Bayesian insights into fractional flow reserve-guided complete revascularization versus culprit-only percutaneous coronary intervention in patients with myocardial infarction
The recent meta-analysis by Ezenna et al. [
1] addresses a crucial question: Does revascularization reduce the incidence of composite adverse events compared with culprit-only PCI? Their comprehensive study, including five randomized controlled trials [
[2],
[3],
[4],
[5],
[6]], suggests a significant effect compared to control. However, the frequentist framework is often limited when employing a random effects model to pool a small number of studies, especially with a high heterogeneity [
7].
To delve deeper into this topic, we utilized a random-effects meta-analysis model within a Bayesian framework, focusing on a composite outcome that included all-cause mortality, myocardial infarction, stroke, and repeat revascularization (
Table 1). This approach updates prior beliefs using current data to establish a posterior distribution. In our primary analysis, we employed a vague prior for the overall effect and an informative prior for the between-study heterogeneity parameter. Additionally, we conducted sensitivity analyses by fitting models with various priors to assess whether our choices significantly influenced the results or our conclusions [
8].
Table 1. Definition of composite adverse events in all included studies.