INTRODUCTION: MELD is a satisfactory predictor of 3-month mortality in patients with cirrhosis. However, sub-populations of patients at increased hazard of mortality, but not well represented by MELD do exist. For instance, exception scores for hepatocellular carcinoma (HCC) were created to more adequately represent risk. The management of patients with a low MELD despite a significant complication burden remains a challenge. Therefore, the study aims to identify patient characteristics that augment risk particularly in the low MELD stratum.

METHODS: We performed a retrospective, single-center study of all patients diagnosed with cirrhosis (ICD-9) between January 2009 and November 2014, irrespective of candidacy for liver transplantation. The dataset was extracted from a transplant datamart derived from an enterprise data warehouse comprised of electronic health record and social security death master file sources. Patient survival stratified by MELD and MELD velocity (defined as change in MELD over units of time) were analyzed using Kaplan-Meier and log-rank methodology. Competing risk models using Cox proportional hazards and boosted regressions were utilized to examine predictors of patient survival. Interaction among risk factors were examined using stratified analysis and the construction of interaction terms within the regression models.

RESULTS: The dataset includes 6779 patients with cirrhosis, 51% were female, 57% are white, 848 (13%) were listed for LT, 361 (5%) received a LT or SLK. 705 cirrhotic patients died. The strongest predictor of mortality was MELD velocity (OR 50; 6.95-369), compared to MELD on first presentation (OR 1.07, 1- 1.15), episodes of bacteremia (OR 12.9; 3.55-47.3), age (OR 1.09, 1.01-1.16) and portal hypertension (OR 6.7; 1.39-31.7). Findings were consistent in various sub-populations of cirrhotics (listed, not listed, low MELD scores).

CONCLUSION: MELD velocity is the strongest early predictor of mortality in patients with cirrhosis. Targeted interventions such as living donor liver transplant or deceased donor liver transplant with a high DRI graft aimed at patients with dynamic changes in MELD despite relatively low absolute scores might mitigate their higher hazard of mortality. Future modification of liver allocation policy may be necessary to address this vulnerable patient population.

« Back to 2015 American Transplant Congress