BACKGROUND : Prediction of future lung function will enable the identification of individuals at high risk of developing chronic obstructive pulmonary disease (COPD), but the trajectory of lung function decline greatly varies among individuals. We developed and validated an individualized prediction model of lung function trajectory and risk of airflow limitation in the general population.
METHODS : Data were obtained from the Framingham Offspring Cohort, which included 4,167 participants ≥20 years of age and had ≥2 valid spirometry assessments. The primary outcome was pre-bronchodilator forced expiratory volume at 1 second (FEV1); the secondary outcome was the risk of airflow limitation (defined as FEV1/forced vital capacity < Lower Limit of Normal). We developed mixed-effects regression models for individualized prediction and employed a machine-learning algorithm to determine essential predictors. The model was validated in two large, independent multi-center cohorts (N=2,075 and 12,913, respectively).
RESULTS : With 20 common predictors, the model explained 79% of variation in FEV1 decline in the derivation cohort. In two validation datasets, the model had low error in predicting FEV1 decline (root mean square error range: 0.18-0.22L) and high discriminative power in predicting risk of airflow limitation (C statistic range: 0.86-0.87). We implemented this model in a freely accessible Web-based application, which allows prediction based on flexible sets of predictors (http://resp.core.ubc.ca/ipress/FraminghamFEV1).
CONCLUSIONS : The individualized predictor is an accurate tool to predict long-term lung function trajectories and risk of airflow limitation in the general population. This model enables identifying individuals at higher risk of COPD, whom can be targeted for preventive therapies.
Chen Wenjia, Sin Don D, FitzGerald J Mark, Safari Abdollah, Adibi Amin, Sadatsafavi Mohsen
COPD, FEV(1), FEV(1)/FVC, airflow limitation, lung function, predictive modelling