Performance of contemporary surgical risk scores for mitral valve surgery

BACKGROUND: Risk stratification for mitral valve repair or replacement (MVR) is important in the decision-making for treating several mitral valve disease but is rarely studied. We compared the prognostic utility of EuroSCORE, EuroSCORE II, and Society of Thoracic Surgeons (STS) Score for MVR. METHODS: The three scores were retrospectively calculated for consecutive patients undergoing isolated MVR at Auckland City Hospital during 2005-2012 and their discrimination and calibration for mortality and morbidities assessed. RESULTS: There were 408 patients (mitral valve repair 48.1% and replacement 51.9%) followed-up for 6.0 ± 2.6 years. The operative mortality was 2.5%. Mean EuroSCORE, EuroSCORE II, and STS Score were 7.6%, 3.4%, and 3.5%. C-statistics were 0.844, 0.817, and 0.850 for operative mortality. Hosmer-Lemeshow test p values were 0.076, 0.541, and 0.306, and Brier scores 0.0246, 0.0035, and 0.0075, respectively, for operative mortality. The numerically highest c-statistic for predicting complications include EuroSCORE for return to the operating room (c = 0.673); EuroSCORE II for stroke (c = 0.669) and mediastinitis (c = 0.801); and STS for renal failure (c = 0.828), ventilation >24 hours (c = 0.789), and composite morbidity (c = 0.732). The individual STS complication models for MVR had a numerically higher c-statistic only for stroke (c = 0.737). CONCLUSIONS: All scores discriminated mortality and most morbidities after MVR, although EuroSCORE over-estimated operative mortality. The STS Score was the best overall predictor of mortality and morbidity in the MVR cohort.