ABSTRACT
BACKGROUND: Previous studies have observed inconsistent associations between coronavirus disease 2019 (COVID-19) and heart failure (HF), but these studies were prone to bias based on reverse causality and residual confounding factors. We aimed to investigate genetic liability between COVID-19 and heart failure using a bidirectional Mendelian randomization study. METHODS: The causal relationship between COVID-19 (including COVID-19, hospitalized COVID-19 compared with the general population, and severe COVID-19) and HF are determined by using a bidirectional Mendelian randomization analysis. We drew on summary statistics from the largest HF genome-wide association study (GWAS) meta-analysis on individuals of European ancestry, which included 47,309 HF patients and 930,014 controls. The inverse variance weighted (IVW), an adaption of the Egger regression (MR-Egger), the weighted median, and weighted model were conducted for the Mendelian randomization analysis to estimate a causal effect. To confirm the stability, we performed a "leave-one-out" approach for the sensitivity analysis. RESULTS: Genetically predicted severe COVID-19 was not significantly associated with the risk of HF [odds ratio (OR), 1.003; 95% confidence interval (CI), 0.969-1.037; p = 0.867]. The IVW demonstrated that there was no association between genetically hospitalized COVID-19 infection and HF risk [OR, 1.009; 95% CI, 0.939-1.085; p = 0.797]. There was no evidence to support the association between genetically determined COVID-19 and the risk of HF [OR, 1.066; 95% CI, 0.955-1.190; p = 0.253]. In addition, genetically predicted HF was also not causally associated with COVID-19 [OR, 1.162; 95% CI, 0.824-1.639; p = 0.393]. MR-Egger analysis indicated no evidence of directional pleiotropy. CONCLUSION: The current bidirectional Mendelian randomization analysis overcomes the limitations of observational studies. Our findings indicated that there is no causal association between COVID-19 and HF.