ICARUS score for predicting peri-procedural bleeding in patients undergoing percutaneous coronary intervention with cangrelor

Background: Tools for precise prediction of bleeding risk in patients undergoing percutaneous coronary intervention (PCI) with cangrelor are lacking. Methods: Consecutive patients undergoing PCI and treated with cangrelor in 7 centers were retrospectively enrolled. The primary endpoint was Bleeding Academic Research Consortium (BARC) BARC 2, 3, or 5 bleeding 48 h after PCI. Predictors of BARC 2-5 bleeding were identified in a derivation cohort and combined into a numerical risk score. Discrimination and calibration were assessed in the derivation and validation cohorts. A threshold to define high bleeding risk (HBR) was identified and its diagnostic accuracy was compared with that of currently recommended bleeding risk scores. Results: 1071 patients undergoing PCI with cangrelor were included. Fifty-four patients (5 %) experienced a BARC 2-5 bleeding, of whom 24 (44 %) from the access site. Age >= 75 years (odds ratio [OR] 2.58, 95 % confidence interval [CI] 1.21-5.48, p = 0.01), acute coronary syndrome at presentation (OR 8.14, 95 % CI 2.28-52, p = 0.01), and femoral access (OR 6.21, 95 % CI 2.71-14, p < 0.001) independently predicted BARC 2-5 bleeding at 48 h after PCI. The three items were combined to form a new risk score, the ICARUS score, showing good discrimination in both the derivation (area under the curve [AUC] 0.78) and internal validation (AUC 0.77) cohorts, and excellent calibration. An ICARUS score > 9 points accurately identified patients at HBR, showing better discrimination than other risk scores. Conclusions: A risk score based on age, clinical presentation and access site, predicts the risk of periprocedural bleeding in patients receiving cangrelor (ClinicalTrials.gov ID: NCT05505591).