Community-based prediction models of cardiovascular events, acute exacerbations and all-cause mortality in individuals with chronic obstructive pulmonary disease: a systematic review and meta-analysis on behalf of the International Cardiovascular and Respiratory Alliance.

Joseph, T.; Raveendra, K.; Haris, M.; Kirupananthan, J.; Aslam, A.; Mircescu, A.; Bhardwaj, A.; Wong, A.; Nadarajah, R.; Price, DB.; Bhutani, M.; Gale, C.

Community-based prediction models of cardiovascular events, acute exacerbations and all-cause mortality in individuals with chronic obstructive pulmonary disease: a systematic review and meta-analysis on behalf of the International Cardiovascular and Respiratory Alliance.

All Authors

Joseph, T.

Raveendra, K.

Haris, M.

Kirupananthan, J.

Aslam, A.

Mircescu, A.

Bhardwaj, A.

Wong, A.

Nadarajah, R.

Price, DB.

Show 2 more

LTHT Author

Joseph, Tobin
Haris, Mohammad
Nadarajah, Ramesh
Gale, Christopher

LTHT Department

Cardio-Respiratory
Cardiology

Publication Date

2026

Item Type

Journal Article
Systematic Review
Meta-Analysis

Subject

PULMONARY DISEASE, CHRONIC OBSTRUCTIVE , HOSPITALISATION , CARDIOVASCULAR DISEASES , DISEASE PROGRESSION , RISK ASSESSMENT

Abstract

INTRODUCTION: Preventable morbidity and mortality from chronic obstructive pulmonary disease (COPD) accrue from major adverse cardiovascular events (MACEs) and acute exacerbations of COPD (AECOPD). The study aims to summarise models for the prediction of these cardiopulmonary events in community-based settings. METHODS: We searched for studies of multivariable models derived, validated or augmented for the prediction of cardiopulmonary events in COPD and used community-based data sources using MEDLINE and Embase from inception through 10 April 2025. Discrimination measures for the model with C-statistic data from >=3 cohorts were pooled by Bayesian meta-analysis, and heterogeneity and risk of bias assessments were undertaken. RESULTS: No models were identified that predicted cardiopulmonary events in COPD using community-based data. Of the 71 models included, 5 predicted cardiovascular events, 32 predicted AECOPD and 30 predicted all-cause mortality. None were eligible for meta-analysis for the prediction of cardiovascular events or AECOPD. For all-cause mortality, age, dyspnoea and airflow obstruction-surprise question (ADO-SQ) (0.763, 95% CI 0.533 to 0.942) and body mass index, airflow obstruction, dyspnoea score and exercise capacity (BODE) (0.753, 95% CI 0.583 to 0.907) demonstrated good prediction performance, while ADO (0.638, 95% CI 0.443 to 0.827) demonstrated adequate prediction performance. The risk of bias was high for 57.9% of studies, and none had clinical utility evaluated. CONCLUSIONS: Despite the high burden of MACE and AECOPD, there is an absence of community-based models that predict this composite outcome. Models to identify individuals with COPD at high risk of cardiopulmonary events could enable targeted clinical intervention. PROSPERO REGISTRATION NUMBER: CRD420251026275.