Abstract 19685: Liver Dysfunction as Reflected by MELD (the Model For End Stage Liver Disease) Score Could Better Predict Postoperative Outcome in Patients Undergoing Cardiac Surgery than Child-Turcotte-Pugh Classification

Abstract

Background: Liver dysfunction increases postoperative morbidity and mortality following cardiac surgery. Child-Turcotte-Pugh (CTP) classification had been used in the evaluation of liver function. More recently, the composite model for end stage liver disease (MELD) score has become recognized as a more robust and objective scoring system than CTP classification. However, the predictive power of CTP and MELD scores for postoperative outcome in patients undergoing cardiac surgery has not yet been fully elucidated. Methods and Results: Postoperative outcome and preoperative liver function estimated by the CTP classification or the MELD scores were investigated in 1856 consecutive patients (69.3% male, age 66. 8±12.2 years) undergoing cardiac surgery between 2010 and 2013 at our institution. The proportions of patients in CTP class A, B and C were 91.8% (n=1704), 8.1% (n=151) and 0.05% (n=1); and mean MELD scores in patients in each CTP class were 9.4, 15.5 and 36.7, respectively. A MELD cutoff value of 12 can predict postoperative mortality with a sensitivity and specificity of 78.7 and 63.9%, respectively, with an area under the curve (AUC) from the ROC analysis of 0.75248 (p<0.0001). Child class B/C predicts morality with a sensitivity and specificity of 23.7 and 92.2%. High MELD score (>12) was associated with longer intubation time (33.6±5.1 vs. 12.2±2.7 hours, p<0.0001), longer ICU (4.8±0.3 vs. 2.2±0.2 days, p<0.0001) and total-hospital stay (34.5±1.0 vs. 22.8±0.6 days, p<0.0001) and more incidence of renal replacement therapy requirement (46.0 vs. 4.4%, p<0.0001) than those with low MELD scores (<12). The predictive strength of the MELD, MELD component variables and CTP class (B/C vs. A) were quantitatively compared based on AUC by using the DeLong test, which revealed that MELD score was the strongest predictor for morality (pairwise comparison, all p<0.05). Conclusions: Assessment of liver dysfunction using the MELD scoring system is useful for predicting postoperative morality and morbidity following cardiac surgery. MELD scoring system was a stronger predictor for morality than CTP classification.