Sunburn, sunscreens, and phenotypes: some risk factors for cutaneous melanoma in southern Brazil

Background The risk factors for cutaneous malignant melanoma have been studied in populations from numerous countries around the world. There are no published studies on the risk factors for this malignancy in Brazil, the largest country in South America. Methods A case-control study of all melanoma patients attending a university hospital in Porto Alegre, Brazil, was conducted over a 3-year period from 1995 to 1998. Phototype, hair and eye color, solar habits, history of sunburn, use of sunscreens, and the number of nevi were evaluated through a questionnaire and full body skin examination. Bivariate analysis and a logistic regression model were used to evaluate the data. Results One hundred and three malignant melanoma patients and 206 matched controls were enrolled in the study. The female to male ratio was 2 : 1. Light phototypes were more prone to the development of cutaneous melanoma. Although stronger in the bivariate analysis, in the logistic regression model, phototypes I or II and ephelides emerged only as moderate risk factors; light eye color and light hair color were not independently significant, with adjusted odds ratios (OR) close to zero. Commonly acquired nevi (CAN) showed a significant and strong effect in the bivariate analysis only when the "30 or more" category was compared to baseline. In the logistic regression model, the presence of a large number of CAN showed an association with increased levels of risk, although these findings did not reach classical significance. Dysplastic or atypical nevi seemed to contribute more strongly, although still with a moderate excess of relative risk. When the use of sunscreens was compared to no use at all, it appeared to show progressive protection as the solar protection factor (SPF) increased. Only SPF15 or greater (SPF15+) showed strong and significant protection when compared to baseline. Physical measures offered a weaker level of protection. Nevertheless, there was a significant increase in the risk of melanoma for those with a large number of sunburn episodes. It was found that 30 or more alleged episodes of sunburn showed a very strong OR of 11.4 (95% confidence interval, 2.6-50.5), the most significant in the study. Conclusions Phototypes I and II, freckles, a large number of acquired nevi, dysplastic nevi, and inadequate photoprotection appeared as risk factors with moderate strength for cutaneous malignant melanoma in the studied population. The color of the eyes and hair showed a very weak statistical significance as a risk factor. Sunscreens showed progressive significance corresponding to an increase in SPF, the best scores in statistical protection being achieved in users of sunscreens with SPF15 or greater. Frequent sunburn episodes appeared as the most important risk factor associated with malignant melanoma in this sample of the white population in southern Brazil.