Machine learning-based prediction of adverse events following an acute coronary syndrome (PRAISE): a modelling study of pooled datasets.

BACKGROUND: The accuracy of current prediction tools for ischaemic and bleeding events after an acute coronary syndrome (ACS) remains insufficient for individualised patient management strategies. We developed a machine learning-based risk stratification model to predict all-cause death, recurrent acute myocardial infarction, and major bleeding after ACS. METHODS: Different machine learning models for the prediction of 1-year post-discharge all-cause death, myocardial infarction, and major bleeding (defined as Bleeding Academic Research Consortium type 3 or 5) were trained on a cohort of 19 826 adult patients with ACS (split into a training cohort [80%] and internal validation cohort [20%]) from the BleeMACS and RENAMI registries, which included patients across several continents. 25 clinical features routinely assessed at discharge were used to inform the models. The best-performing model for each study outcome (the PRAISE score) was tested in an external validation cohort of 3444 patients with ACS pooled from a randomised controlled trial and three prospective registries. Model performance was assessed according to a range of learning metrics including area under the receiver operating characteristic curve (AUC). FINDINGS: The PRAISE score showed an AUC of 0·82 (95% CI 0·78-0·85) in the internal validation cohort and 0·92 (0·90-0·93) in the external validation cohort for 1-year all-cause death; an AUC of 0·74 (0·70-0·78) in the internal validation cohort and 0·81 (0·76-0·85) in the external validation cohort for 1-year myocardial infarction; and an AUC of 0·70 (0·66-0·75) in the internal validation cohort and 0·86 (0·82-0·89) in the external validation cohort for 1-year major bleeding. INTERPRETATION: A machine learning-based approach for the identification of predictors of events after an ACS is feasible and effective. The PRAISE score showed accurate discriminative capabilities for the prediction of all-cause death, myocardial infarction, and major bleeding, and might be useful to guide clinical decision making. FUNDING: None.

Evaluation of optimal medical therapy in acute myocardial infarction patients with prior stroke.

BACKGROUND: Treatment of acute myocardial infarction (AMI) patients with prior stroke is a common clinical dilemma. Currently, the application of optimal medical therapy (OMT) and its impact on clinical outcomes are not clear in this patient population. METHODS: We retrieved 765 AMI patients with prior stroke who underwent percutaneous coronary intervention (PCI) during the index hospitalization from the international multicenter BleeMACS registry. All of the subjects were divided into two groups based on the prescription they were given prior to discharge. Baseline characteristics and procedural variables were compared between the OMT and non-OMT groups. Mortality, re-AMI, major adverse cardiovascular events (MACE), and bleeding were followed-up for 1 year. RESULTS: Approximately 5% of all patients presenting with AMI were admitted to the hospital for ischemic stroke. Although the prescription rate of each OMT medication was reasonably high (73.3%-97.3%), 47.7% lacked at least one OMT medication. Patients receiving OMT showed a significantly decreased occurrence of mortality (4.5% vs 15.1%, p < 0.001), re-AMI (4.2% vs 9.3%, p = 0.004), and the composite endpoint of death/re-AMI (8.6% vs 20.5%, p < 0.001) compared to those without OMT. No significant difference was observed between the groups regarding bleeding. After adjusting for confounding factors, OMT was the independent protective factor of 1-year mortality, while age was the independent risk factors. CONCLUSIONS: OMT at discharge was associated with a significantly lower 1-year mortality of patients with AMI and prior stroke in clinical practice. However, OMT was provided to just half of the eligible patients, leaving room for substantial improvement. CLINICAL TRIAL REGISTRATION: NCT02466854.

Impact of renin-angiotensin system blockade on the prognosis of acute coronary syndrome based on left ventricular ejection fraction.

INTRODUCTION AND OBJECTIVES: For patients with acute coronary syndrome (ACS) treated with percutaneous coronary intervention (PCI), it is unclear whether angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB) are associated with reduced mortality, particularly with preserved left ventricular ejection fraction (LVEF). The goal of this study was to determine the association between ACEI/ARB and mortality in ACS patients undergoing PCI, with and without reduced LVEF. METHODS: Data from the BleeMACS registry were used. The endpoint was 1-year all-cause mortality. The prognostic value of ACEI/ARB was tested after weighting by survival-time inverse probability and after adjustment by Cox regression, propensity score, and instrumental variable analysis. RESULTS: Among 15 401 ACS patients who underwent PCI, ACEI/ARB were prescribed in 75.2%. There were 569 deaths (3.7%) during the first year after hospital discharge. After multivariable adjustment, ACEI/ARB were associated with lower 1-year mortality, ≤ 40% (HR, 0.62; 95%CI, 0.43-0.90; P=.012). The relative risk reduction of ACEI/ARB in mortality was 46.1% in patients with LVEF ≤ 40%, and 15.7% in patients with LVEF> 40% (P value for treatment-by-LVEF interaction=.008). For patients with LVEF> 40%, ACEI/ARB was associated with lower mortality only in ST-segment elevation myocardial infarction (HR, 0.44; 95%CI, 0.21-0.93; P=.031). CONCLUSION: The benefit of ACEI/ARB in decreasing mortality after an ACS in patients undergoing PCI is concentrated in patients with LVEF ≤ 40%, and in those with LVEF> 40% and ST-segment elevation myocardial infarction. In non-ST-segment elevation-ACS patients with LVEF> 40%, further studies are needed to assess the prognostic impact of ACEI/ARB.

Prevalence and outcome of patients with cancer and acute coronary syndrome undergoing percutaneous coronary intervention: a BleeMACS substudy.

BACKGROUND: The prevalence and outcome of patients with cancer that experience acute coronary syndrome (ACS) have to be determined. METHODS AND RESULTS: The BleeMACS project is a multicentre observational registry enrolling patients with acute coronary syndrome undergoing percutaneous coronary intervention worldwide in 15 hospitals. The primary endpoint was a composite event of death and re-infarction after one year of follow-up. Bleedings were the secondary endpoint. 15,401 patients were enrolled, 926 (6.4%) in the cancer group and 14,475 (93.6%) in the group of patients without cancer. Patients with cancer were older (70.8±10.3 vs. 62.8±12.1 years, P<0.001) with more severe comorbidities and presented more frequently with non-ST-segment elevation myocardial infarction compared with patients without cancer. After one year, patients with cancer more often experienced the composite endpoint (15.2% vs. 5.3%, P<0.001) and bleedings (6.5% vs. 3%, P<0.001). At multiple regression analysis the presence of cancer was the strongest independent predictor for the primary endpoint (hazard ratio (HR) 2.1, 1.8-2.5, P<0.001) and bleedings (HR 1.5, 1.1-2.1, P=0.015). Despite patients with cancer generally being undertreated, beta-blockers (relative risk (RR) 0.6, 0.4-0.9, P=0.05), angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (RR 0.5, 0.3-0.8, P=0.02), statins (RR 0.3, 0.2-0.5, P<0.001) and dual antiplatelet therapy (RR 0.5, 0.3-0.9, P=0.05) were shown to be protective factors, while proton pump inhibitors (RR 1, 0.6-1.5, P=0.9) were neutral. CONCLUSION: Cancer has a non-negligible prevalence in patients with acute coronary syndrome undergoing percutaneous coronary intervention, with a major risk of cardiovascular events and bleedings. Moreover, these patients are often undertreated from clinical despite medical therapy seems to be protective. Registration:The BleeMACS project (NCT02466854).

BleeMACS: rationale and design of the study.

BACKGROUND: Bleeding events after an acute coronary syndrome have a negative impact on prognosis. Available risk scores are limited by suboptimal accuracy, prediction of only in-hospital events and absence of patients treated with new antiplatelet agents in the current era of widespread use of percutaneous coronary intervention. DESIGN: The BleeMACS (Bleeding complications in a Multicenter registry of patients discharged after an Acute Coronary Syndrome) project is a multicenter investigator-initiated international retrospective registry that enrolled more than 15 000 patients discharged with a definitive diagnosis of acute coronary syndrome and treated with percutaneous revascularization. The primary end point is the incidence of major bleeding events requiring hospitalization and/or red cell transfusion concentrates within 1 year. An integer risk score for bleeding within the first year after hospital discharge will be developed from a multivariate competing-risks regression. CONCLUSION: The BleeMACS registry collaborative will allow development and validation of a risk score for prediction of major bleeding during follow-up for patients receiving contemporary therapies for acute coronary syndrome.