42 British rournal of Ophthalmology 1995; 79: 42-45 Placebo controlled trial of fusidic acid gel and

oxytetracycline for recurrent blepharitis and Br J Ophthalmol: first published as 10.1136/bjo.79.1.42 on 1 January 1995. Downloaded from rosacea

D V Seal, P Wright, L Ficker, K Hagan, M Troski, P Menday

Abstract topical and systemic anti-staphylococcal A prospective, randomised, double blind, antibiotics in addition to anti-inflammatory partial crossover, placebo controlled trial therapy. has been conducted to compare the Fusidic acid has been in clinical use since performance of topical fusidic acid gel 1962 and is particularly effective against (Fucithalmic) and oral oxytetracycline as staphylococci. It has a steroid structure but no treatment for symptomatic chronic glucocorticoid activity. A new gel preparation blepharitis. Treatment success was containing 1% microcrystalline fusidic acid judged both by a reduction in symptoms (Fucithalmic) has recently been shown to be and clinical examination before and after effective for treating acute bacterial conjunc- therapy. Seventy five per cent of patients tivitis and for reducing the staphylococcal lid with blepharitis and associated rosacea flora before surgery.5-8 were symptomatically improved by Oxytetracycline has been selected for the fusidic acid gel and 500/0 by oxytetra- treatment of patients with chronic blepharitis cycline, but fewer (35%/o) appeared to as it has both anti-inflammatory and anti- benefit from the combination. Patients staphylococcal properties.9 10 It has been with chronic blepharitis of other aetiolo- demonstrated to be effective in a controlled gies did not respond to fusidic acid gel but trial in ocular rosacea for comeal infiltrates and 25% did benefit from oxytetracycline and conjunctival hyperaemiaII and in non-ocular 300/0 from the combination. Our results rosacea for papular disease.12 These beneficial demonstrate the need to investigate effects have also been demonstrated with doxy- patients with blepharitis for concomitant cycline. 13 rosacea as they respond well to targeted We have studied the subjective and objective therapy. responses of patients with chronic blepharitis (Br_J Ophthalmol 1995; 79: 42-45) of various aetiologies treated with fusidic acid and oxytetracycline in a randomised, con- trolled prospective way. http://bjo.bmj.com/ Chronic blepharitis is a common problem but is difficult to manage because various aetiolo- gies are involved. Universal treatment cannot Materials and methods therefore be applied and, until now, specific Patients suffering from chronic blepharitis therapy has not been identified for each sub- were offered the opportunity to participate in group. Symptomatic relief remains the major this trial. The majority had taken part in a therapeutic contribution. previous study, assessing the role of staphylo- on October 2, 2021 by guest. Protected copyright. There is controversy over the role of inflam- coccal CMI in this condition, when their mation as distinct from infection, but in some symptoms and signs were tabulated in detail.2 patients the two may be interrelated. This study was designed as a prospective, Tennent Institute of Ophthalmology, Seborrhoeic disease and meibomian dysfunc- randomised, double blind, placebo controlled, Western Infirmary, tion are considered primarily inflammatory.' partial crossover group comparison trial of the 38 Church Street, Atopic keratoconjunctivitis, characterised by effect of fusidic acid and oxytetracycline to Glasgow GIl 6NT D V Seal heavy Staphylococcus aureus colonisation, give symptomatic relief. A similar protocol appears inflammatory rather than infective. for assessing symptoms and signs was used Institute of Folliculitis and ulcerative blepharitis in as before.2 Informed, written consent was Ophthalmology and 'staphylococcal' lid disease usually represent obtained from all patients with chronic Moorfields Eye Hospital, City Road, infection caused by S aureus which may be blepharitis entering the study. Exclusion London EClV 2PD complicated by cell mediated marginal kera- criteria were known hypersensitivity to fusidic P Wright titis.2 3 acid, oxytetracycline, or benzalkonium L Ficker K Hagan We have demonstrated enhanced cell chloride (preservative in Fucithalmic), simul- M Troski mediated immunity (CMI) to S aureus, but not taneous wearing of contact lenses, female to coagulase negative staphylococci (CNS), patients of child bearing potential or those who Leo Laboratories Ltd, Longwich Road, among patients with symptomatic blepharitis. were pregnant or breast feeding, concurrent Princes Risborough, We have found that coagulase negative staphy- use of prescribed anti-infective drugs, other Aylesbury, Bucks lococci represent normal commensal flora.2 4 ophthalmic complications, or severe renal HP17 9RR P Menday The pathogenic role of bacteria isolated from impairment. patients with chronic blepharitis remains con- The drugs used in this study were Correspondence to: Miss L Ficker. troversial and it is not surprising that there is Fucithalmic, a topical preparation of 1% continued debate over its treatment. Thera- in a Accepted for publication fusidic acid carbomer gel made isotonic by 5 September 1994 peutic approaches have included lid hygiene, adding mannitol, buffered to pH 5 5, and Placebo controlled trial offusidic acid gel and oxytetracycline for recurrent blepharitis and rosacea 43

preserved with benzalkonium chloride; eyelash loss, seborrhoea, telangectasia, Fucithalmic placebo with similar gel but lack- meibomianitis, chronic inflammation of the lid ing the fusidic acid; oxytetracycline tablets margin, superficial punctate corneal epitheli- containing 250 mg oxytetracycline and oxy- opathy, and marginal keratitis. In addition, Br J Ophthalmol: first published as 10.1136/bjo.79.1.42 on 1 January 1995. Downloaded from tetracycline placebo - a matching tablet but patients were examined for the presence of lacking the oxytetracycline. Drugs were dis- rosacea with excessive blushing, thickening patched every 2 months to patients by the and vascularisation of facial skin, rhinophyma, pharmacy so that they were unaware whether papular and pustular erythema, conjunctival they were entering a placebo or active treat- hyperaemia and discharge, telangectasia of the ment phase. lid margin, chalazia, and signs of keratitis. Lid The daily dosage for fusidic acid and microbiology was not performed in this study, placebo gel was one application of one drop of as the aim was to determine the symptomatic gel 12 hourly to the inferior fornix and to the response to each of the two antibiotic treat- eyelid margins, and for oxytetracycline and ment regimens. placebo tablets it was one tablet taken 12 Data were analysed for the comparative hourly with 100 ml of water, at least 1 hour efficacy of the three active treatment regimens before meals. (fucidin, oxytetracycline, or the combination) The maximum duration of therapy was 8 relative to the two placebo phases and to one months, broken into four 2 month phases. All another using Friedman's test, followed by the patients entering the study received topical Mann Whitney U test, comparing magnitude placebo gel and placebo tablets for the first 2 of changes. This was based on the patients' month phase. In the following 2 month phase, subjective assessment of symptoms with each a randomised 50% of patients received fusidic change of treatment and the investigators' acid gel and oxytetracycline tablets, while the assessment of signs, before and after treat- other 50% received either fusidic acid gel and ment. The x2 with Yates's correction was used placebo tablets or oxytetracycline tablets and to test differences in proportions (for example, placebo gel. This was followed by a third phase responders/non-responders). All tests were two of 2 months in which all patients received tailed and statistical significance was defined as placebo medication again. In the final 2 month p<0 05. phase, the 50% of patients who had originally received a combination of both fusidic acid and oxytetracycline were randomly selected so Results that equal numbers would receive either Sixty one patients with chronic blepharitis Fucithalmic gel and placebo tablets or oxy- were recruited into this clinical trial of whom tetracycline tablets and placebo gel. The 43 (71%) completed both placebo and an remaining 50% ofpatients received a combina- active treatment phase. Thirty four patients tion ofboth fusidic acid gel and oxytetracycline completed all four phases, seven completed tablets. three phases, and two completed two phases. The trial flow diagram, for each phase, is Of 43 patients included in the analysis, 23 http://bjo.bmj.com/ illustrated below: were male and 20 were female with a mean age Placebo (Phase 1) Placebo of 56; three patients had unilateral blepharitis, Fucithalmic+ Fucithalmic or the remainder being bilateral. Twenty seven oxytetracycline (Phase 2) oxytetracycline patients were recruited from a previous study Placebo (Phase 3) Placebo of chronic blepharitis, when 116 were tested Fucithalmic or Fucithalmic+ for CMI to S of whom 11 had given aureus,2 on October 2, 2021 by guest. Protected copyright. oxytetracycline (Phase 4) oxytetracycline an enhanced CMI response. Sixteen addi- Patient compliance was expected to be high, tional patients were recruited with chronic since blepharitis is a chronic disease and blepharitis but were not tesL%., for enhance- patients are usually well motivated to partici- ment. No serious adverse reactions were pate in treatment regimens. On entry to the reported. study, inclusion criteria, based on previous Of 43 patients analysed, 18 (44%) were assessment of symptoms and signs,2 and exclu- thought to have concomitant non-ocular sion criteria were checked while demographic rosacea of mild degree (erythema and blushing data, concurrent disease, and medication were of facial skin), one of moderate, and one of recorded. severe degree. Four patients were thought to The following symptoms were recorded by have seborrhoea of mild degree. Nine patients the patients on diary score cards before the (23%) suffered from atopy, seven of mild, and start of the trial and every 2 months thereafter, two of moderate degree. immediately before commencing the next Analysis of symptoms, recorded by patients treatment phase: burning, itching, grittiness, before the trial, gave an overall mean score of soreness, dryness, blurring of vision, pain, dis- 23 (range 0 to 32) out of a possible 104 for charge, morning stickiness, redness, watering, both eyes together, with mean values of 22.7 photophobia, and aggravation of symptoms in and 22-6 respectively for those with and with- smoky or hot environments. These symptoms out enhanced CMI to S aureus. Analysis of were graded in severity; 0 absent, 1 mild, 2 signs gave a mean score of 9 4 out of a possible moderate, and 3 severe. 60 for each eye; mean scores for enhanced and The patients were examined by the investi- non-enhanced patients were 10-3 and 9 1 gators, at the start of the study and after each respectively with no statistical difference. active treatment phase, for signs of chronic Within this patient group folliculitis was blepharitis including folliculitis, collarettes, observed in only one patient. 44 Seal, Wright, Ficker, Hagan, Troski, Menday

Table 1 Change in symptoms ofblepharitis with change in treatment (placebo to active which the aetiology is unproved, and the phase, selfscored on diary cards) response frequently assessed on symptomatic % Improvement grounds, a placebo controlled study is required Improved No change Worse Br J Ophthalmol: first published as 10.1136/bjo.79.1.42 on 1 January 1995. Downloaded from Ros/Non-R Ros/Non-R Ros/Non-R Ros/Non-R to establish efficacy of antibacterial therapy. Placebo to fusidic acid (n= 18) 6/0 1/1 1/9 75/0 Furthermore, a double blind design is essential Placebo to oxytetracycline (n=22) 5/3 2/2 3/7 50/25 to remove the possibility of both investigator Placebo to fusidic acid+ oxytetracycline (n=34) 6/5 3/5 8/7 35/30 and patient bias. This study which was double blind and Ros=patient with rosacea; Non-R=patient without rosacea. placebo controlled required each patient to receive placebo treatment for a 2 month period Initial analysis did not take account of the before each of two active treatment phases, confounding factor of rosacea. Comparison of when changes in symptoms and signs were symptomatic responses to each antibiotic was investigated. The study design also allowed for not significant (p=0 8). Comparison of the crossover between combination and single drug magnitude of changes for fusidic acid plus therapy and should be a useful model for future oxytetracycline versus fusidic acid or oxytetra- trials of drug therapy in blepharitis. A third 2 cycline alone, and for fusidic acid versus oxy- month period of placebo treatment at the end tetracycline, was not significant. Comparison of the trial would have been useful, as it would of 'matched pairs' in 30 patients who com- have allowed an evaluation ofthe effect ofwith- pleted all four phases was performed for fusidic drawing active therapy and the detection of any acid versus fusidic acid plus oxytetracycline for relapse that occurred. The use of a self assess- 12 patients, and for oxytetracycline versus ment scheme removed the bias that can occur fusidic acid plus oxytetracycline in 18 others, with interrogation by an investigator. but no significant difference was found. Forty seven per cent of patients in this trial Scoring of symptoms was also analysed for all suffered from rosacea, compared with 27% in patients known to have either enhanced or the previous study.2 This was unexpected but non-enhanced CMI to S aureus, but no signifi- has allowed investigation oftherapy for rosacea cant differences were found. associated blepharitis within a randomised, Symptoms were then analysed for patients controlled trial. Analysis ofthe 20 patients with with and without rosacea following active rosacea has shown that topical fusidic acid treatment (Table 1). Seventy five per cent of alone was most beneficial in relieving symp- those with rosacea improved with fusidic acid toms (75%). This may reflect recently alone (X2 test with Yates's correction 7-6, described in vitro immunosuppressive effects p<0 01), 50% with oxytetracycline, but only of fusidic acid which inhibited T cell prolifera- 35% with the combination. The reverse tion similarly to cyclosporin A.15 occurred in patients without rosacea; 90%/o Results for oxytetracycline in the rosacea were worse with fusidic acid alone but 25% group are similar to those of Bartholomew who improved with oxytetracycline. found that it produced a higher number of The number of patients observed with remissions (19 of 35) compared with the group http://bjo.bmj.com/ improvement or no change at the end of each given placebo (11 of 35).11 These authors active treatment period compared with their demonstrated that non-specific signs, includ- own baseline, are given in Table 2. Separate ing blepharitis, responded well to oxytetra- results are given for those patients with and cycline but found no permanent change in without enhancement of CMI to S aureus; 19 conjunctival and corneal vascularisation. Their of 20 treatment episodes in 11 patients with findings of a high remission rate in one third of enhancement gave improvement (or no their patients using placebo demonstrates that on October 2, 2021 by guest. Protected copyright. change) on antibiotic therapy compared with spontaneous remissions can occur. Others 23 of 30 treatment episodes in 16 patients have found comparable results in rosacea of without enhancement. This was not statisti- skin without ophthalmic signs when treating cally significant. with oral and topical metronidazole in double blind placebo controlled trials'6; the explana- tion is unknown but suppression of cell medi- Discussion ated immunity has been suggested. In Open, randomised, comparative clinical trials addition, one other controlled trial of papular are often employed when comparing the rosacea ofthe skin demonstrated a comparable efficacy of different antibacterial treatments - effect between oxytetracycline and ampicillin, for example, in conjunctivitis.14 However, in both significantly better than placebo.'2 chronic conditions, such as blepharitis in Doxycycline has been shown to be beneficial in ocular rosacea, as a once daily dose, but the study was open and uncontrolled.13 These Table 2 Analysis ofsigns in blepharitis (placebo to active phase, scored by controlled studies, and our own patients, con- ophthalmologist) firm the uncontrolled findings of Brown'7 18 Improved or no change/worse (% better or no change) that oral oxytetracycline gave symptomatic in ocular rosacea patients Placebo to Placebo to Placebo tofusidic improvement fusidic acid oxytet acid +oxytet although half relapsed on drug withdrawal. An association between ocular rosacea Non-enhanced (n= 16) 5/2 6/2 12/3 (77) CMI enhanced (n=1 1)* 5/0 5/0 9/1 (95) and lid infection with S aureus, including CMI not known (n=16)t 4/0 5/1 9/1 (90) blepharitis, marginal ulceration, and other been pursued for 50 *8/11 had rosacea; t40% expected to have enhanced CMI for S aureus2. corneal complications, has CMI=cell mediated immunity; oxytet=oxytetracycline. years.'9 Brown demonstrated succinctly that Placebo controlled trial offfusidic acid gel and oxytetracycline for recurrent blepharitis and rosacea 45

lid colonisation with S aureus often occurred in well as to systemic oxytetracycline but patients with rosacea. They found that such responded less well to the combination. In colonisation did not change with oxytetracy- contrast, symptoms of other patients with Br J Ophthalmol: first published as 10.1136/bjo.79.1.42 on 1 January 1995. Downloaded from cline treatment on a qualitative basis although blepharitis did not respond to topical fusidic they did not investigate quantitatively.17 acid. Analysis of signs has suggested greater Lempert20 has shown that chalazia of the lids improvement with fusidic acid or oxytetracy- are more common in rosacea (64%) compared cline in those patients with enhancement to with a control group (13%). Others have S aureus, although this was not proved described granulomatous disease of the lids in statistically. This suggests that before com- rosacea in both the black2l and white popula- mencing treatment for blepharitis, it should be tion,22 the latter suggesting similarities of established whether there is coexisting rosacea, histological lesions with mycobacterial and particularly mild and limited to excessive syphilitic disease. We have observed in this blushing, and treatment targeted accordingly. study that eight out of 11 patients with This trial was conducted under clinical trials exemption and enhanced CMI to S aureus has rosacea. This was approved by Moorfields Hospital ethics committee. The study was conducted according to the Declaration of Helsinki observation is worthy of further study. as adopted by the 18th World Medical Assembly (1965) and as The relation between rosacea and CMI revised by the World Medical Assemblies Tokyo (1975) and Venice (1983). enhancement to S aureus remains speculative We are grateful to Mr V Andrews, principal pharmacist, but both appear frequently to coexist; in the Moorfields Eye Hospital for distributing placebo and active treatments to our patients on a randomised basis, and to previous study this enhancement was associ- Mr IA Mackie for advice. ated with folliculitis.2 In addition, rosacea lids are frequently colonised with S aureus, in a 1 McCulley JP, DoughertyJM, Deneau DG. Classification of chronic blepharitis. Ophthalmology 1982; 89: similar manner to the atopic individual but in 1173-80. those in which this CMI is not enhanced.3 2 FickerL, Ramakirshman M, Seal DV, Wright P. Role of cell-mediated immunity to staphylococci in blepharitis. While it is not understood why patients with AmJOphthalmol 1991; 111: 473-9. rosacea and others develop CMI enhancement 3 Ficker L, Seal DV, Wright P. Staphylococcal blepharitis (Chapter 58). In: Pepose J, Holland G, Wilhelmus K, eds. to S aureus, the continued presence of S aureus Ocular infection and immunity. Chicago: Mosby, 1994. on the lid margin of the enhanced patient can 4 Seal D, FickerL, Wright P. Role of coagulase-negative staphylococci in chronic blepharitis. Microb Ecol Health be expected to contribute to a chronic inflam- Dis 1992; 5: 69-75. matory reaction of granulomatous type. This, 5 Hvidberg J. Fusidic acid in acute conjunctivitis. Acta in turn, would benefit from both the anti- Ophthalmol 1987; 65: 43-7. 6 Dirsdal M. Fucithalmic in acute conjunctivitis. Acta inflammatory effect of oxytetracycline (and Ophthalmol 1987; 65: 129-33. possibly acid and metronidazole) as 7 Taylor B, Burd EM, TabbaraKF. Comeal and intra-ocular fusidic penetration of topical and subconjunctival fusidic acid. well as the antibacterial effect of topical fusidic BrJOphthalmol 1987; 71: 598-601. acid which reduces S aureus colonisation of 8 Taylor PB, TabbaraKF, Burd EM. Effect of preoperative fusidic acid on the normal eyelid and conjunctival lids. Failure of systemic oxytetracycline to bacterial flora. BrJ Ophthalmol 1988; 72: 206-9. eradicate lid colonisation by S aureus in 9 Dilly PN, Mackie IA. Distribution of tetracycline in the 18 conjunctiva of patients on long-term systemic doses. rosacea17 is expected. The observation that BrJOphthalmol 1981; 69: 25-8. http://bjo.bmj.com/ the combination of fusidic acid and oxytetra- 10 Dougherty JM, McCulley JP, Silvany ME, Meyer DR. The role of tetracycline in chronic blepharitis. Invest cycline was less frequently effective than either Ophthalmol Vis Sci 1991; 32: 2970-5. 11 Bartholomew RS, Reid Cheesborough alone was surprising (Table 1). It was not BJ, MJ, Macdonald M, Galloway NR. Oxytetracycline in the treatment of statistically significant, however, and warrants ocular rosacea: a double-blind trial. BrJ Ophthalmol 1982; further investigation particularly with reference 66: 386-8. 12 Marks R, Ellis J. Comparative effectiveness of oxytetra- to therapy for systemic symptoms. cycline and ampicillin in rosacea. Lancet 1971; ii: 1049-52. Patients with blepharitis but without rosacea 13 Frucht-Pery J, Chayet AS, Feldman ST,LinS, Brown SI. The effect of doxycycline on ocular rosacea. AmJf on October 2, 2021 by guest. Protected copyright. did not benefit symptomatically from topical Ophthalmol 1989; 107: 434-5. fusidic acid and there was less benefit from 14 Trimethoprim-Polymyxin B Sulphate Ophthalmic Ointment Study Group. Trimethoprim-polymyxin B oxytetracycline. This reflects the different sulphate ointment versus chloramphenicol ophthalmic aetiologies of blepharitis in these patients. ointment in the treatment of bacterial conjunctivitis. JAntimicrob Chemother 1989; 23: 261-6. Coagulase negative staphylococci were 15 Bendtzen K, Diamant M, Faber V. Fusidic acid, an unlikely to have been pathogenic in either immunosuppressive drug with functions similar to cyclosporin A. Cytokine 1990; 2: 423-9. group, based on our previous work.24 16 Nielsen PG. Treatment of rosacea with 1% metronidazole Treatment of blepharitis has in the past been cream. A double-blind study. Br J Dermatol 1983; 108: 327-32. based on laboratory tests demonstrating 17 BrownS, Shahinian L. Diagnosis and treatment of ocular oxytetracycline inhibition of lipase production rosacea. Ophthalmology 1978; 85: 779-86. 18 Jenkins MS, Brown SI, LempertSL, WeinbergRJ. Ocular by coagulase negative staphylococci.'0 rosacea. Am J Ophthalmol 1979; 88: 618-22. Suppression of this bacterial enzyme in meibo- 19 Wise G. Ocular rosacea. Am J Ophthalmol 1943; 26: 591-609. mianitis'0 is still theoretical and controlled 20 Lempert SL, Jenkins MS, Brown SI. Chalazia and rosacea. clinical trials need to be performed. ArchOphthalmol 1979; 97: 1652-3. 21 Browning DJ, Rosenwasser G, Lugo M. Ocular rosacea in In conclusion, we have found that symptoms blacks.JfAm Ophthalmol 1986; 101: 441-4. of rosacea associated blepharitis frequently 22 Patrinely JR, Font RL, Anderson RL. Granulomatous acne rosacea of the eyelids. Arch Ophthalmol 1990; 108: responded well to topical fusidic acid gel as 561-3.