Optimal Hemodynamics and Risk of Severe Outcomes Post-Left Ventricular Assist Device Implantation.

Abstract

Data on pre-implant hemodynamic optimization before continuous flow left ventricular assist device placement (CF-LVAD) with respect to patient-centered outcomes remain limited. Consecutive patients undergoing CF-LVAD implant between 2007 and 2017 were identified. Hemodynamic variables, trends, and laboratory studies were evaluated pre-LVAD implant in a logistic regression model to identify predictors of the primary composite endpoint: the need for right ventricular assist device therapy, the requirement for hemodialysis at 90 days, and 30-day mortality. Multivariate modeling identified three variables significantly associated with the primary endpoint: right ventricular stroke work index (RVSWI), right atrial pressure (RAP), and blood urea nitrogen (BUN); all immediately pre-LVAD, p < 0.01. Optimal dichotomization points were 500 mmHg*ml*m-2, 12 mmHg, and 40 mg/dL. The three-component model identified an AUC of 0.77 (p < 0.0001) for the composite endpoint. Optimization of 2/3 parameters, 1/3, and 0/3 was associated with odds ratios of 3.5 (95% CI, 1.1-11.7), 7.2 (95% CI, 2.1-24.2), and 20.6 (95% CI, 5.3-80.6), respectively, relative to those patients who were fully optimized (3/3 parameters). The number of optimized parameters was also associated with 1-year overall survival (p = 0.02). Low RVSWI, high RAP, and high BUN were independently associated with adverse outcomes after the CF-LVAD implant, demonstrating a stepwise association with severe postimplant adverse events.