Influence of genes, nevi, and sun sensitivity on melanoma risk in a family sample unselected by family history and in melanoma-prone families.

BACKGROUND: Few family studies have investigated the effects of genetic, environmental, and host factors on melanoma risk, and most have been restricted to high-risk families. We assessed the role of these factors on melanoma risk in two types of families: families ascertained through melanoma probands but unselected by family history and melanoma-prone families.
METHODS: Data on pigmentary traits, nevus phenotypes, exposure to sun, and reactions to sunlight were collected from 295 families unselected by family history and 53 melanoma-prone families. We modeled melanoma risk using a logistic regressive model incorporating the effect of a melanoma-predisposing gene, familial dependence, and potential risk factors (e.g., pigmentary traits, nevus phenotypes, history of sun exposure, skin reactions to sunlight). Maximum-likelihood estimates of the parameters of the regressive model were used to compute odds ratios associated with each risk factor and age-specific melanoma risk depending on the genotype at the melanoma-predisposing gene and the effects of risk factors. All statistical tests were two-sided.
RESULTS: In the families unselected by family history, there was statistically significant evidence (P<.001) for a dominant gene, with melanoma risk reaching 0.49 and 0.67 by age 80 years in male and female gene carriers, respectively. Melanoma risk was statistically significantly influenced by total nevi (odds ratio of hazard function [OR] = 5.81, 95% confidence interval [CI] = 3.47 to 8.99), sun exposure (OR = 5.37, 95% CI = 4.44 to 6.36), and sunburn interacting with the gene (OR = 26.31, 95% CI = 7.56 to 99.22 in gene carriers and OR = 1.67, 95% CI = 1.36 to 2.03 in noncarriers). Twenty of the 53 melanoma-prone families had cosegregating mutations in CDKN2A, a gene known to be associated with melanoma. In these 53 families, three risk factors in addition to CDKN2A mutations increased melanoma risk: dysplastic nevi (OR = 2.32, 95% CI = 2.08 to 2.58), total nevi (OR = 1.99, 95% CI = 1.61 to 2.20) and sunburn (OR = 5.16, 95% CI = 4.82 to 5.52).
CONCLUSIONS: Together, a melanoma-predisposing gene (identified as being CDKN2A in melanoma-prone families), number of nevi and/or dysplastic nevi, and sun-related covariates influence melanoma risk in both families unselected by family history and melanoma-prone families.