Validation of Three Platelet Function Tests for Bleeding Risk Stratification During Dual Antiplatelet Therapy Following Coronary Interventions

BackgroundAlthough low platelet reactivity (LPR) is commonly detected during bleeding, a validated threshold for reliable DAPT bleeding risk stratification is lacking. We tested the diagnostic utility of 3 conventional platelet-activity assays to define the predictive value (if any) of LPR for bleeding. HypothesisWe hypothesized whether one of these tests be better than any others for predicting bleeding events. MethodsPatients (n=800) following drug-eluting stent implantation received DAPT. Bleeding was assessed by Bleeding Academic Research Consortium (BARC) classification and events were collected for 1 year after stenting. Platelet reactivity was measured by light transmittance aggregometry (LTA), VerifyNow, and multiple electrode aggregometry (MEA). The LPR values for bleeding event stratification were defined as 15% for LTA, 139 PRU for VerifyNow, and 25 U for MEA. ResultsBleeding events occurred in 18 patients (2.3%). All tests distinguished LPR as an independent predictor for bleeding by univariate analysis ([HR]: 5.00, 95% [CI]: 1.8-14.0, P = 0.002 for LTA; HR: 21.3, 95% CI: 6.2-73.0, P < 0.0001 for VerifyNow; and HR: 7.4, 95% CI: 2.2-25.5, P = 0.002 for MEA). Multivariate analysis revealed that only VerifyNow (HR: 11.5, 95% CI: 2.9-45.7, P < 0.0004) remained an independent predictor for bleeding. However, the specificity (81.5%, 60.2%, and 81.7%, respectively) and sensitivity (61.1%, 83.3%, and 83.2%, respectively) of all 3 tests were quite low. ConclusionsAmong 3 conventional platelet-activity assays, VerifyNow was better than LTA or MEA for triaging future bleeding risks. However, all 3 tests failed to reliably predict future bleeding.