Abstract
Background Dysfunction of the small airways is a precursor of COPD but is not detectable on standard spirometric testing until significant destruction has occurred. A proportion of COPD patients have a forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) <0.7 which is greater than the lower limit of normal (LLN), when adjusted for their age and sex. It is not understood whether this group of patients, known as “discordant COPD”, are representative of “early COPD” or overdiagnosis.
Methods We sought to characterise discordant COPD (disCOPD) using radiology, lung function, serum biomarkers, activity monitoring and quality-of-life scores, comparing with COPD patients with an FEV1/FVC <0.7 and <LLN and healthy, age-matched controls.
Results Six out of eight serum biomarkers were significantly different in the disCOPD group versus healthy controls, as were the scores of all four quality-of-life questionnaires. Activity monitoring revealed similar levels of sedentary time between the disCOPD group and concordant COPD (conCOPD). Computed tomography analysis showed less involvement of small airway dysfunction and emphysema in the disCOPD group versus conCOPD.
Conclusions Collectively, our findings support the hypothesis that disCOPD is a clinically relevant phenomenon that represents a pre-COPD state. Identification of such patients is important for early intervention and management before progression to fully established COPD.
Shareable abstract
Discordant COPD (FEV1/FVC <0.7, >LLN) is a clinically relevant phenomenon that represents an early COPD state, rather than being overdiagnosis of COPD, as demonstrated by CT analysis, serum biomarker levels, activity monitoring and quality of life scores. https://bit.ly/3WG8wYT
Introduction
COPD is a progressive respiratory condition that affects nearly 2% of the UK adult population, and causes significant morbidity and mortality [1]. It typically affects older adults, frequently with a history of tobacco smoking, commonly causing dyspnoea, wheeze, a cough that may be productive and frequent exacerbations. Small airway dysfunction (SAD) is a precursor and central feature of COPD. The small airways are known as the “silent” zone, as changes such as airway remodelling, mucus plugging and immune cell infiltration occur in the absence of clinical symptoms [2]. Changes in standard lung function only become detectable once around 75% of small airways have been destroyed [3]. Identifying patients who may have SAD but have not yet progressed to fully established COPD has important implications for intervention and management of these “pre-COPD” patients. The Global Initiative for COPD 2023 report identified these pre-COPD patients as a priority for further research, specifically the identification of patients that later develop COPD and assessment of interventions that may slow or halt this process [4]. While computed tomography (CT) guided measurement of lung volumes has been shown to predict progression to COPD, the feasibility and cost-effectiveness of this is questionable [5, 6].
Diagnosis of COPD is typically based on clinical presentation and traditional spirometry showing airflow obstruction. However, spirometry is a poor tool in the early stages of the disease and can result in unreliable measurements [4]. Additionally, there is increasing evidence of age-related decline in the forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio that is used to diagnose COPD, in the absence of clinically significant respiratory disease [7]. As such, there is a proportion of “discordant” COPD patients, whose FEV1/FVC ratio is <0.7, but not less than the lower limit of normal (LLN) when adjusted for their age and sex. This may lead to overdiagnosis of the healthy elderly population. Another group is recognised, referred to as preserved ratio impaired spirometry (PRISm) who have a preserved ratio (>0.7) but impaired spirometry such as reduced FEV1 or FVC. This group represents up to 12% of the population and is independently associated with morbidity and mortality [8]. Investigating discordant and PRISm patients through other methods, such as CT scans, other measures of lung function and biomarkers is crucial to help characterise the early stages of the disease and differentiate early COPD changes which require follow-up from age-related changes. There is currently data on CT changes in PRISm patients, as well increasing evidence on the use of biomarkers in characterising SAD [9, 10].
We conceived this study to investigate if “discordant COPD” was more likely to represent disease or overdiagnosis, by conducting more detailed phenotyping of SAD and emphysema, hypothesising that either of these problems could exist prior to the development of spirometrically defined COPD. By using a mixed methodology of radiology, lung function, biomarkers and quality-of-life (QoL) measurements, we aimed to complete a comprehensive assessment of lung pathology and its impact on these patients.
Methods
Study design
We used an observational cohort study design, to assess the degree and nature of SAD in COPD cases and healthy, age-matched controls. The study was ethically approved (IRAS ID: 253739) and all subjects gave informed consent.
Study population
COPD patients and healthy controls were recruited from a multi-site UK National Health Service (NHS) trust in Birmingham, UK. All patients were over the age of 60 years. Control patients were recruited from previous research databases such as the Healthy Elders cohort [11]. The COPD patients were divided into concordant (FEV1/FVC <0.7 and <LLN; conCOPD) and discordant (FEV1/FVC <0.7 but >LLN; disCOPD). Importantly this is not the same as PRISM, where FEV1/FVC is >0.7. The control patients were confirmed to have no obstruction on their spirometry (FEV1/FVC >0.7). Exclusion criteria included inability to understand or speak English, and presence of clinically significant lung disease other than COPD. COPD patients were excluded if they had an exacerbation within the prior 6 weeks. Data on comorbid diseases were collected from the medical record.
Lung function
Pulmonary function measurements were made during the baseline visit. Spirometry, gas transfer and static lung volumes (Ultima PFT Series, Medical Graphics, UK) and forced oscillatory technique (Resmon Pro, Intermedical, UK) were measured by a clinical scientist in a respiratory physiology laboratory. Lung clearance index (5%) was obtained using a nitrogen washout technique (Ultima PFT Series, Medical Graphics, UK). Global Lung Function Initiative (GLI) (2012) reference equations were used to determine % predicted and LLN. % predicted values account for age and sex, thus had potential to offset any eventual differences in these between groups.
Symptoms and QoL
All patients completed four health-related QoL questionnaires; the St George's Respiratory Questionnaire (SGRQ) for impact of respiratory disease on health; the SF-36 for a measure of self-reported health status; the Baseline Dyspnoea Index (BDI) for impact of dyspnoea on daily life; and the COPD Assessment Test (CAT) for the impact of COPD symptoms on daily life.
Physical activity monitoring
All subjects were given a GT3X Actigraph accelerometer (Pensacola, Florida, USA) to wear for 1 week, to provide information about their levels of physical activity. The accelerometer was worn on the right hip, and participants were asked to remove the device for sleeping and water-based activities. The GT3X was set to record accelerations in 10-s epochs, which were subsequently converted into activity counts, and interpreted to determine the frequency, intensity and duration of physical activity.
Data collected by the GT3X was analysed using the Actilife Software (Version 6). For inclusion in the analysis, all participants were required to wear the GT3X for ≥10 h, on ≥4 days. Non-wear time was defined as 60 min of zero counts, with a spike tolerance of 2 min [12]. Time spent sedentary, and in light, moderate and vigorous intensity physical activity, were determined using accelerometer cut-points used in previous research with adults (i.e. sedentary time ≥100 counts·min−1, light-intensity physical activity 101–2019 counts·min−1, moderate-to-vigorous intensity physical activity ≥2020 counts·min−1) [13].
Biomarkers
Blood samples were drawn and centrifuged at 1000×g for 10 min to obtain serum. Serum aliquots were stored at −80°C until further required. Blood biomarkers identified in our previous literature review (CC-16, RAGE, SPD, MMP-8, MMP-9, MMP-12, PAI-1 and CCL-18) were measured using commercially available enzyme-linked immunosorbent assay (ELISA) kits or a Luminex®-based platform [10] (details in supplementary material).
CT scans
COPD patients had CT scans performed to assess for the presence of emphysema and SAD. The study protocol required inspiratory and expiratory scanning. Non-contrast CT imaging was performed in the caudo-cranial direction while patients were supine, fully inhaled and exhaled, with image acquisition using the CT system (SOMATOM go, Siemens Healthineers, Germany). All scans were reviewed by a radiologist for clinical abnormalities prior to quantitative analysis. Parametric response mapping was performed, according to classification criteria described by Vasilescu et al. [14]. For full details of CT protocol, see supplementary material.
Statistical analysis
Statistical analysis was conducted using R and significance assumed at p<0.05. Descriptive statistics were calculated for patient characteristics, including biomarkers, lung function, CT parameters, symptoms and the amount of time spent doing different activity levels for each of the three groups. Comparison of variables between health, discordant and concordant COPD were performed using ANOVA, Kruskal–Wallis or multiple analysis of variance tests. Categorical data were analysed via Fisher's exact test. Data are shown as mean±sd for normally distributed data and median (IQR) for non-normally distributed data.
Results
64 patients were recruited from December 2019 to October 2022. Significant pauses to recruitment and study assessments occurred during the COVID-19 pandemic. Patient characteristics are shown in table 1. Most clinically diagnosed COPD patients were taking treatment, namely short-acting β-agonist (SABA) alone (17.1%), long-acting muscarinic antagonist (LAMA) or LABA (31.4%), or long-acting β-agonist (LAMA)/LABA/inhaled corticosteroid (28.6%), with >95% of conCOPD patients taking regular treatment. Amongst disCOPD patients, regular treatment was far less universal. Fostair (steroid and LABA) and Trimbow (steroid and LABA/LAMA), inhalers likely to penetrate the small airways, were used in 21% disCOPD and 29% conCOPD. The proportion of patients with clinically significant comorbidities that may affect study outcome measures (e.g. cardiac disease, which can impact shortness of breath, and orthopaedic issues, which can impact physical activity) did not differ between disCOPD and conCOPD (51% and 43% respectively; p=0.407). Pulmonary hypertension was not reported in the medical history for any included subject.
Subject characteristics and pulmonary function parameters for healthy controls, discordant COPD (disCOPD) and concordant COPD (conCOPD)
Lung function
Most patients had mild COPD, as demonstrated by FEV1 >80% predicted; this was our aim in recruitment, to best identify aspects relevant to early diagnosis. ConCOPD patients were younger (p=0.002) with mean transfer factor of the lung for carbon monoxide (TLCO) and KCO <80% predicted (both p<0.001) and elevated total lung capacity and functional residual capacity (both p=0.001) implying that emphysema and hyperinflation are confined to this group. Established SAD markers were more aberrant in COPD versus health, with lower FEV1/FVC, lower MMEF25–75, raised X5Hz and raised R5-19 on forced oscillation technique (all p≤0.04). Notably, disCOPD exhibited differences in oscillometry from health, and higher lung clearance index (p=0.005).
Symptoms and QoL outcomes
Four questionnaires were completed that captured QoL measurements for both respiratory-based outcomes (CAT, SGRQ and BDI) and general health status (SF36), shown in table 2.
Quality-of-life scores
COPD patients reported worse impact on QoL than the control patients in all questionnaires. A sequential trend was observed across all scores, whereby the median value worsened (higher or lower, depending on individual score interpretation) with disCOPD being worse than healthy, and conCOPD being worse than disCOPD. No significance was observed between the disCOPD and conCOPD groups for any questionnaire.
Physical activity monitoring
27 of the 64 patients in the study returned activity monitors with successful recordings (healthy controls n=14, disCOPD n=4, conCOPD n=9). This is due to a combination of failure to retrieve the devices, loss to follow-up and device failure. Due to the low numbers in each group, only descriptive analysis could be undertaken.
On average, participants wore the accelerometer for 13.0±0.89 h each day. In all three groups, the majority of accelerometer wear time was spent engaged in sedentary time (healthy control=73%, disCOPD=72%, conCOPD=81%). Lower proportions of time spent in moderate-to-vigorous physical activity were seen in COPD overall (healthy control=5%, disCOPD=3%, conCOPD=2%), but light-intensity physical activity seemed to be maintained by disCOPD patients, compared to healthy controls (healthy control=16%, disCOPD=21%, conCOPD=17%).
Biomarkers
A total of eight biomarkers were measured in the serum of 55 subjects: 18 healthy controls, 13 with disCOPD and 24 with conCOPD. Six out of eight biomarkers displayed significant difference between the three groups. Both PAI-1 and CCL18 were significantly higher in the serum of disCOPD versus healthy controls (p=0.02 and p=0.0006, respectively). PAI-1, CCL18 and MMP-8, -9 and -12 were all significantly higher in the conCOPD compared to the healthy controls (p=0.003, p=0.0009, p=0.0005, p<0.0001 and p=0.01, respectively). CC16 was significantly lower in the conCOPD group versus healthy controls (p=0.006). The only significant difference observed between the COPD groups was for MMP-12, where serum concentration was significantly higher in the conCOPD group versus disCOPD (p=0.02).
CT scans
35 patients with COPD were scheduled to have a CT scan of whom 22 patients had the required inspiratory and expiratory CT scan, eight patients had a clinical inspiratory CT scan only, four patients declined a CT scan, and one patient died before CT scanning could be performed. No patients had interstitial disease or significant pulmonary vascular abnormalities on imaging.
Quantitative CT measures between discordant and concordant COPD groups were compared (figures 1, 2 and 3). A decrease in total pulmonary volume from inspiration to expiration phase in the disCOPD group was noted (figure 1). Both groups showed a similar range of PRM Norm, with the median value lower in the conCOPD group (figure 2). In addition, we observed a lower proportional PRMEmph and PRMfSAD in disCOPD patients (figure 3). However, the difference in quantitative lung morphology between disCOPD and conCOPD groups was not statistically significant (total pulmonary volume at the peak of inspiration and expiration, p=0.14; PRMNorm, p=0.24; PRMEmph, p=0.25; PRMfSAD, p=0.19). See also supplementary table S1.
Lung volumes from computed tomography (CT) imaging. Box-and-whisker plots comparing the total pulmonary volume measured in litres in discordant COPD (disCOPD) and concordant COPD (conCOPD) groups using parametric response mapping as a quantitative CT predictive measure. Mann–Whitney tests were performed to compare two individual groups: a) total pulmonary volumes at the peak of inspiration; b) total lung volume throughout the entire lung at the peak of expiration.
Gas trapping measured by parametric response mapping (PRM). Box-and-whisker plots comparing the proportional normal pulmonary function (PRM Norm) in discordant COPD (disCOPD) and concordant COPD (conCOPD) groups using PRM as a quantitative computed tomography predictive measure. PRM data shown as the percentage of the total lung volume. Mann–Whitney tests were performed to compare two individual groups.
Box-and-whisker plots comparing the proportional emphysematous change and functional small airways disease (SAD) in discordant COPD (disCOPD) and concordant COPD (conCOPD) groups using parametric response mapping as a quantitative computed tomography predictive measure. Mann–Whitney tests were performed to compare two individual groups: a) proportional emphysematous extent change; b) proportional functional SAD.
Discussion
These results suggest that disCOPD may be a clinically significant disease, not a form of overdiagnosis. disCOPD patients had symptoms, physical activity, quantitative radiology and biomarker concentrations more closely resembling conCOPD than health. It emphasises the importance of global assessment of the patient, especially when there is high clinical suspicion of COPD, or a COPD-like illness.
Clinical diagnosis of COPD, defined by a diagnostic label having been given to the patient by their primary care provider, did not result in treatment for all disCOPD individuals – prevalence of inhaled treatment was near universal in the case of conCOPD but approaching 50% in the disCOPD group. This may reflect the fact that a COPD diagnosis should be physiologically confirmed, and in the absence of this (in this case FEV1/FVC >LLN) arguably those individuals did not meet current guidelines for a COPD diagnosis and thus management. However, the more detailed clinical data suggested a significant symptom burden from all questionnaires, implying that disCOPD patients may require treatment in the same manner as conCOPD.
There was no evidence of emphysema or hyperinflation on lung function in disCOPD patients, despite the presence of SAD, highlighting the potential for misdiagnosis of early COPD due to silent SAD. While formal statistical analysis of accelerometer data was not possible, descriptive estimates suggested similar levels of sedentary time across all groups, with the conCOPD group demonstrating the highest time spent sedentary. There was evidence of some preserved light-intensity physical activity in the disCOPD group when compared to conCOPD patients; there could be potential to further improve levels of light-intensity physical activity if regular inhaled therapy was more common. This again suggests that the disCOPD group are perhaps a precursor to the conCOPD group, and that their physical activity levels may reduce further as the disease progresses. A randomised, placebo-controlled trial investigated the efficacy of inhaled nitrous oxide (iNO) in improving physical activity in pulmonary fibrosis patients [15]. Using actigraphy as a measure of physical activity, the trial showed an increase in moderate-to-vigorous activity in those receiving iNO versus placebo.
Studies showing the benefit of inhaled bronchodilators, in the absence of airflow obstruction, have not shown clear benefit for physical activity levels, though evidence is limited [16].
Blood biomarkers were also indicative of a pathophysiological burden in disCOPD. A broad range of serum concentrations were observed in the disCOPD cohort, with a distinct population existing above or below the median value (depending on direction of significance). This population may represent an “at-risk” group of patients that may progress to developing conCOPD. We measured eight biomarkers which have evidence of involvement in SAD [10]. CC16 is expressed largely by club cells in the small airways and is believed to have a protective role within the airways. In a longitudinal cohort, low serum CC16 was associated with accelerated lung function decline [17]. Our results were consistent with this; decreasing levels of serum CC16 from healthy controls to disCOPD, and further in conCOPD, were observed. Similarly, an inverse association with MMP-8 and MMP-9 with lung function was observed. This trend is widely acknowledged in previous COPD studies and is likely reflective of increasing extracellular matrix degradation [18, 19]. More specifically, MMP concentrations have shown a clear association with radiologically confirmed SAD [20]. Serum levels of PAI-1 increased with worsened FEV1; PAI-1 is a serine protease inhibitor, associated with several inflammatory conditions, considered to contribute to the pathophysiology of SAD [21, 22]. CCL18 is predominantly secreted by tissue-resident alveolar macrophages and increased serum concentrations are likely proportionate to numbers of the innate immune cells present within the airway [23]. Serving as a pro-inflammatory chemokine, our finding that CCL18 is significantly elevated in the COPD groups is suggestive of increased airway inflammation.
In COPD, CT is the method used for imaging and diagnosing the key morphological changes, including SAD and emphysema [20]. Quantitative approaches help us to overcome subjectivity in sub-classifying the extent and severity of emphysema and gas trapping related more to SAD than emphysema [24, 25, 26]. Vasilescu et al. [14] validated the capability of noninvasive PRM CT biomarkers and showed their ability to identify SAD and emphysema morphology. Pompe et al. [27] added to the value of using PRM and observed that PRMfSAD and PRMEmph biomarkers were associated with clinical parameters of lung function test, allowing them to identify the presence and severity of COPD. Consistent with our other results, we demonstrated that PRMfSAD and PRMEmph were less impaired in disCOPD than conCOPD patients. This fits with disCOPD being a pre-COPD state, where radiological evidence of disease is present. Disease progression models in COPD suggest that most patients progress from SAD and emphysema to large airway changes over time, rather than the opposite [28].
This study was limited by lower than anticipated recruitment, which is related to the overlap with the COVID-19 pandemic UK peaks. Numbers in this study were lower than in other studies of discordant COPD [29, 30]. Similarly, this limited our ability to appropriately age- and sex-match our cohorts; for lung function and oscillometry measures, we report % predicted values, which account for these confounders. We chose not to stratify by severity of COPD partly for this reason, but also because the distribution of lung function demonstrated that most patients had mild disease (the target population for recruitment to the study). We also only report cross-sectional data, whereas longitudinal studies of spirometry have been able to give evidence of temporal changes [8]. Our ethics committee did not allow CT scanning of healthy controls, although disCOPD patients were still able to be compared with conCOPD.
Despite these absences, a major strength of this study is the breadth of investigations used to investigate our stated aim. Rather than relying on one single measure, we have used symptoms, physiology, radiology and serum biochemistry to holistically assess the clinical relevance of disCOPD, and overall have found a cohesive message: disCOPD patients appear to have similar features to conCOPD, including when compared with healthy controls. This is applicable to the interpretation of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2024 report, which advocates the use of a 0.7 FEV1/FVC threshold for COPD diagnosis but expresses reservations about overdiagnosis in the elderly, and to UK national guidance, which does not mention the LLN [31, 32]. Our data support the use of the 0.7 threshold in clinical practice, and suggests that, rather than representing overdiagnosis, disCOPD spirometry with symptoms represents disease status.
Use of the 0.7 threshold is likely to make spirometry interpretation easier for the non-specialist and identify COPD at an earlier stage. Identifying the disease at an early stage is likely to have multiple benefits for patients: aside from starting pharmacology earlier, they will be able to access COPD and exacerbation management plans, smoking cessation guidance, physiotherapy where indicated and annual review policies. These approaches all aim to slow progression of the disease, and by accessing them earlier, a greater proportion of lung function may be preserved. Future research could assess asymptomatic patients with this spirometry pattern over time, to see what proportion develop symptoms and FEV1/FVC <LLN. Assessing relation of this potential “pre-COPD” state to risk factors for fixed airflow obstruction, such as family history of airways disease, might also be of interest.
Conclusion
Our data concur with GOLD and suggest that symptomatic patients with FEV1/FVC ratio <0.7, but above the LLN, are not overdiagnosed with COPD, but instead represent a clinically significant subset of the disease. When paired with a high clinical suspicion of COPD, such spirometry should trigger holistic investigation, including further lung function and radiology.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material 00264-2024.SUPPLEMENT
Footnotes
Provenance: Submitted article, peer reviewed.
Ethics statement: The study was ethically approved (Integrated Research Application System identifier 253739) and all subjects gave informed consent.
Conflict of interest: None declared.
Support statement: This study was supported by an investigator-initiated grant from Chiesi; the funder had no input to study design, analysis or publication. A.M. Turner is also funded by the National Institute for Health and Care Research (NIHR) Midlands Patient Safety Research Collaboration, as well as NIHR Efficacy and Mechanism Evaluation, and NIHR Health Technology Assessment. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received March 21, 2024.
- Accepted May 9, 2024.
- Copyright ©The authors 2024
This version is distributed under the terms of the Creative Commons Attribution Licence 4.0.
References
