Statistics from Altmetric.com
Chronic obstructive pulmonary disease (COPD) is an important disease from a public health perspective, with a number of preventable occupational, environmental and personal risk factors. The Global Initiative for Obstructive Lung Disease (GOLD) was implemented to raise awareness of COPD and to improve the prevention and treatment of this lung disease.1 A concern has been raised regarding use of the criterion “forced expiratory volume in 1 s/forced vital capacity (FEV1/FVC) <0.70” in the definition of GOLD stage I which may lead to overdiagnosis of COPD, particularly in older individuals.2–5
To address this controversy, Mannino and colleagues assessed COPD-related hospitalisations and mortality among 5201 individuals aged 65 years and older who had participated in the Cardiovascular Health Study.6 The authors concluded that study participants they termed “potentially overdiagnosed” (those with FEV1/FVC <0.70 who also had an FEV1/FVC ratio above or equal to the lower limit of normal (LLN)) were more likely to die and to have COPD-related hospitalisations during the 11 year follow-up period than “normals” (those with FEV1/FVC ⩾0.70 and FVC >80% predicted who were asymptomatic). Based on this finding, the authors suggested that using the LLN criterion may miss some older individuals who could benefit from intervention.
This conclusion was based on the analysis where all “potentially overdiagnosed” participants classified as GOLD stages I–IV were grouped together and compared with the “normals”, with a resulting mortality hazard ratio (HR) of 1.3 (95% CI 1.1 to 1.5). However, as noted above, the controversy focuses on individuals classified in GOLD stage I. Persons with FEV1/FVC <0.70 and FEV1/FVC ⩾LLN within GOLD stage I represent the majority (72%) of the “potentially overdiagnosed” individuals2 who would theoretically benefit from further clinical evaluation and treatment through adherence to the GOLD criteria. The adjusted mortality HR for this subgroup was 1.1 (95% CI 0.96 to 1.3) compared with the “normal” group.6 The magnitude of the HR is quite small and the confidence interval includes 1.0. This contrasts with the significantly raised adjusted mortality HR of 1.4 (95% CI 1.1 to 1.7) for the GOLD stage I subgroup with FEV1/FVC <LLN. The study results therefore do not substantiate a significant increase in all-cause mortality among the GOLD stage I subgroup that is “potentially overdiagnosed”.
The authors report a small but statistically significant increase in the risk of hospitalisation with mention of COPD in the GOLD stage I subgroup of “potentially overdiagnosed” individuals compared with the “normal” group. However, individuals with respiratory symptoms but normal lung function (ie, GOLD stage 0 in the paper) had a similar increase in hospitalisations, both before and after adjustments. Because the “potentially overdiagnosed” subgroup included individuals with symptoms consistent with COPD, the observed increases in hospitalisations with mention of COPD could be attributable to the presence of respiratory symptoms rather than to any lung function abnormalities. The authors could address this confounding of lung function differences by symptoms by re-analysing the hospitalisations after excluding individuals with symptoms in the “potentially overdiagnosed” subgroups.
An additional concern with the study relates to the potential for diagnostic bias. The authors indicate that they analysed hospitalisations “with mention of COPD” but do not indicate whether or not these hospitalisations were attributable to COPD. Because the entire study population underwent initial spirometric testing, it is possible that some of the treating physicians were aware of the spirometry report and may therefore have been more likely to have mentioned COPD on discharge for those patients with an FEV1/FVC ratio <0.70, even if COPD did not actually contribute to the hospitalisation. The authors could address this potential for diagnostic bias by re-analysing the results, restricting the outcome of interest to hospitalisations for which COPD or other respiratory illness was the primary discharge diagnosis.
Adherence to the FEV1/FVC <0.70 criterion in GOLD stage I will potentially identify an additional 5.4 million individuals (58%) in the US population as having COPD compared with use of the LLN criterion.2 This large group of individuals will experience anxiety as well as financial costs, but we do not believe there is sufficient documentation of a potential benefit.6 We agree with others who recommend reconsideration of the FEV1/FVC ratio <0.70 as a criterion for identifying mild COPD.2–5
This contribution was prepared entirely by US Federal Government employees as part of their official duties, is called a “US Government work” and is in the public domain in the USA. The findings and conclusions in this letter have not been formally disseminated by the National Institute for Occupational Safety and Health (NIOSH) and should not be construed to represent any agency policy or determination.
Competing interests: None.