This report updates a prior systematic review and applies a Bayesian network meta-analysis to assess comparative outcomes of P2Y12 inhibitors after acute coronary syndromes (ACS). The update added two trials to the prior search, yielding 19 randomized controlled trials (RCTs) with a combined sample of 60,619 patients. The analysis aims to quantify comparative efficacy and safety while explicitly presenting uncertainties and probabilities of clinically meaningful differences.
The primary efficacy endpoint was a composite MACE outcome defined as all-cause mortality, a recurrent non-fatal myocardial infarction, or non-fatal stroke. The primary safety endpoint was the trial-reported major bleeding event as reported in the included RCTs.
The Bayesian NMA used a primary binomial complementary log-log model with a log(time) offset and random effects. The authors report a comprehensive statistical workflow including prior and posterior predictive checks, convergence diagnostics, formal model comparisons, sensitivity analyses, and probability summaries across a pre-specified range of practical equivalence (ROPE). The ROPE was defined as hazard ratio (HR) 0.90–1.11 to identify regions of likely clinical equivalence versus meaningful benefit or harm.
Nineteen RCTs totalling 60,619 patients were included. The preprint states that data and code used for the analysis are publicly available on the authors’ GitHub repository. The authors declare no competing interests and note that only previously published data were used.
For the MACE composite endpoint, the Bayesian posterior estimates and probability statements were as follows:
Prasugrel versus clopidogrel: posterior HR 0.87 (95% credible interval [CrI] 0.76–1.03). The analysis reports a 70% posterior probability that prasugrel confers a clinically relevant efficacy benefit (HR < 0.9) compared with clopidogrel.
Prasugrel versus ticagrelor: posterior HR 0.89 (95% CrI 0.73–1.12) with a reported 54% probability that prasugrel provides a clinically relevant efficacy benefit relative to ticagrelor.
Ticagrelor versus clopidogrel: posterior HR 0.98 (95% CrI 0.82–1.18). Importantly, 76% of the posterior probability mass for this comparison lies within the ROPE (HR 0.90–1.11), indicating a high probability of practical equivalence between ticagrelor and clopidogrel for MACE in this synthesis.
These results emphasize probabilistic statements rather than binary declarations of statistical significance, quantifying the likelihood of benefit versus equivalence.
For the primary safety endpoint of study-reported major bleeding, the Bayesian analysis quantified posterior hazards and probabilities of meaningful harm (defined as HR > 1.1):
Ticagrelor versus clopidogrel: posterior HR 1.26 (95% CrI 1.01–1.63) with an 86% posterior probability of meaningful bleeding harm (HR > 1.1) compared with clopidogrel.
Prasugrel versus clopidogrel: posterior HR 1.10 (95% CrI 0.88–1.26) with a reported 51% probability of meaningful bleeding harm (HR > 1.1).
These estimates quantify the trade-off between efficacy and bleeding risk across agents and highlight probabilistic assessments of harm.
By reporting probabilities across a ROPE (HR 0.90–1.11), the Bayesian framework explicitly identifies ranges where treatments are likely to be practically equivalent. The finding that a large portion of the posterior probability for ticagrelor versus clopidogrel lies within the ROPE underscores that, for MACE, clinically meaningful differences may be absent despite point estimates near unity. The Bayesian presentation also allows direct probability statements about clinically relevant benefit (HR < 0.9) or harm (HR > 1.1), which the authors use to characterize uncertainty and practical decision-relevant ranges.
The authors contrast their Bayesian results with a recent frequentist network meta-analysis that concluded prasugrel provided the optimal efficacy-safety balance. In this Bayesian update, despite a large evidence base of 19 RCTs and roughly 60,000 patients, the analysis revealed persistent uncertainties and identified probable regions of equivalence among P2Y12 inhibitors that the prior frequentist synthesis may have underappreciated. The authors emphasize that probabilistic summaries and ROPE-based interpretations change the framing from a single "best" agent to a nuanced depiction of likely benefit, harm, and equivalence.
The preprint notes that only previously published trial data were used. The authors declare no competing interests. They state that trial registration and reporting guidelines were addressed as applicable. The data and code supporting the Bayesian NMA are available at the referenced GitHub repository, as reported in the source.
This Bayesian reanalysis provides clinicians and guideline developers with probability-based assessments of comparative efficacy and bleeding risk among P2Y12 inhibitors after ACS and highlights where uncertainty and practical equivalence remain despite a large randomized evidence base.