ABSTRACT
Objective@#To evaluate the association of smoking with the risk of ankylosing spondylitis (AS) using a Mendelian randomization (MR) approach.@*Methods@#A total of 16 383 186 AS-associated single nucleotide polymorphisms (SNPs), 378 smoking initiation associated SNPs and 126 lifetime smoking score-associated SNPs were collected from three large-scale genome-wide association studies (GWAS). The association of smoking phenotypes with the risk of AS was examined using inverse-variance weighted (IVW) with AS as a outcome variable, smoking initiation and lifetime smoking score as exposure factors and SNPs with strong associations with smoking as instrumental variables, and sensitivity analyses were performed with maximum likelihood-based method, MR pleiotropy residual sum and outlier (MR-PRESSO) test and MR-Egger regression analysis.@*Results@# A 33.5% increased risk of AS was found among genetically predicted smokers relative to non-smokers (OR=1.335, 95%CI: 1.059-1.682), and an increase in predicted lifetime smoking by per standard deviation resulted in a 101.4% increased risk of AS (OR=2.014, 95%CI: 1.341-3.024). The maximum likelihood-based method and MR-PRESSO test showed consistent correlated effect estimations and MR-Egger regression analysis identified no evidence of pleiotropy.@*Conclusion@#It is genetically predicted that smoking is associated with an increased risk of AS.