Predictive Value of the Model for End-Stage Liver Disease Score in Patients Undergoing Left Ventricular Assist Device Implantation

Abstract

Axial flow left ventricular assist device (LVAD) implantation is an effective therapy for patients with advanced heart failure. As the preoperative hepatic and renal function play a critical role in determining adverse events after LVAD implantation, we analyzed the predictive role of the model for end-stage liver disease (MELD) score in determining in-hospital mortality after surgery. One hundred twenty-six patients underwent implant of an LVAD at our institution. Their individual preoperative MELD scores and perioperative total blood product usage (TBPU) were calculated. As LVAD implant as a reoperation is known to influence postoperative bleeding and mortality independently, the patients were divided into group I (first cardiac surgery) and group II (reoperative surgery). Group I: LVAD implantation was performed in 68/126 (54%) patients as their first cardiac surgery. The mean MELD score was 16.3 ± 6. Median TBPU for this group was 20.7 (0, 135) units. Inhospital mortality/30-day mortality was 4/68 (5.8%). Increasing MELD score (c-statistic = 0.88) and TBPU were found to be predictors of early mortality. An increasing MELD score was associated with more TBPU (p < 0.01) with a 10.9 ± 3 TBPU increase per a 10 unit rise in the MELD score. Group II: Of the 126 patients, 58 (46%) underwent LVAD implantation as a reoperation. Mean MELD score for these patients was 16 ± 5. Inhospital mortality/30-day mortality in this group was 12% and median TBPU was 30 (4,153) units. The MELD score was not predictive of inhospital mortality in these patients (p = 0.97). The MELD score is predictive of early mortality in patients undergoing LVAD implantation as their first cardiac surgery. Use of this score to select patients for LVAD implantation may be appropriate.