Prioritization for liver transplantation

Abstract

Objective: to determine the threshold MELD scores when prioritizing for liver transplantation. Materials and methods. We conducted a cohort study of 350 patients who were waitlisted for liver transplantation between 2015 and 2020. Results. A logistic regression model was used to identify the independent predictors of liver transplantation waitlist mortality. MELD scores and serum albumin at the time of listing were significant predictors of mortality (p = 0.001 and p = 0.004, respectively). Their predictive values were confirmed using ROC (Receiver Operating Characteristic) analysis. The area under the ROC curve (AUC) was 0.883 [95% confidence interval (CI) 0.828–0.939; p < 0.001] for MELD, and 0.841 [95% CI 0.775–0.907; p < 0.001] for serum albumin. Mortality odds ratio was 3.7778, 95% CI (1.619–7.765) provided that the listing MELD score was ≥25. Mortality odds ratio was 2.979 (95% CI 1.63–5.95) provided that the listing serum albumin concentration was ≤30.1 g/L. With a threshold MELD score of 25, there were significant differences between patient survival when comparing patient cohorts with MELD ≥25 and with MELD ≤25 (Log-rank, p < 0.0001). Conclusion. The MELD model has a high predictive ability in prioritization of waitlisted candidates for liver transplantation. The threshold MELD score and mortality predictors were determined. There were significant differences between patient survival among patient cohorts with MELD ≥25 and with MELD ≤25.