Bayesian Reanalysis of Revascularization Strategies for Left Main Disease: Coronary Artery Bypass Grafting vs Percutaneous Intervention.

Abstract

The preferred revascularization strategy for left main disease between coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) remains controversial, given discordant conclusions from randomized and observational evidence for mortality. Accordingly, we performed a bayesian reanalysis of the meta-analysis by Sabatine and coworkers to assess the probability of effectiveness using different assumptions. Data were extracted from the meta-analysis of 4 trials by Sabatine and coworkers. The primary outcome was all-cause mortality; secondary outcomes included myocardial infarction, stroke, repeated revascularization, and composite outcomes. Bayesian methods estimated 5-year hazard ratios and absolute risk reductions using priors ranging from no prior preference to data-informed assumptions. For all-cause mortality, bayesian analysis indicated a higher probability of lower mortality with CABG, exceeding 80% across priors. Under the minimally informative prior, the hazard ratio was 1.1 (95% credible interval, 0.91-1.33), and the absolute risk reduction was 0.88% (95% credible interval: -0.92% to 2.76%). Under the downweighted historical prior, the hazard ratio was 1.40 (1.25-1.56) and the absolute risk reduction 4.61% (3.06%-6.11%). CABG also showed highly probable benefits for myocardial infarction, repeated revascularization, and composite outcomes. Conversely, stroke favored PCI (hazard ratio, 0.83 [0.58-1.19]; absolute risk reduction -0.40% [-1.40% to 0.62%]). This reanalysis suggests a higher probability that CABG is associated with reduced mortality, myocardial infarction, repeated revascularization, and composite outcomes, whereas PCI offers a lower risk of stroke. Interpretation of mortality was prior dependent, whereas other outcomes were largely prior agnostic.