Although heavy drinking has well documented negative effects on the brain, some epidemiologic studies have suggested modest cognitive benefits from moderate alcohol intake, defined as up to two alcoholic beverages per day. It has been difficult to assess the effect of alcohol consumption on cognition accurately because of the potential for confounding in retrospective studies, as well as variation in the amount of alcohol found in different types of beverages. Using a cohort of selected participants in the longitudinal Nurses’ Health Study, Stampfer and colleagues examined the relationship over time between moderate alcohol consumption and scores on standardized measures of cognitive function.

The study population consisted of women who were 70 years or older in 1995. All had enrolled in the Nurses’ Health Study in 1976 and subsequently completed six dietary questionnaires between 1980 and 1998, including questions about frequency and type of alcohol use. Study participants were classified into three groups based on self-reported average daily alcohol consumption: those who did not drink, those who drank up to 15 g (approximately one drink) daily, and those who drank 15 to 30 g (one to two drinks) per day. Participants were excluded if they reported consuming more than two drinks daily, if consumption varied by more than one drink daily on different questionnaires, or if they reported use of antidepressant medications.

Of the 11,102 eligible participants, 51 percent drank no alcohol, 44 percent consumed less than one drink daily, and 5 percent consumed one to two drinks daily. All participants underwent standardized telephone assessments of cognitive status and verbal memory in 1995 and 1997 by specially trained nurses who were blinded to the women’s drinking habits. Mean scores for cognitive tests in each group were calculated using linear regression models to control for potential confounders such as age, education level, and physical activity; medical problems such as diabetes and hypertension; and use of hormone therapy, nutritional supplements, and tobacco.

Nondrinkers and women who consumed one to two drinks daily had similar average cognitive scores. Women who consumed less than one drink daily had higher average scores, and their risk of being classified as cognitively impaired (based on a predefined cutoff score) was 20 percent less than that of nondrinkers. These women also were less likely than nondrinkers to experience a decline in cognitive scores after two years. The type of alcoholic beverage consumed did not affect the results.

The authors conclude that women who consume less than one drink daily have a decreased risk of cognitive decline in later life. They discuss possible mechanisms for a protective effect of low-level alcohol consumption, including modifying lipid profiles and preventing subsequent vascular dementia. Noting that they may not have accounted for all possible confounders, however, the authors advise caution in recommending moderate alcohol consumption as preventive medicine.