Background: This study evaluates cardiopulmonary bypass (CPB) as a predictor of 30-day postoperative mortality and modifies Parsonnet and Euro SCORE models accordingly to develop a new model. Methods: Information of 1920 consecutive patients who underwent elective and emergent surgery in our center was collected. Parsonnet and Euro SCORE model parameters in addition to 81 variables including perioperative information gathered. Following statistical analysis by R software a new model considering CPB under the name of Iranian model was designed. Parsonnet and Euro SCORE models were recalibrated and CPB variable was entered. Data validation was performed in 40 consecutive patients. Results: P value of our five predictor models including Iranian, Parsonnet (P) and modified Parsonnet (MP), Euro SCORE (ES) and modified Euro SCORE (MES) models were < 2e-16. Iranian model has a lower overestimation of mortality (0.4375) and its area under curve (AUC) was higher (0.9537). AUCof P,MP, ES andMESmodels were 0.9551, 0.9841, 0.8659 and 0.9465 respectively. Overestimation of early post operative mortality of P, MP, ES and MES models were 0.6483, 0.5271, 0.6267 and 0.5056 respectively. Conclusions: This study confirmed that CPB as a variable is a predictor of mortality and is applicable in risk stratification models. CPB increases AUC and decreases Overestimation of mortality. Iranian model as the first CPB dependent mortality prediction model has more accurate mortality estimation in respect to other models.

Keywords