An 8-year follow-up of 450 sons of alcoholic and control subjects.

BACKGROUND Between 1978 and 1988, 453 sons (age range, 18-29 years) of alcoholic and control subject were evaluated for their level of reaction (LR) to alcohol. This article presents the results of the 8.2-year follow-up of 450 of these men. The three goals were (1) to attempt to replicate results of the follow-up of the first 223 subjects, (2) to evaluate the potential impact of the quantity and frequency of drinking at the time of the original study on the relationship between LR and alcoholic outcome (ALC), and, most importantly, (3) to test if the relationship between family history (FH) and ALC might be mediated by LR in a subset of the sample. METHODS Face-to-face structured follow-up interviews were carried out with the subjects and separately with an additional informant, and blood samples, as well as urine specimens, were obtained for determination of state markers of heavy drinking and drug toxicology screens. RESULTS First, the rate of development of DSM-III-R abuse and dependence on alcohol was 14.1% and 28.6%, respectively, for family history positive (FHP) subjects, compared with 6.6% and 10.8%, respectively, for family history negative (FHN) men. Second, neither consideration of the quantity nor the frequency of drinking at the time of the original study, nor their combination, effectively diminished the relationships between LR and ALC. Third, among men who drank and demonstrated the 15% highest and lowest scores of LR at about the age of 20 years (ie, 30% of the relevant population), the correlation between FH and ALC was greatly reduced when LR was considered, but the correlation between LR and ALC was not greatly diminished when the impact of FH was evaluated. CONCLUSIONS In this sample of moderately functional white men, the development of alcoholism occurred in relationship to an FH of alcoholism, but alcohol abuse or dependence was unrelated to prior psychiatric disorders. For this group, LR at the age of 20 years was associated with future alcoholism in a manner that was independent of the drinking practices at the time of the original study. At least among those men with clearly high and low LR scores, these data are consistent with the conclusion that LR might be a mediator of the alcoholism risk.