TY - JOUR T1 - Evaluating the effects of alcohol and tobacco use on cardiovascular disease using multivariable Mendelian randomization JF - bioRxiv DO - 10.1101/757146 SP - 757146 AU - Daniel B. Rosoff AU - George Davey Smith AU - Nehal Mehta AU - Toni-Kim Clarke AU - Falk W. Lohoff Y1 - 2019/01/01 UR - http://biorxiv.org/content/early/2019/09/06/757146.abstract N2 - Alcohol and tobacco use, two major modifiable risk factors for cardiovascular disease (CVD), are often consumed together. Using large publicly available genome-wide association studies (results from > 940,000 participants), we conducted two-sample multivariable Mendelian randomization (MR) to simultaneously assess the independent effects of alcohol and tobacco use on CVD risk factors and events. We found genetic instruments associated with increased alcohol use, controlling for tobacco use, associated with increased high-density-lipoprotein-cholesterol (HDL-C), decreased triglycerides, but not with coronary heart disease (CHD), myocardial infarction (MI), nor stroke; and instruments for increased tobacco use, controlling for alcohol use, associated with decreased HDL-C, increased triglycerides, and increased risk of CHD and MI. Exploratory analysis found associations with HDL-C, LDL-C, and intermediate-density-lipoprotein metabolites. Consistency of results across complementary methods accommodating different MR assumptions strengthened causal inference, providing strong genetic evidence for the causal effects of modifiable lifestyle risk factors on CVD risk. ER -