Assessment of risk prediction using edmonton frail scale and european system for cardiac operative risk evaluation II among older patients undergoing coronary artery bypass graft surgery in a tertiary care hospital in India

Background: The risk assessment for outcomes of older people undergoing cardiac surgery employ scales such as the Euro-Score II, and STS (Society for Thoracic Surgeons), which use clinical and laboratory data. Some studies have suggested a lower accuracy in older patients. Frailty assessment, using functional parameters, has shown promise in this age group. The aim of this study is to compare the validity of risk prediction of Euro-score II, with the Edmonton Frail Scale (EFS), in older patients undergoing elective coronary artery bypass grafting (CABG). Methods: This was a prospective, observational study of a cohort of patients above 60 years scheduled for elective CABG in a single centre. The patients were graded on the Euro-Score II scale and the EFS scales. The primary outcome of 30th day mortality, and the secondary outcome of immediate post-operative complications during hospitalization were recorded. Results: A total of 487 patients were recruited. The mean age was 68 years. Male subjects comprised 81.1% of the study group. Classification of risk as per the EFS placed 76.3 % as low risk, 23.4% as intermediate, and none were considered to be high risk. The EuroScore II classification placed 86% in the intermediate and high risk groups. The AUC in the ROC (receiver operator curve) for the EFS was 0.793 and for the and EuroScoreII it was 0.752. The 30th day mortality threshold fit occurred at 5/6 score for both EuroScore II and EFS. Euroscore- II sensitivity/specificity was 66.7%/75.1% respectively. The EFS had a sensitivity of 66.7% and a specificity of 77.1%. The ROC curves for the secondary outcomes were not significant. Conclusion: Both scales are of modest value in predicting short-term mortality in older patients, and require further refinements for improving clinical decision-making in the individual patient.