Prospective evaluation of ventricular assist device risk scores' capacity to predict cardiopulmonary exercise parameters.

Abstract

Risk scores for left ventricular assist device (LVAD) therapy are known to predict morbidity and adverse events in addition to mortality. This study evaluates the capacity of popular LVAD risk scores to predict cardiopulmonary exercise parameters. Adult patients undergoing continuous flow LVAD implantation were prospectively followed. Five risk scores were calculated before implantation: Model for End-stage Liver Disease (MELD), MELD excluding international normalized ratio (MELD-XI), MELD including sodium (MELD-Na), HeartMate2 Risk Score (HMRS) and Destination Therapy Risk Score (DTRS). Cardiopulmonary exercise tests (CPETs) were performed before and after implantation; peak oxygen consumption (vO2max), the lowest ventilation to carbon dioxide output ratio (vE/vCO2) and exercise time were measured. Ninety-two patients were implanted during the study period; of these, 30 patients completed preimplantation and postimplantation CPETs (CPET cohort). The mean preimplantation and postimplantation CPET dates were 29 ± 10 days before and 109 ± 5 days following implantation. CPET parameters significantly improved after implantation (P < 0.05). In multivariate analysis, MELD, MELD-XI, MELD-Na and HMRS independently predicted both preimplantation and postimplantation vE/vCO2, while MELD-Na and HMRS were also independent predictors of preimplantation and postimplantation vO2max, respectively. Four preimplantation LVAD risk scores (HMRS, MELD, MELD-Na and MELD-XI) independently predict important cardiopulmonary exercise parameters such as vE/vCO2 and vO2 max in LVAD therapy. Out of these 4 risk scores, MELD-Na and HMRS appear to be the best predictors of preimplantation and postimplantation CPET parameters, respectively.