Predicting outcome after cardiac surgery in patients with cirrhosis: A comparison of Child-Pugh and MELD scores

Abstract

Background & Aims: This study aims to quantify the risk of cardiac surgery in patients with cirrhosis. Methods: Records of all adult patients with cirrhosis undergoing cardiac surgery using cardiopulmonary bypass at the Cleveland Clinic (Cleveland, OH) from January 1992 to June 2002 were analyzed for any relationship of Child-Pugh class and/or score and Model for End-Stage Liver Disease (MELD) score with outcome measures of hepatic decompensation and death during the first 3 months after surgery. Results: Forty-four patients underwent coronary artery bypass grafting (16 patients), valve surgery (16 patients), a combination of the 2 procedures (10 patients), or pericardiectomy (2 patients). Twelve patients (27%) developed hepatic decompensation, and 7 patients (16%) died. Proportions of hepatic decompensation were 3 of 31, 8 of 12, and 1 of 1 patients, and death, 1 of 31, 5 of 12, and 1 of 1 patients in Child-Pugh classes A, B, and C, respectively. The association of hepatic decompensation and mortality with Child-Pugh class, Child-Pugh score, and MELD score was significant ( P < 0.005). Areas under the receiver operating characteristic curves for mortality were similar for Child-Pugh (0.84 ± 0.09) and MELD scores (0.87 ± 0.09). A cutoff Child-Pugh score >7 was found to have a sensitivity and specificity of 86% and 92% for mortality, with a negative predictive value of 97% (95% confidence interval [CI], 83–99) and positive predictive value of 67% (95% CI, 31–91), respectively. However, a similar cutoff value for MELD score could not be established. Conclusions: Child-Pugh score and/or class and MELD score are significantly associated with hepatic decompensation and mortality after cardiac surgery using cardiopulmonary bypass in patients with cirrhosis. Such surgery can be conducted safely in patients with a Child-Pugh score ≤7. Patients with a Child-Pugh score ≥8 have a significant risk for mortality.