A Bayesian network meta-analysis of 19 randomized controlled trials encompassing 60,619 patients undergoing PCI after acute coronary syndrome finds that prasugrel holds only a 70% probability of meaningful efficacy benefit over clopidogrel for MACE reduction (HR 0.87, 95% CrI 0.76–1.03) and just 54% over ticagrelor. Critically, 76% of the posterior probability for ticagrelor versus clopidogrel falls within a pre-specified zone of practical equivalence. On bleeding risk, ticagrelor carries an 86% probability of meaningful harm versus clopidogrel, while prasugrel shows 51%.
This analysis directly challenges a recent frequentist NMA that declared prasugrel the optimal agent — a conclusion that statistical framing may have overstated. Frequentist approaches yield binary significance thresholds; Bayesian methods instead express findings as probability distributions, surfacing the genuine uncertainty lurking behind point estimates. For practicing cardiologists and patients, the implication is substantial: the three dominant antiplatelet agents may perform more similarly than guideline confidence suggests, making bleeding risk tolerance, cost, and individual patient factors more decisive than headline efficacy differences. The finding that clopidogrel — the oldest, cheapest, and most studied option — remains plausibly equivalent to its newer rivals will be clinically provocative. Limitations include indirect comparisons across heterogeneous trial populations and variable bleeding definitions. As a preprint posted on medRxiv and not yet peer-reviewed, these conclusions require independent scrutiny before influencing prescribing practice. Still, this represents a methodologically sophisticated, potentially paradigm-shifting reassessment of a high-stakes cardiovascular decision.