Discordance Between Facial Wrinkling and the Presence of Basal Cell Carcinoma

OBSERVATION Discordance Between Facial Wrinkling and the Presence of Basal Cell Carcinoma

Rebecca C. C. Brooke, MRCP; Sophie A. Newbold, MB, ChB; Nicholas R. Telfer, FRCP; Christopher E. M. Griffiths, MD, FRCP

Background: During routine surgical treatment of basal completed a questionnaire about sun exposure. Despite cell carcinomas (BCCs), we observed an apparent in- being older (P=.03), patients with a BCC were found to verse relationship between the presence of a BCC and sig- have a lower mean grade of wrinkling than controls nificant wrinkling of the face. To ascertain the veracity (P=.001). Using logistic regression, increasing grade of of this observation, we performed a clinical and ques- wrinkling was associated with a progressive reduction in tionnaire-based case-referent study. risk of developing a BCC.

Observation: One hundred eighteen successive white Conclusion: Mechanisms responsible for the produc- patients (mean±SD age, 71.9±9.5 years) attending the tion of facial wrinkles may either be separate to or miti- hospital for surgical treatment of a BCC and 121 control gate against the development of a BCC of the face. (no skin cancer) patients (mean±SD age, 69.1±10.8 years) were assessed for grade of facial wrinkling using a previously validated photonumeric scale of photoaging and Arch Dermatol. 2001;137:751-754

HILE conducting regression and adjusting for age, sex, and clinical studies on smoking history, an increase in grade of facial photoaging wrinkling was associated with a progres- and during surgi- sive reduction in the likelihood of devel- cal treatment of fa- oping a BCC. As the frequency of grade 2 Wcial basal cell carcinomas (BCCs), we ob- wrinkling alone was small (n=9), grades served that patients who develop this form 2 and 3 were combined to form the refer- of skin cancer appeared to have smoother, ent or comparison group. In this study, no less wrinkled facial skin than those who subjects had a wrinkle grade of 0 or 1. do not (Figure 1). Patients with a BCC There was some evidence of a threshold did not seem to develop the deep coarse effect, with the maximum protective ef- facial wrinkling commonly designated as fect being observed at wrinkling grade 5 one of the markers of significant sun ex- (Table). Multivariate analysis indicated posure. Another observation seemingly in that those subjects with a wrinkling of favor of this hypothesis is the apparent grade 5 were 90% less likely to have a BCC scarcity of BCCs on the heavily sun- than were subjects with a lower amalgam- exposed nape of the neck but the pres- ated wrinkle grade—grades 2 and 3. ence of characteristic deep wrinkling at this When analyzed by Fitzpatrick pho- site—cutis rhomboidalis nuchae. Thus, we totype, after adjusting for age, compared performed a case-referent study to deter- with those who always tan (Fitzpatrick mine if patients with a BCC on the face had grade IV), those with lower grades (tans a significantly lesser degree of facial wrin- slowly or not at all) were more likely to kling compared with those in the same age develop a BCC (grade II or III: odds ratio, range who did not have a BCC. 2.7 [95% confidence interval, 1.1-4.9]; and grade I: odds ratio, 1.7 [95% confidence RESULTS interval, 0.7-4.0]).

From the Dermatology Centre, Despite being older (P=.03, t test), pa- COMMENT Hope Hospital, The University tients with a BCC had a lower mean grade of Manchester, Manchester, of wrinkling than controls (P=.001, Mann- The results of this study indicate that England. Whitney test) (Figure 2). Using logistic BCCs are more likely to occur on facial

(REPRINTED) ARCH DERMATOL / VOL 137, JUNE 2001 WWW.ARCHDERMATOL.COM 751

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 The clinical hallmarks of high cumulative sun exposure in whites are coarse wrinkles, actinic keratoses, PATIENTS AND METHODS telangiectasia, and actinic lentigines (“age spots”), features known alternatively as photoaging. In individual A clinical and questionnaire-based case-referent study cases, the features of photoaging often differ despite was performed within the setting of a dermatology equivalent sun exposure; for instance, some people have tertiary referral center in Manchester, England. We wrinkles predominantly while others have relatively studied all identified white patients (n=118 [50 men smooth facial skin with telangiectasia.7 Indeed, it ap- and 68 women]) older than 50 years (mean±SD age, pears that individuals with cutis rhomboidalis nuchae of- 71.9±9.5 years) who presented to the Dermatology ten have smooth unwrinkled faces. English et al8 dem- Centre at Hope Hospital, Manchester, during a onstrated that on traditionally wrinkled but less BCC 3-month period in 1999. Patients were attending the prone sites—the back of the hand and the nape of the hospital either for surgical treatment of a BCC (thus neck—wrinkling as assessed by cutaneous microtopog- giving histological verification of the diagnosis) or for a first diagnosis. Control patients were 121 white raphy or a clinical severity scale, respectively, showed patients (59 men and 62 women) older than 50 years moderate agreement with reported sun exposure. How- (mean±SD age, 69.1±10.8 years) who had never had ever, their study did not assess facial wrinkling. Photo- a BCC and who were outpatients or inpatients un- aging also occurs in nonwhite skin but with different phe- der the care of other departments within the hospi- notypes—actinic lentigines predominate in Far East tal. All identified patients in any department able to Asians,9 whereas wrinkles can be a feature of long-term answer the questionnaire and give informed con- sun exposure in black skin. sent were included to minimize selection bias. There Relatively little is known about the cellular biologi- were no other exclusion criteria. cal features of sun-induced wrinkling, although it is ac- All subjects completed the same questionnaire cepted that this may be due in part to loss of extracellu- and were assessed for extent of wrinkling by a single lar matrix components subjacent to the dermoepidermal observer (S.A.N.) using a previously validated pho- 10 1 junction. Such components include type I, III, and VII tonumeric assessment scale of facial wrinkling. This 11-13 scale gives a range of severity for facial wrinkles from collagens and fibrillin —a cumulative loss resulting 0 (no wrinkling) to 8 (severe wrinkling). Although from a combination of decreased synthesis and in- this design could result in observer bias, the alter- creased breakdown from sun-induced activation of ma- native of using high-quality accurate photographs of trix metalloproteinase production.14 each study participant suitable for third-person blind- There is an increasing body of evidence to indicate ing was considered financially unfeasible. The ques- that the relationship between sun exposure and the sub- tionnaire asked for demographic details; known con- sequent development of a BCC is nonlinear.5 The chro- founders (eg, radiation treatment for acne); childhood nicity of sun exposure is directly linked to the subse- arsenic ingestion as “tonics”; other skin problems and treatments for these, such as UV irradiation; smok- quent development of a squamous cell carcinoma. By ing history; and occupational history. No patient or contrast, the intermittency of intense sun exposure ap- control self-declared any prior cosmetic procedure pears to be a stronger determinant of BCC risk than an (dermabrasion, laser resurfacing, or a chemical peel equivalent amount of sun exposure delivered over a longer for facial photoaging). Skin phototype was estab- period.5 lished for all participants using the Fitzpatrick clas- Others15 have found that BCC and squamous cell sification,2 which categorizes the sensitivity of white carcinoma of the skin are more likely to occur in people skin to UV radiation into 4 groups. Average sun ex- who tan with difficulty, equating to Fitzpatrick photo- posure was included in the questionnaire but not con- types I to III. Our observations are in keeping with this trolled for because the relationship between sun ex- but identify a particular nonwrinkled phenotype within posure and the development of a BCC, although still causal, is less well defined than for that of squa- this group that more likely presents with a BCC. Al- mous cell carcinoma.3-6 There is thought to be a pla- though we studied BCC, as this is the more prevalent non- teau in the risk for a BCC at higher cumulative doses melanoma skin cancer, it is possible that squamous cell of UV radiation. In addition, recall bias can be intro- carcinoma may also occur more commonly on non- duced when trying to assess lifetime sun exposure, wrinkled as opposed to wrinkled sun-exposed skin. While particularly for a group of patients who are aware that acknowledging the potential for observer bias, as one ob- they have developed a carcinoma thought to be etio- server carried out all the photonumeric assessments of logically related. patients and controls, this study strongly suggests that All statistical analyses were carried out using Stata the underlying mechanisms responsible for the produc- software, version 6.0 (Stata Corp, College Station, tion of facial wrinkles may be different from those pro- Tex). ducing BCC. Intriguingly, particularly as BCCs are rare or comparatively rare on the nape of the neck and back of the hand, this could also imply that wrinkling pro- tects against the development of a BCC. An explanation skin that, although photoaged, is relatively unwrinkled. for this could lie in the hypothesis that there are per- This observation implies that the mechanisms respon- haps 2 forms of the collagen repair process in the pap- sible for producing wrinkles associated with photoag- illary dermis following sun exposure—a nonfibrotic re- ing are different from those involved in the develop- sponse, with loss of collagen leading to a wrinkled clinical ment of BCCs. phenotype and relatively few BCCs; and a fibrotic re-

(REPRINTED) ARCH DERMATOL / VOL 137, JUNE 2001 WWW.ARCHDERMATOL.COM 752

Figure 1. Two examples of basal cell carcinomas occurring on a background of nonwrinkled but photoaged and telangiectatic skin: in a 72-year-old man (A) and in a 55-year-old man (B).

50 Patients With Basal Odds Ratios and 95% Confidence Intervals (CIs) for Grade 45 Cell Carcinoma (n=118) of Facial Wrinkling and Development of Basal Cell 40 Control Patients Without Basal Carcinoma, Adjusted for Age, Sex, and Smoking History Cell Carcinoma (n=121) 35 Odds Ratio (95% CI) 30

25 Variable Univariate Analysis Multivariate Analysis 20 Age 1.0 (1.0-1.1) 1.1 (1.0-1.1) % of All Patients 15 Sex Male 1.0* 1.0* 10 Female 1.3 (0.8-2.2) 2.2 (1.2-4.1) 5 Smoked 0 Never 1.0* 1.0* 12345678 Ever 1.0 (0.6-1.6) 1.5 (0.8-2.9) None Grade of Wrinkling Severe Wrinkle grade 2 and 3 (n = 96) 1.0* 1.0* Figure 2. Distribution of wrinkling severity among all patients. Facial 4 (n = 56) 0.9 (0.4-1.7) 0.4 (0.2-1.0) wrinkling is less for patients with basal cell carcinoma compared with control patients without basal cell carcinoma. 5 (n = 30) 0.2 (0.1-0.6) 0.1 (0.1-0.2) 6 (n = 24) 0.3 (0.1-0.8) 0.1 (0.1-0.3) 7 (n = 24) 0.5 (0.2-1.2) 0.1 (0.1-0.4) sponse, with preservation of collagen resulting in a clini- 8 (n = 11) 0.2 (0.1-1.0) 0.1 (0.1-0.2) cal phenotype with few if any wrinkles but a tendency to develop BCCs. One speculative mechanism that could *Referent group. underlie these processes is predicated on a role for transforming growth factor ␤,16 which stimulates collagen for- Whatever the pathomechanisms that link clinical mation, thereby facilitating wrinkle repair, but is locally phenotype and BCCs, this study has provided some clues, angiogenic and immunosuppressive, leading to telangi- through clinical observation, to factors that determine a ectasia and growth of BCCs, respectively. predisposition to nonmelanoma skin cancer of the face.

(REPRINTED) ARCH DERMATOL / VOL 137, JUNE 2001 WWW.ARCHDERMATOL.COM 753

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 Accepted for publication January 31, 2001. squamous cell carcinomas of the skin. Br J Cancer. 1996;73:1447-1454. This study was supported by the Academic Dermatol- 7. Griffiths CE. The clinical identification and quantification of photodamage. Br J Dermatol. 1992;127(suppl 41):37-42. ogy Fund, Hope Hospital, The University of Manchester, 8. English DR, Armstrong BK, Kricker A. Reproducibility of reported measure- Manchester, England. ments of sun exposure in a case-control study. Cancer Epidemiol Biomarkers Corresponding author and reprints: Christopher E. M. Prev. 1998;7:857-863. Griffiths, MD, FRCP, Dermatology Centre, Hope Hospital, 9. Goh SH. The treatment of visible signs of senescence: the Asian experience. Br J Dermatol. 1990;122(suppl 35):105-109. The University of Manchester, Salford, Manchester M6 8HD, 10. Contet-Audonneau JL, Jeanmarie C, Pauly G. A histological study of human wrinkle England (e-mail: [email protected]). structures: comparison between sun-exposed areas of the face, with or without wrinkles, and sun-protected areas. Br J Dermatol. 1999;140:1038-1047. 11. Talwar HS, Griffiths CE, Fisher GJ, Hamilton TA, Voorhees JJ. Reduced type I REFERENCES and III procollagens in photodamaged adult human skin. J Invest Dermatol. 1995; 105:285-290. 1. Griffiths CE, Wang TS, Hamilton TA, et al. A photonumeric scale for the assess- 12. Craven NM, Watson RE, Jones CJ, Shuttleworth CA, Kielty CM, Griffiths CE. Clini- ment of cutaneous photodamage. Arch Dermatol. 1992;128:347-351. cal features of photodamaged human skin are associated with a reduction in col- 2. Fitzpatrick TB. Soleil et peau. J Med Esthet. 1975;2:33-34. lagen VII. Br J Dermatol. 1997;137:344-350. 3. Vitasa BC, Taylor HR, Strickland PT, et al. Association of nonmelanoma skin can- 13. Watson RE, Griffiths CE, Craven NM, Shuttleworth CA, Kielty CM. Fibrillin-rich cer and actinic keratosis with cumulative solar ultraviolet exposure in Maryland microfibrils are reduced in photoaged skin: distribution at the dermal-epidermal watermen. Cancer. 1990;65:2811-2817. junction. J Invest Dermatol. 1999;112:782-787. 4. Hunter DJ, Colditz GA, Stampfer MJ, et al. Risk factors for basal cell carcinoma 14. Fisher GJ, Datta SC, Talwar HS, et al. Molecular basis of sun-induced premature in a prospective cohort of women. Ann Epidemiol. 1990;1:13-23. skin ageing and retinoid antagonism. Nature. 1996;379:335-339. 5. Kricker A, Armstrong BK, English DR, Heenan PJ. Does intermittent sun expo- 15. Kricker A, Armstrong BK, English DR. Sun exposure and non-melanocytic skin sure cause basal cell carcinoma? a case-control study in Western Australia. Int cancer. Cancer Causes Control. 1994;5:367-392. J Cancer. 1995;60:489-494. 16. Schmid P, Itin P, Rufli T. In situ analysis of transforming growth factors-␤ (TGF- 6. Rosso S, Zanetti R, Martinez C, et al. The multicentre south European study ␤1, TGF-␤2, TGF-␤3) and TGF-␤ type II receptor expression in basal cell carci- “Helios,” II: different sun exposure patterns in the aetiology of basal cell and nomas. Br J Dermatol. 1996;134:1044-1051.

News and Notes

The Annales de Dermatologie/Annals of Dermatology, published continu- ously since 1869, is one of the most long-standing dermatology journals in the world. In accordance with the journal’s editorial policy, the most pertinent pa- pers from French-speaking dermatologists are published exclusively in French. To share its publications with the English-speaking community, the Annales de Dermatologie/Annals of Dermatology initiated, in 1998, free Internet access (www.e2med.com/ad) to the English version of the most inter- esting articles. All English versions are full articles available to colleagues consulting PubMed or our Internet site and can be accessed free of charge. English-speaking dermatologists worldwide will certainly appreciate read- ing the latest data available from their French-speaking colleagues.

(REPRINTED) ARCH DERMATOL / VOL 137, JUNE 2001 WWW.ARCHDERMATOL.COM 754