Investigation of the Causal Relationship Between Alcohol Consumption and COVID-19: A Two-Sample Mendelian Randomization Study

Abstract

AbstractAssociation between alcohol intake and Coronavirus disease 2019 (COVID-19) risk has been explored in several observational studies, but the results are still controversial. These associations may be biased by reverse causation or confounded by other environmental exposures. To avoid potential biases, we used Mendelian randomization (MR) method to evaluate whether alcohol intake is the causal risk factor for COVID-19. Two-sample MR analyses were performed utilizing summary data from the UK Biobank with 38,984 COVID-19 patients and 1,644,784 control participants. Both inverse-variance weighted (IVW) and genetic risk score (GRS) methods were applied to estimate the relationship including COVID-19 vs. general population, hospitalized COVID-19 vs. not hospitalized COVID-19, hospitalized COVID-19 vs. general population, and severe COVID-19 vs. general population. Additionally, we conducted various sensitivity analyses to evaluate the impact of assumptions on the findings and ensure the robustness of the results. Using 80 single nucleotide polymorphisms as instrumental variables, we found that alcohol intake was not significantly associated with the occurrence of COVID-19 in both IVW and GRS methods (IVW: beta = 0.0372; 95% CI − 0.1817 to 0.2561; P = 0.74; GRS: beta = 0.0372, 95% CI − 0.1737 to 0.2481, P = 0.73). Furthermore, similar results were also observed in comparison hospitalized COVID-19 with not hospitalized COVID-19 (IVW: beta = − 0.3625; 95% CI − 1.4151 to 0.6900; P = 0.50; GRS: beta = − 0.3625, 95% CI − 1.3633 to 0.6383, P = 0.48), hospitalized COVID-19 with general population (IVW: beta = − 0.1203; 95% CI − 0.5997 to 0.3591; P = 0.62; GRS: beta = − 0.1203, 95% CI − 0.5352 to 0.2946, P = 0.57), and severe COVID-19 with general population (IVW: beta = 0.2963; 95% CI − 0.3682 to 0.9607; P = 0.38; GRS: beta = 0.2963, 95% CI − 0.3240 to 0.9166, P = 0.35). Besides, the heterogeneity and sensitivity tests suggested absence of bias due to pleiotropy. Our results highlight no evidence to support the causal role of alcohol consumption in COVID-19 risk. Further large-scale prospective studies are warranted to replicate our findings.