Association of Obstructive Sleep Apnea With Risk of Myocardial Infarction: A Multivariable Mendelian Randomization and Mediation Analysis

BACKGROUND: Observational studies have suggested an association between obstructive sleep apnea (OSA) and myocardial infarction (MI), but whether this relationship is causal or largely reflects shared risk factors remains unclear. METHODS AND RESULTS: We performed a 2-sample Mendelian randomization (MR) analysis to evaluate the causal effect of OSA on MI. Summary statistics for OSA were obtained from FinnGen, and MI data were obtained from the UK Biobank, with external validation using CARDIoGRAMplusC4D. Mediation MR was used to assess 13 potential mediators, and a 6-step multivariable MR framework was applied to estimate the direct effect of OSA after sequential adjustment for potential confounders. Reverse MR was conducted to test possible reverse causality. Genetically predicted OSA liability was associated with increased MI risk (odds ratio [OR] per log-OR increase, 1.0024 [95% CI, 1.0010-1.0039]; P=0.001). Body mass index (BMI) was the strongest mediator, explaining 35.94% of the association (P=0.030), whereas systolic blood pressure (SBP) showed minimal mediation (0.28%; P=0.678). In stepwise multivariable MR, the OSA-MI association was attenuated after adjustment for BMI and SBP (P=0.156), suggesting partial confounding by shared cardiometabolic risk. In a model including SBP and atrial fibrillation (AF), AF remained independently associated with MI (P=0.004), whereas OSA showed only a marginal direct effect (P=0.050). Reverse MR found no evidence that MI influenced OSA risk. CONCLUSIONS: These findings support a causal association between OSA and MI and suggest that this relationship may be mediated in part through obesity-related and arrhythmia-related pathways. AF may represent an important intermediate component of OSA-related cardiovascular risk beyond traditional hemodynamic factors. Keywords: obstructive sleep apnea; myocardial infarction; Mendelian randomization; mediation analysis; obesity.