Developing and validating the CE-MACE model to predict one-year major adverse cardiovascular events post COPD exacerbation using routine healthcare data
File(s)
Accepted version
Author(s)
Wang, Ye
Ioannides, Anne
Kallis, Constantinos
Quint, Jennifer
Type
Journal Article
Abstract
Background
Patients with chronic obstructive pulmonary disease (COPD) are at an elevated risk of cardiovascular events, particularly following acute exacerbations. Existing prediction models underestimate their risk. We developed and validated the CE-MACE model to predict one-year major adverse cardiovascular event (MACE) risk following a COPD exacerbation.
Methods
Using electronic health records from the UK Clinical Practice Research Datalink (CPRD) Aurum database, we included patients aged ≥40-years-old with moderate or severe COPD exacerbations. The outcome was fatal or non-fatal MACE (acute coronary syndrome, arrhythmia, heart failure, or ischemic stroke). Cause-specific hazard models were used to estimate the coefficient of predictors. Model performance was assessed by Nagelkerke’s R², Harrell’s C statistic, calibration-in-the-large, calibration slope, and decision curve analysis. Internal-external cross validation was used to evaluate model performance across nine geographic regions. A risk score was derived from the original coefficients.
Results
A total of 338,981 patients were included. The overall one-year cumulative incidence rate of MACE following COPD exacerbation was 5.04% (4.96–5.12%). Six predictors were retained: age, exacerbation severity, MACE history, mMRC dyspnoea scale, hypertension, and diabetes. In internal-external cross validation, the pooled estimation for Nagelkerke’s R² was 3.120% (2.945–3.295%), Harrell’s C statistic was 0.752 (0.746–0.759). Calibration-in-the-large was -0.006 (-0.008–-0.004) and calibration slope equal to 0.999 (0.977–1.020). The model suggested clinical utility with higher net benefit than default strategies across the risk thresholds range from 2% to 18%. Model- driven risk stratification demonstrated significantly different cumulative incidence rates among these risk categories. Sensitivity analysis revealed that the CE-MACE model is generalisable to those patients with any history of COPD exacerbation.
Conclusion
The CE-MACE highlights the burden of cardiovascular multimorbidity in patients with COPD. The model could help precisely identify those at high risk in clinical practice and promote integrated multidisciplinary healthcare for these patients.
Patients with chronic obstructive pulmonary disease (COPD) are at an elevated risk of cardiovascular events, particularly following acute exacerbations. Existing prediction models underestimate their risk. We developed and validated the CE-MACE model to predict one-year major adverse cardiovascular event (MACE) risk following a COPD exacerbation.
Methods
Using electronic health records from the UK Clinical Practice Research Datalink (CPRD) Aurum database, we included patients aged ≥40-years-old with moderate or severe COPD exacerbations. The outcome was fatal or non-fatal MACE (acute coronary syndrome, arrhythmia, heart failure, or ischemic stroke). Cause-specific hazard models were used to estimate the coefficient of predictors. Model performance was assessed by Nagelkerke’s R², Harrell’s C statistic, calibration-in-the-large, calibration slope, and decision curve analysis. Internal-external cross validation was used to evaluate model performance across nine geographic regions. A risk score was derived from the original coefficients.
Results
A total of 338,981 patients were included. The overall one-year cumulative incidence rate of MACE following COPD exacerbation was 5.04% (4.96–5.12%). Six predictors were retained: age, exacerbation severity, MACE history, mMRC dyspnoea scale, hypertension, and diabetes. In internal-external cross validation, the pooled estimation for Nagelkerke’s R² was 3.120% (2.945–3.295%), Harrell’s C statistic was 0.752 (0.746–0.759). Calibration-in-the-large was -0.006 (-0.008–-0.004) and calibration slope equal to 0.999 (0.977–1.020). The model suggested clinical utility with higher net benefit than default strategies across the risk thresholds range from 2% to 18%. Model- driven risk stratification demonstrated significantly different cumulative incidence rates among these risk categories. Sensitivity analysis revealed that the CE-MACE model is generalisable to those patients with any history of COPD exacerbation.
Conclusion
The CE-MACE highlights the burden of cardiovascular multimorbidity in patients with COPD. The model could help precisely identify those at high risk in clinical practice and promote integrated multidisciplinary healthcare for these patients.
Date Acceptance
2026-02-17
Citation
European Respiratory Journal
ISSN
0903-1936
Publisher
European Respiratory Society
Journal / Book Title
European Respiratory Journal
Copyright Statement
Copyright © 2026 Copyright Owner. This is the author’s accepted manuscript made available under a CC-BY licence in accordance with Imperial’s Research Publications Open Access policy (www.imperial.ac.uk/oa-policy)
License URL
Publication Status
Accepted