Search   |  Advanced

Research Summary

No Association between Moderate Alcohol Intake and Improved Cognitive Function Seen in a Large Cohort Study Using Innovative Methods

Most prospective observational studies have shown that moderate alcohol use is associated with slightly better cognitive function, but there is always concern about confounding from other lifestyle factors (i.e., the better function being a result of something unrelated to drinking). A “Mendelian randomization study” in a cohort of almost 7000 men aged 50+ in China used aldehyde dehydrogenase-2 (ALDH2) genotype as an “instrumental variable” to decrease the likelihood that the observed association between alcohol consumption* and cognitive function** would be due to some other factor (ALDH2 genotype would be expected to be related to drinking but not to cognitive function).

  • Presence of the ALHD2 genotype was strongly associated with higher alcohol consumption but explained only 3% of the variance in use.
  • Alcohol consumption (either from reported intake or genotype testing) was not associated with delayed 10-word recall score or MMSE score.

*Consumption categories included never user, former user, occasional user (amount not defined but drinking on <1 day per week), moderate user (≤210 g per week), and heavy user (>210 g per week).
**Cognitive function was assessed via delayed 10-word recall score in 4707 participants and by Mini-Mental State Examination (MMSE) score in 2284 participants.


This study showed little effect of reported alcohol intake on cognitive function. It is unfortunate that the authors did not use measures of cognitive functioning shown to be more accurate (e.g., the Montreal Cognitive Assessment) or measures that adjust for education and socioeconomic status (e.g., the Wechsler Adult Intelligence Scale). In addition, the primary beverage consumed in the cohort was rice wine, which contains no polyphenols. Although Mendelian randomization techniques are designed to offer unbiased estimates of effect, the instrumental variable used should have a strong correlation with the outcome (i.e., alcohol consumption); in this study, it did not. As stated by the authors, causality should be verified in a variety of settings using different kinds of evidence, including experimental or genetic studies, rather than relying on observational studies. R. Curtis Ellison, MD


Au Yeung SL, Jiang CQ, Cheng KK, et al. Evaluation of moderate alcohol use and cognitive function among men using a Mendelian randomization design in the Guangzhou Biobank Cohort Study. Am J Epidemiol. February 1, 2012 [Epub ahead of print]. doi:10.1093/aje/kwr462.