Including body composition in MELD scores improves mortality prediction among patients awaiting liver transplantation.

ABSTRACT

BACKGROUND & AIMS: The Model for End-stage Liver Diseases (MELD) is widely accepted for prioritizing candidates awaiting liver transplantation (LT). However, MELD scores do not reflect the severity of the nutritional or functional status of patients with cirrhosis. METHODS: This retrospective study analyzed data from 173 patients who were waitlisted for LT at our institution between April 2006 and December 2016. By including skeletal muscle mass, muscle quality and visceral adiposity evaluated using plain computed tomography imaging in MELD scores, we developed body composition-MELD (BC-MELD), and investigated its impact on the prediction of mortality among patients awaiting LT. RESULTS: The equation generated using Cox regression analysis was as follows BC-MELD = MELD score + 3.59 × low SMI + 5.42 × high IMAC + 2.06 × high VSR. (IMAC, intramuscular adipose tissue content; SMI, skeletal muscle mass index; VSR, visceral-to-subcutaneous adipose tissue area ratio). The median BC-MELD score was 17.4 and the area under the receiver operating characteristic curve (AUC) revealed a cut-off BC-MELD score of 21.4 (AUC = 0.835, P < 0.001, sensitivity 87.5%, specificity 70.7%). Waitlist mortality in patients with high BC-MELD was significantly higher in all tested cohorts (P < 0.001) and among patients with lower conventional MELD scores (<15) (P < 0.001). The discriminatory power was significantly better for BC-MELD than MELD scores (AUC; 0.835 vs. 0.732, P = 0.001 for 3-month, AUC; 0.765 vs. 0.671, P = 0.002 for 6-month, AUC; 0.716 vs. 0.615, P < 0.001 for 12-month, AUC; 0.636 vs. 0.584, P = 0.014 for overall mortality). CONCLUSIONS: BC-MELD is the first to include not only muscularity but also visceral adiposity. It predicted waitlist mortality more accurately than the conventional MELD score. A new allocation system based on BC-MELD might lead to better outcomes for patients with cirrhosis awaiting LT.