Abstract
Background The highest burden of chronic obstructive pulmonary disease (COPD) occurs in low- and middle-income countries. Low-cost oral medications, if effective, could enable affordable, accessible COPD treatment.
Methods In this randomised, three-arm, double-blind, double-dummy, placebo-controlled study conducted in 37 centres in China, symptomatic patients with moderate to very severe COPD were randomised 1:1:1 to placebo twice daily plus placebo once daily, low-dose theophylline 100 mg twice daily plus placebo once daily or low-dose theophylline 100 mg twice daily plus low-dose oral prednisone 5 mg once daily for 48 weeks. The primary end-point was annualised exacerbation rate.
Results 1670 subjects were randomised and 1242 completed the study (1142 with acceptable data at week 48). Subjects (75.7% male) had a mean age of 64.4 years, with mean±sd baseline post-bronchodilator forced expiratory volume in 1 s (FEV1) 1.1±0.4 L (42.2% predicted) and St George's Respiratory Questionnaire (SGRQ) score 45.8±20.1. There were negligible differences between annualised exacerbation rates across the three treatments: 0.89 (95% CI 0.78–1.02) on theophylline plus prednisone, 0.86 (95% CI 0.75–0.99) on theophylline plus placebo and 1.00 (95% CI 0.87–1.14) on placebo. The rate ratio for theophylline plus prednisone versus pooled theophylline plus placebo and placebo was 0.96 (95% CI 0.83–1.12), for theophylline plus placebo versus placebo was 0.87 (95% CI 0.73–1.03; p=0.101) and for theophylline plus prednisone versus placebo was 0.90 (95% CI 0.76–1.06; p=0.201). Secondary outcomes of hospitalisations, FEV1, SGRQ and COPD Assessment Test score showed no statistically significant difference between treatment arms. Serious adverse events other than exacerbations were <2% and did not differ between treatment arms.
Conclusions Low-dose theophylline alone or in combination with prednisone did not reduce exacerbation rates or clinically important secondary end-points compared with placebo.
Abstract
This large, rigorously conducted RCT showed that the combination of low-dose theophylline and prednisone did not affect exacerbation rate in patients with moderate to severe COPD in China https://bit.ly/2KSQ2BK
Introduction
The global burden of chronic obstructive pulmonary disease (COPD) is rising and the majority of this burden falls in low- and middle-income countries (LMICs) [1, 2]. The mainstay of pharmacological treatment for COPD is the early and sustained use of bronchodilators, and the later introduction of inhaled corticosteroids (ICSs) for exacerbating patients. The current Global Initiative for Chronic Obstructive Lung Disease (GOLD) strategy recommends short-acting β-agonists (SABAs) and long-acting bronchodilators (long-acting β-agonists (LABAs) and long-acting muscarinic antagonists (LAMAs)) alone or in combination for symptomatic COPD, in all GOLD stages [3]. Theophylline is also of benefit; however, due to its potential toxicity when used in bronchodilating doses, it is not recommended as step-up therapy from SABAs unless inhaled long-acting bronchodilators are not available or affordable. This is the case in many LMICs, where theophylline is currently the main treatment for COPD in addition to short-acting bronchodilators [4]. ICSs play an important role in reducing exacerbations, especially in frequent exacerbators and people with worse airflow obstruction, but have minimal or no effect on airflow as measured by forced expiratory volume in 1 s (FEV1) or lung function decline, leading to the view that COPD is a corticosteroid-resistant disease [5, 6].
In LMICs, most inhaled medications, including ICSs, are neither affordable nor accessible for the majority of patients with COPD [4, 7] and affordable oral therapies are frequently used. In vitro and in vivo studies show that corticosteroids and theophylline, both in low doses, have synergistic and clinically useful anti-inflammatory effects in COPD [8, 9]. The molecular mechanisms for this effect suggest that it occurs through theophylline increasing the activity of the nuclear enzyme histone deacetylase-2 (HDAC2), which is reduced in COPD cells, thus preventing the anti-inflammatory effect of corticosteroids [10, 11], an effect separate from the mechanism of bronchodilation [12]. Small clinical studies have suggested that low-dose theophylline, at levels below those which cause bronchodilatation, can reverse corticosteroid insensitivity in COPD [9, 13]. One small study has shown an effect for low-dose theophylline on FEV1 as well as exacerbations [14].
A recently published randomised controlled trial (RCT), Theophylline with Inhaled Corticosteroids (TWICS), examined this effect in COPD patients taking ICSs [15]. Treatment for 1 year with low-dose theophylline did not reduce the number of COPD exacerbations compared with placebo when added to their maintenance ICS and inhaled bronchodilator therapy, although in a post hoc analysis there was a significant reduction in severe (hospitalised) exacerbations in the theophylline group. The study design did not assess adherence to ICS, so it is not clear if patients were taking adequate amounts to enable this hypothesis to be tested.
The aim of the present study, the Theophylline And Steroids in COPD Study (TASCS) RCT, was to compare the efficacy and safety of low-dose theophylline and oral prednisone given daily for 48 weeks versus low-dose theophylline alone or placebo alone, on time to first exacerbation and annualised rate of exacerbations in patients with moderate to very severe COPD.
Methods
Study design
TASCS was a 48-week, randomised, three-arm, double-blind, double-dummy, placebo-controlled study in symptomatic patients with moderate to very severe COPD, conducted in 37 centres in China. The study comprised a washout/run-in period of 2–4 weeks followed by a treatment phase.
Study population
Eligible patients were male or female, aged 40–80 years, with diagnosed, stable moderate to very severe COPD (FEV1/forced vital capacity (FVC) <0.70 and FEV1 <70% predicted). They had either a smoking history of ≥10 pack-years or biomass exposure assessed by a standard exposure questionnaire of >10 years [16]. Although patients were not required to have had a treated COPD exacerbation in the year prior to screening, investigators were encouraged to consider such patients as eligible. Patients were ineligible if an exacerbation occurred within 4 weeks of screening. They were also excluded if they were prescribed domiciliary oxygen, had coexistent illness precluding participation in the study or suggesting a life expectancy <1 year, pulmonary resection, or current asthma. In relation to potential risk of treatment, patients were excluded if they had known sensitivity to, or intolerance of, theophylline, clinical evidence of chronic liver disease, or transaminase or γ-glutamyl transferase elevation >1.5× upper limit of normal, or random blood glucose level >8 mmol·L−1. Permitted medications included regular inhaled LAMA and/or LABA therapy, and short-acting anticholinergic and SABA inhaled rescue medication. Maintenance ICSs, oral corticosteroids, parenteral corticosteroids and oral syrups or other formulations containing theophylline were not permitted during the study.
Intervention
The TASCS study compared 1) placebo plus placebo, 2) low-dose theophylline plus placebo and 3) low-dose theophylline plus low-dose oral prednisone. Symptomatic patients with moderate to very severe COPD were randomised 1:1:1 to placebo twice daily plus placebo once daily, theophylline 100 mg twice daily plus placebo once daily or slow-release theophylline 100 mg twice daily plus prednisone 5 mg once daily for 48 weeks.
Study procedures
Potential participants were identified through hospital outpatient and inpatient clinics at 37 centres in China, and undertook a screening visit 2–4 weeks before the randomisation visit. During the run-in, all prohibited medications, specifically theophylline-containing medications and ICSs, were ceased, while LABAs, LAMAs and SABAs were continued. The run-in was undertaken to ensure stability and patients' acceptance of the changes prior to randomisation. Pre- and post-bronchodilator spirometry was performed at the screening visit to confirm eligibility. Vital signs, spirometry, COPD Assessment Test (CAT) score [17], medical history including bone fractures, medication, demographics, routine pathology and St George's Respiratory Questionnaire (SGRQ) were also recorded prior to randomisation [18]. Study visits took place at 12-week intervals for 48 weeks, with spirometry, SGRQ, CAT score and diary card review as well as medication return and dispensing of new medication. Symptoms attributable to corticosteroid and theophylline toxicity were recorded at each patient visit.
Adherence with study treatment was assessed by return of study drug at clinic visits and by diary card recording of episodes where study drug was discontinued for ≥5 days, with number of days discontinued and reason for discontinuation.
Outcomes
The primary outcome was the number of COPD exacerbations per participant in 48 weeks, annualised as a rate per patient per year. A COPD exacerbation was defined as symptomatic deterioration in COPD symptoms (cough, sputum production or dyspnoea) requiring treatment with antibiotics, initiation of a course of systemic corticosteroids, hospitalisation or a combination of these. Exacerbations were graded as mild if they required symptomatic treatment with inhaled bronchodilators only, moderate if managed with antibiotics and/or oral corticosteroids, and severe if the exacerbation also resulted in emergency department presentation or hospitalisation. Secondary outcomes included time to first severe exacerbation leading to hospitalisation or death, health status using SGRQ and CAT scores, and pre- and post-bronchodilator FEV1, FVC and FEV1/FVC ratio.
A subgroup of 75 patients based at three major centres had blood samples taken at baseline, 12 and 48 weeks for serum theophylline, which was analysed blind in a central laboratory after completion of the study. Morning serum cortisol (08:00–10:00 h) was measured in a subgroup of 91 patients at 50 weeks. The 50-week time-point (2 weeks following completion of the 48-week study drug administration) was chosen to assess adrenal recovery and the safety of administration of prednisone 5 mg daily for 48 weeks [19].
Trial oversight
The trial protocol and informed consent procedures were approved centrally by the Human Research Ethics Committee of the University of Sydney (Sydney, Australia) and by the institutional review board at each study site. Patient information and consent forms, questionnaires, and diary cards were forward- and back-translated. Participant information statements and consent forms were approved by each ethics committee and formatted in accordance with their own guidelines and requirements. All patients provided written informed consent prior to undertaking any study-specific procedures. The study was performed in accordance with the Declaration of Helsinki and Good Clinical Practice Guidelines.
The trial was conducted under the direct supervision of the principal investigators (N.B. and C.R.J.) and registered at ClinicalTrials.gov with identifier number NCT02261727.
Randomisation and masking
Treatment randomisation was stratified by smoking status and clinical site, using a secure web-based randomisation system and centrally administered in agreement with Standard Operating Procedures on randomisation at The George Institute for Global Health (Sydney, Australia). Study drug manufacture, packaging and labelling was undertaken by approved Chinese suppliers and distributors.
Statistical methods
All analyses were performed on an intention-to-treat basis.
To compare COPD rates, we fitted multivariable negative binomial regressions with a pre-specified covariate set which included treatment arm, hospital region, history of smoking, sex, exacerbation in the year prior to randomisation and post-bronchodilator FEV1 % pred at screening. For time-to-event outcomes we used proportional hazard Cox regressions and for continuous secondary outcomes we assessed the difference across treatments by ANCOVA.
Pre-specified subgroup analyses for the primary outcome included age (<65 and ≥65 years), sex, smoking status (current smoker, ex-smoker and never-smoker), biomass exposure, on oral steroid at baseline (yes/no), COPD exacerbation in last 12 months (yes/no), SGRQ score (<45 versus ≥45), CAT score (<20 versus ≥20), FEV1 % pred thresholds (<50% and ≥50%) and eosinophils (<0.30 versus ≥0.30, <0.20 versus ≥0.20 and <0.15 versus ≥0.15×109 cells·L−1).
To overcome the potential issue of multiple testing among three treatment groups, the primary comparison was pre-specified as the theophylline plus prednisone group versus the pooled combination of the other two groups (theophylline plus placebo and placebo). We specified a hierarchical process where, if the primary comparison was found to be statistically significant, separate comparisons between the two treatment groups versus the placebo group would be conducted. No multiple testing adjustments were performed, although we critically assessed any p-value <0.05. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary NC, USA). Further statistical information is included in the supplementary material.
Results
Participants commenced enrolling in the study in June 2014 and the last patient completed the study on May 14, 2018. In total, 2325 patients were screened, 1670 were randomly assigned to study treatment and 1242 completed the study (a 26% withdrawal rate) (figure 1). Out of 2325 patients screened, 665 were not eligible to proceed to randomisation at baseline. Failure to meet spirometric requirements was the principal reason for not meeting eligibility criteria: either due to post-bronchodilator FEV1 >70% predicted or FEV1/FVC >0.7, or inability to perform reproducible spirometry meeting American Thoracic Society/European Respiratory Society reproducibility standards. Protocol noncompliance was the other major cause, primarily physician decision regarding participant's difficulty with trial adherence or failure to return for randomisation visit within 4 weeks of screening.
CONSORT diagram. Study participant flow from screening, run-in, randomisation, all visits and study completion. ITT: intention-to-treat; PP: per-protocol.
There were no clinically significant differences between the treatment groups with regard to baseline demographic characteristics, COPD history or treatment (table 1). The mean±sd age of participants was 64.4±8.0 years, of whom 75.7% were male. Current smokers had a mean±sd 42.7±23.9 pack-year history and comprised 19.9% of the total, with ex-smokers 53.3% and never-smokers 26.8%. Of the total population, 38.4% of participants had a biomass exposure history, and 17.1% had dust and fume exposure >10 years. Mean±sd baseline post-bronchodilator FEV1 was 1.1±0.4 L (42.2% predicted) and acute bronchodilator reversibility was 0.10±0.04 L. 47% of patients reported having an acute exacerbation of COPD requiring treatment with an antibiotic, systemic steroids or both in the previous 12 months.
Baseline characteristics of the patients in the intention-to-treat population
Annualised exacerbation rates across the three treatment arms were similar: 0.89 (95% CI 0.77–1.02) on theophylline plus prednisone, 0.86 (95% CI 0.75–0.99) on theophylline plus placebo and 1.00 (95% CI 0.87–1.14) on placebo at week 48 (table 2). There was no statistical difference in the rate ratio of exacerbations in the comparison between theophylline plus prednisone versus pooled theophylline plus placebo and placebo, which was 0.96 (95% CI 0.83–1.12; p=0.6084). The rate ratio for theophylline plus placebo versus placebo was 0.87 (95% CI 0.73–1.03; p=0.101) and for theophylline plus prednisone versus placebo was 0.90 (95% CI 0.76–1.06; p=0.201). Time to first COPD exacerbation did not differ between the treatment arms (table 3 and figure 2).
Annualised chronic obstructive pulmonary disease (COPD) exacerbation rate in each study arm at 48 weeks in the intention-to-treat and per-protocol populations
Time to first (mild, moderate or severe) chronic obstructive pulmonary disease (COPD) exacerbation in the intention-to-treat population#
a) Time to first chronic obstructive pulmonary disease (COPD) event (i.e. a COPD exacerbation defined as a symptomatic deterioration in COPD symptoms (cough, sputum production or dyspnoea) requiring treatment with antibiotics, initiation of a course of systemic corticosteroids, hospitalisation or a combination of these). b) Time to first mild COPD event (i.e. a COPD exacerbation requiring symptomatic treatment with inhaled bronchodilators only). c) Time to first moderate COPD event (i.e. a COPD exacerbation managed with antibiotics and/or oral corticosteroids). d) Time to first severe COPD event (i.e. a COPD exacerbation which also resulted in emergency department presentation or hospitalisation).
A per-protocol analysis was undertaken for the annualised exacerbation rate (primary outcome). The per-protocol population included all participants who completed the study within an 8-week window of week 48 and who were not withdrawn or discontinued due to adverse events (4.2%). Other reasons for exclusion from the per-protocol population were death (4.0%), protocol noncompliance including major protocol deviations or treatment adherence <60% (8.1%), loss to follow-up (38.6%), investigator decision (1.6%), withdrawal based on patient decision (41.2%), or ineligible based on the protocol entry criteria (2.7%). There were no significant differences between the three treatment groups for total, mild, moderate or severe annualised exacerbation rates in the per-protocol population (table 2).
Secondary outcomes of hospitalisations, FEV1, SGRQ and CAT showed negligible differences between treatment arms (table 4)). In view of the lack of difference between treatment arms we did not undertake the pre-specified subgroup analyses based on age, sex, smoking status, exacerbation history, SGRQ and CAT score for the primary outcome. Rate ratios by clinically relevant subgroups did not identify significant predictors of treatment effect (figure 3). Mean FEV1, SGRQ and CAT scores were numerically better during and at the end of the study in all arms than at baseline, although statistical analysis was not undertaken for these comparisons.
Secondary outcomes in each of the three study arms in the intention-to-treat population at 48 weeks: change compared with baseline
Forest plot showing rate ratio (95% CI) by subgroup. COPD: chronic obstructive pulmonary disease; SGRQ: St George's Respiratory Questionnaire; CAT: COPD Assessment Test; FEV1: forced expiratory volume in 1 s. Model adjusted for hospital region, history of smoking, sex, exacerbation in the year prior to randomisation and post-bronchodilator FEV1 % pred at screening.
Medication compliance, estimated based on pill return at each clinic visit, was >80% in >85% of subjects. Random serum theophylline levels and morning serum cortisol levels at 50 weeks were consistent with treatment allocation, the mean theophylline levels being significantly higher in the two theophylline-containing arms than the placebo arm (supplementary table S1) and the mean morning cortisol being lower in the prednisone-containing arm compared with the nonprednisone arms (supplementary table S2).
Patients were more likely to complete the study on the theophylline plus prednisone treatment arm compared with the other two study arms combined, but not comparing the arms individually (supplementary table S2). The number of adverse events was low, with an expected incidence across the treatment arms (table 5).
Adverse events by treatment group in the intention-to-treat population
Discussion
In this large RCT in China, we showed no difference in the annualised rate of exacerbations on combined low-dose theophylline and prednisone versus placebo or low-dose theophylline alone. Nor was there any benefit for this combination in our secondary outcomes, which included quality of life and lung function. Our findings indicate that the treatment combination of low-dose theophylline and prednisone does not confer benefit over low-dose theophylline alone and that low-dose theophylline alone provided no clinical benefits compared with placebo.
The patients in our study had moderate to very severe COPD based on their lung function and exacerbation history, although we could not allocate a GOLD grading for GOLD A–D [3] as we did not collect a history of hospitalised exacerbations in the previous 12 months. The mean baseline CAT score of 18.1 and baseline SGRQ score of 45.8 suggest that patients in China record a lower impact of COPD for their severity of airflow obstruction than is the case in many other countries in COPD studies.
Although there was no evidence of benefit of low-dose theophylline plus prednisone in this study, on average the TASCS participants improved their COPD status during the study year. Although the mean rate of exacerbations reported in the prior year (0.7) was slightly lower than the rate recorded during the study year (0.9–1.0), the prior year rate was based on self-report only, and likely subject to recall bias and under-reporting. All pre-specified secondary end-points (lung function and health status outcomes) improved between study commencement and completion in all treatment arms. The consistency of this response suggests that the patients gained benefit from clinical trial participation, even if not from the study drug interventions, although we cannot exclude the possibility that patients who dropped out of the study were sicker and this helped improve the scores of all outcomes studied.
It is possible that low-dose theophylline does improve corticosteroid responses at a cellular level in COPD, but does not restore HDAC2 levels or activity adequately enough to produce a clinically evident benefit [20]. On the basis of the random blood tests in TASCS, we can conclude that the patients in the prednisone-containing arm on the whole were adherent to the intervention and we do not consider poor adherence was an issue in TASCS. Even 2 weeks after ceasing the study and daily administration of prednisone, randomly selected patients still had lower mean morning cortisol levels than the patients in the other two study arms [19]. Additionally, study medication adherence was high and although it dropped slightly during the study, at week 48 on the basis of pill returns, 88% of participants were >80% adherent with the medication.
There are some limitations to our study, although we consider the findings are consistent, robust and support our conclusions for no effect of these interventions. There was a 26% withdrawal rate in this study, with a higher rate of withdrawals occurring earlier in the trial. Patients requested withdrawal if they felt the treatment was not helping them and, in particular, if they suffered an exacerbation. This was most evident at the first study visit after treatment commencement, as can be seen in supplementary figure S1 at the 12-week visit, when the greatest number of withdrawals occurred.
During the study, owing to slow recruitment, we reduced the power of the study from 90% to 80% in order to reduce our target study population. Estimation of study sample size based on 80% power is the case for several published studies of COPD exacerbations [21, 22]. Even so, the TASCS results do not suggest any trend to a difference between treatments that would become statistically significant if we had maintained 90% power by having a higher number of study participants.
Many of the patients in this study were from rural towns in China located a significant distance from major centres and some patients travelled 2–3 h to reach the study centre. As a result, severe exacerbations requiring oral corticosteroids and/or antibiotics were often identified by hospital presentations and the proportion of hospitalised exacerbations is relatively high in TASCS compared with some other COPD study populations [23–25]. Conversely, we consider that a proportion of milder exacerbations were not diagnosed, treated or recorded. While it is recognised that mild COPD exacerbations often go unrecognised by patients and clinicians, in China we believe this is even more likely to be the case. The mean exacerbation rate in TASCS of 0.9 per patient per year almost certainly underestimates the true exacerbation rate for the study population.
In conclusion, in this large RCT comparing low-dose theophylline and prednisone with placebo or low-dose theophylline alone, there was no demonstrable effect on exacerbation rate, lung function or COPD related quality of life.
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 ERJ-03338-2020.Supplement
Supplementary tables ERJ-03338-2020.Tables
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-03338-2020.Shareable
Acknowledgements
We thank all the investigators in our 37 participating centres, their research assistants who managed the TASCS study at each site (principal investigators and sites are listed in the supplementary material) and all the TASCS study participants, many of whom travelled long distances. We thank the TASCS study monitors and Project Managers at George Clinical (Beijing, China), statisticians Qiang Li and Kris Rogers of The George Institute for Global Health (Sydney, Australia) for their support during the conduct of the trial, and our database managers at George Clinical (Sydney, Australia) who assisted in data cleaning, resolution of queries, and finalising tables and figures.
Footnotes
This article has an editorial commentary: https://doi.org/10.1183/13993003.04564-2020
This article has supplementary material available from erj.ersjournals.com
This study is registered at ClinicalTrials.gov with identifier number NCT02261727. Individual de-identified participant data are available on application to the corresponding author. The protocol and statistical analysis plan are available on the Centre for Open Science website: https://osf.io/39qkm
Author contributions: C.R. Jenkins, N. Berend, P.J. Barnes and B. Celli contributed to the study concept and design, data interpretation, and final approval of the manuscript. A. Martin contributed to implementation of the study, data collection, cleaning, interpretation and final approval of the manuscript. A. Scaria, G-L. Di Tanna and T. Bradbury contributed to statistical analysis and interpretation, and final approval of the manuscript. F-Q. Wen, N-S. Zhong, J-P. Zheng and A. Martin contributed to the study execution, data interpretation, review and final approval of the manuscript. C.R. Jenkins wrote the primary manuscript, and all authors revised and contributed to its final review. C.R. Jenkins had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Conflict of interest: C.R. Jenkins reports personal fees from Boehringer Ingelheim and Chiesi, grants, personal fees and nonfinancial support from GlaxoSmithKline, personal fees and nonfinancial support from AstraZeneca, Novartis and Sanofi Genzyme, outside the submitted work.
Conflict of interest: F-Q. Wen has nothing to disclose.
Conflict of interest: A. Martin has nothing to disclose.
Conflict of interest: P.J. Barnes reports grants and personal fees for consultancy and advisory board work from AstraZeneca, grants and personal fees for advisory board work and lectures from Boehringer Ingelheim, personal fees for advisory board work and lectures from Novartis and Teva, personal fees for advisory board work from Pieris and Epi-Endo, outside the submitted work.
Conflict of interest: B. Celli reports grants and research facilities from AstraZeneca, personal fees for consultancy and scientific committee work from GlaxoSmithKline, personal fees for consultancy from Boehringer Ingelheim, Sanofi-Aventis, Menarini, Chiesi and Pulmonx, outside the submitted work.
Conflict of interest: N-S. Zhong has nothing to disclose.
Conflict of interest: J-P. Zheng has nothing to disclose.
Conflict of interest: A. Scaria has nothing to disclose.
Conflict of interest: G-L. Di Tanna is a former employee of Amgen.
Conflict of interest: T. Bradbury reports receiving a top-up scholarship funded by GlaxoSmithKline, outside the submitted work.
Conflict of interest: N. Berend reports grants from the National Health and Medical Research Council, Australia, State Key Laboratory of Respiratory Disease and Guangzhou Institute of Respiratory Disease China, and West China Hospital, Chengdu, China, during the conduct of the study; and a salary from GlaxoSmithKline, outside the submitted work.
Support statement: This work was supported by the National Health and Medical Research Council, Australia; State Key Laboratory of Respiratory Diseases and Guangzhou Institute of Respiratory Diseases, Guangzhou, China; and West China Hospital, Chengdu, China. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received September 1, 2020.
- Accepted November 21, 2020.
- Copyright ©ERS 2021. For reproduction rights and permissions contact permissions{at}ersnet.org
References
