Abstract No. 279 Performance of The new MELD 3.0 Score Compared to MELD, MELD-Na, and Child-Pugh Scores for Predicting Mortality After TIPS in Cirrhosis

To compare the prognostic abilities of the new Model for End-Stage Liver Disease 3.0 score (MELD 3.0), relative to MELD, MELD-Na and the Child-Pugh scoring system in patients with cirrhosis who underwent transjugular intrahepatic portosystemic shunt (TIPS) creation to enhance our understanding of patient selection for TIPS. A total of 358 patients (aged 25-80 years; mean, 55.0 years) with cirrhosis who underwent TIPS creation at a single institution from September 2005 to May 2022 were retrospectively reviewed. Inclusion criteria were cirrhosis and portal hypertensive complications, available lab work within 45 days of TIPS, and available follow up data for mortality. Patients without cirrhosis were excluded. The MELD, MELD-Na, MELD 3.0, and Child–Pugh scores were evaluated as predictors of 30-day, 3-month, and 1-year mortality after TIPS using receiver operating curves (ROC) to examine the area under the curve (AUC). Of the four grading systems, MELD 3.0 demonstrated the greatest sensitivity and specificity for predicting survival at all three time points: 30-day (cutoff ≥19; AUC 0.77, 95% confidence interval [CI] 0.682-0.857), 3-month (cutoff ≥18; AUC 0.755, 95% CI 0.683-0.827), and 1-year (cutoff ≥18; AUC 0.716, 95% CI 0.651-0.781). These results were significantly superior to those of the MELD (cutoffs for 30-day, 3-months, and 1-year mortality: ≥16, ≥16, and ≥15 respectively; AUCs: 0.685, P=0.018; 0.687, P=0.018; and 0.645, P=0.003), MELD-Na (cutoffs: ≥19, ≥16, and ≥19 respectively; AUCs: 0.689, P=0.013; 0.688, P=0.008; and 0.658, P=0.006), and Child-Pugh scores (AUCs for 30-day and 3-months mortality: 0.644, P=0.006 and 0.644, P=0.02 respectively). The new MELD 3.0 score demonstrated greater sensitivity and specificity for predicting mortality after TIPS compared to MELD, MELD-Na, and Child-Pugh scoring systems, representing a more clinically relevant scoring system for patient selection.