A Drink a Day for Women’s Mental Health

Abstract & Commentary

Comment by Mary Elina Ferris, MD, Clinical Associate Professor, University of Southern California; Associate Editor, Internal Medicine Alert.

Synopsis: Moderate consumption of alcohol in women (about 1 drink daily) was associated with better cognitive scores at 2-year average follow-up in women aged 70 to 81 in the Nurses’ Health Study compared to nondrinkers, while excessive drinkers did not show any association with either improvement or decline.

Source: Stampfer MJ, et al. N Engl J Med. 2005:352:245-253.

The Nurses’ Health Study enrolled 121,700 US female registered nurses aged 30-55 years in 1976, and has followed them every 2 years since with written questionnaires about lifestyle and health, along with dietary habits added in 1980. This article reports on a study on cognitive function begun in 1995 on participants aged 70 or older who were not institutionalized nor had a stroke (21,202 women).

A new strategy of telephone interviews to measure cognitive function was used, and responses were completed for 93% of the 12,480 women identified after exclusions for antidepressant use and fluctuating patterns of alcohol use. Approximately half were nondrinkers, 44% drank up to one glass daily, and 5% more than one glass. Baseline and follow-up interviews varying from 1.3 to 5.5 years were performed (average, 2-year follow-up). Data on alcohol consumption were collected from the most recent written questionnaire before the baseline interview for beer, red and white wine, and liquor. The accuracy of the questionnaire was validated by a smaller study using weekly dietary records lasting 3 months in which participants weighed or measured all their food and drinks.1

The telephone interview was modeled on the Mini-Mental State Exam with added standardized tests of immediate and delayed recall, verbal fluency, digit span backward, and 10-word list for verbal memory. Nurse interviewers were blinded to the participants’ drinking status or the study’s hypothesis. The telephone interview was also validated in another study, which featured both telephone and in-person interviews and found a correlation of 0.81, and confirmed the natural rate of cognitive decline to be similar in that study and this present one.

Results without adjustment showed slightly better cognitive scores for drinkers of up to one drink daily (<15 grams alcohol) compared to non-drinkers. However, the relative risk was more significant after adjustments for potential confounding factors of age, education, and other variables determined by regression models. No difference was found among the types of alcoholic beverage consumed, nor for diabetes. Final results suggested a decrease in the risk of cognitive impairment of 20% for consumption of up to 1 drink/day over the average 2-year follow-up of the study.


Previous research from this study group has suggested that moderate alcohol consumption may add protection against cardiovascular disease in women,2 and this large study now suggests further benefit to mental function. Stampfer and colleagues have validated their research strategies using dietary questionnaires and telephone interviews, but it is impossible to control for all variables in an observational study. Older persons who drink may be generally in better health than nondrinkers, and of course mental function is always hard to accurately measure.

There is ample evidence that excessive alcohol intake is deleterious to multiple body organs, including the brain, and in fact hidden alcoholism among elderly women is a serious concern often undetected by physicians. The accompanying editorial warns that alcohol consumption can be a double-edged sword with as many hazards as benefits for our patients. A substantial residue of uncertainty remains with research in this area, requiring more long-term studies to confirm a preventive benefit. 

Moderate alcohol consumption may prevent cardiovascular disease and cognitive decline because it increases HDL cholesterol and reduces fibrinogen and other thrombotic factors, which reduce small thrombi in both heart and brain vessels. Although it’s too early to promote widespread increased alcohol consumption for elderly women, we can feel reassured that low-to-moderate use may help, rather than harm, our patients.


1. Willett WC et al. Amer J Epidemiology. 1985;122: 51-65.

2. Stampfer MJ, et al. N Engl J Med. 1988;319:267-273.