Causal relationship between Brugada syndrome and electrocardiogram traits: A bidirectional Mendelian randomization study

Abstract

Introduction

Observational studies have suggested associations between Brugada syndrome (BrS) and electrocardiograms traits. Nonetheless, the causal relationships remains uncertain in observational studies. This study aims to investigate the causal relationships between BrS phenotypic risk and electrocardiogram traits using Mendelian randomization (MR) analysis and colocalization analysis.

Methods

MR analysis was performed to investigate the causal relationships between BrS phenotype risk and electrocardiogram traits (P wave duration, PR interval, QRS wave duration, ST segment duration, T wave duration, QT interval, heart rate (HR) and heart rate variability). The genetic instruments for BrS (number of cases = 12,821) were obtained from the latest GWAS. GWAS summary data of electrocardiogram traits were obtained from the MRC-IEU and GWAS catalog databases. The causal relationships were obtained through MR methods, and sensitivity analyses (e.g. Cochran's Q test, MR-PRESSO). Furthermore, the causal relationships were evaluated whether they were driven by one linkage disequilibrium using colocalization analysis.

Results

We found that there are positive causal relationships between BrS phenotypic risk and P wave duration, PR interval, QRS wave duration and QT interval, respectively (IVW_P: β = 1.238, 95 % CI = 0.857–1.619, P<0.001; IVW_PR: β = 2.199, 95 % CI = 1.358–3.039, P<0.001; IVW_QRS: β = 0.157, 95 % CI = 0.115–0.198, P<0.001; IVW_QT: β = 0.593, 95 % CI = 0.391–0.796, P<0.001), and there is a negative causal relationship between BrS phenotypic risk and heart rate (IVW_HR: β = −0.023, 95 % CI = −0.03 ∼ −0.015, P<0.001). Additionally, there are bidirectional causal relationships between BrS phenotypic risk and P wave duration and PR interval, respectively (IVW_P: OR = 1.217, 95 % CI = 1.118–1.325, P<0.001; IVW_PR: OR = 1.02, 95 % CI = 1.008–1.032, P = 0.001). Furthermore, colocalization analysis identified that the causal relationships between BrS phenotype risk and P wave duration, PR interval and QRS wave duration were driven by rs6790396, rs6801957 and rs6801957, respectively.

Conclusions

Bidirectional causal relationships were identified between BrS phenotypic risk and P wave duration and PR interval, respectively. There were positive causal relationships between BrS phenotypic risk and QRS wave duration and QT interval, respectively, and there is a negative causal relationship between BrS phenotypic risk and heart rate.