The evolving epidemiological literature, including improved methodology for assessing causality in observational studies, is raising doubts about whether moderate alcohol consumption has a protective effect on health.