Longitudinal alcohol consumption in early adulthood has been studied, but reports from later adulthood are scarce.

Finding longitudinal trends in alcohol consumption using prospective birth cohort data.

To investigate trends in self-reported alcohol consumption in adulthood.

In the Northern Finland Birth Cohort 1966 (N=5,621) alcohol use was studied with questionnaires at ages 31 and 46. Participants were classified into steady drinkers, increasers and decreasers based on the change in consumption (g/day). Multinomial regression analysis was conducted with educational, marital and employment baseline statuses and their changes, and baseline alcohol use as factors influencing the change in alcohol consumption.

70% of alcohol users were classified as steady drinkers, 10% as decreasers and 20% as increasers. In all, daily alcohol consumption rose over 30%, particularly among women. The unemployed, singles and low-educated consumed the most (P<0.01). Being a divorced male (OR 1.5; 95%CI: 1.0-2.1) or a long-term unemployed female (1.6; 1.0-2.6) predicted increase in alcohol use. The probability of decrease was higher among single men (OR 1.6; 1.0-2.4) and women (2.8; 1.7-4.4) at 31y, among men entering a relationship (1.9; 1.2-3.1), and among divorced women (2.4; 1.4-3.8) and entering a relationship (2.1; 1.3-3.5).

Alcohol usage among middle-aged Northern Finns is rather stable and increase is more exhibited than decrease. Gender differences in predictors existed: changes in relationship status predicted decrease in usage in women, whereas in men divorce predicted increase in usage and a new relationship predicted decrease. Long-term unemployment predicted increase only in women.