scholarly journals Toward Dynamic Risk Prediction of Outcomes After Coronary Artery Bypass Graft: Improving Risk Prediction With Intraoperative Events Using Gradient Boosting

Author(s):  
Makoto Mori ◽  
Thomas J.S. Durant ◽  
Chenxi Huang ◽  
Bobak J. Mortazavi ◽  
Andreas Coppi ◽  
...  

Background: Intraoperative data may improve models predicting postoperative events. We evaluated the effect of incorporating intraoperative variables to the existing preoperative model on the predictive performance of the model for coronary artery bypass graft. Methods: We analyzed 378 572 isolated coronary artery bypass graft cases performed across 1083 centers, using the national Society of Thoracic Surgeons Adult Cardiac Surgery Database between 2014 and 2016. Outcomes were operative mortality, 5 postoperative complications, and composite representation of all events. We fitted models by logistic regression or extreme gradient boosting (XGBoost). For each modeling approach, we used preoperative only, intraoperative only, or pre+intraoperative variables. We developed 84 models with unique combinations of the 3 variable sets, 2 variable selection methods, 2 modeling approaches, and 7 outcomes. Each model was tested in 20 iterations of 70:30 stratified random splitting into development/testing samples. Model performances were evaluated on the testing dataset using the C statistic, area under the precision-recall curve, and calibration metrics, including the Brier score. Results: The mean patient age was 65.3 years, and 24.7% were women. Operative mortality, excluding intraoperative death, occurred in 1.9%. In all outcomes, models that considered pre+intraoperative variables demonstrated significantly improved Brier score and area under the precision-recall curve compared with models considering pre or intraoperative variables alone. XGBoost without external variable selection had the best C statistics, Brier score, and area under the precision-recall curve values in 4 of the 7 outcomes (mortality, renal failure, prolonged ventilation, and composite) compared with logistic regression models with or without variable selection. Based on the calibration plots, risk restratification for mortality showed that the logistic regression model underestimated the risk in 11 114 patients (9.8%) and overestimated in 12 005 patients (10.6%). In contrast, the XGBoost model underestimated the risk in 7218 patients (6.4%) and overestimated in 0 patients (0%). Conclusions: In isolated coronary artery bypass graft, adding intraoperative variables to preoperative variables resulted in improved predictions of all 7 outcomes. Risk models based on XGBoost may provide a better prediction of adverse events to guide clinical care.

Circulation ◽  
2014 ◽  
Vol 130 (suppl_2) ◽  
Author(s):  
Li Sun ◽  
Shu Ding ◽  
Xinwei Feng ◽  
Fangqin Wu ◽  
Pixiong Su ◽  
...  

Introduction: Postoperative delirium (POD) is a common neuropsychiatric complication after coronary artery bypass graft (CABG) surgery and contributes to the morbidity and mortality. Although the pathogenesis of POD is largely unknown, surgery induced systemic inflammation might play a role. We examined the predictive values of serum interieukin-6 (sIL-6) level in predicting the occurrence of POD. Methods: Ninety nine patients undergoing the first time elective isolated CABG were consecutively recruited. All enrolled patients were assessed for POD twice a day for 5 days using the Confusion Assessment Method for Intensive Care Unit, and divided into either POD or non-POD groups. IL-6 levels were measured by ELISA (R&D Human IL-6 Quantikine) before operation (T0) and 6 (T6), 12 (T12) and 18 (T18) hours after operation. One-way repeated ANOVA was used to test the differences in trends and patterns of sIL-6 changes over time between the two groups. Logistic regression and ROC were performed to determine the predictive values of IL-6. Results: POD developed in 42.4% (42 of 99) patients. Among them, 90.4% POD occurred within the first 2 days after CABG. Patients with POD were older and more likely to be female. The sIL-6 levels were significantly higher in POD group than those in non-POD group with the highest levels at T6 and T12 (Figure 1). One-way ANOVA for repeated measures revealed interaction effects between time and group (with p values of <0.001 and 0.019). Logistic regression showed that sIL-6 level ≥491.37 pg/mL at T18 was a strong predictor of POD (OR=4.706, 95%CI 1.998-11.082, p<0.001) even after adjusting for age and female gender (OR=10.220, 95%CI 2.465-42.365, p=0.001), with a sensitivity and specificity of 66.7% and 68.4%, respectively. Conclusions: Higher sIL-6 level is an independent predictor of POD, and the sIL-6 level above 491.37pg/mL at 18 hours after CABG has a higher specificity in predicting the occurrence of POD.


2019 ◽  
Vol 8 ◽  
pp. 204800401986212 ◽  
Author(s):  
Redoy Ranjan ◽  
Dipannita Adhikary ◽  
Sabita Mandal ◽  
Sanjoy Kumar Saha ◽  
Kamrul Hasan ◽  
...  

Introduction European System for Cardiac Operative Risk Evaluation (EuroSCORE) was developed to identify patients who may have a greater postoperative risk for adverse effects following adult cardiac surgery. This study evaluated the discriminatory potential of using the EuroSCORE system in predicting the early, as well as late, postoperative outcomes following coronary artery bypass graft surgery in Bangladesh. Methods A total of 865 patients who underwent isolated coronary artery bypass graft surgery were evaluated with the EuroSCORE risk scoring system. Moreover, we also compared the discriminatory potentials between the EuroSCORE II and the original logistic EuroSCORE. Results Operative mortality was best predicted by EuroSCORE II (area under the curve (AUC) 0.863, Brier score 0.030) compared to the original logistic EuroSCORE (AUC 0.849, Brier score 0.033). However, the overall expected-to-observed mortality ratio for EuroSCORE II was 1.1, whereas the observed ratio for the original logistic EuroSCORE was 1.7. EuroSCORE II was predictive of an intensive care unit stay of five days or more (AUC 0.786), prolonged inotropes use (AUC 0.746), stroke (AUC 0.646), de novo dialysis (AUC 0.810), and low output syndrome (AUC 0.715). Moreover, a high EuroSCORE II quintile significantly predicted the risk for late mortality (p < 0.0001). Conclusions EuroSCORE has an important role in predicting the early, as well as late, postoperative outcomes following coronary artery bypass surgery. However, the performance of EuroSCORE II is significantly better than the original logistic EuroSCORE in predicting postoperative morbidity and mortality after isolated coronary artery bypass graft surgery among Bangladeshi patients.


2006 ◽  
Vol 55 (5) ◽  
pp. 451
Author(s):  
Seung Ho Joo ◽  
Byoung Wook Choi ◽  
Jae Seung Seo ◽  
Young Jin Kim ◽  
Tae Hoon Kim ◽  
...  

2013 ◽  
Vol 2 (6) ◽  
Author(s):  
M. Hadadzade ◽  
S. Forouzania ◽  
S. Mirhoseini ◽  
H. Peighambari ◽  
N. Naserzade ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document