Drinking habits and executive functioning: a propensity score-weighted analysis of 78,832 adults

Excessive alcohol intake compromises cognitive functioning. At the same time, moderate alcohol consumption is reported to protect against Alzheimers disease among elderly. Little is known about dose-dependent effects of alcohol consumption on higher-order cognitive functioning among generally healthy adults. Here, we applied propensity weighted analyses to investigate associations between habitual drinking patterns and executive functioning in the general population. A community sample of N=78,832 Dutch adults (age 18-65, 40.9% male) completed the Ruff Figural Fluency Task of executive functioning (range 1-165), and self-reported their past month consumption of alcoholic beverages on a food frequency questionnaire. Participants were stratified according to drinking level (abstinent [22.0%], occasional [<2.5 g/day, 21.4%], light [2.5-14.9 g/day, 42.9%], moderate [15-29.9 g/day, 11.4%], or heavy [>30 g/day, 2.3%]) and binge-drinking (yes [10.6%] vs. no [89.4%]). Groups were equivalised using multinomial propensity score weighing based on demographic, socioeconomic, health-related and psychosocial factors influencing drinking behaviour. Compared to abstinent participants, task performance was better among light drinkers ({beta}[95% CI]=0.056[0.033-0.078] or +1.3 points, p<0.0001) and moderate drinkers ({beta}[95% CI]=0.111[0.079-0.143] or +2.5 points, p<0.0001), but not among occasional drinkers ({beta}[95% CI]=0.018[-0.006-0.043], p=0.1432) or heavy drinkers ({beta}[95% CI]=0.075[-0.009-0.158], p=0.0791). No difference was found between binge-drinkers and non-binge-drinkers ({beta}[95% CI]=0.032[-0.002-0.066], p=0.0654). In conclusion, we observed better executive functioning in light-moderate drinkers compared to abstainers, but not in occasional drinkers, heavy drinkers or binge-drinkers. This non-linear association is reminiscent of the dose-dependent effect of alcohol consumption on cardiovascular risk. Further studies may determine whether cardiovascular, inflammatory and/or other somatic factors mediate the association between moderate drinking and higher-order cognitive functions. Although analyses were adjusted for observed factors that influence drinking behaviours, we cannot exclude a contribution from unobserved residual confounding.