Incremental Value of Platelet Reactivity over a Risk Score of Clinical and Procedural Variables in Predicting Bleeding after Percutaneous Coronary Intervention via the Femoral Approach: Development and Validation of a New Bleeding Risk Score

Background - Growing evidence suggests that platelet reactivity (PR) may predict bleeding. We investigate the incremental value of PR in predicting bleeding after percutaneous coronary intervention (PCI) via the femoral approach over a validated bleeding risk score (BRS) of clinical and procedural variables. Methods and Results - A total of 800 patients undergoing elective PCI via the femoral approach were included. PR was measured before PCI with the VerifyNow P2Y12 assay and low PR was defined as a P2Y12 reaction unit value ≤178. Calculation of the BRS included the following: age, sex, intra-aortic balloon pump, glycoprotein IIb/IIIa inhibitors, chronic kidney disease, anemia, and low-molecular-weight heparin within 48-hour pre-PCI. A new risk score including low PR (BRS-PR) was developed and validated in an independent cohort of patients (n=310). Bleeding events at 30 days after PCI were defined according to the thrombolysis in myocardial infarction, Randomized Evaluation in PCI Linking Angiomax to Reduced Clinical Events (REPLACE)-2, and Bleeding Academic Research Consortium criteria. Both BRS and PR showed high discriminatory power for bleeding (area under the curve [AUC] >0.7 for all definitions). Discriminatory power of BRS-PR (AUC=0.809 for thrombolysis in myocardial infarction bleeding; AUC=0.814 for Bleeding Academic Research Consortium class ≥2 bleeding; AUC=0.708 for Bleeding Academic Research Consortium class ≥3 bleeding; and AUC=0.813 for REPLACE-2 bleeding) was significantly higher than that of BRS alone (P<0.001 for all bleeding definitions). In the validation set, BRS-PR showed higher discriminatory power for thrombolysis in myocardial infarction bleeding than BRS alone (AUC=0.788 versus 0.709; P=0.036). Conclusions - PR has incremental predictive value on bleeding events after elective PCI via the femoral approach over a validated risk score of clinical and procedural variables. A risk score including PR yields significantly better prognostic performance compared with the original BRS.