Concavity of the maximal expiratory flow–volume curve, and incidence of COPD and respiratory symptoms: a population-based cohort study

Background Concavity of the maximal expiratory flow–volume curve is widely regarded as an early indicator of obstructive airways disease. However, its discriminatory accuracy for respiratory outcomes has remained poorly defined. We aimed to examine the discriminatory accuracy of concavity of the maximal expiratory flow–volume curve versus post-bronchodilator (BD) forced expiratory volume in 1 s/forced vital capacity (FEV 1 /FVC) as predictors of incidence of COPD and respiratory symptoms. Methods Spirometry was performed on a subset of the Tasmanian Longitudinal Health Study cohort at age 45 years, and incidence of COPD and respiratory symptoms were prospectively monitored over the next 8 years (n=852). Central and peripheral concavity were assessed using a published algorithm based on post-BD forced expiratory flow at 50% of the FVC (FEF 50% ) and FEF 75% , respectively. Optimal thresholds were determined using the unweighted Youden Index for COPD incidence. Results Among participants without COPD at age 45 years, central and peripheral concavity were greater in those who developed COPD by age 53 years than in those who did not (mean difference: +20%, 95% CI 12–28, and +15%, 95% CI 7–23, respectively). Central concavity above the optimal threshold (27%) had a sensitivity of 70% and specificity of 79% for COPD incidence, while peripheral concavity above the optimal threshold (47%) had a sensitivity of 79% and specificity of 50%. Excess central and peripheral concavity were associated with an increased odds of developing wheeze and exertional dyspnoea over the 8-year follow-up. Post-BD FEV 1 /FVC was more sensitive and specific for COPD incidence than the concavity indices, but was not associated with incident respiratory symptoms. Conclusion Concavity indices were more useful for assessing future risk of respiratory symptoms but had lower discriminatory accuracy for COPD incidence compared to post-BD FEV 1 /VC.