Services on Demand
On-line version ISSN 1806-4841
An. Bras. Dermatol. vol.79 no.4 Rio de Janeiro July/Aug. 2004
CLINICAL, LABORATORY AND THERAPEUTIC INVESTIGATION
Risk markers and risk factors for actinic keratosis and basal cell carcinoma: a case-control study*
Fernando Passos da RochaI; Ana M. B. MenezesII; Hiram Larangeira de Almeida JuniorIII; Elaine TomasiIV
IMasters Degree Student
IITitular Professor of Pneumology - Federal University of Pelotas
IIIAdjunct Professor of Dermatology - Federal University of Pelotas and Catholic University of Pelotas (Masters degree course in Health and Behavior)
IVAssistant Professor of Qualitative Methodology - Catholic University of Pelotas
BACKGROUND: Incidence of cutaneous neoplasms
is increasing worldwide and there is little information from South Brazil about
its markers and premalignant lesions.
OBJECTIVES: To identify the risk factors and dermatological risk markers for cutaneous malignancies.
METHODS: A case-control study nested in a population-based prevalence survey was performed in adults aged >50 living in the urban area of Pelotas. In the prevalence study, a questionnaire was applied to identify possible malignant or premalignant lesions and these subjects were examined by two doctors (gold standard). A sub-sample of the persons who answered No to the questions was also sent to the Ambulatory for the same procedure. The sample was constituted of 288 persons: 74 cases and 214 controls. The number and the type of efflorescence was evaluated at the interview with the doctors. The questionnaire also investigated socioeconomic level, behavioral variables and skin examination. The measure of effect used was the Odds Ratio (OR); and to control confounding factors, logistic regression, with hierarchical pattern, was utilized.
RESULTS: In the multivariate analysis, the following variables remained significant: age >80 years (OR = 10.21), white skin (OR = 4.85), blond or red hair (OR = 3.69) and risk markers: solar elastosis (OR = 4.35), cutis rhomboidalis nuchae (OR = 2.88) and more than 10 melanoses on the back of the hands (OR = 6.0).
CONCLUSION: Elderly individuals, with fair skin and hair, solar elastosis, cutis rhomboidalis nuchae and a great number of melanoses on the hands, are at high risk for actinic keratosis and basal cell carcinoma.
Key words: epidemiology; risk factors; skin neoplasms.
The expression "skin cancer epidemic" has been frequently used since the incidence of this disease has been increasing and now involves from 900,000 to 1,200,000 people a year, in the USA.1 The cost to the health insurance companies for the treatment of nonmelanoma cutaneous tumors (NMCT), in the years 1994 and 1995, surpassed 500 million dollars.2
In Canada, the incidence is also growing, the most frequent being basal cell carcinomas (BCC) and squamous cell carcinomas in areas exposed to the solar light, and the most common melanomas are in the trunk of men and lower limbs of women.3
Likewise in Brazil, the rates of those lesions have been growing,4 mainly in the South, where there is a prevalence of colonization by white-skinned European immigrants.
Despite the low mortality, NMCT are responsible for important sequels following surgical treatment, as well as an onus for the health systems.5 Melanomas present very high morbimortality in the more advanced phases, although they are not as prevalent as NMCT.3
The early diagnosis of skin cancer is an important strategy for reducing the treatment costs and improving the prognostic of the disease. The identification and treatment of precursor lesions, as well as dysplastic melanocytic nevus and actinic keratoses, are also fundamental to lower the incidence of malignant lesions.
To improve the effectiveness of the campaigns to detect malignant cutaneous lesions, people at greater risk for developing the disease should be identified and encouraged to participate in these campaigns.6
Risk factors, such as fair skin, blue or green eyes, blond or red hair, time of exposure to the sun, propensity to sunburn and solar sensitivity,7 have been associated to a greater risk for the development of NMCT; the presence of dysplastic nevi is associated to melanomas.8
The presence of telangiectasia in the face, solar elastosis and freckles considered risk markers for NMCT.9
Maia, in work carried out in Brazil, identified risk factors for NMCT to be skin types I and II according to the Fitzpatrick classification, activity in rural areas and presence of solar lesions.10
Similarly, Zanetti and cols., in Europe, found that color of the skin, eyes, and hair is strongly associated to greater risk of developing NMCT.11
In study in England by Lear and cols., risk factors of 1.66 were observed for skin type I, 1.61 for people with red or blond hair, 1.71 for individuals with green or blue eyes, and 2.36 for people of high social class, probably due to intermittent exposure to solar radiation.5
The objective of the present study was to identify risk factors and risk markers for premalignant and malignant cutaneous lesions in a population aged 50 years or over in a city located in the South of Brazil.
MATERIAL AND METHODS
A case-control study was performed nested in a prevalence study done on a population-based sample in the city of Pelotas, State of Rio Grande do Sul, Brazil. Details for the calculation of sample size and methodology of the prevalence (or transversal) study, from which the cases and controls for the current study originate, have been published in another article.12
In the prevalence investigation, a questionnaire was applied in 2,112 homes, among adults aged >50 years (n = 1,292), with general demographic questions (skin color, eye color, hair color, sensitivity of the skin to the sun and age); socioeconomic (income and education); life habits in general (time of exposure to the sun at work, tobacco and alcohol usage), as well as specific subjects on the appearance of skin lesions within the preceding six months. Individuals who answered yes to the specific questions underwent a skin examination. A sub-sample that answered no was also examined, four census sections were selected at random using a raffle and all individuals of the same age group were examined. The results of the prevalence of premalignant and malignant lesions, and of sensitivity, specificity, predictive value and accuracy of the questionnaire for screening the lesions have also already been published.12
At skin exam, markers of prolonged exposure to solar light were documented (elastosis in the face, cutis rhomboidalis nuchae, cervicofacial solar poikiloderma and solar melanosis in the face and in the back of hands, the latter were counted). cutis rhomboidalis nuchae, a variant of solar elastosis, was evaluated independently, since it is a very well-defined entity and easily detected during skin exam.
People with malignant lesions were referred for surgical resection. Premalignant lesions were treated with cryotherapy. Individuals that failed to attend the exam were seen in their homes by the doctors, in order to reduce losses and bias.
Univariate, bivariate and multivariate analysis by regression logistics was performed, according to a hierarchical model (Chart 1). All the variables were entered in the model according to hierarchical levels, which enabled an adjustment for confounding factors. For each level only the variables with p <0.20 were retained and the remainder were excluded. In the final model, those variables that reached a significance of p <0.05 were considered significant.
The variable total "hours of solar exposure " was constructed from the answer to how many daily working hours were done in the sun, multiplied by the number of years of work. Due to the small number of people with red-hair, these were grouped together with blond hair. The same procedure was done for people with green eyes, who, in the multivariate analysis, were considered together with blue-eyed individuals.
For quality control of the field work, 10% of the interviews were repeated by the team of supervisors.
Of the 1,292 individuals interviewed, 200 answered yes to the questions relative to the appearance of skin lesions in the preceding six months. Of these, 186 were examined and 45 presented premalignant lesions and nine basal cell carcinomas, totaling 54 people as true positive (the percentile of losses was 4.9%).
No malignant melanocytic lesions or squamous cell carcinomas were found.
A sub-sample (102 people) of those that answered no (1,092) to the screening questions were also examined and among these 18 premalignant lesions and two BCCs were diagnosed, totaling 20 false-negatives.
Thus, a total of 74 cases of premalignant and malignant cutaneous lesions were detected, and 214 controls were used for the study.
The sample size of this case-control study (74 cases and 214 controls) reached the 95% level of significance, power of 80%, prevalence of 5% in the non solar exposed and 17% in the exposed individuals, for variables with an Odds Ratio (OR) of four. For variables with OR equal to three and other parameters maintained in the model, the sample size reached the 90% confidence level and power of 75%. The losses and refusals amounted to 7%.
The demographic and behavioral variables are presented in table 1. The greatest percentage of cases was in the 60 to 69-year-old (39.2%) group, and, among the controls, from 50 to 59 years (48.6%). Regarding sex, females prevailed in the cases (66.2%) and controls (59.8%). The distribution of income, consumption of alcohol in the preceding 30 days and smoking was similar for cases and controls; the group with lower educational level was larger among the controls (26.6%) than the cases (19%).
Table 2 presents the distribution of dermatological characteristics for the sample and solar exposure.
Most of the sample was constituted by white people. There was a predominance, among the cases, of people with blond or red hair and blue or green eyes. A larger number of people in the cases (68.9%) reported reddening of the skin when exposed to the sun, in relation to the controls (49.3%).
The analysis of solar exposure at work revealed close percentages for cases and controls (59.5% and 55.1%, respectively). While for total hours of solar exposure during their life, a higher percentage was observed for the cases (37.8%) than for the controls (24.8%).
With regard to the risk markers for premalignant or malignant cutaneous lesions there was a clear difference between cases and controls, with higher prevalence of markers among the cases.
The results of the bivariate analysis for the demographic and behavioral variables are shown in table 3, and dermatological characteristics and solar exposure in table 4. The first column shows the crude OR with the 95% confidence interval (CI 95%); in the second column, adjusted OR is presented for sex and age, with the respective 95% confidence interval and p value.
The risk for premalignant and malignant cutaneous lesions was directly proportional to age, in that at age 80 years or over, the probability of acquiring the disease was approximately six times greater than among those in the 50 to 59-year-old group. In the analysis adjusted for sex, this risk continued. There were no statistical differences in terms of sex in the crude analysis, nor in the age-adjusted analysis.
The variables of income, education, smoking and alcohol consumption did not present an association with the appearance of the lesions.
Table 4, presents the results of the bivariate analysis of the dermatological characteristics and shows that white-skinned people were about six times more prone to premalignant and malignant cutaneous lesions, in relation to non-whites, a value that reached 7.6 when controlled for age and sex.
In the crude analysis, blond or red-haired individuals showed an OR of approximately five in relation to dark-haired individuals; the same occurred with blue-eyed (OR = 4.1) in relation to brown-eyed subjects. Those that had difficulty in tanning when exposed to the sun, presented an approximately two times greater probability of having a malignant or premalignant lesion than those that tan easily, both in the crude and in the adjusted analysis.
Exposure to the sun at work did not show an association with the studied outcome. It should be underscored that those that reported having worked for more time under solar exposure had a 60 to 80% greater risk for the disease, although the risk only reached borderline statistical significance (p = 0.05).
In the crude analysis of all the risk markers (Table 4) premalignant and malignant cutaneous lesions were significantly associated with: solar elastosis (OR = 6.5); cutis rhomboidalis nuchae (OR = 3.6); solar poikiloderma (OR = 1.7); six or more melanosis lesions in the face (OR = 3.6); and 10 or more lesions in the back of the hands (OR = 8.5). When adjusted for sex and age, the melanoses, in the face and in the back of the hands, presented a lower OR, such that the presence of melanoses in the face had a statistical significance of 0.06.
In table 5 the results of the multivariate analysis can be observed, after control of the confounding factors, according to a hierarchical model (Chart 1). All the variables that presented a value of p <0.20, were retained in the model and the others were excluded. All the biological variables entered in the analysis together, as an independent chain, after adjusting for other factors (age, skin color, color of the eyes, hair color and solar sensitivity). Of these, except for sex, all were kept in the model. As age increased, the possibility of acquiring premalignant and malignant cutaneous lesions increased; people over 80 years of age had a 10 times greater probability of having the disease than younger subjects. White-skinned Individuals showed an approximately five times greater risk for lesions than those with dark skin. The risk for those with light hair was 3.7 times higher than for those with dark hair. The variables of eye color and solar sensitivity presented p >0.05.
The socioeconomic variables (first level, since the biological variables were in an independent chain) were excluded from the analysis for having p > 0.20.
Following this, the variables of the second hierarchical level were entered into the model: smoking and consumption of alcohol, and exposure to the sun, and none remained in the model, due to a lack of statistical significance.
Three of the five risk markers, at the third level of analysis, remained significant after adjustment for all the variables located in higher levels or in the same level. These were: solar elastosis (OR 4.4); cutis rhomboidalis nuchae (OR 2.9); and over 10 melanoses in the back of the hands (OR six). The probability of having premalignant and malignant cutaneous lesions was directly proportional to the number of melanoses in the back of the hands. Solar poikiloderma and melanosis in the face did not remain in the model.
In summary the results of the multivariate analysis of the risk factors for premalignant and malignant cutaneous lesions, after control for confounding factors, were: advanced age, white skin, light hair, presence of solar elastosis, cutis rhomboidalis nuchae and high number of melanoses in the back of the hands.
The high morbidity of NMCT and high mortality from melanoma, in the advanced phases, represent an important public health problem. Prevention and early diagnosis, through knowledge of their risk factors and markers, are fundamental to reduce morbimortality.
The literature points to several risk factors for these diseases, such as skin type, eye color and light hair, sensitivity of the skin to the sun,9,11 time of exposure to the sun13 and sunburns in infancy.1
Marks, in Australia, has reported that the presence of solar keratosis is directly related to the development of epidermoid carcinoma, though the same does not occur in BCC.7 The presence of keratoses, associated to genetic propensity and cumulative exposure to the sun, can be an important marker for the development of this neoplasia.14
Green and Battistutta, in Australia, found that the presence of keratoses is a risk factor for two types of neoplasia (squamous and basal cell carcinoma). People with one to five solar keratoses in the face had increasing risk of developing BCC3,9, and the presence of more than five solar keratoses in the face presented a risk of 4.2 for epidermoid carcinoma.9
Sober found solar keratoses were risk markers for the development of both forms of NMCT.15
Solar elastosis was also identified by Green and Battistutta to be a risk marker for BCC and for epidermoid carcinoma.9 Melanoses on the back of the hands only showed an association with BCC when the number of lesions exceeded 20,16 the same was not observed for squamous carcinoma.9
In the present study no cases of squamous cell carcinoma or of melanoma were found and consequently it was not possible to evaluate any association between the risk markers and these types of cancer.
Nevertheless, the risk markers studied herein (cutis rhomboidalis nuchae, elastosis, poikiloderma and solar melanosis) were significantly associated with actinic keratoses and BCC in the crude analysis. In the multivariate analysis, the following risk markers remained significant: solar elastosis, cutis rhomboidalis nuchae and melanosis in the back of the hands. The presence of more than 10 melanoses conferred a six times greater risk for the studied outcome.
As for the association between malignant and premalignant skin lesions and certain risk factors, the findings corroborate the literature with regard to skin type, eyes and hair color.9,10,16 OOther risk factors, such as time of exposure in the sun at work and during leisure, sensitivity of the skin, family history of cutaneous cancer, sunburns in childhood and socioeconomic condition, are controversial findings in the literature.13,17
The case-control study by Maia,10 in Brazil, did not demonstrate an association between BCC and the following risk factors: color of the eyes and hair, family history of skin cancer, presence of freckles in adolescence and time of exposure to the sun. Skin type (Fitzpatrick classification types I, II and III, when associated to a history of sunburn), rural workers and presence of three or more cutaneous risk markers were associated to BCC in the multivariate analysis. The author proposes two hypotheses for the lack of statistically significant findings: small size of the sample and ethnic aspects. However, the sample size does not seem to be a plausible explanation, since the author studied 259 cases with 518 controls. The most probable reason is due to the fact that most of the population studied by Maia originated from areas of North and Northeast Brazil, where the ethnic composition is different.
A multicentric type case-control study, performed in Europe, has pointed to a high risk for people with blond or red hair, people with light eyes, a previous history of sunburns in childhood and people that suffer burns without ever becoming suntanned.11 In England, Lear has demonstrated that skin type 1, red or blond hair, green or blue eyes and high social class were associated with malignant and premalignant skin lesions, while no such association was found with working in the sun or smoking.5
In a case-control study in Canada, where most of the population lives between 490 and 540 degrees latitude and are of European descent, the most significant risk factor found for BCC was relative to rural work (p <0.001), while other risk factors, such as eye color, light hair, skin sensitivity and time of exposure in the sun, presented slightly less significance.16
The results of this study have shown that biological variables, except for sex, were strongly associated with the outcome, in the crude analysis. When adjusted for sex and age, the risks increased, probably due to the negative confounding factor caused by age.
The bivariate analysis demonstrated the following risks, in decreasing order: white skin (OR 7.6); red or blond hair (OR = 6.3); advanced age (OR = 6.1); blue eyes (OR = 4.3); and sensitivity of the skin to the sun (OR = 2.4).
There was no significant association between the outcome and socioeconomic variables or use of tobacco and alcohol, in spite of there being an increased risk among people with a higher educational level. It is possible that socioeconomic factors are associated to this disease (either risk or protection) due to the type of work or leisure habits.
Concerning the variable solar exposure, it is necessary to underscore the limitations in the collection of this information in the present study. The information gathered referred only to exposure to the sun while at work and leisure activities were not investigated. Besides, the variable presents an important memory bias, that hinders its correct quantification. Nevertheless, on analyzing total hours of exposure to the sun at work, it was observed that, in the category of greater solar exposure, the risk reaches 1.9 with a statistical significance of 0.05. In the multivariate analysis, when controlling this variable for all the others of the same and higher levels, it was no longer statistically significant. The literature reports different findings in relation to this risk factor.13 Gallagher, for instance, found no association between solar exposure and BCC, although a greater risk was found for people with sunburns in childhood and adolescence.17 Green and Battistutta, in Australia, only found an association with solar exposure at work, and not during leisure nor among people with a history of multiple sunburns in infancy.9
As age progresses, people with light skin and hair, with presence of skin lesions resulting from chronic solar exposure, such as solar elastosis, cutis rhomboidalis nuchae and a great number of melanoses in the hands, have a greater risk of developing premalignant and malignant cutaneous lesions. Thus these cutaneous markers should be used in the detection of risk groups by prevention campaigns.
1. Miller DL, Weinstock MA. Nonmelanoma skin cancer in the United States: incidence. J Am Acad Dermatol 1994;30(5Pt 1):774-8. [ Links ]
2. Housman TS, Feldman SR, Williford PM, et al. Skin cancer is among the most costly of all cancers to treat for the Medicare population. J Am Acad Dermatol 2003; 48: 425-9. [ Links ]
3. Gallagher RP, Ma B, McLean DI, Yang CP, Ho V, Carruthers JA, et al. Trends in basal cell carcinoma, squamous cell carcinoma, and melanoma of the skin from 1973 through 1987. J Am Acad Dermatol 1990;23(3 Pt 1):413-21. [ Links ]
4. Ministério da Saúde, Instituto Nacional de Câncer. Estimativa da incidência e mortalidade por câncer no Brasil. Brasília: INCA; 1999. [ Links ]
5. Lear JT, Tan BB, Smith AG, Bowers W, Jones PW, Heagerty AH, et al. Risk factors for basal cell carcinoma in the UK: case-control study in 806 patients. J R Soc Med 1997;90(7):371-4. [ Links ]
6. Rhodes AR. Public education and cancer of the skin. What do people need to know about melanoma and nonmelanoma skin cancer? Cancer 1995;75(2 Suppl):613-36. [ Links ]
7. Marks R, Rennie G, Selwood T. The relationship of basal cell carcinomas and squamous cell carcinomas to solar keratoses. Arch Dermatol 1988;124:1039-42. [ Links ]
8. Kelly JW, Rivers JK, MacLennan R, Harrison S, Lewis AE, Tate BJ. Sunlight: a major factor associated with the development of melanocytic nevi in australian schoolchildren. J Am Acad Dermatol 1994;30(1):40-8. [ Links ]
9. Green A, Battistutta D. Incidence and determinants of skin cancer in a high-risk australian population. Int J Cancer 1990;46(3):356-61. [ Links ]
10. Maia M, Proença NG, Moraes JC. Risk factors for basal cell carcinoma: a case control study. Rev Saúde Pública 1995;29(1):27-37. [ Links ]
11. Zanetti R, Rosso S, Martinez C, Navarro C, Schraub S, Sancho-Garnier H, et al. The multicentre south European study 'Helios'. I: Skin characteristics and sunburns in basal cell and squamous cell carcinomas of the skin. Br J Cancer 1996;73(11):1440-6 [ Links ]
12. Rocha FP, Menezes AMB, Almeida Junior HL, Tomasi E. Especificidade e sensibilidade de rastreamento para lesões cutâneas pré-malignas e malignas. Rev Saúde Pública 2002; 36: 101-6 [ Links ]
13. Rosso S, Zanetti R, Martinez C, Tormo MJ, Schraub S, Sancho-Garnier H, et al. The multicentre south european study "Helios II": differente sun exposure patterns in aetiology of basal cell and squamous cell carcinomas of the skin. Br J Cancer 1996;73:1447-54. [ Links ]
14. Salasche SJ. Epidemiology of actinic keratoses and squamous cell carcinoma. J Am Acad Dermatol 2000;42(1 Pt 2):S4-S7. [ Links ]
15. Sober AJ, Burstein JM. Precursors to skin cancer. Cancer 1995;75(2 Suppl):645-50. [ Links ]
16. Hogan DJ, To T, Gran L, Wong D, Lane PR. Risk factors for basal cell carcinoma. Int J Dermatol 1989;28(9):591-4. [ Links ]
17. Gallagher RP, Hill GB, Bajdik CD, Fincham S, Coldman AJ, MacLean DI, et al. Sunlight exposure, pigmentary factors, and risk of nonmelanocytic skin cancer. Arch Dermatol 1995;131(1):157-63. [ Links ]
Hiram Larangeira de Almeida Jr.
Dermatologia - Faculdade de Medicina
Av. Duque de Caxias 250
96.030-002 - Pelotas - RS
Received on September 29, 2003.
Approved by the Consultive Council and accepted for publication on March 12, 2004
* Work done in the Post-graduation Course on Epidemiology, Faculty of Medicine, Federal University of Pelotas.