Risk factors that predicted problem drinking in Danish men at age thirty

The Danish Longitudinal Study of Alcoholism utilized a prospective, high-risk research paradigm to identify putative markers of adult male alcoholism from a comprehensive database that began with the birth of the subject and extended over three decades. This article focuses on measures antedating abusive drinking that predicted lifetime alcohol abuse/dependence at age 30 years.

The original 330 subjects of this study were drawn from a large Danish birth cohort (N = 9,125) born between 1959 and 1961. The sample included 223 sons of treated alcoholic fathers (high-risk group) and 107 matched sons whose biological fathers had no record of treatment for alcoholism (low-risk group). This sample has been thoroughly investigated with a variety of methods representing multiple domains that included perinatal records, pediatric records, school records, teacher ratings, school physician records and a series of structured interviews and psychometric tests at ages 19-20 and 30 years. The present analysis focuses on the degree to which premorbid differences between the high- and low-risk groups later predicted lifetime drinking problems at age 30 (n = 241).

As expected lifetime alcohol abuse/dependence by age 30 was reported significantly more often in the high-risk group. Of the 394 premorbid variables tested, 68 were found to distinguish the high- from the low-risk group before any subjects had developed a drinking problem. Of these 68 variables, 28 (41%) were also associated with DSM-III-R alcohol abuse/dependence at age 30. These 28 putative markers were reduced to 12 that were entered into a multiple regression analysis to search for the most powerful unique predictors of alcoholism. Four of the 28 putative markers were independently associated with problem drinking at age 30: low birth weight, number of life crises in childhood, ratings of childhood unhappiness and antisocial personality disorder. The regression model accounted for 46% of the drinking outcome variance. A father's alcoholism by itself no longer independently contributed to the prediction of his son's drinking and with one exception, did not systematically interact with the putative markers to facilitate the prediction of alcohol dependence at age 30.

Risk itself. which significantly predicted problem drinking at age 30, was not uniquely associated with the development of alcoholism in adulthood. These findings, rather, provide broad support for the biopsychosocial model of alcoholism, especially for those models that emphasize the cumulative influence over time of internal and external variables in biologically vulnerable individuals.