A Retrospective Study of the Role of Perioperative Serum Albumin and the Albumin–Bilirubin Grade in Predicting Post-Liver Transplant Length of Stay

Abstract

Introduction

Serum albumin’s association with liver transplant outcomes has been investigated with mixed findings. This study aimed to evaluate perioperative albumin level, independently and as part of the albumin–bilirubin (ALBI) grade, as a predictor of post-liver transplant hospital and intensive care unit (ICU) length of stay (LOS).

Methods

Adult liver-only transplant recipients at our institution from September 2011 to May 2019 were included in this retrospective study. Repeat transplants were excluded. Demographic, laboratory, and hospital course data were extracted from an institutional data warehouse. Negative binomial regression was used to assess the association of LOS with ALBI grade, age, BMI, ASA score, Elixhauser comorbidity index, MELD-Na, warm ischemia time, units of platelets and cryoprecipitate transfused, and preoperative serum albumin.

Results

Six hundred and sixty-three liver transplant recipients met inclusion criteria. The median preoperative serum albumin was 3.1 [2.6–3.6] g/dL. The median postoperative ICU and hospital LOS were 3.8 [2.4–6.8] and 12 [8–20] days, respectively. Preoperative serum albumin predicted hospital but not ICU LOS (ratio .9 [95% confidence interval (CI) .84–.99], P = .03, hospital LOS vs ratio .92 [95% CI 0.84–1.02], P = .10, ICU LOS). For patients with MELD-Na ≤ 20, ALBI grade-3 predicted longer hospital and ICU LOS (ratio 1.40 [95% CI 1.18–1.66], P < .001, hospital LOS vs ratio 1.62 [95% CI 1.32–1.99], P < .001, ICU LOS). These associations were not significant for patients with MELD-Na > 20.

Conclusions

Serum albumin predicted post-liver transplant hospital LOS. ALBI grade-3 predicted increased hospital and ICU LOS in low MELD-Na recipients.