Background The impact of pulmonary rehabilitation (PR) on survival in patients with fibrotic interstitial lung disease (ILD) is unknown. Given the challenges conducting a large randomised controlled trial, we aimed to determine whether improvement in 6-minute walk distance (6MWD) was associated with better survival.
Methods This retrospective, international cohort study included patients with fibrotic ILD participating in either inpatient or outpatient PR at 12 sites in 5 countries. Multivariable models were used to estimate the association between change in 6MWD and time to death or lung transplantation accounting for clustering by centre and other confounders.
Results 701 participants (445 men and 256 women) with fibrotic ILD were included. The mean±SD ages of the 196 inpatients and 505 outpatients were 70±11 and 69±12 years, respectively. Baseline/changes in 6MWD were 262±128/55±83 m for inpatients and 358±125/34±65 m for outpatients. Improvement in 6MWD during PR was associated with lower hazard rates for death or lung transplant on adjusted analysis for both inpatient (HR per 10 m 0.94, 95% CI 0.91 to 0.97, p<0.001) and outpatient PR (HR 0.97, 95% CI 0.95 to 1.00, p=0.042). Participation in ≥80% of planned outpatient PR sessions was associated with a 33% lower risk of death (95% CI 0.49% to 0.92%).
Conclusions Patients with fibrotic ILD who improved physical performance during PR had better survival compared with those who did not improve performance. Confirmation of these hypothesis-generating findings in a randomised controlled trial would be required to definitely change clinical practice, and would further support efforts to improve availability of PR for patients with fibrotic ILD.
- idiopathic pulmonary fibrosis
- interstitial fibrosis
- pulmonary rehabilitation
Data availability statement
Data are available upon reasonable request. Data are available upon reasonable request to the corresponding author.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is the key question?
Do patients with fibrotic interstitial lung disease (ILD) who improve their 6-minute walk distance (6MWD) after pulmonary rehabilitation (PR) have a better survival than those who fail to improve their 6MWD?
What is the bottom line?
The increase in 6MWD after inpatient or outpatient PR is associated with an improved survival up to 3 years after PR, suggesting that a successful PR has the potential to improve survival in patients with fibrotic ILD.
Why read on?
This is the largest multinational real-world cohort of PR participants with fibrotic ILD, providing insight into the magnitude of benefit from PR in this population, with a responder analysis showing a relationship between the improvement of physical performance and survival.
Fibrotic interstitial lung diseases (ILDs) typically lead to progressive dyspnoea, fatigue and decrease in physical performance. Survival is poorest in patients with idiopathic pulmonary fibrosis (IPF), with only slightly more favourable mortality rates in other fibrotic ILDs such as fibrotic hypersensitivity pneumonitis (HP), connective tissue disease (CTD)-associated ILD and unclassifiable ILD.1 2 Antifibrotic and immunomodulatory medications can slow disease progression in many ILD subtypes, but their impact on exercise performance or survival has not been clearly established.3–6
Comprehensive patient-tailored pulmonary rehabilitation (PR) aims to improve physical and psychological deficits in patients with fibrotic ILD using structured exercise training and patient education. PR is delivered either in an outpatient setting up to several times per week, or in a specialised inpatient PR clinic with multiple daily sessions.7 Outpatient PR improves short-term exercise capacity, muscle strength, dyspnoea, fatigue, physical activity level and body composition in patients with fibrotic ILD,8–15 while inpatient PR similarly improves physical performance and quality of life.16–18 The sustainability of these short-term benefits beyond 1 year from PR remains controversial, and long-term effects are insufficiently studied.19–21
Despite improving both patient-reported and functional outcomes, PR remains poorly accessible and underused in most regions,22 23 with a multinational study reporting only 1.2% of patients with COPD having access to a PR programme.24 Patients with fibrotic ILDs likely have further access limitations to PR considering that the evidence for its effectiveness is less robust,7 with no studies demonstrating a survival benefit in this population. However, generating such evidence is impractical given the size of the required studies combined with the suggestion from many experts that PR should be standard of care based on previous data.7 25
Given the challenges of conducting a large randomised controlled trial, the primary aim of this multicentre international retrospective cohort study was to determine whether improvement in 6-minute walk distance (6MWD) during PR was associated with subsequent survival. Our expectation was that confirming this association would strengthen the argument that PR improves mortality in fibrotic ILD, thus providing further evidence of clinical utility and greater impetus to improve access to this therapy.
Study design and patient population
For this retrospective cohort study, 12 sites in Canada, the USA, Australia, Germany and Switzerland collected data from consecutive patients with fibrotic ILD of any type who had participated in their first inpatient or outpatient PR since 2000. Programme completion was not a requirement for inclusion. Two sites provided only inpatient PR, eight provided only outpatient PR and two provided both programmes. Patients with systemic disease (eg, sarcoidosis and CTD-ILD) were excluded due to the frequently predominant musculoskeletal impairment in this population. Patients with a greater extent of emphysema than fibrosis on chest CT were also excluded.
Patients participated in standardised, individually tailored inpatient or outpatient PR programmes. Participants of inpatient PR usually stay for 2–4 weeks in a specialised PR clinic, where they typically attend two to three sessions of exercise training per day for 5–6 days of the week, with additional psychological support and educational sessions.16 17 A typical outpatient PR programme provides two to three sessions per week of structured exercise and education in an outpatient setting for a total of 6–12 weeks.9 11 Both inpatient and outpatient exercise sessions included aerobic training and/or interval training with a gradual symptom-limited increase in workload. Resistance training covered all large muscle groups and typically consisted of three sets with 8–12 repetitions. All centres provided supplemental oxygen if needed to maintain a peripheral oxygen saturation ≥88% during training. Educational sessions typically covered information on the pathophysiology of lung diseases, oxygen and medication use, symptom control, coping mechanisms and self-management, with additional counselling on smoking cessation, nutrition and psychological support if needed.
Measurements and outcomes
Individual centres collected data from prospectively maintained medical records and local patient databases. Baseline demographics, ILD diagnoses, key comorbidities, specific medications and pulmonary function test results were entered into a standardised data collection sheet, as well as programme-specific information such as PR modality (inpatient vs outpatient), duration and number of planned sessions. Additional patient-specific information on PR delivery and outcomes was also recorded, including number of attended sessions, use of supplemental oxygen during and/or after PR, and dates of death, lung transplantation, or loss to follow-up.
Participants performed a 6-minute walk test (6MWT) at the start and end of PR according to standardised protocols.26 27 Per cent predicted 6MWD was calculated based on the equations proposed by Enright and Sherrill.28 The minimal clinically important difference (MID) for change in 6MWD in ILD ranges from 22 m to 41 m depending on the estimation approach, with 30 m being the most accepted threshold. For some analyses, participants were further stratified by baseline 6MWD ≥/<350 m.29 The primary outcome was time from the start of PR until the date of death, lung transplantation or censoring. Follow-up time was censored at the time when participants were last known to be alive, which includes participants who were lost to follow-up and participants who were known to be alive at closure of data collection. Secondary outcomes were 1-year, 2-year and 3-year survival after PR.
Participant and PR programme characteristics are reported as number (per cent), mean±SD or median (IQR), stratified by inpatient or outpatient PR participation. The association of change in 6MWD with baseline characteristics was assessed by independent t-tests for categorical variables and by Spearman or Pearson correlation for continuous variables depending on the distribution of the data. To estimate the contribution of PR modality and individual centres to 6MWD variability, linear mixed models fitted with restricted maximum likelihood were used (outcome: 6MWD, random effects: PR modality and centre). The intraclass correlation from these models is the proportion of variance explained by the random effect (eg, between centre variance/total variance).
Cox proportional hazards models with mixed effects accounting for clustering by centre were used to estimate the impact of change in 6MWD on the composite outcome of time to death or lung transplantation. Survival up to 3 years was analysed using logistic regression models with mixed effects. ORs and HRs with 95% CIs are reported per 10 m or 10%-predicted change in 6MWD as applicable, with this rescaling used to improve interpretability of effect estimates. In addition to the random effect for centre, fixed effects were added to adjust for prespecified potential confounders (age, sex, baseline 6MWD). Kaplan-Meier curves with p values from log-rank tests were plotted from survival estimates in specific groups. These groups included change in 6MWD ≥30 m vs <30 m in all inpatient and outpatient PR participants and stratified by baseline 6MWD ≥/<350 m, and participants with adherence to ≥80% vs <80% of PR sessions. A two-sided significance level of p<0.05 was considered statistically significant. Data were analysed using R V.3.6.0 (R Foundation for Statistical Computing, Vienna, Austria).
A total of 701 patients with fibrotic ILD were included from the 12 participating sites (figure 1 and table 1), with 4 sites providing inpatient PR (196 participants) and 10 sites providing outpatient PR (505 participants). Due to incomplete records in the PR programmes, 33% of participants were labelled as having ‘fibrotic ILD’ without a specific diagnosis. In those with a specific diagnosis, 64% had IPF, 7% had fibrosing HP, 6% had idiopathic non-specific interstitial pneumonia, 14% had unclassifiable ILD and 9% had other specific ILDs. IPF was more common in those undergoing inpatient PR. Prednisone was more frequently used in participants attending inpatient PR than outpatient PR (42% vs 14%). Antifibrotic medication was used in 36 (30%) inpatient and 45 (31%) outpatient participants with a diagnosis of IPF. Mycophenolate mofetil and azathioprine were used in 19 inpatient (10%) and 26 outpatient participants (7%). Long-term oxygen therapy was prescribed in 54% of inpatient and 27% of outpatient participants. Baseline 6MWD was 262±128 m (54±25 %-predicted) for inpatient participants and 358±125 m (77±28 %-predicted) for outpatient participants (figure 1A). Linear mixed models demonstrated that 12% of the variance in baseline 6MWD was explained by the individual centre and 16% by the modality of PR (inpatient vs outpatient).
Programme characteristics and adherence
Median duration of inpatient and outpatient PR was 3 and 8 weeks, respectively. On average, inpatient programmes had 10.6 and outpatient programmes 2.3 planned sessions per week. More than 80% of sessions were completed by 74% of outpatient PR participants, with participation not routinely assessed for inpatient programmes. The 80 participants (11%) who lacked a follow-up 6MWT had similar demographic characteristics, but slightly poorer baseline pulmonary function (eg, mean FVC 65% vs 68% and mean diffusing capacity of the lung for carbon monoxide (DLCO) 39% vs 47%) and 6MWD (mean 269 vs 339 m) compared with those who completed follow-up (online supplemental table S1).
Outcomes of PR
The mean change in 6MWD was larger following inpatient (55±83 m) compared with outpatient PR (34±65 m), with 65% and 52% of inpatient and outpatient PR participants having a ≥30-metre increase in 6MWD, respectively (table 2 and figure 1B). Higher baseline 6MWD was the strongest factor associated with smaller change in 6MWD during PR (inpatient PR r=−0.30 and outpatient PR r=−0.16, both p<0.001). Other variables associated with change in 6MWD are described in online supplemental table S2. Only 1% of the variance in change in 6MWD during PR was explained by the individual centre and 3% by the modality of PR (inpatient vs outpatient).
During the median (IQR) follow-up time of 25 (10–52) months after PR, 267 patients with fibrotic ILD died, and 33 received a lung transplant. One-year and 2-year survival were 79% (74%–86%) and 63% (55%–71%) after inpatient PR, and 88% (95% CI 86% to 91%) and 77% (95% CI 73% to 81%) after outpatient PR (online supplemental figure S1). Variables associated with longer survival after PR included younger age, female sex, never smoking, higher body mass index, higher FVC and DLCO %-predicted, no supplemental oxygen dependency, higher baseline 6MWD, shorter duration of inpatient PR, prednisone treatment and availability of 6MWT at PR completion (online supplemental table S3).
Survival after inpatient PR
Improvement in 6MWD during inpatient PR was associated with lower hazard rates for death or lung transplant on unadjusted analysis (HR 0.97, 95% CI 0.94 to 0.99, p=0.02). This association was stronger after adjustment for baseline 6MWD (HR 0.94, 95% CI 0.91 to 0.97, p=0.007), and in a multivariable model that included baseline 6MWD as well as age and sex as fixed effects (HR 0.94, 95% CI 0.91 to 0.97, p<0.001; table 3). Improvement in 6MWD %-predicted was similarly associated with survival after inpatient PR (online supplemental table S5). Similar findings were also observed for prediction of 1-year, 2-year and 3-year survival (online supplemental table S4).
Risk of death for participants exceeding various thresholds of change in 6MWD during PR is shown in figure 2A. Stratifying by the generally accepted MID of 6MWD, risk of death was one-third lower in inpatient PR participants who increased their 6MWD by ≥30 m compared with the remainder (HR 0.66, 95% CI 0.42 to 1.02, p=0.06; Figure 3A). Further stratifying by baseline 6MWD, the subgroup with baseline 6MWD ≥350 m and increase in 6MWD ≥30 m had a 68% lower hazard for early mortality compared with the subgroup with baseline 6MWD <350 m and increase in 6MWD <30 m (HR 0.32, 95% CI 0.15 to 0.69, p=0.003; Figure 3C).
Survival after outpatient PR
Improvement in 6MWD during outpatient PR was only associated with lower hazard rates for death or lung transplant after adjustment for negative confounding by baseline 6MWD (HR 0.97, 95% CI 0.95 to 0.99, p=0.03), with unchanged findings in a multivariable model that included baseline 6MWD as well as age and sex as fixed effects (HR 0.97, 95% CI 0.95 to 1.00, p=0.042; table 3). Corresponding findings for the association of change in 6MWD %-predicted with survival after outpatient PR are reported in online supplemental table S5. Similarly, improvement in 6MWD during outpatient PR was only associated with 1-year, 2-year and 3-year survival after accounting for baseline 6MWD, age and sex (online supplemental table S4).
Risk of death for participants exceeding various thresholds of change in 6MWD during PR is shown in figure 2B. Risk of death was similar in PR participants stratified by change in 6MWD at ≥30 m vs <30 m (figure 3B). Further stratification by baseline 6MWD ≥ vs <350 m discriminated survival rates after outpatient PR. For example, the subgroup with baseline 6MWD ≥350 m and increase in 6MWD ≥30 m has a 59% lower hazard for early mortality compared with the subgroup with baseline 6MWD <350 m and increase in 6MWD <30 m (HR 0.41, 95% CI 0.26 to 0.63, p<0.001; Figure 3D). Patients who participated in ≥80% of planned outpatient PR sessions had a lower risk of death than patients completing <80% of sessions (HR 0.67 (95% CI 0.49 to 0.92), p=0.01, figure 4).
A sensitivity analysis for time from completion of inpatient or outpatient PR to lung transplant or death did not reveal any change in the findings. Similarly, a modelling approach with the addition of statistically driven confounders (inpatient PR: IPF diagnosis, long-term oxygen therapy, PR duration; outpatient PR: ever smoker) resulted in unchanged findings (data not shown).
In the absence of randomised controlled trials investigating the impact of PR on survival in patients with fibrotic ILD, we conducted a study to evaluate whether indirect evidence exists to support this potential benefit. Using this large international cohort, we confirmed our a priori hypothesis that improvement in 6MWD following PR is associated with a lower risk of death in the subsequent 3 years. The relationship between larger increase in 6MWD during PR and lower risk of death after inpatient and outpatient PR was robust to adjustment for potential confounders. Overall, these findings support the hypothesis that PR has the potential to improve survival in patients with fibrotic ILD.
There are only a few small randomised controlled trials investigating the effect of outpatient PR on physical performance, with one study reporting no change in 6MWD on average,15 and most previous studies reporting mean improvement in 6MWD ranging from 29 m to 70 m.8 12 13 19 30–32 The only randomised controlled trial of inpatient PR included 24 patients and reported a mean difference in 6MWD of 61 m in favour of the PR group.16 Our study provides the most precise real-world estimate of benefit from PR by analysing a large cohort of more than 700 patients with fibrotic ILD participating in PR, including patients located in 12 centres, 5 countries and 3 continents. Participants improved their 6MWD by an average of 55 m after inpatient PR and 34 m after outpatient PR, with a smaller increase in participants with a higher 6MWD at PR initiation. A likely explanation for this finding is the ceiling effect of 6MWT, which has previously been demonstrated and discussed in this and other contexts.11 12 17 28
Beyond the expected increase in 6MWD during PR, we demonstrate that the magnitude of this improvement is associated with greater survival that persists up to 3 years after PR. Although our results were robust to adjustment for multiple potential confounders and consistent across multiple centres and types of PR, the causality of this relationship cannot be definitively established without a randomised controlled trial. Given the challenges in conducting such a trial, the evidence reported in this study provides what may remain as the strongest evidence of the link between PR success and survival in patients with fibrotic ILD. It is hoped that this evidence provides further justification to increase the resources for and the availability of PR so that more patients can benefit from this intervention.
There are several potential reasons for the stronger association between improvements in 6MWD with survival in inpatient versus outpatient PR. For example, patients with fibrotic ILD who are admitted for inpatient PR may have recently suffered from acute worsening of their disease that led to hospitalisation and rapid deconditioning. An improvement of 6MWD in this setting reflects successful recovery from this ‘acute’ deconditioning during PR, which is likely a good prognostic sign. In other patients, the deterioration leading to inpatient PR admission may be due to irreversible disease progression. These PR participants probably have less benefit from PR, little or no increase in 6MWD, and a worse long-term prognosis. In contrast, outpatient PR likely includes a more homogeneous population of patients with less severe disease and a more gradual functional decline. These patients experience a more consistent, but less dramatic improvement in 6MWD during PR, with a weaker overall association of this improvement with long-term survival compared with inpatient PR participants. Examining these possibilities requires a detailed understanding of pre-PR disease status and trajectory, which was beyond the scope of our study. An additional possibility is that there are programme-related differences between inpatient and outpatient PR that impact change in 6MWD and its association with survival; however, testing this would require more direct head-to-head prospective comparison of these two modalities, which is likely impractical.
We found that 11% of participants had missing 6MWD at the end of outpatient PR, with these patients having more severe disease and most of these patients dropping out of PR early. Although having complete data would be ideal, we importantly found that patients who complete more than 80% of outpatient PR sessions had a lower hazard for early mortality. Although there is potential for significant confounding (eg, a generally healthier lifestyle may correlate with greater adherence to PR and other interventions), this emphasises the importance of tailoring PR to the specific needs of each patient and identifying potential reasons for non-adherence. PR maintenance programmes should similarly be refined for patients with ILD, and future research is needed to develop and implement tools facilitating adherence to ongoing supervised exercise training in this population. A specific ILD diagnosis was not available in one-third of patients, although previous studies have suggested that patients with different fibrotic ILDs respond similarly to PR,30 and the missing diagnoses are thus unlikely to significantly impact our main findings. Lastly, the most important limitation of this study is its retrospective nature, which limits establishment of cause and effect, specifically for our suggestion that PR improves survival in patients with fibrotic ILD. A multinational randomised controlled trial would be needed to resolve this issue, but this is likely not practicable given the substantial expense of such a trial and the lack of clinical equipoise for PR in patients with ILD.7 25
In summary, this large multinational retrospective cohort confirms a clinically relevant mean improvement in physical performance during inpatient and outpatient PR in patients with fibrotic ILD. The most successful PR participants had the lowest mortality in the subsequent years, which supports the hypothesis that PR has the potential to improve survival in patients with fibrotic ILD. These hypothesis-generating findings would ideally be confirmed in a randomised controlled trial, but we hope that in the meantime this study will further motivate efforts to improve availability of PR for patients with fibrotic ILD.
Data availability statement
Data are available upon reasonable request. Data are available upon reasonable request to the corresponding author.
Patient consent for publication
Ethics committees of all individual centres approved the study based on the decision of the main centre's ethical approval (UBC-PHC Research Ethics Board, ID H17-01997).Approval for this study was obtained from each local research ethics board.
Contributors SAG, SAH and CJR designed the study. SAG, SAH, MKS, PB, LB, AEH, JB, NH, JW, NM, MK, RG, IJ, BT, KAJ, SAM, KDB, JSS, KS, DA, SPB, JM, VF, CG, PGC and CJR contributed to data acquisition. SAG and CJR contributed to analysis of the data and drafting of the manuscript. All authors critically revised and finally approved the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.