Development of Isolated Coronary Artery Bypass Grafting (CABG) Mortality Risk Stratification Model: An Initiative Model for South Asian Countries and its Comparison with International Risk Stratification Models

Abstract Background: Risk models of isolated coronary artery bypass grafting (CABG) have been reported from several series, in Western countries. The Society of Thoracic Surgeons (STS) National Adult Cardiac Database (NCD) and European System for Cardiac Operative Risk Evaluation (EuroSCORE) have contributed much to this field. To develop mortality Risk stratification models of isolated coronary artery bypass grafting (CABG) using a structured prospective database in order to assess and improve the quality of cardiac surgical care. Data was extracted from the Tabba Heart Institute (THI) Cardiac Surgery registry database, to develop a South Asian risk stratification model for expected mortality of isolated CABG procedure, and compared it with international risk stratification models. Methods: The study population consisted of 5324 adult patients who have undergone isolated CABG procedure between January 1, 2011, and June 30, 2016 at Tabba Heart Institute (THI). The structured database was then divided into developmental and validation subsets. Model developmental dataset consisted of 4793 patients and was used for the construction of the risk model whereas the validation dataset consisted of 531 patients and was used for testing and validation of the risk model. A number of factors associated with mortality suggested by the panel of cardiac surgeons and biostatisticians were taken into account in logistic regression analysis. The outcome of the risk model was mortality. Model discrimination ability was tested using the area under the receiver operating characteristic curve (C index). Model calibration was tested by the Hosmer-Lemeshow test. Results: Of 4793 subjects, the mean age was 57.7±8.8 years, with 17.5% females, 54.2% had diabetes mellitus, and the mean preoperative creatinine was 1.09±0.45 (mg/dl), whereas for New York Heart Association class II, III and IV were 21.1%, 23.2% and 2.7% respectively, with 31.5% urgent and 2.6% emergent and salvage. The observed mortality rate was 2.4% in the model development dataset. Logistic regression analysis revealed that the influential risk factor that were found to statistically significant were Age (OR 1.054; CI,= 1.017-1.093; p