Variation and covariates of the number of benign nevi in adolescents.

Melanocytic nevi of diameter greater than or equal to 2 mm were counted on most of the skin surface of 349 adolescents aged 14-15 years of European race or ethnicity in Dunedin, New Zealand. Total counts are described by means of a form of Poisson-error log-linear modeling suitable for data showing unexplained variation (NE Breslow, Appl Statist 1984;33:38-44). There were marked interpersonal variation in the number of nevi; only some was attributable to observed factors. The mean and median counts were 23.8 and 18 nevi, respectively. The estimated ratio of the number of nevi for females compared with males was 0.7 (95% confidence interval (CI) 0.6-0.8). Greater amounts of sunbathing were associated with greater numbers of nevi. Hair and eye color, socioeconomic status, and sunburn history did not show statistically significant effects. Time since menarche and shaving status also showed no effects. Lack of suntan was associated with lower counts. Freckling was positively correlated with higher counts; the severe freckling group had an estimated ratio of 1.9 (95% CI 1.3-2.8) compared with those with no or very few freckles. The results are consistent with the hypothesis that ultraviolet radiation exposure from recreational sun exposure positively influences the total burden of nevi in normal subjects. Comparison with other epidemiologic studies suggests that the typical ultraviolet radiation dose-nevus yield curve might be steeper in males than females. Unexplained variation of nevus count may reflect heterogeneity of constitutional factors not yet measured in epidemiologic studies.