Abstract
This meta-analysis of observational studies aimed at estimating the overall prevalence of overdiagnosis and overtreatment in subjects with a clinical diagnosis of Chronic Obstructive Pulmonary Disease (COPD). MedLine, Scopus, Embase and Cochrane databases were searched, and random-effect meta-analyses of proportions were stratified by spirometry criteria (Global Initiative for COPD (GOLD) or Lower Limit of Normal (LLN)), and setting (hospital or primary care). Forty-two studies were included. Combining the data from 39 datasets, including a total of 23,765 subjects, the pooled prevalence of COPD overdiagnosis, according to the GOLD definition, was 42.0% (95% Confidence Interval (CI): 37.3-46.8%). The pooled prevalence according to the LLN definition was 48.2% (40.6-55.9%). The overdiagnosis rate was higher in primary care than in hospital settings. Fourteen studies, including a total of 8183 individuals, were included in the meta-analysis estimating the prevalence of COPD overtreatment. The pooled rates of overtreatment according to GOLD and LLN definitions were 57.1% (40.9-72.6%) and 36.3% (17.8-57.2%), respectively. When spirometry is not used, a large proportion of patients are erroneously diagnosed with COPD. Approximately half of them are also incorrectly treated, with potential adverse effects and a massive inefficiency of resources allocation. Strategies to increase the compliance to current guidelines on COPD diagnosis are urgently needed.