Reverse causal effect of atrial fibrillation on 17 site-specific cancer risk: A Mendelian randomization study

Abstract

Background: Bidirectional association between atrial fibrillation (AF) and cancer was reported by observational investigations. Additional study is warranted to investigate the causal effects of AF on risk of cancer. Methods: This study was a summary-level Mendelian randomization (MR) analysis. Genetic instrument for AF was developed from a genome-wide association study (GWAS) meta-analysis for AF including 537,409 European ancestry individuals including 55,144 AF cases. The outcome data for risk of 17 site-specific cancer from a previous GWAS meta-analysis of the UK Biobank (48,961/359,825 case/controls) and Genetic Epidemiology Research on Adult Health and Aging (16,001/50,525 case/controls) cohorts including European ancestry individuals was investigated. Inverse variance weighted method was the main MR method, supported by pleiotropy-robust sensitivity analysis including MR-Egger regression and penalized weighted median method. Results: The causal estimates indicated that AF was causally linked to higher risk of cancers of lung, breast, cervix, endometrium, and melanoma. MR-Egger test for directional pleiotropy indicated absence of a pleiotropy in the identified causal estimates and MR-Egger regression and median-based methods provided similarly significant findings. On the other hand, the genetic predisposition of AF was significantly associated with lower risk of esophagus/gastric cancer, but possibility of a directional pleiotropy remained in the association. Conclusions: AF is a causal factor for certain types of cancer. Appropriate cancer screening should be suggested in clinical guidelines for AF patients. Future trial is necessary to confirm whether appropriate management of AF may reduce the risk of cancer which is a major cause of deaths in AF patients.