European System for Cardiac Operative Risk Evaluation II and Liver Dysfunction

Background: The importance of liver dysfunction in predicting mortality in patients undergoing cardiovascular surgery is an important topic due to the general desire to improve current risk scores such as EUROSCORE II (European System for Cardiac Operative Risk Evaluation), with EUROSCORE III being currently under development. The model for End-Stage Liver Disease (MELD) Score has already proven its utility in predicting outcomes for patients undergoing abdominal, cardiovascular or urological surgery. In the present study, we want to see its usefulness in proving the postoperative mortality in patients undergoing coronary artery bypass surgery. Methods: This was a retrospective study, and it included 185 patients, with 93 survivors being randomly chosen from a total of 589 surviving patients using age, emergency and the weight of cardiac procedures as criteria to match the 92 deceased patients during hospitalization in the postoperative period who underwent coronary artery bypass grafting (CABG) alone or CABG and other concomitant cardiovascular interventions during a 10-year period of time. We calculated for all these patients, at the time of admission, the MELD Score and EUROSCORE II, and we analyzed the predictive performance of the two scores and their constituents. Results: In the multivariable model, patients with a MELD Score ≥ 5.54 had a 2.38-fold increased risk of death (95% C.I.: 1.43-3.96, p = 0.001), while those with a EUROSCORE ≥ 10.37 had a 8.66-fold increased risk of death (95% C.I.: 3.09-24.29, p < 0.001). After combining the two scores, the conditional scenario achieved a high overall accuracy of 84.32% (p < 0.001) in predicting mortality. Conclusions: Patients with a MELD Score ≥ 5.54, had good sensitivity and a very good specificity in terms of mortality prediction, but the conditional scenario, leveraging both risk scores, i.e., the MELD Score and EUROSCORE, offers the highest utility in terms of enhancing mortality prediction regarding these patients.