The Asian system for cardiac operative risk evaluation for predicting mortality after isolated coronary artery bypass graft surgery (ASCORE-C).
2020
OBJECTIVE The mortality of coronary artery bypass graft surgery (CABG) in Asian patients predicted by Western population-based risk models is not ideal. We aimed to develop a risk scoring system based on a multiracial Asian patient population to predict early-stage mortality. METHODS A total of 2495 patients who underwent isolated CABG in Singapore from 2009 to 2015 were included to develop a regional scoring system: the regional Asian System for Cardiac Operative Risk Evaluation in CABG (ASCORE-C). Predictors were identified via a stepwise multiple logistic regression to construct models for 30-day, 90-day, and 1-year cumulative mortality prediction. The optimal cut-offs of ASCORE-C risk strata were determined by the multiway χ2 and Interaction Detection decision tree. The performance of ASCORE-C was compared with European System for Cardiac Operative Risk Evaluation score (EuroSCORE) II and Society of Thoracic Surgeons (STS) score by observed/estimated mortality ratio. RESULTS The observed postoperative 30-day, 90-day, and 1-year cumulative mortality following isolated CABG were 2.6%, 3.3%, and 4.6%, respectively. The ASCORE-C to predict these three events were constructed with age (>65 years), Malay ethnicity, congestive heart failure, abnormal heart rhythm, aortic atherosclerosis, estimated glomerular filtration rate, peripheral vascular disease, critical preoperative status, and emergency surgery. Tested by three different validation datasets, receiver operating characteristic curve of ASCORE-C model prediction performed well with the Hosmer-Lemeshow test. The ASCORE-C was found to have better predictive accuracy than EuroSCORE II and STS score in CABG mortality prediction, especially for the high-risk patients. CONCLUSION The ASCORE-C system is reliable for early-stage CABG mortality prediction in a multiracial Asian population.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
30
References
0
Citations
NaN
KQI