Comparative effectiveness of prasugrel and ticagrelor among patients undergoing contemporary PCI: insights from the BMC2 registry

Abstract Background The ISAR-REACT 5 trial found acute coronary syndrome (ACS) patients treated with prasugrel had lower incidence of death, myocardial infarction (MI), or stroke compared to those treated with ticagrelor. European guidelines give preference to prasugrel over ticagrelor in ACS patients undergoing percutaneous coronary intervention (PCI). Purpose To compare incidence of MI, stroke, and death at one year for ACS and non-ACS patients who undergo PCI in a large statewide U.S. registry and were discharged on ticagrelor or prasugrel. Methods The study included all ACS patients who underwent PCI from April 2018 to January 2024, were initiated on prasugrel or ticagrelor, and were enrolled in both Medicare and the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2), a clinical registry of all PCI cases at non-federal hospitals in Michigan, U.S. Exclusion criteria mirrored ISAR-REACT 5, such as patients on dialysis and with stroke history. The primary outcome was MI, transient ischemic attack (TIA)/stroke, or death at one year, identified through linked Medicare administrative claims. A secondary analysis of non-ACS patients initiated on prasugrel or ticagrelor was also performed. 3:1 (ticagrelor:prasugrel) propensity matching on 27 patient and procedural covariates was used to adjust for non-random antiplatelet selection, and Cox proportional hazards models were fitted adjusting for all 27 covariates. Results There were 10,924 patients identified, of whom 9,182 (84%) were initiated on ticagrelor and 6,764 (62%) had an ACS indication for PCI. Patients initiated on ticagrelor were more likely to be older (72 years vs. 69 years), have urgent/emergent PCI (66% vs. 57%), present with cardiogenic shock (3.1% vs. 1.2%), and have pre-procedural hemoglobin of 12.0 g/dL (21% vs. 16%) (p.001 for all). After propensity matching, no significant baseline differences remained with an absolute standardized difference less than 10% on all covariates. Propensity matching for the ACS patients included 2,840 and 938 in the ticagrelor and prasugrel cohorts, respectively, and for non-ACS patients it included 2,281 and 769 in the ticagrelor and prasugrel cohorts, respectively. In the propensity-matched analysis for ACS patients, there was no significant difference in cumulative incidence of MI, TIA/stroke or death at one year for ticagrelor compared to prasugrel (13.1% vs. 12.4%, adjusted hazard ratio (aHR) = 1.10; p = 0.25). One year incidence for the three individual outcomes were also not significant (Figure 1). For non-ACS patients, there was also no difference in the one year cumulative incidence of the composite outcome (6.3% vs. 6.4%, aHR = 0.98; p = 0.87) (Figure 2). Conclusion In a different population than ISAR-REACT 5; namely, older North American patients with more comorbidities, there was no significant difference in the cumulative incidence of MI, stroke, or death at one year for ACS patients treated with prasugrel or ticagrelor.Figure 1 Figure 2