Hemorrhagic risk scores in hospitalized patients with acute coronary syndrome: can they (only) predict bleeding events?

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Acute coronary syndrome (ACS) is a life-threatening condition and its therapeutic approach increases the risk of important bleeding events which are associated with a worse prognosis. Along with hemorrhagic events, a drop on hemoglobin level not related to bleeding or the development of anemia could have a negative impact on prognosis. Both CRUSADE and PARIS bleeding risk scores are used to evaluate and to stratify the risk of major bleeding in ACS. However their actual predictive value has been questioned and validity of these scores in predicting in-hospital mortality (IHM) is not established. Objectives To evaluate the actual prognostic value of CRUSADE and PARIS bleeding scores in ACS patients during their hospitalization stay. Methods Retrospective single center cohort study including 103 hospitalized patients after an acute ST-segment elevation myocardial infarction (STEMI) regardless of its reperfusion strategy. In-hospital major hemorrhagic events (IHMHE), considered intracerebral hemorrhage, those resulting in hemodynamic compromise or requiring a blood transfusion, were assessed. Data on hemoglobin levels (HL) at hospital admission and at the time of hospital discharge were also collected and a composite endpoint (CE) of IHMHE and a drop in HL  ≥ 3g/dL were elaborated. Both scores were calculated for each patient, its predictive value and their impact on IHM were determined. Results Out of 103 patients enrolled, the median age was 58.15 ± 12.6 years and 85.4% were male. Two IHMHE occurred, twenty patients (19.4%) had anemia at the time of hospital discharge and 16 of these patients (15.5%) were not anemic at the time of hospital admission. Nine (8.7%) patients had a drop in their HL of at least 3g/dL. The five bleeding risk categories defined by CRUSADE investigators were used, with 48 (46.6%) patients in the very low risk category, 9 (8.7%) and 6 (5.8%) in the high and very high risk category, respectively. Hospitalization length stay was 5.6 ± 4.1 days with an overall in-hospital mortality (IHM) of 5.8%. Receiver operating characteristic curve (ROC) analysis showed that CRUSADE score had an excellent discriminatory power for the CE (AUC 0.927, 95% CI 0.854-1.000) and the PARIS score had an acceptable discriminatory value (AUC 0.775, 95% CI 0.616-0.935). Both CRUSADE and PARIS bleeding scores also had prognostic value in evaluating IHM (AUC 0.929, 95% CI 0.856-1.000 and AUC 0.788, 95% CI 0.634-0.942, respectively). No specific and independent predictors of IHMHE were found, neither related to individual characteristics nor to therapeutic approach.   Conclusion The presenting study showed that CRUSADE and PARIS scores still have discriminatory power to assess CE and to assess IHM in ACS patients. Their addition to stratification tools could be of interest.