The MELD 3.0 policy significantly improved 12-month deceased donor liver transplantation rates for adult HCC patients compared to the pre-policy era (62.2% vs 53.3%, p<0.001).
Cohort (n=12,479)
Yes
Does the MELD 3.0 policy improve deceased donor liver transplantation access and reduce socioeconomic disparities in adult HCC patients?
The MELD 3.0 policy improved overall access to deceased donor liver transplantation for HCC patients but failed to resolve persistent socioeconomic disparities.
Absolute Event Rate: 62.2% vs 53.3%
p-value: p=<0.001
Abstract Background: The MELD 3.0 policy, implemented by UNOS on July 13, 2023, updated the liver allocation scoring system to better reflect 90-day mortality risk. While intended to improve transplant equity, questions remain about its impact on socioeconomic disparities. We examined how MELD 3.0 affected DDLT access across socioeconomic groups and whether it narrowed or widened existing disparities. Methods: We analyzed 12,479 adult HCC patients in the 21-month periods before (n=6,154) and after (n=6,325) MELD 3.0 implementation using UNOS data. Competing risks analysis with Aalen-Johansen estimators assessed DDLT rates accounting for death/deterioration. Gray's test evaluated socioeconomic disparities in DDLT access by race/ethnicity, insurance, education, employment, citizenship, gender, and region. Era-specific censoring used listing date as time origin. Results: MELD 3.0 implementation significantly improved DDLT rates at 12 months: 53.3% pre-policy vs. 62.2% post-policy (+8.9 percentage points, p0.001). Socioeconomic disparities showed mixed patterns. Race/ethnicity disparities persisted and slightly widened: Asian patients improved from 49.0% to 54.7%, while White patients increased from 55.1% to 64.3%. Black patients rose from 55.1% to 59.5%, and Hispanic patients from 46.1% to 64.8%, with differences remaining significant (pre p0.001, post p0.001). Education disparities remained minimal: less than college vs. college graduates had similar rates pre-policy (53.0% vs. 53.2%) and post-policy (62.5% vs. 62.1%), with significance reflecting sample size rather than meaningful differences (p0.001 both eras). Insurance disparities showed mixed results: private insurance maintained advantage over public insurance, though significance weakened (p=0.015 pre, p=0.043 post). Citizenship disparities showed a slight reduction in absolute difference (US vs. non-US: 53.4% vs. 51.0% pre; 62.3% vs. 60.7% post), while statistical significance strengthened (p=0.047 pre, p=0.010 post). Employment and regional disparities persisted: post-policy, employed patients had slightly higher access, and substantial regional variation continued (p0.001 both eras). Gender differences remained non-significant. Conclusions: MELD 3.0 improved overall DDLT access for HCC patients but had varied effects on socioeconomic disparities. Education disparities remained minimal, while race/ethnicity, insurance, employment, citizenship, and regional differences continued. Although MELD 3.0 enhanced transplant access, persistent disparities highlight the need for additional interventions targeting socioeconomic determinants beyond allocation score refinements. Citation Format: Tehyun Phillip Eom, Junho Song, Hyungjune Ku, Sungsu Park, Minkwan Kim, Sunghan Kim, Sangsoo Lee, Heekyoo Kim, Seoyeong Ku, Hyewon Kim, Amy Choi, Changmin Jo. MELD 3.0 policy and evolving socioeconomic disparities in liver transplantation abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2361.
Eom et al. (Fri,) conducted a cohort in Hepatocellular carcinoma (HCC) requiring liver transplantation (n=12,479). MELD 3.0 policy vs. Pre-MELD 3.0 policy was evaluated on Deceased donor liver transplantation (DDLT) rates at 12 months (p=<0.001). The MELD 3.0 policy significantly improved 12-month deceased donor liver transplantation rates for adult HCC patients compared to the pre-policy era (62.2% vs 53.3%, p<0.001).