Alcohol and high-density lipoprotein cholesterol: causal inference from diverse study designs.

The association between reported alcohol intake and plasma high-density lipoprotein (HDL) cholesterol concentration is examined in an effort to establish whether it was a cause-and-effect basis. A cross-sectional descriptive study of several populations reveals a strong and consistent dose-response pattern: Social drinkers have mean HDL cholesterol levels that are higher than those of teetotalers by as much as 33%. Cross-sectional analyses in another epidemiological study reveal the association to be independent of potential confounding factors such as smoking and body weight, and longitudinal analyses suggest that it is also not a result of certain unmeasured sources of confounding. A small experiment reveals a 15% reduction in HDL cholesterol levels among social drinkers who abstain from alcohol from a 2-week period. The evidence supports the conclusion that alcohol habits are probably one of the determinants of plasma HDL cholesterol level. A clarification of the relevance of this phenomenon to clinical medicine awaits future clinical efforts.