Objective: Prediction of operative risk in adults undergoing cardiac surgery remains a challenge. The European System for Cardiac Operation Risk Evaluation II (EuroSCORE II) and The Society of Thoracic Surgeons score are most commonly used in clinical settings but are not calibrated minimally invasive cardiac surgery interventions (MICS). Alternative risk scoring system is with its simplicity provide by ACEF score (only three variables present, simple additive score counting).
Methods: We sought to test discrimination power and calibration of the above mentioned scores for minimally invasive subset of patients. We have identified patients who underwent cardiac surgery operation with minimally invasive approach from 2007 to 2016 from prospective cardiac surgical database in a single institution. Additional variables were included if necessary for STS score, EuroSCORE II and ACEF score calculation. Discriminatory power was assessed with the C index. Calibration was evaluated with calibration curves and associated statistics.
Results: A total of 2747 patients were identified from main database. There were actual 27 (1.4%) hospitals deaths. The mean STS score predicted mortality were 1.2 ± 1.1 %. Discriminatory power was uniformly good (for STS Mortality: area under curve was 0.87; 95% confidence interval, 0.81 - 0.93). The mean EuroSCORE II predicted mortality were 2.4 ± 3.4 %. Discriminatory power was uniformly good but inferior to that of STS (for EuroSCORE II Mortality: area under curve was 0.84; 95% confidence interval, 0.77 - 0.92). The mean ACEF predicted mortality were 2.4 ± 2.3 %. Discriminatory power was uniformly good but inferior to that of STS and ESII (for ACEF Mortality: area under curve, 0.72; 95% confidence interval, 0.63 - 0.82).
STS score was underestimating mortality(p < 0.01), EuroSCORE II and ACEF were overestimating mortality (respectively, p < 0.01 and p < 0.01, see Figure 1).
Conclusions: The STS score has very good discrimination power for MICS subset of patients. However, it is not calibrated for the same subset. EuroSCORE II has also good discrimination power, but not calibrated for the MICS subset of patients. ACEF score, very simple to calculate, performs with decent discrimination power but not calibrated for the MICS subset of patients. No algorithm seems well calibrated for accurate risk estimation, although STS overall predicted mortality rate was very close real mortality in MICS subset of patients.