Optimal spirometry thresholds for the prediction of chronic airflow obstruction: a multinational longitudinal study

Abstract

IntroductionChronic airflow obstruction is key for COPD diagnosis but strategies for its early detection are limited. We aimed to define the optimal z-score thresholds for spirometry parameters to discriminate chronic airflow obstruction incidence.MethodsThe Burden of Obstructive Lung Disease study is a multinational cohort study. Information on respiratory symptoms was collected and pre- and post-bronchodilator spirometry was performed at baseline. Eighteen study sites were followed up with repeat measurements after a median of 8.4 years. We converted lung function measurements into z-scores using the Third National Health and Nutrition Survey reference equations. We used the Youden index to calculate the optimal z-score thresholds for discriminating chronic airflow obstruction incidence. We further examined differences by smoking status.ResultsWe analysed data from 3057 adults (57% females, mean age: 51 years at baseline). Spirometry parameters were good at discriminating chronic airflow obstruction incidence (AUC 0.80–0.84), while respiratory symptoms performed poorly. The optimal z-score threshold was identified for pre-bronchodilator FEV1/FVC <-1.336, equivalent to the 9th percentile (sensitivity: 78%, specificity: 72%). All z-score thresholds associated with a lower post-bronchodilator FEV1/FVC and greater odds of chronic airflow obstruction at follow-up. The risk of chronic airflow obstruction was slightly greater for current smokers and, to some extent, never smokers with a pre-bronchodilator FEV1/FVC less than the 9th/10th percentiles at baseline, particularly among males.ConclusionsSpirometry is better than respiratory symptoms at predicting chronic airflow obstruction incidence. A pre-bronchodilator FEV1/FVC <9th/10th percentiles, particularly among current smokers, could suggest early airflow obstruction or pre-COPD.