A comprehensive evaluation of the longitudinal association between alcohol consumption and inflammation: Multiverse and vibration of effects analyses

Abstract

Background: Moderate alcohol consumption appears to be associated with reduced inflammation compared to abstinence. However, there are concerns about the impact of common variations in researcher-defined data processing and analysis parameters on this relationship. In light of this, we performed comprehensive multiverse and vibration of effects analyses to evaluate the robustness of the alcohol–inflammation association.Methods: Using the 1970 British Birth Cohort Study, the relationship between alcohol consumption in early/mid-adulthood (ages 34 and 42) and levels of inflammation marker high-sensitivity C-reactive protein (hsCRP) at age 46 was assessed in 3101 people. Linear regression compared low-to-moderate consumption and consumption above various international drinking guidelines with an ‘abstinent’ reference. Research parameters of interest related to: definitions of drinking and reference groups; alcohol consumption measurement year; outcome variable transformation; and breadth of covariate adjustment. Results were graphically depicted, variance decomposed by parameters, and specifications within parameters compared.Results: Primary analyses were limited to those universes with occasional drinkers as reference. All universes resulted in lower levels of inflammation amongst low-to-moderate drinkers compared to occasional drinkers (1st percentile effect: -0.21; 99th percentile effect: -0.04). Estimates comparing above-guidelines drinking with occasional drinkers were less definitive (1st percentile effect: -0.26; 99th percentile effect: 0.43). Breadth of covariate adjustment, and measurement year and national guidelines used to classify drinking groups accounted for considerable effect variance. Conclusions: The association between low-to-moderate drinking and lower hsCRP levels appears robust to common variations in researcher-defined parameters. The association between above-guidelines drinking and hsCRP levels is less definitive.