Mendelian randomization analyses of smoking and Alzheimer's disease in Chinese and Japanese populations.

Abstract

Previous epidemiological studies have reported controversial results on the relationship between smoking and Alzheimer's disease (AD). Therefore, we sought to assess the association using Mendelian randomization (MR) analysis. We used single nucleotide polymorphisms (SNPs) associated with smoking quantity (cigarettes per day, CPD) from genome-wide association studies (GWAS) of Japanese population as instrumental variables, then we performed two-sample MR analysis to investigate the association between smoking and AD in a Chinese cohort (1,000 AD cases and 500 controls) and a Japanese cohort (3,962 AD cases and 4,074 controls), respectively. Genetically higher smoking quantity showed no statistical causal association with AD risk (the inverse variance weighted (IVW) estimate in the Chinese cohort: odds ratio (OR) = 0.510, 95% confidence interval (CI) = 0.149-1.744, p = 0.284; IVW estimate in the Japanese cohort: OR = 1.170, 95% confidence interval CI = 0.790-1.734, p = 0.434). This MR study, for the first time in Chinese and Japanese populations, found no significant association between smoking and AD.