Background Observational studies suggest smoking, cannabis use, alcohol consumption, cannabis use, and substance use disorders (SUDs) may play a role in the susceptibility for respiratory infections and disease, including coronavirus 2019 (COVID-2019). However, causal inference is challenging due to comorbid substance use.
Methods Using genome-wide association study data of European ancestry (data from >1.7 million individuals), we performed single-variable and multivariable Mendelian randomization to evaluate relationships between smoking, cannabis use, alcohol consumption, SUDs, and respiratory infections.
Results Genetically predicted lifetime smoking was found to be associated with increased risk for hospitalized COVID-19 (odds ratio (OR)=4.039, 95% CI 2.335-6.985, P-value=5.93x10-7) and very severe hospitalized COVID-19 (OR=3.091, 95% CI, 1.883-5.092, P-value=8.40x10-6). Genetically predicted lifetime smoking was also associated with increased risk pneumoniae (OR=1.589, 95% CI, 1.214-2.078, P-value=7.33x10-4), lower respiratory infections (OR=2.303, 95% CI, 1.713-3.097, P-value=3.40x10-8), and several others. Genetically predicted cannabis use disorder (CUD) was associated with increased bronchitis risk (OR=1.078, 95% CI, 1.020-1.128, P-value=0.007).
Conclusions
We provide strong genetic evidence showing smoking increases the risk for respiratory infections and diseases even after accounting for other substance use and abuse. Additionally, we provide find CUD may increase the risk for bronchitis, which taken together, may guide future research SUDs and respiratory outcomes.