Y. Juo1, N. Satou1, R. Shemin1, P. Benharash1 1University Of California – Los Angeles,Cardiac Surgery,Los Angeles, CA, USA
Introduction:
The Society of Thoracic Surgeons Predicted Risk of Mortality (STS PROM) Score is developed specifically for calculation of short-term postoperative mortality and morbidity risks. It is a composite score from 35 variables encompassing demographics, baseline comorbidities, severity/complexity of primary cardiac disease, and anatomic/technical factors for specific procedures. The calculation of PROM involves a special online calculator that is beyond the immediate reach of most surgeons in a clinical setting. Here we seek to evaluate the efficacy of simpler prediction scores such as CHADS2 and CHA2DS2-VASC in comparison with PROM score.
Methods:
All adult patients undergoing cardiac surgery at a single institution from 2008 to 2015 were identified using the institutional Society of Thoracic Surgeons database. Patients receiving transplant, ECMO and ventricular assist devices were excluded. An area under the curve (AUC) analysis was performed to evaluate the discriminatory abilities of STS PROM, CHADS2, and CHA2DS2-VASC scores in predicting postoperative mortality, complications, major adverse cardiac events (MACE), acute renal failure, and prolonged length of stay.
Results:
4,232 patients underwent major cardiac surgery during the study period. Most common procedures were valve replacement (n=1,356, 32.04%) and coronary artery bypass grafting (n=875, 20.67%). 30-day mortality occurred in 118 (2.79%) patients while postoperative morbidity occurred in 1,686 (39.84%) patients, including 678 (16.02%) MACE, 145 (3.43%) acute renal failures, and 482 (11.39%) prolonged length of stays. The only outcome in which STS PROM Score demonstrated a larger AUC than either CHA2DS2-VASC or CHADS2 was postoperative mortality (0.82 vs 0.65 vs 0.57, p=0.0489). There was no significant AUC difference when comparing STS PROM Score against either CHA2DS2-VASC or CHADS2 when used in prediction of postoperative complications (0.60 vs 0.57 vs 0.57, p=0.48), MACE (0.55 vs 0.53 vs 0.54, p=0.60), acute renal failure (0.74 vs 0.66 vs 0.67, p=0.37), or prolonged length of stay (0.65 vs 0.67 vs 0.64, p=0.72).
Conclusion:
STS PROM score was superior to both CHADS2 and CHA2DS2-VASC scores in prediction of postoperative mortality. There was no significant difference in the predictive value between the three prediction scores for other postoperative morbidities Furthermore, none of the scoring systems provided satisfactory discriminatory power for the prediction of complications in view of most AUC values approximating 0.5. Further investigation is required for a more parsimonious and effective morbidity prediction score.
