Preoperative Predictors for Length of ICU Admission after Ventricular Assist Device (VAD) Implantation As Bridge to Transplant—A Single Centre Experience

Introduction: VAD are increasingly used as a bridge to cardiac transplant. Duration of postoperative ICU stay (iLOS) is a major determinant of VAD implantation cost and may predict subsequent morbidity and mortality. We sought to identify preoperative factors that predict iLOS and mortality. Methods: A retrospective review of all VAD implants at our institution. Preoperative variables analysed: days from admission to VAD, diabetes, redo sternotomy, univentricular vs. biventricular VAD, VAD generation (first vs. third), age, gender, body mass index (BMI), serum albumin, haemoglobin, white cell count (WCC), platelet count, GFR. Fisher's exact test used for categorical and Student's T test for continuous variables. Pearson's correlation used to analyse correlation between continuous variables. Results: Fifty-three VAD implanted (41 first and 12 third generation). Three patients died within 48 h of initial admission due to acute operative complications and were excluded from iLOS analysis. Mean age: 43.8 years (range 15–67). Mean days from admission to VAD: 15.5 (range 0–145). Mean overall iLOS: 24.7 days (range 4–103). There was a significantly longer iLOS for first generation vs. third generation VAD (32.8 vs. 21.3; p = 0.04). Weak correlation (r = 0.44, p = 0.01) between length of hospital stay and days to VAD implant. No significant association between age, gender, diabetes, redo sternotomy, days to implantation, univentricular vs. biventricular, BMI, serum albumin, haemoglobin, WCC, platelets or GFR. Twenty two percent mortality for first vs. 0% third generation (p < 0.001). Conclusion: Only use of third generation VAD predicts reduced mortality and iLOS.