. Informed consent was obtained from hypohidrotic (anhidrotic) ectodermal dysplasia (ARHED). Filled the family members who participated in the study. Affected symbols represent affected subjects. Open symbols represent individuals (IV-2 and IV-3) of the family (Fig. 1a) showed unaffected individuals. Double lines are representative of the characteristic features of HED, including fine and curly consanguineous unions. (b) Clinical findings in hypohidrotic sparse hair, absent eyebrows and eyelashes, absence of axillary ectodermal dysplasia. Phenotypic appearance of an affected individual and pubic hair, conical teeth, diminished sweating, dry and (IV-3) having curly hair on the scalp, saddle nose, prominent lips and thin skin, protruding prominent lips, hyperpigmentation of absent eyelashes and eyebrows. The same individual (IV-3) with the skin around the eyes and mouth and saddle-shaped nose sparse and thin hair on scalp. (Fig. 1b). The autosomal recessive mode of inheritance and clinical features of the patients, compatible with HED, led us to screen the candidate genes reported earlier to be respon- family (Fig. 2a). This in-frame deletion mutation was not sible for HED. To search for an underlying mutation in the identified in a panel of 200 chromosomes of Punjabi Pakistani EDAR gene, the entire coding portion (12 exons) and intron– population to whom this family belongs and the mutation exon boundaries of the EDAR gene was sequenced in two was not identified outside the family. affected (IV-2 and IV-3) and four normal individuals (III-3, To date 20 pathogenic mutations including 14 missense, III-4, IV-4, IV-5) of the family (Fig. 1a). Amplified polymer- one nonsense, three splice sites and two deletions in the EDAR ase chain reaction (PCR) fragments from genomic DNA were gene have been reported. Thirteen of the missense mutations sequenced directly in an ABI Prism 310 automated DNA were identified in the two functional domains of the EDAR sequencer (PE Applied Biosystems, Foster City, CA, U.S.A.). protein: extracellular and death domains. 2,6–8 DNA sequence To amplify the 438-bp PCR fragment containing exon 5 of analysis of the EDAR gene in the family, presented here, revea- the EDAR gene the following primers were used: 5¢-GGAAACT led a third novel deletion mutation (399_404delGGTCTG) in GAGTGGACAGAGC-3¢ (intron 4, sense) and 5¢-ACTCAAGGCT exon 5 of the EDAR gene. This deletion mutation is located in CAGATGTGGC-3¢ (intron 5, antisense). the extracellular domain, spanning from amino acid residues The sequence analysis of exon 5 of the EDAR gene from 27–187 of the EDAR protein [http://www.expasy.org/uniprot/ affected individuals revealed a novel 6-bp in-frame deletion Q9UNE0 (accessed 10 Mar 2007)]. The mutation results mutation starting at nucleotide position 399 (c.399_404del- in skipping of the three amino acids methionine, valine and GGTCTG) (Fig. 2c). This deletion was present in the heterozy- cysteine and insertion of a new amino acid isolucine gous state in obligate carriers III-3 and III-4 (Fig. 2b). The (p.M133_C135delinsI) in one of the ‘cysteine-rich regions’ mutation was not identified in two normal individuals of the of the extracellular domain. The skipping of cys135 in the
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 209
(a) References 1 Kere J, Strivastava AK, Montonen O et al. X-linked anhidrotic (hypohidrotic) ectodermal dysplasia is caused by mutation in a novel transmembrane protein. Nat Genet 1996; 13:409–16. 2 Monreal AW, Ferguson BM, Headon DJ et al. Mutations in the human homologue of mouse dl cause autosomal recessive and dominant hypohidrotic ectodermal dysplasia. Nat Genet 1999; 22:366–9. 3 Headon DJ, Overbeek PA. Involvement of a novel TNF receptor (b) homologue in hair follicle induction. Nat Genet 1999; 22:370–4. 4 Headon DJ, Emmal SA, Ferguson BM et al. Gene defect in ectoder- mal dysplasia implicates a death domain adapter in development. Nat Genet 2001; 414:913–16. 5 Morlon A, Munnich A, Smahi A. TAB2, TRAF6 and TAK1 are involved in NF-kappaB activation induced by the TNF-receptor, Edar and its adaptator Edaradd. Hum Mol Genet 2005; 14:3751–7. (c) 6 Shimomura Y, Sato N, Miyashita A et al. A rare case of hypohidrotic ectodermal dysplasia caused by compound heterozygous mutations in the EDAR gene. J Invest Dermatol 2004; 123:649–55. 7 Naeem M, Muhammad D, Ahmad W. Novel mutations in the EDAR gene in two Pakistani consanguineous families with auto- somal recessive hypohidrotic ectodermal dysplasia. Br J Dermatol 2005; 15:346–50. 8 Chassaing N, Bourthoumieu S, Cossee M et al. Mutations in EDAR account for one-quarter of non-ED1-related hypohidrotic ectoder- mal dysplasia. Hum Mutat 2006; 27:255–9. Fig 2. Sequence analysis of the EDAR gene mutation. DNA sequence of 9 Smith CA, Farrah T, Goodwin RG. The TNF receptor superfamily exon 5 of the EDAR gene from (a) a control individual, (b) a heterozygous of cellular and viral proteins: activation, costimulation, and death. carrier, and (c) a homozygous (affected) individual. The bar above the Cell 1994; 76:959–62. wild-type sequence in (a) represents the sequence that is deleted in the 10 Banner DW, D’Arcy A, Janes W et al. Crystal structure of the sol- homozygous state in the affected individual (c). uble human 55 kd TNF receptor-human TNF beta complex: impli- cations for TNF receptor activation. Cell 1993; 73:431–45. deletion mutation (399_ 404delGGTCTG) abolishes a disulfide bridge that exists between cys135 and cys148 (http://www. Conflicts of interest: none declared. expasy.org/uniprot/Q9UNE0) (which is likely to produce unstable protein), changes the secondary structure of the protein and ⁄or alters the interaction site for ligand binding. EDAR, like other TNF-like receptors, have elongated structures by a scaffold of disulfide bridges. These disulfide bonds form ‘cysteine-rich domains’ (CRDs) that are the basic hallmark Mondor’s phlebitis after using tadalafil of the TNF receptor superfamily.9 These elongated chains of TNF receptors fit in the interacting ‘grooves’ of their ligand DOI: 10.1111/j.1365-2133.2007.07948.x trimers.10 Our finding extends the body of evidence that supports the significance of the EDAR signalling pathway, SIR, A 45-year-old caucasian male physician came to our clinic which is now recognized as a powerful regulator of skin because of a tender, flesh-coloured subcutaneous cord-like development and postnatal remodelling. structure localized in the coronal sulcus, circling the penis and partly extending onto the proximal shaft. The patient reported Acknowledgments that the lesion had appeared approximately 24 h after conjugal sexual intercourse, and had remained stable and asymptomatic We sincerely wish to thank the family members for their par- for 4 weeks. A similar episode had happened about 5 months ticipation. The work presented here was funded by Higher earlier, 2 days after conjugal sexual intercourse, with complete Education Commission (HEC), Islamabad, Pakistan. Muhammad resolution in 2 weeks of abstinence. On both occasions, Tariq is supported by an indigenous PhD fellowship from he had consumed, 2 h before sexual intercourse, 20 mg of HEC, Islamabad, Pakistan. tadalafil to improve erectile performance and duration. Physical examination did not reveal other abnormalities or Department of Biochemistry, Faculty of Biological M. TARIQ regional lymphadenopathy. Complete blood count and Sciences, Quaid-i-Azam University, Islamabad, Pakistan N. WASIF urinalysis were within normal limits; serology for human Correspondence: Wasim Ahmad. W. AHMAD immunodeficiency virus and syphilis (including VDRL and E-mail: [email protected] FTA-ABS) was negative.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 210 Correspondence
We diagnosed Mondor’s phlebitis and suggested temporary References sexual and pharmacological abstinence. Complete restitutio 1 Rosen T, Hwong H. Sclerosing lymphangitis of the penis. J Am Acad ad integrum was obtained in 3 weeks, without any therapy. Dermatol 2003; 49:916–18. 1 Sclerosing lymphangitis of the penis (other suggested 2 Pedersen T. Circular indurated lymphangitis of the penis. Ugeskr names are nonvenereal sclerosing lymphangitis, circular Laeger 1979; 141:2603–4. indurated lymphangitis,2 benign transient lymphangiectasis, 3 Findlay GH, Whiting DA. Mondor’s phlebitis of the penis. Mondor’s phlebitis3 and lymphangiofibrosis thrombotic occlu- A condition miscalled ‘non-venereal sclerosing lymphangitis’. sive4), first described by Hoffman in 1923 as ‘gonorrheal Clin Exp Dermatol 1977; 2:65–7. pseudochancre’5 and subsequently reported by the same 4 Marsch WC, Stuttgen G. Sclerosing lymphangitis of the penis: a lymphangiofibrosis thrombotic occlusive. Br J Dermatol 1981; author in 1938 as ‘nonvenereal plastic lymphangitis of the 104:687–95. 6 coronal sulcus of the penis with circumscribed edema’, is a 5 Hoffmann E. Vortauschung primarer syphilis durch gonorrheische benign pathology of the superficial dorsal penile vein, docu- lymphangitis (gonorrheischer pseuprimaraffekt). Munchen Med Wschr mented in sexually active men aged 18–66 years.1 This disease 1923; 70:1167–8. is clinically characterized by the onset, within 24–48 h after 6 Hoffmann E. Ubernicht venerische plastiche lymphangitis im sexual intercourse, of a firm cord-like swelling of the penile sulcus coronaries penis mit um schriebenem Oden. Derm Z 1938; coronal sulcus, rarely extending to the proximal shaft, asymp- 78:24–7. 7 Kumar B, Narang T, Radotra BD, Gupta S. Mondor’s disease of tomatic or with mild to moderate pain and discomfort during penis: a forgotten disease. Sex Transm Infect 2005; 81:480–2. 1,2 erection. Although multiple aetiologies have been proposed, 8 Agrawal SK, Singal A, Pandhi D. Mondor’s phlebitis of penis the cause of penile Mondor’s disease still remains unclear. following recurrent candidal balanoposthitis. Int J Dermatol 2005; Venereal and nonvenereal causes, including irritation from 44:83–4. menstrual blood, tuberculosis, circumferential scarring from 9 Tanii T, Hamada T, Asay Y, Yorifuji T. Mondor’s phlebitis of the circumcision and mechanical trauma have been theoretically penis: a study with factor VIII related antigen. Acta Derm Venereol postulated, but have not yet received sufficient clinical and 1984; 64:337–40. 10 Kalsi JS, Kell PD. Update on oral treatments for male erection scientific support.1 With regard to the origin of the disease dysfunction. J Eur Acad Dermatol Venereol 2004; 18:267–74. (lymphatic vs. venous), many studies have evidenced the predominant involvement of the veins, through the striking Conflicts of interest: none declared. histological feature of plump endothelial cells with positive staining for CD31 and CD34;7 however, anatomical and histo- logical differentiation between venous and lymphatic vessels of the penis is really difficult. The emissary veins, arising from the erectile tissue, perfor- ate tunica albuginea, join a series of circumflex vessels curving Cutaneous angiokeratoma and venous around the shaft and lead to the deep dorsal vein; circumflex malformations in a Hispanic-American patient vessels are anatomically prone to stretching and torsion, even- with cerebral cavernous malformations tually leading to oedema in a circumferential shape.7–9 Tadalafil is a potent, reversible and selective inhibitor of DOI: 10.1111/j.1365-2133.2007.07978.x phosphodiesterase-5 (PDE-5), an enzyme found in trabecular smooth muscle and able to catalyse the degradation of cyclic SIR, Familial cerebral cavernomas, or cerebral cavernous mal- guanosine monophosphate (cGMP), leading to elevated cyto- formations (CCM), is a disease characterized by the develop- solic calcium concentration and smooth muscle contraction. ment of CCM (cavernomas) of the brain. These malformations Tadalafil and other PDE-5 inhibitors act by blocking the above comprise abnormally enlarged, thin-walled, vascular channels biochemical pathway and, consequently, promoting erection.10 that form raspberry-like angiomas in the brain separated from In our patient, postulating an anatomic variation in the the surrounding neural parenchyma by connective tissue. Three venous arcade as the basic defect,9 the consumption of tadalafil forms of autosomal dominant CCM have been found to result might have caused an unusual, extraordinary mechanical com- from heterogeneous mutations in CCM genes (CCM1, CCM2, pression of the emissary veins, through a prolonged over- CCM3). These genes encode the proteins KRIT1 (Krev-1 ⁄rap1a expansion of the corpora cavernosa, working as a precipitating interacting protein), MGC4607 and PDCD10 (programmed cell factor. In our opinion, this hypothesis could also explain the death 10), respectively.1–3 The exact role of CCM proteins in absence of further recurrence of the disease in our patient after cerebrovascular development is still being elucidated, but they successive sexual intercourse without tadalafil consumption. seem to be important to both angiogenesis and vascular remodelling.3,4 Institute of Dermatology, University of Messina, C. GUARNERI In patients with CCM, the phenotypic expression of caver- A.O.U. ‘G. Martino’, Via Consolare Valeria, F. GUARNERI nomas varies with age and other unknown factors. The disease Gazzi, 98125 Messina, Italy may be clinically silent in up to 15% of patients with CCM1 E-mail: [email protected] mutations, and symptoms typically develop in the third to
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 211
fifth decades of life with the CCM causing seizures, haemor- The patient’s family history was significant for a daughter rhage or focal neurological deficit.5,6 with multiple cavernous malformations of the brain and a Studies involving Hispanic and Mexican-American patients grandson who died at age 25 years due to intracerebral with CCM have revealed a common 2105CfiT mutation in haemorrhage from CCM. Neither had a history of skin find- the CCM1 (now KRIT1) gene, indicating a relatively recent ings, according to the family. Further investigation of the founder effect.7 We report a Hispanic-American patient with patient’s medical history revealed that a cutaneous venous this ‘common Hispanic mutation’, who exhibited CCM and malformation had been excised from his right arm in 1994. multiple cutaneous vascular malformations. Neurological evaluation for his chronic headaches and dizzi- In August 2003, a 77-year-old Hispanic man presented to ness in both 1992 and 1997 included magnetic resonance our dermatology clinic for evaluation of an asymptomatic imaging findings consistent with multiple cavernomas within nodule on his left hand that had been slowly growing for the brain. 1 year. Physical examination revealed a 1 cm diameter blue- Genetic testing confirmed our suspicions that this family black hyperkeratotic nodule on the left thenar eminence suffers from familial cerebral cavernomas. Specifically, the (Fig. 1a). A discrete, 8 mm diameter, blue mass was excised patient and his daughter tested positive for the ‘common at surgery. Histological examination revealed dilated blood- Hispanic’ 2105CfiT mutation in the CCM1 (KRIT1) gene. filled vessels in the papillary dermis with overlying epidermal Follow-up examination of our patient led to the discovery of hyperkeratosis consistent with an angiokeratoma (Fig. 1b). another venous malformation just superior to the right elbow At the 2-month follow-up visit, the patient was found to in March 2004. have a 2 cm bluish, mobile subcutaneous mass in his right Cutaneous vascular malformations have been reported in forearm (Fig. 2a). A red-purple, firm, lobulated soft tissue patients with multiple CCM, although never in the context of mass measuring 1.7 · 1.1 · 0.4 cm was excised (Fig. 2b,c). the common Hispanic mutation.5,6,8–10 In a survey of 173 Histological examination demonstrated large, dilated vascular patients with CCM from 57 French families with familial cere- channels lined by flat endothelium in the reticular dermis bral cavernomas, Labauge et al.5 identified 10 cases of localized (Fig. 2d). A diagnosis of cutaneous venous malformation was cutaneous hyperkeratotic vascular malformations in four fam- made. ilies with CCM and cutaneous venous malformations. All of the skin lesions were vascular malformations of predominantly capillary type with a venular component in association (a) with an overlying hyperkeratotic epidermis. Clinically, they appeared as crimson-coloured, irregularly shaped plaques measuring up to several centimetres. The nine patients studied genetically did not share the same CCM1 mutation.5 This could be interpreted as circumstantial evidence that the presence of skin manifestations may reflect differences in disease pene- trance, and do not result from a specific mutation. Hyperkeratotic cutaneous capillary-venous malformations (HCVM) were also reported in the setting of CCM by Eerola et al.9 They screened five families with CCM and found one family with two affected boys who manifested HCVM on the limbs. The family was found to have a CCM1 mutation as well. Similar lesions were reported in a Chinese family with a novel mutation in CCM1.10 Ostlere et al.8 reported a single patient with eruptive angio- (b) keratomas and cerebral cavernomas. Angiokeratomas represent ectasias of blood vessels of the superficial dermis. The hist- ology and clinical appearance of HCVM is similar to that of solitary angiokeratoma with perhaps a deeper venous compon- ent to the neoplasm. These two diagnoses are more similar then dissimilar and may represent different stages in the evo- lution of one entity. The incidence, prevalence and biology of cutaneous vascular malformations in patients with CCM with the common His- panic mutation are areas for future investigation. It would also be interesting to investigate the relationship between Fig 1. (a) Nodule on the left thenar eminence. (b) Histology cutaneous involvement and prognosis. The identification of consistent with angiokeratoma (haematoxylin and eosin; original multiple cutaneous vascular lesions in a Hispanic-American magnification · 100.) patient may prove to be a simple way to screen for CCM
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 212 Correspondence
(a) (b)
(c) (d)
Fig 2. (a) Right arm mass. (b) Excisional biopsy. (c) Gross appearance of mass. (d) Histology consistent with venous malformation (haematoxylin and eosin; original magnification · 40.) families, and perhaps should provoke further investigation of 4 Seker A, Pricola KL, Guclu B et al. CCM2 expression parallels that of their family history for strokes or neurological symptoms. CCM1. Stroke 2006; 37:518–23. 5 Labauge P, Laberge S, Brunereau L et al. Hereditary cerebral cavern- ous angiomas: clinical and genetic features in 57 French families. Departments of Dermatology and *Neurology, B.J. ZLOTOFF Socie´te´ Franc¸aise de Neurochirurgie. Lancet 1998; 352:1892–7. University of New Mexico Health Science R.H. BANG 6 Denier C, Labauge P, Brunereau L et al. Clinical features of cerebral Center, Albuquerque, NM 87131-5231, R.S. PADILLA cavernous malformations patients with KRIT1 mutations. Ann Neurol U.S.A. L. MORRISON* 2004; 55:213–20. E-mail: [email protected] 7 Gunel M, Awad IA, Finberg K et al. A founder mutation as a cause of cerebral cavernous malformation in Hispanic Americans. N Engl J Med 1996; 334:946–51. References 8 Ostlere L, Hart Y, Misch KJ. Cutaneous and cerebral haemangiomas associated with eruptive angiokeratomas. Br J Dermatol 1996; 135:98– 1 Sahoo T, Johnson EW, Thomas JW et al. Mutations in the gene 101. encoding KRIT1, a Krev-1 ⁄rap1a binding protein, cause cerebral 9 Eerola I, Plate KH, Spiegel R et al. KRIT1 is mutated in hyper- cavernous malformations (CCM1). Hum Mol Genet 1999; 8:2325–33. keratotic cutaneous capillary-venous malformation associated with 2 Craig HD, Gunel M, Cepeda O et al. Multilocus linkage identifies cerebral capillary malformation. Hum Mol Genet 2000; 9:1351–5. two new loci for a mendelian form of stroke, cerebral cavernous 10 Chen DH, Lipe HP, Qin Z et al. Cerebral cavernous malformation: malformation, at 7p15-13 and 3q25.2-27. Hum Mol Genet 1998; novel mutation in a Chinese family and evidence for heterogeneity. 7:1851–8. J Neurol Sci 2002; 196:91–6. 3 Bergametti F, Denier C, Labauge P et al. Mutations within the programmed cell death 10 gene cause cerebral cavernous malfor- mations. Am J Hum Genet 2005; 76:42–51. Conflicts of interest: none declared.
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 Correspondence 213
study were using individual, concurrent, uncontrolled topical therapies. In addition, patient compliance as assessed by study Folate with methotrexate: big benefit, medication counts was higher in the placebo arm (90% vs. questionable cost 78%). Finally, and most importantly, the baseline PASI scores of both groups were low and therefore susceptible to pro- DOI: 10.1111/j.1365-2133.2007.08005.x portionally large random fluctuations. It is our experience that
1 PASI scores < 10 are unreliable in terms of measuring and SIR, In their recent report Salim et al. present a randomized, comparing disease severity. Furthermore, the dose of folic double-blind, placebo-controlled exploratory study investigat- acid used in this study was higher than the 1 mg daily2 or ing the impact of folic acid supplementation on the efficacy of 5–10 mg weekly3 often recommended for the reduction of methotrexate (MTX) in the treatment of psoriasis. There is MTX side-effects. For the above reasons, it is not clear that abundant evidence that concomitant folate supplementation folic acid use has a clinically significant impact on MTX with either folic or folinic acid can reduce the incidence of efficacy in psoriasis, and future large-scale studies are needed. MTX-related side-effects and allows for greater drug tolerabil- We agree with the authors that even if folate does reduce ity.2,3 Importantly, multiple studies in patients with rheuma- the efficacy of MTX, the associated reduction in MTX toxicity toid arthritis have failed to show a difference in MTX efficacy could provide a net benefit to folic acid supplementation. when folate is added. The present study is a well-designed, In light of the well-established benefits of folate supple- albeit small trial (n = 22) comparing the addition of folic acid mentation and debatable effects on efficacy, folate supple- vs. placebo to the drug regimen of patients taking MTX for mentation is still warranted when treating psoriasis patients stable psoriasis. The mean Psoriasis Area and Severity Index with MTX. (PASI) score rose from 6Æ4 at baseline to 10Æ8 after 12 weeks of therapy in the folic acid group, with the largest change Ronald O. Perelman Department of Dermatology, I. BROWNELL occurring at week 3 when one subject presented with a New York University School of Medicine, B.E. STROBER ‘marked flare’ of psoriasis. The placebo arm mean PASI 560 First Avenue, TCH 158, score was relatively stable, moving from 9Æ8to9Æ2 after New York, NY 10016, U.S.A. 12 weeks. The observed difference in mean PASI change Correspondence: Bruce Strober. between the two groups was statistically significant E-mail: [email protected] (P <0Æ05). The authors conclude that daily supplementation of 5 mg folic acid reduces the efficacy of low-dose MTX in the treatment of psoriasis. References While the trends in PASI scores were different between 1 Salim A, Tan E, Ilchyshyn A, Berth-Jones J. Folic acid supplementa- thetreatment groups in this trial, the clinical relevance of these tion during treatment of psoriasis with methotrexate: a randomized, small changes is unclear. Furthermore, the observed differ- double-blind, placebo-controlled trial. Br J Dermatol 2006; ences in disease severity could be due to multiple potential 154:1169–74. confounding factors. The demographics, MTX treatment his- 2 Strober BE, Menon K. Folate supplementation during methotrexate tory and baseline disease severity were notably different therapy for patients with psoriasis. J Am Acad Dermatol 2005; 53:652–9. 3 Whittle SL, Hughes RA. Folate supplementation and methotrexate between the two study arms. Specifically, the placebo arm had treatment in rheumatoid arthritis: a review. Rheumatology 2004; a younger mean age (54 vs. 60 years), higher mean base- 43:267–71. line PASI score (9Æ8 vs. 6Æ4), longer mean duration of prior MTX use (99 vs. 47 months) and higher mean cumulative Conflicts of interest: none declared. MTX dose (4Æ2 vs. 2Æ1 g). Furthermore, all patients in the
Language: English News and Notices Contact: Katrin Steinmann, Graz, Austria DOI: 10.1111/j.1365-2133.2007.08080.x Telephone: +43-699-11081567 Fax: +43-316-696110 International Short Course on Dermoscopy Email: [email protected] Date: July 17–21, 2007 Website: http://www.meduni-graz.at/dermoscopy/ Venue: Department of Dermatology, Medical University of Description: This course is for residents in dermatology Graz, Auenbruggerplatz 8, A-8036 Graz and for dermatologists from universities or private practice Organizer: Medical University of Graz, Department of Derma- as well as for physicians or nurses interested in the diagnosis tology of pigmented skin lesions. Type of Event: Course
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214 214 News and Notices
Congress: 8th Annual Meeting of the Austrian Academy of Applicants are asked to apply for funding from the category most Cosmetic Surgery in cooperation with the American and appropriate to the research they wish to undertake. Please note Asian Academy of Cosmetic Surgery applications may only be entered under one category. The Venue and date: Vienna/Austria – Hotel Marriott, September closing date is 28th September 2007 with awards being made in 13–16, 2007 December. Main topics: Safety in Cosmetic Surgery, Facial Surgery, The Research Award: A project grant of up to 2 years duration Anatomy, Laser Therapy, Breast Operations, Liposuction, and maximum funding of £70,000 (£35,000 per annum). Abdominoplasty, Brachioplasty, Leg-Lift, Medico-Cosmetics, The BSF Fellowship: A grant of £60,000 to support a dermato- Hormonal Therapy, Botox, Fillers, New Methods and Technologies logical clinician through one year of research. Applications are Chair: made by a head of department. It is envisaged that the Dr Peter Lisborg, Vienna Medical Academy successful fellow will pursue a career in dermatology. Phone: +43/1/405 13 83 10 The BSF Studentship: Available to any UK based research Fax: +43/1/407 82 74 institution seeking to fund dermatological research at a Ph.D. e-mail: [email protected] student level. The 3-year Studentship is a fixed amount of Information: £75,000 (£25,000 per year). A¨rztezentrale Med.Info In addition the Trustees are pleased to announce the first BSF Helferstorferstrasse 4, A-1014 Wien Clinical Trial. The award will support a multi-centre clinical Phone: (+43/1) 531 16 – 75 trial up to a maximum of £70,000 per year over 4 years – Fax: (+43/1) 531 16 – 61 £280,000 in total. e-mail: [email protected] Application forms and details of each award are available online at www.britishskinfoundation.org.uk or you can contact British Skin Foundation the BSF office, specifying the type of grant you are interested in. 2007 Awards Tel: 020 7391 6341 Call for Grant Applications British Skin Foundation The Trustees of the British Skin Foundation have pleasure in annou- 4 Fitzroy Square ncing that funding is once again available for skin disease research. London WIT 5HQ
2007 The Authors Journal Compilation 2007 British Association of Dermatologists • British Journal of Dermatology 2007 157, pp183–214