Preoperative liver dysfunction influences blood product administration and alterations in circulating haemostatic markers following ventricular assist device implantation.

Abstract

Preoperative liver dysfunction may influence haemostasis following ventricular assist device (VAD) implantation. The Model for End-stage Liver Disease (MELD) score was assessed as a predictor of bleeding and levels of haemostatic markers in patients with currently utilized VADs. Sixty-three patients (31 HeartMate II, 15 HeartWare, 17 Thoratec paracorporeal ventricular assist device) implanted 2001-11 were analysed for preoperative liver dysfunction (MELD) and blood product administration. Of these patients, 21 had additional blood drawn to measure haemostatic marker levels. Cohorts were defined based on high (≥18.0, n = 7) and low (<18.0, n = 14) preoperative MELD scores. MELD score was positively correlated with postoperative administration of red blood cell (RBC), platelet, plasma and total blood product units (TBPU) , as well as chest tube drainage and cardiopulmonary bypass time. Age and MELD were preoperative predictors of TBPU by multivariate analysis. The high-MELD cohort had higher administration of TBPU, RBC and platelet units and chest tube drainage postimplant. Similarly, patients who experienced at least one bleeding adverse event were more likely to have had a high preoperative MELD. The high-MELD group exhibited different temporal trends in F1 + 2 levels and platelet counts to postoperative day (POD) 55. D-dimer levels in high-MELD patients became elevated versus those for low-MELD patients on POD 55. Preoperative MELD score predicts postoperative bleeding in contemporary VADs. Preoperative liver dysfunction may also alter postoperative subclinical haemostasis through different temporal trends of thrombin generation and platelet counts, as well as protracted fibrinolysis.