A Novel Risk Score Predicts Early Right Ventricular Failure after Lvad: A Derivation-validation Multicenter Study

Abstract

Introduction Right ventricular failure (RVF) after LVAD implantation is associated with increased morbidity and mortality. Despite several RVF predictive models, poor performance in external validation cohorts has limited their widespread clinical adoption. Objective To develop a novel RVF predictive model, ascertain its performance in an independent validation cohort, and develop an RVF risk score. Methods Consecutive LVAD patients were prospectively enrolled at the Utah Transplantation Affiliated Hospitals (U.T.A.H) Cardiac Transplant Program (n=477, Derivation cohort). LVAD patients from Inova Heart & Vascular Institute and Henry Ford Medical Center formed the external dataset (n=321, Validation cohort). The primary outcome was early RVF, defined as the need for RVAD or intravenous inotropes for >14 days. The secondary outcome was 3-year all-cause mortality. Multivariable logistic regression analysis was used to develop a predictive model. An RVF risk score was developed using weighted points based on the β-regression coefficients of the multivariable predictors. Results The study included 798 patients, with a mean age of 56y, 84% male, and 30% INTERMACS Profile 1-2. Compared to the derivation cohort, the validation cohort had a higher proportion of African-Americans (37% vs 7%; p 35mg/dL, PA pulse pressure 0.5). The model had a c-statistic of 0.73 ([95% CI:0.67-0.79]; p Conclusion We propose a novel scoring system to predict post-LVAD RVF, achieving high discriminative performance after being tested in distinct and highly heterogeneous populations. This simple predictive tool could impact patient selection and perioperative management of LVAD patients.