BackgroundAtrial fibrillation (AF) is one of the most common form of arrhythmia. Previous studies have shown a link between AF and mental illness. However, the causal relationship between mental illness and AF remains unclear. The purpose of this study was to investigate the bidirectional causal relationship between borderline personality disorder (BPD) and AF.MethodWe used the bidirectional Two-sample Mendelian randomization (TSMR) method to evaluate the causal relationship between BPD and AF. Instrumental variables associated with BPD were derived from a genome-wide association study involving 214,816 Europeans (2,637 cases and 212,179 controls). We then obtained atrial fibrillation data from the GWAS meta-analysis (60,620 cases and 970,216 controls). The TSMR analyses were performed in five methods, namely fixed-effect inverse-variance weighted (IVW) method、random-effect IVW method, MR Egger regression method, Weighted median method and Simple mode method. Several sensitivity analyses are used to test the robustness of positive results.ResultsThe fixed-effect inverse-variance weighted model [Odds ratio (OR), 1.033, 95% confidence interval (CI), 1.011-1.056, P = 0.0031], random-effect inverse-variance weighted model (OR, 1.033; 95%CI, 1.005-1.062; P = 0.0191) and Weighted median (OR, 1.034; 95%CI, 1.002-1.068; P = 0.0394) all showed that genetically predicted BPD was associated with an increased risk of AF. Sensitivity analysis using other MR Methods, including the MR-Egger intercept, MR-Presso method, and leave-one-out analyses, showed that the results were robust. In reverse MR analysis, there was no causal relationship of AF on BPD.ConclusionOur study provides a causal relationship between BPD and AF. This means that patients with BPD should be monitored for the occurrence of AF. Early screening and proper management of BPD may show anti-arrhythmic benefits.