The controversial paper by Stampfer MJ et al looked at moderate consumption and cognitive function in women (NEJM 2005:352:245-253. Using data collected between 1995 and 2001, the authors evaluated cognitive function in 12,480 participants in the Nurses’ Health Study who were 70-81 years old, with follow-up assessments in 11,102 two years later.
Among moderate drinkers, as compared with nondrinkers, the relative risk (RR) of impairment was 0.77 on the test of general cognition (95% CI: 0.67, 0.88) and 0.81 on the global cognitive score (95% CI: 0.70, 0.93). The results for cognitive decline were similar; In this very large study of women aged 70-81, those who consumed up to 15 g/day of alcohol (an average of up to a little over one “typical drink” per day) had about 20% better cognition and memory than nondrinkers. For example, the RR for substantial decline in the cognitive function score over a 2-year period was 0.85 for moderate drinkers (versus nondrinkers), with 95% CI: 0.74, 0.98). There were no significant associations between higher levels of drinking (15 30g/day), but heavier drinkers were not studied. There were no significant differences according to type of beverage. The apoE genotype was also studied, but different genotypes gave similar results. The authors conclude: “Our data suggest that up to one drink per day does not impair cognitive function and may actually decrease the risk of cognitive decline.”
The authors excluded women who had recently changed their alcohol intake (to exclude people who may have changed their intake due to health problems), ex-drinkers, and also the few women reporting more than 30 g/day, so that their focus was on stable moderate drinkers (versus nondrinkers).
The instrument they used to judge cognitive functioning has been validated as giving reliable indices of general cognition and verbal memory. In their study, 51% were nondrinkers, 44% in the 1.0 - 14.9 g/day (up to a little more than one typical drink, which is usually considered to be about 12 g), and 5% (648 subjects) reported 15 - 30 g/day.
The upper limit for alcohol consumption associated with improved cognition might be up to 30 g/day, although the findings in the higher-consumption group were not statistically significant.
These results are similar to many other recent studies showing a decrease in the risk of cognitive dysfunction and dementia for moderate drinkers. While the authors in the present study adjusted for education, they did not include pattern of drinking in their analyses.
The authors point out that cognition normally declines with age. They suggest that the degree of protection from 1.0 - 14.9 g/day of alcohol is the cognitive equivalent of being about one and one-half years younger. The authors conclude “Although the adverse effects of excessive alcohol intake are well known and caution should be exercised in recommending even moderate alcohol intake, our results combined with those of other studies suggest that women who consume up to one drink per day have less cognitive impairment and better cognitive function than nondrinkers.”
An editorial by Evans DA, Bienias JL et al counters Stampfer’s results, stating that demonstrating definitively that the changes in cognition seen in this study were actually attributable to alcohol consumption is impossible, since this was an observational study and not a trial. However, the investigators realized this and they described approaches they used in attempting to deal with such a limitation. The observed effect of cognitive change to alcohol, could be confounded by the fact that healthier people may be the ones drinking, so the effects being attributed to alcohol are really due to their general good health. In response to this, it can be pointed out that (1) Stampfer et al adjusted their analyses for a number of social and health characteristics that strongly relate to cognitive function and, more importantly, that such adjustments had little effect on their estimates; (2) that in the follow-up analyses the investigators adjusted for the baseline cognitive values and still demonstrated a decreased risk of dysfunction to be associated with drinking; (3) that the analyses excluded people who changed their intake in the years preceding the study, so persons who may have been beginning to decrease their intake because of increasing dementia were not included, (4) that similar effects were seen for all beverages (whereas social and educational factors tend to differ by beverage), and (5) most importantly, that most previous large studies from markedly different populations around the world have also found protection against cognitive decline from moderate drinking. All of these factors strongly suggest that the direction of the observed effect is, in all probability, correct.
R. Curtis Ellison, MD, Institute on Lifestyle & Health at Boston University comments ‘I personally feel that there is very strong evidence from observational studies, animal experiments, and even certain objective measures in humans (from arterial function studies, brain scans, etc.) that there is no harm in terms of many of the problems of aging from moderate alcohol use, and strong suggestions of benefit for coronary disease, ischemic stroke, and (from this and many other studies) even cognitive decline. There are theoretical worries that any encouragement to the elderly to drink moderately will lead many to become abusers of alcohol, but I await evidence of the extent that this occurs among patients being encouraged to drink by their doctors. While being cautious about recommending alcohol use to the general public makes sense, I am less concerned about a physician who knows the medical and social history of an individual patient encouraging him or her to consume small amounts of alcohol, given no contraindications. “Do no harm” should dictate physicians’ practices, but doing harm can occur from sins of omission as well as sins of commission.’