Influence of ascertainment strategy on finding sex differences in genetic estimates from twin studies of alcoholism.

Twin studies have yielded contradictory findings about sex differences in genetic influences on the etiology of alcoholism. Studies based on population registers or epidemiological samples have yielded similar estimates of heritability (50-60% of the total variance) for males and females. In contrast, studies of twins identified through treatment settings have found sizeable genetic contributions to alcoholism in males but usually negligible heritabilities for females. We investigated this discrepancy by applying a "simulated" treatment ascertainment strategy to data on alcohol-related disorders collected by structured interviews with a population-based sample of adult twins aged 18-56 years from the Mid-Atlantic Twin Registry. Structural models were used to estimate heritabilities for two definitions of treatment, and these estimates were compared with those obtained from the population-based sample. In males, heritability estimates were similar across sampling methods, but the treatment ascertainment methods yielded higher estimates of common environmental influences. For females, heritability estimates based on a broad definition of treatment were similar to those obtained by using the random ascertainment design. However, estimates based on sampling women who had been in alcohol-treatment programs were (nonsignificantly) lower than those obtained with the other methods. These results provide partial support for the hypothesis that differences in sampling method may account for differences in heritability estimates for alcoholism among studies of female twins. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 96:754-761, 2000.