PRECISE-DAPT score for bleeding risk prediction in patients on dual or single antiplatelet regimens: insights from the GLOBAL LEADERS and GLASSY

Abstract

Aims The five-item PRECISE-DAPT, integrating age, haemoglobin, white-blood-cell count, creatinine clearance, and prior bleeding, predicts bleeding risk in patients on dual antiplatelet therapy (DAPT) after stent implantation. We sought to assess whether the bleeding risk prediction offered by the PRECISE-DAPT remains valid among patients receiving ticagrelor monotherapy from 1 month onwards after coronary stenting instead of standard DAPT and having or not having centrally adjudicated bleeding endpoints.Methods and results The PRECISE-DAPT was calculated in 14 928 and 7134 patients from GLOBAL LEADERS and GLASSY trials, respectively. The ability of the score to predict Bleeding Academic Research Consortium 3 or 5 bleeding was assessed and compared among patients on ticagrelor monotherapy (experimental strategy) or standard DAPT (reference strategy) from 1 month after drug-eluting stent implantation. Bleeding endpoints were investigatorreported or centrally adjudicated in GLOBAL LEADERS and GLASSY, respectively. At 2 years, the c-indexes for the score among patients treated with the experimental or reference strategy were 0.67 [95% confidence interval (CI): 0.63-0.71] vs. 0.63 (95% CI: 0.59-0.67) in GLOBAL LEADERS (P = 0.27), and 0.67 (95% CI: 0.61-0.73) vs. 0.66 (95% CI: 0.61-0.72) in GLASSY (P = 0.88). Decision curve analysis showed net benefit using the PRECISEDAPT to guide bleeding risk assessment under both treatment strategies. Results were consistent between investigator-reported and adjudicated endpoints and using the simplified four-item PRECISE-DAPT.Conclusion The PRECISE-DAPT offers a prediction model that proved similarly effective to predict clinically relevant bleeding among patients on ticagrelor monotherapy from 1 month after coronary stenting compared with standard DAPT and appears to be unaffected by the presence or absence of adjudicated bleeding endpoints.