Is it possible to improve the accuracy of EuroSCORE?

OBJECTIVE: We derived a new risk-scoring method by modifying some of the risk factors included in the EuroSCORE algorithm. METHODS: This study includes 3613 patients who underwent cardiac surgery at the Vaasa Central Hospital, Finland. The EuroSCORE variables, along with modified age classes ( or = 80 years), eGFR-based chronic kidney disease classes (classes 1-2, class 3 and classes 4-5) and the number of cardiac procedures, were entered into the regression analysis. RESULTS: An additive risk score was calculated according to the results of logistic regression by adding the risk of the following variables: patients' age classes (0, 2, 4 and 6 points), female (2 points), pulmonary disease (3 points), extracardiac arteriopathy (2 points), neurological dysfunction (4 points), redo surgery (3 points), critical preoperative status (8 points), left ventricular ejection fraction (> 50%: 0; 30-50%: 2 and < 30%: 3 points), thoracic aortic surgery (8 points), postinfarct septal rupture (9 points), chronic kidney disease classes (0, 3 and 6 points), number of procedures (1: 0; 2: 2 and 3 or more: 7 points). The modified score had a better area under the receiver operating characteristic curve (additive: 0.867; logistic: 0.873) than the EuroSCORE (additive: 0.835; logistic: 0.840) in predicting 30-day postoperative mortality. The modified score, but not EuroSCORE, correctly estimated the 30-day postoperative mortality. CONCLUSION: EuroSCORE still performs well in identifying high-risk patients, but significantly overestimates the immediate postoperative mortality. This study shows that the score's accuracy and clinical relevance can be significantly improved by modifying a few of its variables. This institutionally derived risk-scoring method represents a modification and simplification of the EuroSCORE and, likely, it would provide a more realistic estimation of the mortality risk after adult cardiac surgery.