BackgroundHyperglycemia in the neohepatic phase of liver transplantation (LT) tends to decrease toward completion of the surgical procedure. Refractory hyperglycemia in the neohepatic phase (RH) is influenced by multiple perioperative factors and may be connected to posttransplant outcomes. We attempted to demonstrate the relationship of RH to posttransplant outcomes and to establish a predictive model for RH in living donor liver transplantation (LDLT). MethodsPerioperative data of 211 patients who underwent LDLT from 2009 and 2012 were reviewed, including declines in the blood glucose levels during the neohepatic phase. Perioperative variables including the posttransplant model for end-stage liver disease (MELD) score until day 30 were compared between patients with normal declines in blood glucose and patients with RH. Selected variables after intergroup comparisons were examined by means of multivariate logistic regression to establish a predictive model for RH occurrence. ResultsThe mean blood glucose decline was 22.3 ± 31.5 mg/dL during the neohepatic phase, and 84 of 203 patients (41.4%) had no decline in blood glucose. In intergroup comparisons, preoperative factors associated with RH included sex, Child-Pugh-Turcotte class, MELD score, emergency, liver enzymes, and graft-to-recipient weight ratio. During surgery, surgical time, serum lactate, and arterial pH were associated with RH. After surgery, the RH group showed slower recovery of the MELD score (15.2 versus 11.9 days) and higher MELD scores until day 10 (P < .05). After the multivariate analysis, recipient sex, emergency, surgical time (≤9 h), and the final intraoperative serum lactate level (≥5.0 mmol/L) were included in the predictive model for RH. ConclusionsRH was associated with delayed functional recovery of the liver graft in LT. Recipient sex, emergency, surgical time, and the final intraoperative serum lactate level were identified as predictors of RH. Close monitoring of intraoperative blood glucose in LDLT may be an early prognostic indicator.