VALIDATION OF THE NEWLY DEVELOPED MODEL OF END-STAGE LIVER DISEASE 3.0 IN PATIENTS ON LIVER TRANSPLANTATION WAITLIST IN KOREA

Abstract

Background: The Model of End-Stage Liver Disease (MELD) score is used in Korea to allocate liver grafts. Currently, MELD 3.0, a new predictive model that combines albumin were newly proposed. This study investigated the performance of the MELD 3.0 score for the prediction of 3-month survival in Korea compared with other model including original MELD score.

Methods: Validation was performed in 2,153 patients listed for liver transplantation (LT) from 2 tertiary hospitals in Korea. Discrimination (c-index, iAUC) and calibration analysis were performed between MELD 3.0, MELD-Na, original MELD, and Child-pugh score.

Results: Patients were 52.95 ± 11.17 years old and had an original MELD score of 18.70 ± 9.65. Alcohol (39.9%) and chronic hepatitis B (38.5%) were two main etiology. MELD 3.0 with albumin showed best discrimination (c-index 0.738, iAUC 0.719) than MELD-Na (c-index 0.730, iAUC 0.707), original MELD (c-index 0.718, iAUC 0.687), and Child-pugh score (c-index 0.689, iAUC 0.688) for predicting 3-month overall survival. MELD 3.0 without albumin also showed the similar performance as MELD 3.0 with albumin, and the difference between the two scoring systems was not significant. In addition, MELD 3.0 with albumin reclassified a net of 24.4% of quintile to a higher MELD tier, particularly in women with high MELD score (>20).

Conclusion: Compared with the previous allocation system, the MELD 3.0 score was significantly better at predicting short term prognosis in patient with LT waitlist in Korea. Moreover, MELD 3.0 is expected to overcome the sex disparity of the previous scoring system.