Comparison of five major airflow limitation criteria to identify high-risk individuals with COPD: a contemporary population-based cohort

Background Different airflow limitation criteria are often used to diagnose COPD. We investigated head-to-head whether Global Initiative for Chronic Obstructive Lung Disease (GOLD) (FEV1/FVC <0.70) and four lower limit of normal (LLN) (FEV1/FVC <LLN) criteria to diagnose airflow limitation differ in identifying individuals at risk of COPD exacerbations and mortality. Methods 108 246 individuals aged 20-100 years randomly selected from the general population were followed from 2003 through 2018 to determine risk of COPD exacerbations, respiratory mortality and all-cause mortality. LLN criteria used equations from Global Lung Initiative (GLI), National Health and Nutrition Examination Survey (NHANES), European Community for Steel and Coal (ECSC) and Copenhagen City Heart Study (CCHS)/Copenhagen General Population Study (CGPS). Results Prevalence of airflow limitation was 17% for GOLD, 8.6% for GLI, 10% for NHANES, 8.2% for ECSC and 14% for CCHS/CGPS. During 14.4 years follow-up, we observed 2745 COPD exacerbations, 762 respiratory deaths and 10 338 all-cause deaths. Comparing individuals with versus without airflow limitation, HRs for COPD exacerbations were 17 (95% CI 14 to 20) for GOLD, 21 (18 to 24) for GLI, 20 (17 to 23) for NHANES, 21 (18 to 24) for ECSC and 18 (16 to 21) for CCHS/CGPS. Corresponding HRs for respiratory mortality were 3.7 (3.1 to 4.3), 6.4 (5.6 to 7.5), 5.7 (4.9 to 6.6), 6.2 (5.3 to 7.2) and 4.5 (3.9 to 5.2), and for all-cause mortality 1.5 (1.4 to 1.5), 1.9 (1.8 to 2.0), 1.8 (1.7 to 1.9), 1.9 (1.8 to 2.0) and 1.7 (1.6 to 1.7), respectively. Differences in Harrell's C were minute for these outcomes; nonetheless, Harrell's C was slightly higher for LLN criteria compared with GOLD for mortality outcomes. Conclusions The prevalence of airflow limitation ranged from 8% to 17% using GOLD and four different LLN criteria; however, identified individuals with the five different criteria had similar risk of COPD exacerbations and mortality.