Alcohol consumption and mortality. II. Studies of male populations

This is the second of a set of three papers evaluating drinking status and mortality risk. Analysis of eight general population surveys of men evaluated all-cause mortality rates by drinking pattern.

Raw data from three studies of youth and five studies of adults were evaluated. Logistic regression models controlled for confounding characteristics. Meta-analysis combined study results.

Drinking pattern was alternatively defined by quantity, frequency and volume of drinking. Final models included drinking pattern (as well as abstinence in the youth models and long-term abstainers and former drinkers in adult models), age and other confounding variables. Models also evaluated interactions of age and, respectively, long-term abstinence and former drinking.

No evidence was found for the hypothesis that abstinence is associated with greater mortality risk than light drinking. In the youth samples, abstainers had a lower risk of dying than those drinking less than 15 times per month. One study of the adult samples showed a significant age by former drinker interaction; this did not alter the lack of association of former drinking with mortality risk or the homogeneity of results across studies for this finding. The most consistent finding was the association of heavy drinking with mortality among youth. Among adults, drinking 43 or more drinks per month and drinking 21 or more times per month were associated with increased mortality risk. Quantity per occasion was not significantly associated with mortality risk among adults.

That frequent drinking was related to mortality risk, whereas heavier quantity was unrelated, is inconsistent with the belief that daily consumption of a few glasses of wine has salutary effects. Empirically, however, this pattern tends to be unusual. Findings were homogeneous across studies lending generalizability to results.