Alcohol consumption and all-cause mortality in older adults in Spain: an analysis accounting for the main methodological issues

Abstract

Background and aims: Observational evidence that light-to-moderate alcohol consumption lowers mortality is questioned because of potential selection biases and residual confounding. We assess the association between alcohol intake and all-cause death in older adults after accounting for those methodological issues.

Methods: Data came from 3045 individuals representative of the non-institutionalized population aged ≥ 60 years in Spain. Participants were recruited in 2008-10, when they reported current and life-time alcohol intake; drinkers were classified as occasional (< 1.43 g/day), light (≥ 1.43 but < 20 g/day for men and ≥ 1.43 but < 10 g/day for women), moderate (≥ 20 but < 40 g/day for men and ≥ 10 but < 20 g/day for women) or heavy (≥ 40 g/day for men and ≥ 24 g/day for women)/binge. Participants were followed-up to 2017 to assess vital status. In analyses, ex-drinkers were removed from the abstainer group and were classified according to their life-time intake to address the 'abstainer bias'. Moreover, analyses were replicated in individuals without functional limitations, and excluded deaths in the first year of follow-up, to address reverse causation. Also, occasional drinkers were used as reference in some analyses to reduce the 'healthy drinker/survivor' bias. Results were adjusted for many covariates to minimize residual confounding.

Results: Compared with never-drinkers, the hazard ratio (95% confidence interval) of mortality for light drinkers was 1.05 (0.71-1.56) and 1.20 (0.72-2.02) in those without functional limitations. Corresponding values for moderate drinkers were 1.28 (0.81-2.02) and 1.55 (0.87-2.75) and for heavy/binge drinkers 1.85 (1.07-3.23) and 2.15 (1.09-4.22). Results were consistent when occasional drinkers were used as reference. Among drinkers without functional limitations, the hazard ratio (95% confidence interval) of mortality per 10 g/day of alcohol was 1.12 (1.02-1.23).

Conclusion: After accounting for potential biases, light-to-moderate drinking among people 60+ years of age appears to have no statistically significant benefit on mortality compared with abstention from alcohol. By contrast, heavy/binge drinking shows a higher death risk compared with abstention from alcohol. Alcohol intake appears to have a positive dose-response with mortality among drinkers.

Keywords: Alcohol; confounding; mortality; older adults; reverse causation; selection bias.