Abstract Background Observational studies have reported associations between non-alcoholic fatty liver disease (NAFLD) and heart failure (HF), but findings remain inconclusive. It remains unclear whether NAFLD and HF share an independent and direct causal relationship. Objective We aimed to investigate the potential causal association between NAFLD and HF using bidirectional two-sample Mendelian randomization (MR). Methods Summary-level data from large-scale genome-wide association studies (GWAS) in European populations were utilized for bidirectional two-sample MR analyses. Inverse variance weighted (IVW) served as the primary analytical method, complemented by four additional approaches: Weighted median, Weighted mode, MR-Egger, and Simple mode. Horizontal pleiotropy was assessed and corrected using MR-pleiotropy residual sum and outlier (MR-PRESSO). Summary-level data for cirrhosis were included as a positive control to enhance validity. Results In the absence of pleiotropy, genetically predicted NAFLD showed no significant effect on HF risk (OR=1.02, 95% CI: 0.99–1.05, P = 0.110). Results from MR-Egger, Weighted median, Weighted mode, and Simple mode aligned with IVW estimates. Sensitivity analyses further supported the absence of causal effects. Conclusion Our two-sample MR analysis does not support a causal association between NAFLD and HF. Previous observational associations may reflect confounding by metabolic disturbances rather than direct causation.Results
Huang et al. (Sat,) studied this question.