Ryrsø et al.

BMC Pulmonary Medicine (2018) 18:154

https://doi.org/10.1186/s12890-018-0718-1

RESEARCH ARTICLE Open Access

Lower mortality after early supervised

pulmonary rehabilitation following COPD-
exacerbations: a systematic review and
meta-analysis
Camilla Koch Ryrsø1,2* , Nina Skavlan Godtfredsen3,4, Linette Marie Kofod5, Marie Lavesen6, Line Mogensen7,
Randi Tobberup8, Ingeborg Farver-Vestergaard9, Henriette Edemann Callesen2, Britta Tendal2, Peter Lange1,10,11
and Ulrik Winning Iepsen1

Abstract
Background: Pulmonary rehabilitation (PR), delivered as a supervised multidisciplinary program including exercise
training, is one of the cornerstones in the chronic obstructive pulmonary disease (COPD) management. We performed a
systematic review and meta-analysis to assess the effect on mortality of a supervised early PR program, initiated during
or within 4 weeks after hospitalization with an acute exacerbation of COPD compared with usual post-exacerbation care
or no PR program. Secondary outcomes were days in hospital, COPD related readmissions, health-related quality of life
(HRQoL), exercise capacity (walking distance), activities of daily living (ADL), fall risk and drop-out rate.
Methods: We identified randomized trials through a systematic search using MEDLINE, EMBASE and Cocharne Library
and other sources through October 2017. Risk of bias was assessed regarding randomization, allocation sequence
concealment, blinding, incomplete outcome data, selective outcome reporting, and other biases using the Cochrane
Risk of Bias tool.
Results: We included 13 randomized trials (801 participants). Our meta-analyses showed a clinically relevant reduction
in mortality after early PR (4 trials, 319 patients; RR = 0.58 (95% CI: [0.35 to 0.98])) and at the longest follow-up (3 trials,
127 patients; RR = 0.55 (95% CI: [0.12 to 2.57])). Early PR reduced number of days in hospital by 4.27 days (1 trial, 180
patients; 95% CI: [− 6.85 to − 1.69]) and hospital readmissions (6 trials, 319 patients; RR = 0.47 (95% CI: [0.29 to 0.75])).
Moreover, early PR improved HRQoL and walking distance, and did not affect drop-out rate. Several of the trials had
unclear risk of bias in regard to the randomization and blinding, for some outcome there was also a lack of power.
Conclusion: Moderate quality of evidence showed reductions in mortality, number of days in hospital and number of
readmissions after early PR in patients hospitalized with a COPD exacerbation. Long-term effects on mortality were not
statistically significant, but improvements in HRQoL and exercise capacity appeared to be maintained for at
least 12 months. Therefore, we recommend early supervised PR to patients with COPD-related exacerbations.
PR should be initiated during hospital admission or within 4 weeks after hospital discharge.
Keywords: Chronic obstructive pulmonary disease, Supervised early pulmonary rehabilitation, Exacerbation of
COPD, Hospital readmissions, Mortality, Systematic review

* Correspondence: camilla.koch.ryrsoe@regionh.dk; http://www.inflammation-

metabolism.dk; http://aktivsundhed.dk
1
The Centre of Inflammation and Metabolism and the Centre for Physical
Activity Research, Rigshospitalet, University of Copenhagen, Blegdamsvej 9,
DK-2100 Copenhagen, Denmark
2
Danish Health Authority, Copenhagen, Denmark
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 2 of 18

Background 4 weeks, in patients hospitalized with a COPD exacerba-

Acute exacerbation in chronic obstructive pulmonary dis- tion compared with usual care. Our primary outcome
ease (AECOPD) is the most common reason for hospital was mortality at the end of PR and at the longest
admission among patients with chronic obstructive pul- follow-up. Secondary outcomes were hospital readmis-
monary disease (COPD) [1]. These events result in higher sion, days in hospital, HRQoL and exercise capacity. We
mortality and lower quality of life [2]. The estimated mor- followed the guidelines of the Grading of Recommenda-
tality rate within 90 days after hospitalization for AECOPD tions Assessment, Development and Evaluation (GRADE)
is approximately 3.6% (1.8–20.4%) while mortality rate dur- Working Group [19] in order to support clinical decision
ing the first 2 years after admission for AECOPD is making in a national Danish setting where only supervised
approximately 31.0% (18.8–45.4%) [3]. The estimated PR programs take place.
30-day and 12-month readmission-rate after AECOPD
hospitalization is approximately 19.2% [4] and 42.3% [5], re- Methods
spectively. Readmission following an AECOPD has a nega- Protocol and registration
tive effect on physical performance by lowering exercise This review was among a series of reviews performed for
capacity, muscle strength and physical activity level, which a guideline developed by the Danish Health Authority.
patients may never fully recover from [6, 7]. Patients with The population, intervention, control intervention (com-
frequent exacerbations may be prone to a more rapid parison) as well as critical and important outcomes
decline in activities of daily living (ADL) and functional (PICO) [20] were determined by the working-group mem-
capacity, which is associated with reductions health-related bers prior to our literature search. The methods and
quality of life (HRQoL) [6]. Repeated exacerbations may review process are a standardized part of the guideline
cause a vicious circle as physical inactivity and low exercise development process within the Danish Health Authority.
capacity are related to a higher risk of hospital readmission, The methods handbook (in Danish) as well as the full
regardless of the COPD severity [8]. guideline and more detail regarding the PICO can be
Pulmonary rehabilitation (PR) has been suggested in accessed at www.sst.dk, the full guideline can also be
AECOPD because of its known beneficial effects on found on https://app.magicapp.org/app#/guideline/2551.
exercise capacity, HRQoL and symptom burden in stable
patients [9, 10]. It should be noted that the evidence in Eligibility criteria
favor of PR in stable COPD is based on studies investi- We considered studies eligible if they compared the ef-
gating supervised PR programs including exercise train- fect of early supervised PR initiated during admission or
ing for 6–12 weeks [11, 12], although at long-term within 4 weeks of hospital discharge (intervention) with
follow-up, adherence to exercise training is low and no early pulmonary rehabilitation/usual care (compari-
effects are not maintained [13]. Likewise, studies have son) in patients admitted and/or having been admitted
shown that early PR, initiated at the beginning of exacer- to hospital with exacerbations of COPD (population).
bation treatment or within 3 weeks of initiation of The PR was defined by a main component of supervised
exacerbation treatment, improves exercise capacity and exercise training but could contain education, smoking
HRQoL along with reductions in hospital readmissions cessation, nutritional support, management in activities
[14] and mortality [15] compared with no PR. Based on of daily living (ADL) and physio-social support.
evidence from randomized controlled trials (RCT), NICE Studies providing inpatient pulmonary rehabilitation
guidelines from 2011 recommended the use of early PR with exercise training was included if rehabilitation were
in patients hospitalized with COPD exacerbations [16]. continued after hospital discharge and/or a comprehen-
Yet, recent concerns have been raised about PR not be- sive rehabilitation program could be documented. Studies
ing safe in AECOPD when initiated during the hospital were excluded if they were not randomized or did not
admission [17]. Based on this new evidence, the 2017 cover the predefined PICO. Our pre-specified outcomes
guideline from the European Respiratory Society (ERS) were evaluated immediately after the end of intervention
and American Thoracic Society (ATS) recommend that and at the longest follow up. Our primary outcome was
PR is not initiated during hospitalization in patients with mortality while secondary outcomes included number of
COPD related exacerbations, but is delayed until after days in hospital, number of COPD related hospital read-
hospital discharge (within 3 weeks) [18]. However, the missions, health related quality of life (HRQoL), exercise
ERS/ATS recommendation was based on both super- capacity (walking tests), activities of daily living (ADL),
vised and unsupervised PR programs, and interestingly, falls and dropout.
the potentially negative effects of early PR were mainly
driven by studies providing unsupervised PR. Information sources and search strategy
Therefore, our aim was to investigate the effect of a A research librarian and search specialist performed the sys-
supervised early PR program, initiated during or within tematic literature search including the following databases:
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 3 of 18

Medline, Embase, Physiotherapy Evidence Database (PE- Study selection

DRO), CINAHL, G-I-N international, NICE, National Clinical guidelines identified in the first search were
Guideline Clearinghouse, Surgical Implant Generation evaluated with the Appraisal of Guidelines for Research
Network, Cochrane Library and OTseeker. The full search and Evaluation instrument version II (AGREE II) by two
strategy is presented in Additional file 1. independent authors and disagreement was resolved
This review is an update of a previous review. First, a through consensus (see Additional file 2). Likewise, sys-
comprehensive search for COPD rehabilitation guide- tematic reviews were assessed with A Measurement
lines and systematic reviews was conducted in July Tool to Asses Systematic Reviews (AMSTAR) by three
2013, yielding a total of 2412 records. In November authors and disagreement was resolved through consen-
2013, a second and more specific search (Medline, sus (see Additional file 3). Based on these assessments
Embase and PEDRO) for RCTs was performed, in we decided to include one clinical guideline [21] and
which 876 additional records were identified. An up- two systematic reviews [14, 15]. From the second search,
dated search for guidelines and systematic reviews was two authors independently evaluated the full text of all
conducted in July 2017, covering the period from July potentially eligible studies and decided whether to in-
2013 to July 2017, where we identified 460 additional clude or exclude each study based on the prespecified
records. The search for primary studies was updated in criteria.
October 2017, covering the period from December
2013 to October 2017. The search resulted in 1187 add- Data analysis and risk of bias
itional records (Fig. 1). All records were screened for Data on participants, study design, interventions and out-
relevant titles or abstracts, while reference lists of comes were extracted from the full-text reports of the in-
included studies were assessed for further eligible cluded studies by two independent authors, using
literature. Covidence (Covidence systematic review software, Veritas

Fig. 1 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram of the article selection processes
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 4 of 18

Health Innovation, Melbourne, Australia. Available at one to four weeks after the inpatient exacerbation treat-
www.covidence.org). Disagreement was resolved through ment and in one study [39] the outpatient rehabilitation
consensus. Each included study was assessed using the was initiated after the “hospital at home” treatment of
Cochrane risk of bias tool [22]. Two independent authors the exacerbation. In four studies [27, 29, 38, 39] the PR
performed the risk of bias assessment, and disagree- consisted of only supervised exercise training, whereas
ment was resolved through discussion and consensus in the remaining nine studies [28, 30–37] PR consisted
(see Additional file 4). of supervised exercise training and education, smoking
We used mean difference (MD) to calculate effect esti- cessation, nutritional support, management in activities
mates for continuous outcomes if the same scale was of daily living (ADL) and physio-social support. Duration
used for a particular outcome. When pooling continuous of the different PR programs was ten days to six months,
outcome data from different scales a standardized mean with training frequencies ranging from two to seven
difference (SMD) was calculated. Rate ratio and relative times a week, and exercise durations of 30–90 min per
risk (RR) was used to calculate effects for dichotomous session. Table S1 in the Additional file 5 shows the
outcomes. Random-effects meta-analyses were performed extensiveness of the PR programs in the included trials.
as we expected variation in population, duration of inter- The participants followed an extensive PR program in
vention, and types of training between the included stud- ten of the included trials [27–31, 33–36, 38]. In the
ies. Review Manager 5.3 software [23] was used for the remaining three studies, the extensiveness of the PR was
statistical analyses and to produce forest plots. Heterogen- deemed as moderate [39], slightly extensive [37], and
eity in the effect estimate was determined using the undescribed [32]. The control group received usual care
I-square (I2) statistic and values below 40% indicated low consisting of optimal medical treatment. There were no
heterogeneity [24]. reported differences in baseline characteristics of pa-
The quality of evidence for each outcome was assessed tients between groups in all of the included studies.
across the included studies as proposed by the GRADE
Working Group [25]. A draft of the grading for each out- Risk of bias within studies
come using the GRADE criteria (i.e., risk of bias, inconsist- Figures 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 and 12 and Additional
ency, indirectness, imprecision, and publication bias) was file 4 shows risk of bias of the included studies. In nine
presented to the working group and the final grading was studies [28, 29, 31, 32, 34–38] the allocation conceal-
reached through discussion and consensus. The full guide- ment was not described, while seven studies [27, 29, 31,
line was then submitted to peer review and public hearing. 36–39] did not report the randomization process. Three
For details on the hearing see www.sst.dk (in Danish). studies [27, 34, 39] blinded the personnel, with only two
of the studies [34, 39] blinding the outcome assessor.
Assessment of PR extensiveness One study [27] was assessed as having a high risk of in-
We assessed the extensiveness of the PR program in the complete outcome data reporting due to a large dropout.
included trials by following the statements and guide- Selective outcome reporting of outcome measures was
lines from BTS [26], ERS/ATS [10]), and as described in detected in one study [34]. No other sources of bias
Puhan et al. [14] (see Additional file 5). were detected. Thus, the quality of evidence from all
studies included was downgraded due to risk of bias
Results (Table 2, Additional file 4).
Study selection
We identified 13 eligible primary RCTs for our analysis. Effect of the intervention
These included a total of 801 participants who were in the We preformed meta-analyses in ten of our predefined
recovery phase of a recent COPD exacerbation. Excluding outcomes. Subgroup analyses were undertaken in order
dropouts (167 participants), 634 participants were included to reveal differences between PR initiated during admis-
in our analysis. Nine of the 13 studies were included in a sion or within one week after discharge and PR initiated
systematic Cochrane review [14]. Figure 1 summarizes the between one and four weeks after discharge from hos-
flow diagram of the two selection processes. pital. For an overview of all the outcomes, our confi-
dence in the estimates and our interpretations see Table
Included studies 2 GRADE Evidence profile.
Table 1 shows the characteristics of the included studies.
In three studies [27–29] patients initiated an inpatient Mortality
PR program within 4 to 8 days of hospital admission. In Total mortality after end of treatment was reported in
one study [30] patients began PR as either in- or outpa- four of the included studies, including 319 randomized
tients and all continued as outpatients, in eight studies participants (early PR: N = 163; usual care: N = 156) [29,
[31–38] the outpatient program was initiated within 32, 33, 37]. A total of 18 events were reported in the
Table 1 Characteristics of the included studies
Reference Country Study Setting, duration Participants Intervention Intervention after Usual care Notes Outcomes Dropouts
design and frequency discharge
Behnke Germany RCT Setting: in- and 46 admitted patients PR consisted of Supervised home- Usual care: Both groups Mortalityb 16 dropouts (8 in
2000 [27] outpatient with AECOPD (mean conventional therapy based training for 6 standard inpatient (intervention and Walking testb PR group and 8 in
Duration: hospital- age: 64–68 years, FEV1: including 30 min of daily mo.: walking care and usual care) were COPD related control group)
based 10 days, 36% predicted). breath exercises with training 3/day at community care supervised by the hospital
home-based Comorbidities: not respirologists and hospital- 125% of the best with respirologists physician. readmissionsb
6 months specified. based training. Exercise 6MWD, health (30 min of daily Dropouta
Frequency: 7/week training consisted of daily check every 2 weeks breathing exercises)
6MWT and 5 self-controlled (mo. 0–3) followed but without
walking sessions at 75% of by phone calls exercise training
the treadmill walking from mo. 3–6.
distance of the respective
day.
Daabis Egypt RCT Setting: outpatient 30 admitted patients PR consisted of patient Outpatient PR Medical treatment. All patients HRQoLa No dropouts
2017 [31] Duration: 8 weeks with AECOPD (mean assessment, exercise received standard Walking reported
Frequency: 3/week age: 58–61 years, FEV1: training (ET), patient treatment with distancea
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154

53–56% of predicted). education including self- optimal medical

Comorbidities: not management of the treatment.
specified. disease, nutrition and
lifestyle issues. Exercise
training consisted of ET
with 30-min of walking at
the intensity of 75% 6MWT
including 30-min of low-
intensity RT.
Deepak India RCT Setting outpatient 60 admitted patients PR consisted of patient Outpatient PR Conventional All patients HRQoLa 4 dropouts
2014 [32] Duration: 12 weeks with AECOPD (mean assessment, exercise treatment without received Walking
age: 59 years, FEV1: 47– testing, exercise training PR. conventional distancea
53% of predicted, 93% (mixture of limb management
men). Comorbidities: strengthening and aerobic consisting of
not specified. activities, tailored to medical treatment.
individual baseline
function), education,
nutrition and psycho-social
rehabilitation.
Eaton New RCT Setting: in- and 97 admitted patients PR consisted of a daily 30- Hospital-based Usual care All patients Walking 13 dropouts (8 in
2009 [28] Zealand outpatient with AECOPD (mean min structured, supervised outpatient program standardized in received usual care distancea PR group and 5 in
Duration: 8 weeks age: 70 years, FEV1: 35– exercise regimen that consisting of 1-h according with the standardized in COPD related control group)
Frequency: 2/week 36% of predicted, 42– included walking and upper sessions of ATS/ERS COPD according with the hospital
45% men). and lower limb supervised exercise guidelines and ATS/ERS COPD readmissionsb
Comorbidities: strengthening exercises. training and standardized guidelines. Dropouta
Measured with educational advises on the
Charlson index (PR sessions (e.g. benefits of exercise
group: 3.1; control: 3.2). coping with and maintaining
dyspnea, daily activities.
management of
ADL, nutritional
advises, airway
clearance).
Kirsten Germany RCT Setting: inpatient 31 admitted patients PR consisted of 6MWT each Inpatient Usual care with All patients Walking testa 2 dropouts (not
1998 [29] Duration: 10 days with AECOPD (mean day and additional 5 supervised walking optimal medical received standard reported in which
Page 5 of 18
Table 1 Characteristics of the included studies (Continued)
Reference Country Study Setting, duration Participants Intervention Intervention after Usual care Notes Outcomes Dropouts
design and frequency discharge
Frequency: 7/week age: 62–66 years, FEV1: walking sessions per day at sessions 5/day. treatment. medical treatment. group)
34–38% of predicted, ≥75% of the respective
90% men). walking distance.
Comorbidities: not
specified.
Ko 2011 China RCT Setting: outpatient 60 admitted patients PR consisted of supervised Supervised Usual care with Both groups were HRQoLb 9 dropouts (5 in
[34] Duration: 8 weeks with AECOPD (mean exercise training including outpatient exercise instructions to seen by the nurse Mortalitya,b PR group and 4 in
Frequency: 3/week age: 73–74 years, FEV1: treadmill, arm cycling, arm training. perform regular specialist at the Walking testb control group) at
41–46% of predicted, and leg strength training at exercise at home baseline Dropouta,b the end of
98% men). 60–70% of VO2max or HRmax (walking and assessment. treatment. 6
Comorbidities: coronary and were advised to muscle stretching dropouts (2 in PR
artery disease, cardiac perform at least 20 min exercise). group and 4 in
arrhythmic, heart home exercises a day. control group) at
failure, hypertension, Education on proper the longest
diabetes. breathing techniques and follow-up.
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154

how to cope with daily

activities.
Ko 2017 China RCT Setting: outpatient 180 admitted patients PR consisted of education Patients are offered Usual care with All patients HRQoLb 38 dropouts (17 in
[33] Duration: 8 weeks with AECOPD (mean (smoking cessation, supervised exercise medical treatment. received standard Mortalitya PR group and 21
(1 year follow up) age: 75 years, FEV1: 42– technique of using training 3/week, if treatment with Walking testb in control group)
Frequency: 3/week 47% of predicted, 94– medications, nutrition, declining they are optimal medical Days in
97% men). dyspnea management, self- offered instructions therapy. hospitala
Comorbidities: management, psychological for self-training,
hypertension, type 2 distress, exercise benefits education, and
diabetes, and strategies, breathing telephone calls.
hyperlipidemia, and sputum-removal
ischemic heart disease, techniques) and individual
heart failure, old physical training program
pulmonary tuberculosis. to perform at home or a
short course of outpatient
PR.
Man England RCT Setting: outpatient 42 admitted patients Supervised multidisciplinary Supervised Usual care with All admitted HRQoFb 8 dropouts (3 in
2004 [35] Duration: 8 weeks with AECOPD (mean PR, 1-h of exercise (aerobic multidisciplinary PR. optimal medical patients received Mortalityb PR group and 5 in
Frequency: 2/week age: 70 years, FEV1: 37– walking and cycling, treatment. standard treatment Walking testb control group)
42% of predicted, 40% strength training for the and home diaries COPD related
men). Comorbidities: upper and lower limb) and which included a hospital
not specified. 1-h of education (with an disease specific readmissionsb
emphasis on self- information pack. Dropouta
management of the
disease, nutrition and
lifestyle issues).
Murphy Ireland RCT Setting: outpatient 31 admitted patients PR consisted of 30–40-min Supervised home- Standard medical All patients Walking testa 5 dropouts (3 in
2005 [39] home-based with AECOPD (mean supervised home-based based exercise treatment without received standard COPD related PR group and 2 in
Duration: 6 weeks age: 65–67 years, FEV1: exercise program, aerobic program. any form of PR medical treatment. hospital control group)
Frequency: 2/week 38–42% of predicted, exercises including step- exercises or lifestyle readmissionsb
65% men). ping up and down a stair, changes advice. Dropouta
Comorbidities: not sitting to stand from a
specified. chair, upper limb strength
exercises with low-impact
elastic band at 3–5 on the
Page 6 of 18
Table 1 Characteristics of the included studies (Continued)
Reference Country Study Setting, duration Participants Intervention Intervention after Usual care Notes Outcomes Dropouts
design and frequency discharge
Borg breathlessness score.
Puhan Switzerland RCT Setting: in- and 36 admitted patients Early inpatient PR within Outpatient PR, Late PR starting 6 Recommended Mortalitya 8 dropouts (4 in
2012 [30] outpatient with AECOPD (mean 2 weeks after exacerbation, exercise training mo. after number of exercise Dropouta PR group and 4 in
Duration: 12 weeks age: 67 years, FEV1: 43– exercise training included included exacerbation, session 24 (ranged control group)
Frequency: 24 46% of predicted, 58% endurance, strength and endurance, exercise training between 18 and
sessions (range 18– men). Comorbidities: calisthenics training in strength and included 36).
36) cardiovascular, addition with education calisthenics training endurance,
endocrine, (e.g. individual action plan, in addition with strength and
musculoskeletal, other. mediational use, exercise at education (as calisthenics training
home, coping with daily described under in addition with
activities, smoking intervention). education.
cessation).
Revitt United RCT Setting: inpatient 28 admitted patients Early PR within 4 weeks of Hospital-based PR. Late PR initiated All patients Dropouta 11 dropouts (3 in
2018 [37] Kingdom Duration: 6 weeks with AECOPD (mean discharge. PR consisted of 7 weeks after received the same control group
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154

Frequency: 2/week age: 66 years; FEV1: individualized aerobic and discharge including PR program. prior to the
1.18 l). Comorbidities: resistance exercises and exercise and program and 8 in
not specified. education on chest education. PR group during
clearance and energy the program)
conservation.
Seymour United RCT Setting: outpatient 60 admitted patients PR consisted of supervised Hospital-led Usual care with All patients were HRQoFb 11 dropouts (7 in
2010 [36] Kingdom (hospital-led) with AECOPD (mean exercise training including a supervised exercise optimal medical provided with Walking testa PR group and 4 in
Duration: 8 weeks age: 65-67 years, FEV1: mixture of limb training. treatment. general information COPD related control group)
Frequency: 2/week 52% of predicted, 45% strengthening and aerobic about COPD and hospital
men). Comorbidities: activities tailored to offered outpatient readmissionsb
hypertension, type 2 individual baseline function appointments with Dropoutb
diabetes, ischemic heart and education session their general
disease. (lasting 2 h). practitioner or
respiratory team.
Troosters Belgium RCT Setting: outpatient 100 patients with PR consisted of 90-min Supervised Usual medical care During exercise Mortalitya 30 dropouts (13 in
2000 [38] Duration: 6 mo (18 AECOPD referred to supervised ET and RT. ET outpatient exercise consisting of training walking testa PR group and 17
mo follow up) outpatient clinic (mean consisting of cycling, training. standard supplemental dropouta,b in control group)
Frequency: 2–3/ age: 60–63 years, FEV1: treadmill walking, and stair community care oxygenwas given at the end of
week 41–43% of predicted, climbing at 60–80% of with respirologist. to maintain oxygen treatment. 21
87% men). initial Wmax during cycle saturation above dropouts (11 in PR
Comorbidities: not ergometer/maximal walking 90%. group and 10 in
specified. speed. RT consisting of control group) at
strength exercises for 5 the longest
muscle groups, 10 reps at follow-up.
60% 1RM.
AECOPD acute exacerbations of chronic obstructive pulmonary disease, COPD chronic obstructive pulmonary disease, CT combined training, ET endurance training, FEV1 forced expiratory volume in 1 s,
HRmax maximum heart rate, HRQoL health related quality of life, RCT randomized controlled trial, 1RM one repetition maximum, RT resistance training, Reps repetitions, VO2max maximal oxygen uptake,
Wmax maximal work load in Watts, 6MWD 6 min walking distance, 6MWT 6 min walking test
a
After end of treatment
b
After longest follow up
Page 7 of 18
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 8 of 18

Fig. 2 The effect of supervised early PR versus usual care on mortality at the end of treatment.

early PR group, whereas 27 events were reported in 34]. Two events were reported in the early PR groups
the usual care group. We found a statistically signifi- while four events were reported in the usual care
cant reduction in mortality favoring PR (RR = 0.58 group. We found no statistical significant difference
(95% CI: [0.35 to 0.98])), with low heterogeneity between groups (RR = 0.55 (95% CI: [0.12 to 2.57])).
(Fig. 2). The quality of evidence was downgraded due Subgroup analysis showed no difference in effect be-
to unclear sequence generation, allocation conceal- tween trials with PR initiated during admission and
ment and blinding together with selective outcome after discharge (P = 0.70) (Fig. 3). Our confidence in
reporting (Table 2). the effect estimate was downgraded due to unclear
Total mortality at longest follow up was reported in sequence generation and allocation concealment to-
three of the included studies, including 127 partici- gether with lack of precision, incomplete outcome
pants (early PR: N = 64; usual care: N = 63) [26, 33, data and selective reporting (Table 2).

Fig. 3 The effect of supervised early PR versus usual care on mortality at the longest follow up
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 9 of 18

Fig. 4 The effect of supervised early PR versus usual care on days in hospital at the end of treatment

Days in hospital estimate showed a decrease in the number of COPD related

One study investigated the effect of early PR on the num- hospital readmissions favoring the early PR (RR = 0.47 (95%
ber of days in hospital after the end of treatment and CI: [0.29 to 0.75])). Low heterogeneity (I2 = 38%) was ob-
stated that early PR led to a statistically reduction of served, and the subgroup analysis showed no difference in
4.27 days (95% CI: [− 6.85 to − 1.69]) in the number of effect between trials with PR initiated during admission and
days in hospitals (Fig. 4). Accordingly, our confidence in after discharge (P = 0.93) (Fig. 5). The quality of evidence
the effect estimate was downgraded due to inclusion of was downgraded due to unclear sequence generation and
only one study (Table 2). allocation concealment together with lack of blinding and
incomplete outcome date (Table 2).
COPD related hospital readmissions
Six studies provided data from 365 participants on the num- Health-related quality of life
ber of COPD related hospital readmissions 3–12 months The St. George’s Respiratory Questionnaire (SGRQ)
from baseline [27, 28, 33, 35, 36, 38]. The pooled effect (scale from 0 to 100, lower is better) were used across

Fig. 5 The effect of supervised early PR versus usual care on COPD related hospital readmissions at the longest follow up
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 10 of 18

Fig. 6 The effect of supervised early PR versus usual care on health-related quality of life at the end of treatment using the St. George’s
Respiratory Questionnaire

studies to assess HRQoL. Two studies were included group compared with the usual care group (Fig. 7). Sub-
and data from 86 participants were pooled in a meta- group analysis showed no difference in effect between
analysis evaluating HRQoL directly after end of early PR trials with PR initiated during admission and after
[31, 32] and showed a statistically and clinically signifi- discharge (P = 0.49). Unclear sequence generation, allo-
cant improvement of 19.43 units on the SGRQ scale cation concealment, blinding and selective outcome
(95% CI: [− 29.09 to − 9.77]) in the early PR group reporting led to downgrading of the confidence in our
compared with the usual care group (Fig. 6) with low effect estimates (Table 2).
heterogeneity. Our confidence in the effect estimate was
downgraded due to unclear sequence generation, alloca- Walking distance
tion concealment, blinding of assessors and incomplete The walking distance (6-Minute Walking Test (6MWT)
outcome data (Table 2). or Shuttle Walking Test (SWT)) after the end of treat-
Four different studies provided data from 323 partici- ment was investigated in eight studies [28, 29, 31, 32,
pants on the effect of early PR on HRQoL 3–12 months 36–39]. Pooling the results (early PR: N = 139; usual
from baseline [33–36] and showed a statistically and care: N = 135) from five trials using 6MWT yielded a
clinically relevant improvement of 8.74 units on the statistically significant mean difference in walking dis-
SGRQ scale (95% CI: [− 12.02 to − 5.45]) in the early PR tance of 76.89 m, favoring early PR (95% CI: [21.34 to

Fig. 7 The effect of supervised early PR versus usual care on health-related quality of life at the longest follow up using the St. George’s
Respiratory Questionnaire
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 11 of 18

Fig. 8 The effect of supervised early PR versus usual care on walking distance at the end of treatment using the 6-Minute Walking Test

132.45]) with high heterogeneity (Fig. 8). The subgroup SWT to evaluate the walking distance after the end of treat-
analysis showed no difference in the effect between PR ment and showed a statistically significant mean difference
initiated during admission and after discharge (P = 1.00). in walking distance of 54.70 m, favoring early PR (95% CI:
However, we found a significant within-group effect of [30.83 to 78.57]). The subgroup analysis showed no differ-
early PR after discharge (Fig. 8). The quality of evidence ence in the effect between PR initiated during admission
was downgraded due to unclear sequence generation, al- and after discharge (P = 0.40). However, we found a signifi-
location concealment, blinding of assessors and incomplete cant within-group effect of early PR during admission and
data together with high risk of inconsistency (Table 2). after discharge (Fig. 9). The quality of evidence was down-
Three trials (early PR: N = 50; usual care: N = 45) used the graded due to unclear sequence generation, allocation

Fig. 9 The effect of supervised early PR versus usual care on walking distance at the end of treatment using the Shuttle Walking Test
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 12 of 18

Fig. 10 The effect of supervised early PR versus usual care on walking distance at the longest follow up using the 6-Minute Walking Test

concealment, blinding of assessors, incomplete outcome to- difference (mean difference: 90.27 m; 95% CI: [− 69.53 to
gether with selective outcome reporting (Table 2). 250.08]) with high heterogeneity (Fig. 10). Subgroup ana-
Three different studies provided data from 217 partici- lysis showed a statistically significant difference between
pants on the effect of early PR on walking distance groups in favor of early PR during admission (P < 0.01)
assessed by 6MWT at 3–12 months from baseline [27, 33, (Fig. 10). Due to unclear sequence generation, allocation
34] and showed no statistically, but a clinically relevant concealment, blinding of assessors, incomplete data and

Fig. 11 The effect of supervised early PR versus usual care on dropout at the end of treatment
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 13 of 18

Fig. 12 The effect of supervised early PR versus usual care on dropout at the longest follow up

selective reporting together with high risk of inconsistency None of the included studies reported results on the
leading to high risk of imprecision the quality of evidence effect of early PR on ADL or the risk of falling.
was downgraded (Table 2).
Discussion
Summary of main findings
Drop-outs The present review summarized the evidence from 13
The effect of early PR on the drop-out rate at the RCTs including 634 participants with an exacerbation of
end of treatment was investigated in eight studies COPD and compared the use of early PR (N = 322) with
providing data from 440 randomized participants usual care (N = 312). Subsequent meta-analysis showed
(early PR: N = 228; usual care: N = 212) [27, 28, 30, that supervised early PR after acute exacerbation of
34, 35, 37–39]. A total of 54 drop-outs were reported COPD reduced mortality and number of days in hospital
in the early PR group, whereas 46 drop-outs were re- together with a reduction in COPD related hospital ad-
ported in the usual care group, with no significant missions and an improvement of HRQoL and exercise
difference between groups (RR = 0.99 (95% CI: [0.71 capacity (walking distance).
to 1.39])) (Fig. 11). The subgroup analysis showed no
difference in the effect between PR initiated during Mortality
admission and after discharge (P = 0.37). Our confi- We found that supervised early PR in patients with ex-
dence in the effect estimate was downgraded due to acerbation of COPD reduced risk of mortality by ~ 42%
unclear sequence generation, allocation concealment, compared with usual care. This finding was based on
blinding of assessors and incomplete data outcome moderate quality of evidence due to methodological is-
together with high risk of inconsistency (Table 2). sues in the included studies and the relatively small
Three different studies provided data from 181 partici- numbers of participants. While similar conclusions have
pants on the effect of early PR on drop-out at 3– been reported in guidelines and systematic reviews in
18 months from baseline (early PR: N = 92; usual care: the past, results from a resent RCT by Greening et al.
N = 89) [34, 36, 38]. A total of 20 drop-outs were re- questioned the beneficial effects by reporting higher
ported in the early PR group, while 18 drop-outs were mortality in the early PR group [15–17]. In this study
reported in the usual care group, with no difference be- authors included patients with COPD related exacerba-
tween groups (RR = 1.05 (95% CI: [0.60 to 1.85])) (Fig. 12, tions during admission and instructed participants in the
Table 2). Subgroup analysis showed no difference in the intervention group to be more physical active the next
effect between trials with PR initiated during admission three months facilitated by technical devices [17]. In
and after discharge (P = 0.30; I2 = 6.5%) (Fig. 12). contrast, the majority of evidence favoring PR in stable
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 14 of 18

Table 2 GRADE Evidence Profile

Supervised early PR versus usual care for patients with acute exacerbation of COPD
Outcome Timeframe Study results and Absolute effect Certainty in the effects Plain text summary
measurements estimates estimates (Quality of evidence)
Usual Early
care PR
Mortality Relative risk 0.58 173 100 Moderate Early pulmonary rehabilitation probably
End of treatment (CI 95% 0.35–0.98) per per Due to serious risk of biasa decreases mortality at the end of treatment
Critical Based on data from 319 1.000 1.000
patients (4 studies)
Difference: 73
fewer per 1.000
(CI 95% 112
fewer - 3
fewer)
Mortality Relative risk 0.55 63 35 Low Early pulmonary rehabilitation may decrease
Longest follow-up (CI 95% 0.12–2.57) per per Due to serious risk of bias mortality slightly at the longest follow-up
a,b
Critical Based on data from 127 1.000 1.000 and serious risk of imprecision
patients (3 studies)
Difference: 28
fewer per 1.000
(CI 95% 55
fewer - 99
more)
Days in hospital Measured by: Days 0.86 4.59 Moderate Early pulmonary rehabilitation probably
End of treatment Lower is better (mean) (mean) Due to serious imprecisionc decreases days in hospital at the end
Important Based on data from 180 of treatment
patients (1 study) Difference: MD
4.27 lower
(CI 95% 6.85
lower - 1.69
lower)
Days in hospital No studies were found that looked at
Longest follow-up number of days in hospital at the
Important longest follow-up
Readmission due to No studies were found that looked at
exacerbation readmission to hospital due to
End of treatment exacerbation at the end of treatment
Important
Readmission due to Rate ratio 0.47 Moderate Early pulmonary rehabilitation probably
exacerbation (CI 95% 0.29–0.75) Due to serious risk of biasa,d decreases readmission to hospital due to
Longest follow-up Based on data from 365 exacerbation at the longest follow-up
Important patients (6 studies)
Health-related quality Measured by: SGRQ Difference: MD Low Early pulmonary rehabilitation may improve
of life Lower is better 19.43 lower Due to serious risk of bias and health-related quality of life at the end
End of treatment Based on data from 86 (CI 95% 29.09 serious risk of imprecisiona,c of treatment
Important patients (2 studies) lower - 9.77
lower)
Health-related quality Measured by: SGRQ Difference: MD Moderate Early pulmonary rehabilitation probably
of life Lower is better 8.74 lower Due to serious risk of biasa,d improves health-related quality of life at
Longest follow-up Based on data from 323 (CI 95% 12.02 the longest follow-up
Important patients (4 studies) lower - 5.45
lower)
Exercise capacity Measured by: SWT (meters) Difference: MD Moderate Early pulmonary rehabilitation probably
End of treatment Higher is better 54.7 more Due to serious risk of biasa,d increases exercise capacity at the end
Important Based on data from 95 (CI 95% 30.83 of treatment
patients (3 studies) more - 78.57
more)
Exercise capacity Measured by: 6MWT Difference: MD Low Early pulmonary rehabilitation probably
End of treatment (meters) 76.89 more Due to serious risk of bias increases exercise capacity at the end
Important Higher is better (CI 95% 21.34 and serious inconsistencya,d,e of treatment
Based on data from 274 more - 132.45
patients (5 studies more)
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 15 of 18

Table 2 GRADE Evidence Profile (Continued)

Supervised early PR versus usual care for patients with acute exacerbation of COPD
Outcome Timeframe Study results and Absolute effect Certainty in the effects Plain text summary
measurements estimates estimates (Quality of evidence)
Usual Early
care PR
Exercise capacity Measured by: SWT (meters) Difference: MD Low Early pulmonary rehabilitation may
Longest follow-up Higher is better 90.27 higher Due to serious risk of bias and increase exercise capacity at the
Important Based on data from 2017 (CI 95% 69.53 serious inconsistency leading longest follow-up
patients (3 studies) lower - 250.08 to serious imprecisiona,b,d,e
higher)
Dropout rate Relative risk 0.99 217 215 Moderate Early pulmonary rehabilitation
End of treatment (CI 95% 0.71–1.39) per per Due to serious risk of biasa,d probably has little impact on
Important Based on data from 440 1.000 1.000 the dropout rate at the end
patients (8 studies) of treatment
Difference: 2
fewer per 1.000
(CI 95% 63
fewer - 85
more)
Dropout rate Relative risk 1.05 202 212 Moderate Early pulmonary rehabilitation
Longest follow-up (CI 95% 0.6–1.85) per per Due to serious risk of biasa,d probably has little impact on
Important Based on data from 181 1.000 1.000 dropout at the longest
patients (3 studies) follow-up
Difference: 10
more per 1.000
(CI 95% 81
fewer - 172
more)
Falls No studies were found that looked at
Longest follow-up falls at the longest follow-up
Important
Activities of daily No studies were found that looked
living at activities of daily living at the
End of treatment end of treatment
Important
Activities of daily No studies were found that looked at
living activities of daily living at the longest
Longest-follow-up follow-up
Important
CI confidence interval, COPD chronic obstructive pulmonary disease, MD middle difference, PR pulmonary rehabilitation, SGRQ St. George’s Respiratory
Questionnaire, SWT Shuttle Walking Test, 6MWT 6 min walking test
Quality of evidence. High quality: We are very confident that the true effect lies close to that of the estimate of the effect; Moderate quality: We are moderately
confident in the effect estimate, the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different; Low
quality: Our confidence in the effect estimate is limited, the true effect may be substantially different from the estimate of the effect
a
Risk of bias: Serious. Unclear/inadequate sequence generation and unclear/inadequate concealment of allocation during randomization process resulting in
potential for selection bias
b
Risk of imprecision: Serious. Wide confidence intervals
c
Risk of imprecision: Serious. Low number of patients
d
Risk of bias: Serious. Inadequate/unclear or lack of blinding of outcome assessors resulting in potential for detection bias
e
Risk of inconsistency: Serious. The magnitude of statistical heterogeneity was high

COPD is based on supervised programs, and therefore of early PR on mortality [14]. Moreover, the review by
we did not include Greening et al. in our review. How- Puhan et al. [14] differs methodologically from the
ever, to assess safety of early PR initiated during the hos- present review, as they included any inpatient and/or
pital admission we performed a subgroup analysis outpatient PR program with no criteria for the compre-
showing no difference between groups rehabilitated dur- hensiveness or supervision of the rehabilitation program.
ing the admission and after discharge. We only included studies of supervised PR programs
Results from this review differ from a previous review similar to what is offered to COPD patients in Denmark,
by Puhan et al. [14] who showed no statistically signifi- which is based on the present large amount of evidence
cant effect of early PR on mortality, but when the au- in favor of supervised PR in stable COPD. This might
thors preformed a subgroup analysis, excluding results explain the lower heterogeneity and greater effects on
from Greening et al. [17], they did find a positive effect mortality in the present review.
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 16 of 18

Hospital length of stay and readmissions affected patients will likely complete or drop-out to the
Moderate-quality evidence showed that supervised early same extent as usual care. As before mentioned, we did
PR significantly reduced the risk of COPD related hos- not include Greening et al. [17], since this study has
pital readmissions at the longest follow up with 53%. In been highly criticized for not offering a sufficiently ex-
addition, the number of days in hospital was reduced by tensive PR programs [45, 46], and interestingly, authors
an average of 4.27 days. Puhan et al. [14] have previously reported a high number of drop-outs. The participants
shown that PR significantly reduced the mean number in the rehabilitation group attended an average of 2.6 su-
of hospital admissions per participant from 1.6 to 0.9 pervised sessions during hospital admission, followed by
during the year following after hospital admission for an mainly unsupervised training after discharge, with a poor
acute exacerbation. Several explanations have been pro- adherence to the home self-management exercise pro-
posed for the substantial effect of PR on hospital readmis- gram (mean of 57.5) [17]. Nevertheless, these results
sion. The main reason is probably that hospitalization suggest that it is important to assess how the PR is deliv-
following an acute exacerbation of COPD leads to signifi- ered. PR programs can differ in many aspects, which may
cant reductions in activity level [6]. It is well known that influence their effectiveness. When assessing the extensive-
the recovery period after an acute exacerbation is long, ness of the PR program; the number of exercise training
even for patients with no subsequent exacerbations [40]. sessions, frequency of exercise training, type of exercise
Thus, PR can be considered an effective intervention for training and supervision of training, as well as
reverting physical inactivity [41] and it has been shown that self-management, education and adherence to the PR pro-
patients who achieved improvement in their daily physical gram need to be considered [26].
activity level after an exacerbation of COPD experienced a In this review ten studies implemented an extensive PR
~ 50% reduction in risk of hospital readmission [42]. program which mostly showed large and consistent effects
on mortality, days in hospital, COPD related hospital
Health-related quality of life and exercise capacity readmissions, HRQoL, and walking distance. The PR pro-
The primary result to support this, in the present review, grams were not exactly similar within the reviewed stud-
are clinically relevant improvements in walking distance ies, but the majority provided either many training
of respectively 76.89 m in 6 min walking distance sessions (more than 16 sessions) [27, 29–31, 33, 34, 38],
(6MWD) and 54.70 m in shuttle walking distance (SWD) programs of long duration (> 12 weeks) [27, 38], or sup-
immediately after early PR and an improvement of ported education [28, 30, 33, 35, 36]. Nevertheless, our
90.27 m in 6MWD at the longest follow up [43], which results show that supervised early PR programs across
are in line with those results from Puhan et al. [14], show- studies with different protocols are effective in patients
ing an improvement of 62.38 m in 6MWD after early PR. with COPD-related exacerbations.
Secondly, we found moderate quality of evidence support-
ing a clinically important improvement in HRQoL imme- Safety
diately after participation of 19.43 units on the SGRQ Currently, the ideal timing of the onset of PR after
scale and an improvement of 8.74 units at the longest fol- AECOPD is highly debated. Based on low-quality of evi-
low up. These effects on HRQoL exceeded the minimal dence, the ERS/ATS Task Force made a conditional
clinically important difference (MCID) for the SGRQ (> recommendation against the initiation of PR during
4-point improvement [44]), and the results are in line with hospitalization since PR initiated during admission was
previous studies showing a large effect of PR on HRQoL found to increase mortality [18]. This conclusion seems
in stable patients with COPD [14]. Although statistically based solely on results from Greening et al. [17], who re-
non-significant, the beneficial effects of early PR versus ported a higher mortality in the unsupervised home-based
usual care on SGRQ at the longest follow up (8.74 units) rehabilitation group at 12 months compared with usual
in present review were close to those observed in stable care group. The difference between groups however, was
COPD patients (6.89 units) [9]. In addition, the present re- not related to the early rehabilitation intervention. Indeed,
view found a greater improvement in HRQoL at the end the per protocol analysis did not show a difference in mor-
of treatment in patients with an exacerbation of COPD tality [17], suggesting that the participants who actually re-
compared with stable COPD patients, which probably is ceived the intervention were not accountable for the
due to the lower baseline during recovery from AECOPD. increased mortality [47]. We did not find any harms of
early supervised PR across 13 RCTs, even when we iso-
Clinical application lated the subgroup that initiated PR during admission.
We found no difference in drop-out rate between partic-
ipants allocated to early PR compared with usual care. Conclusion
Thus, the effects were not driven solely by positive The results of the present review support the substantial
responders to PR, and secondly, the most severely and clinical important benefits of supervised early PR,
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 17 of 18

indicating that this is an effective intervention with the Authors’ contributions

purpose of reducing mortality following a hospitalization All authors contributed to the conception, design, interpretation, drafting,
revising and final approval of the manuscript. CKR, UWI, NSG and LMK
for AECOPD. Our meta-analysis shows that supervised selected the manuscripts for analysis. CKR, HEC and BT performed the data
PR during the recovery period after an AECOPD is su- extraction and the meta-analysis.
perior to usual care in terms of improving prognosis,
Ethics approval and consent to participate
HRQoL and walking distance. Based on moderate to low Not applicable.
quality of evidence, we conclude that supervised early
PR reduces the risk of mortality, COPD-related hospital Consent for publication
readmissions and the number of days in hospital, and Not applicable.
lead to large and clinically relevant improvements in
Competing interests
HRQoL and walking distance. Therefore, we recommend The authors declare that they have no competing interests.
supervised PR to patients with COPD-related exacerba-
tions. PR should be initiated during hospital admission Publisher’s Note
or within 4 weeks after hospital discharge. Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional affiliations.

Additional files Author details

1
The Centre of Inflammation and Metabolism and the Centre for Physical
Activity Research, Rigshospitalet, University of Copenhagen, Blegdamsvej 9,
Additional file 1: Search strategy. The full search strategy from the DK-2100 Copenhagen, Denmark. 2Danish Health Authority, Copenhagen,
systematic multidatabase literature search performed in 2013 and 2017. Denmark. 3Department of Respiratory Medicine, Copenhagen University
(PDF 214 kb) Hospital, Hvidovre, Denmark. 4Department of Clinical Medicine, University of
Additional file 2: AGREE II. A critical group appraisal of: Pulmonary Copenhagen, Copenhagen, Denmark. 5Department of Physiotherapy,
rehabilitation for patients with chronic pulmonary disease (COPD): an Copenhagen University Hospital, Hvidovre, Denmark. 6Department of
evidence-based analysis using the AGREE II Instrument. (PDF 54 kb) Pulmonary and Infectious Diseases, Copenhagen University Hospital,
Additional file 3: AMSTAR (A Measurement Tool to Asses Systematic Nordsjælland, Hillerød, Denmark. 7The Department of the Elderly and
Reviews). An assessment of the methodological quality of the included Disabled, Odense Municipality, Odense, Denmark. 8Department of
systematic reviews. (PDF 10 kb) Gastroenterology, Center for Nutrition and Bowel Disease, Aalborg University
Hospital, Aalborg, Denmark. 9Unit for Psychooncology and Health
Additional file 4: Assessment of the included studies. Characteristics Psychology, Aarhus University Hospital and Aarhus University, Aarhus,
and risk of bias assessment of the included studies. (PDF 329 kb) Denmark. 10Department of Public Health, Section of Social Medicine,
Additional file 5: Table S1. Extensiveness of the PR programs in the University of Copenhagen, Copenhagen, Denmark. 11Medical Department O,
included studies. (PDF 23 kb) Respiratory Section, Herlev and Gentofte Hospital, Herlev, Denmark.

Received: 12 June 2018 Accepted: 4 September 2018

Abbreviations
1RM: One repetition maximum; 6MDT: 6-min walking test; 6MWD: 6-min
walking distance; ADL: Activities of daily living; AECOPD: Acute exacerbation References
in chronic obstructive pulmonary disease; AGREE II: Appraisal of Guidelines 1. Rosenberg SR, Kalhan R, Mannino DM. Epidemiology of chronic obstructive
for Research and Evaluation instrument version II; AMSTAR: A Measurement pulmonary disease: prevalence, morbidity, mortality, and risk factors. Semin
Tool to Asses Systematic Reviews; ATS: American Thoracic Society; BTS: British Respir Crit Care Med. 2015;36(4):457–69.
Thoracic Society; COPD: Chronic obstructive pulmonary disease; CT: Combined 2. Halpin DM, Miravitlles M, Metzdorf N, Celli B. Impact and prevention of
training; ERS: European Respiratory Society; ET: Endurance training; FEV1: Forced severe exacerbations of COPD: a review of the evidence. Int J Chron
expiratory volume in 1 s; GRADE: Grading of Recommendations Assessment, Obstruct Pulmon Dis. 2017;12:2891–908.
Development and Evaluation; HRmax: Maximum heart rate; HRQoL: Health-related 3. Singanayagam A, Schembri A, Chalmers JD. Predictors of mortality in
quality of life; I2: I-square; MCID: Minimal clinically important difference; MD: Mean hospitalized adults with acute exacerbation of chronic obstructive
difference; PICO: Population, intervention, comparison and outcomes; pulmonary disease. Ann Am Thorac Soc. 2013;10(2):81–9.
PR: Pulmonary rehabilitation; RCT: Randomized controlled trials; Reps: Repetitions; 4. Goto T, Faridi MK, Gibo K, Toh S, Hanania NA, Camargo CA Jr, et al. Trends
RT: Resistance training; SGRQ: St. George’s Respiratory Questionnaire; in 30-day readmission rates after COPD hospitalization, 2006-2012. Respir
SMD: Standardized mean difference; SWD: Shuttle walking distance; SWT: Shuttle Med. 2017;130:92–7.
walking test; VO2max: Maximal oxygen uptake; Wmax: Maximal work load in Watts 5. Eriksen N, Vestbo J. Management and survival of patients admitted with an
exacerbation of COPD: comparison of two Danish patient cohorts. Clin
Respir J. 2010;4(4):208–14.
Acknowledgments 6. Pitta F, Troosters T, Probst VS, Spruit MA, Decramer M, Gosselink R.
Birgitte Holm Petersen and Kirsten Birkefoss, search specialists at the Danish Physical activity and hospitalization for exacerbation of COPD. Chest.
Health Authority, are acknowledged for their contributions. 2006;129(3):536–44.
7. Donaldson GC, Wilkinson TM, Hurst JR, Perera WR, Wedzicha JA, et al.
Funding Exacerbations and time spent outdoors in chronic obstructive pulmonary
This study was initiated and financed by the Danish Health Authority. The disease. Am J Respir Crit Care Med. 2005;171(5):446–52.
writhing of the article was supported by the Centre of Inflammation and 8. Seidel D, Cheung A, Suh ES, Raste Y, Atakhorrami M, Spruit MA. Physical
Metabolism and the Centre for Physical Activity Research, Rigshospitalet, inactivity and risk of hospitalisation for chronic obstructive pulmonary
University of Copenhagen, Denmark. disease. Int J Tuberc Lung Dis. 2012;16(8):1015–9.
9. McCarthy B, Casey D, Devane D, Murphy K, Murphy E, Lacasse Y, et al.
Pulmonary rehabilitation for chronic obstructive pulmonary disease.
Availability of data and materials Cochrane Database Syst Rev. 2015;2:CD003793.
The authors declare that all data supporting the findings of this study are 10. Spruit MA, Singh SJ, Garvey C, ZuWallack R, Nici L, Rochester C, et al. An
available within the following articles and its supplementary information files. official American Thoracic Society/European Respiratory Society statement:
Ryrsø et al. BMC Pulmonary Medicine (2018) 18:154 Page 18 of 18

key concepts and advances in pulmonary rehabilitation. Am J Respir Crit 32. Deepak TH, Mohapatra PR, Janmeja AK, Sood P, Gupta M. Outcome of
Care Med. 2013;188(8):e13–64. pulmonary rehabilitation in patients after acute exacerbation of chronic
11. Iepsen UW, Jørgensen KJ, Ringbaek T, Hansen H, Skrubbeltrang C, Lange P. obstructive pulmonary disease. Indian J Chest Dis Allied Sci. 2014;56(1):7–12.
A systematic review of resistance training versus endurance training in 33. Ko FW, Cheung NK, Rainer TH, Lum C, Wong I, Hui DS. Comprehensive care
COPD. J Cardiopulm Rehabil Prev. 2015;35(3):163–72. programme for patients with chronic obstructive pulmonary disease: a
12. Iepsen UW, Jørgensen KJ, Ringbæk T, Hansen H, Skrubbeltrang C, Lange P. randomised controlled trial. Thorax. 2017;72(2):122–8.
A combination of resistance and endurance training increases leg muscle 34. Ko FW, Dai DL, Ngai J, Tung A, Ng S, Lai K, et al. Effect of early
strength in COPD: an evidence-based recommendation based on pulmonary rehabilitation on health care utilization and health status in
systematic review with meta-analyses. Chron Respir Dis. 2015;12(2):132–42. patients hospitalized with acute exacerbations of COPD. Respirology.
13. Ringbaek T, Brondum E, Martinez G, Thogersen J, Lange P. Long-term 2011;16(4):617–24.
effects of 1-year maintenance training on physical functioning and health 35. Man WD, Polkey MI, Donaldson N, Gray BJ, Moxham J. Community
status in patients with COPD: a randomized controlled study. J Cardiopulm pulmonary rehabilitation after hospitalisation for acute exacerbations of
Rehabil Prev. 2010;30(1):47–52. chronic obstructive pulmonary disease: randomised controlled study. BMJ.
14. Puhan MA, Gimeno-Santos E, Cates CJ, Troosters T. Pulmonary rehabilitation 2004;329(7476):1209.
following exacerbations of chronic obstructive pulmonary disease. Cochrane 36. Seymour JM, Moore L, Jolley CJ, Ward K, Creasey J, Steier JS, et al.
Database Syst Rev. 2016;12:CD005305. Outpatient pulmonary rehabilitation following acute exacerbations of COPD.
15. Puhan MA, Gimeno-Santos E, Scharplatz M, Troosters T, Walters EH, Steurer Thorax. 2010;65(5):423–8.
J. Pulmonary rehabilitation following exacerbations of chronic obstructive 37. Revitt O, Sewell L, Singh S. Early versus delayed pulmonary rehabilitation: a
pulmonary disease. Cochrane Database Syst Rev. 2011;10:CD005305. randomized controlled trial - can we do it? Chron Respir Dis. 2018. https://
16. National Institute for Health and Clinical Excellence: Guidance. Chronic doi.org/10.1177/1479972318757469.
Obstructive Pulmonary Disease: Management of Chronic Obstructive 38. Troosters T, Grosselink R, Decramer M. Short- and long-term effects of
Pulmonary Disease in Adults in Primary and Secondary Care: National outpatient rehabilitation in patients with chronic obstructive pulmonary
Clinical Guideline Centre; 2010. disease: a randomized trial. Am J Med. 2000;109(3):207–12.
17. Greening NJ, Williams JEA, Hussain SF, Harvey-Dunstan TC, Bankart MJ, 39. Murphy N, Bell C, Costello RW. Extending a home from hospital care
Chaplin EJ, et al. An early rehabilitation intervention to enhance programme for COPD exacerbations to include pulmonary rehabilitation.
recovery during hospital admission for an exacerbation of chronic Respir Med. 2005;99(10):1297–302.
respiratory disease: randomised controlled trial. BMJ. 2014;349:g4315. 40. Spencer S, Jones PW. Time course of recovery of health status following an
18. Wedzicha JA, Miravitlles M, Hurst JR, Calverley PM, Albert RK, Anzueto A, infective exacerbation of chronic bronchitis. Thorax. 2003;58(7):589–93.
et al. Management of COPD exacerbations: a European Respiratory Society/ 41. Troosters T, Gosselink R, Janssens W, Decramer M. Exercise training and
American Thoracic Society guideline. Eur Respir J. 2017;49(3). pulmonary rehabilitation: new insights and remaining challenges. Eur Respir
19. The Grading of Recommendations Assessment, Development and Rev. 2010;19(115):24–9.
Evaluation (GRADE) Working Group. [Online] http://www. 42. Garcia-Aymerich J, Farrero E, Félez MA, Izquierdo J, Marrades RM, Antó JM.
gradeworkinggroup.org. Accessed 28 Nov 2017. Risk factors of readmission to hospital for a COPD exacerbation: a
20. Guyatt GH, Oxman AD, Kunz R, Atkins D, Brozek J, Vist G, et al. GRADE prospective study. Thorax. 2003;58(2):100–5.
guidelines: 2. Framing the question and deciding on important outcomes. J 43. ATS Committee on Proficiency Standards for Clinical Pulmonary Function
Clin Epidemiol. 2011;64(4):395–400. Laboratories. ATS statement: guidelines for the six-minute walk test. Am J
21. COPD Working Group. Pulmonary rehabilitation for patients with chronic Respir Crit Care Med. 2002;166(1):111–7.
pulmonary disease (COPD): an evidence-based analysis. Ont Health Technol 44. Jones PW. Interpreting thresholds for a clinically significant change in health
Assess Ser. 2012;12(6):1–75. status in asthma and COPD. Eur Respir J. 2002;19(3):398–404.
22. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of 45. Hopkinson NS. What is and what is not post exacerbation pulmonary
Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. rehabilitation? BMJ. 2014;349:g4315.
2011. Availble from http://handbook.cochrane.org. Accessed 28 Nov 2017. 46. Spruit MA, Rochester CL, Pitta F, Goldstein R, Troosters T, Nici L, et al. A 6-
week, home-based, unsupervised exercise training program is not effective
23. Review Manager (RevMan) [Computer program]. Version 5.3. Copenhagen:
The Nordic Cochrane Centre, The Cochrane Collaboration; 2014. in patients with chronic respiratory disease directly following a hospital
admission. BMJ. 2014;g4315:349.
24. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand M, et al.
47. Wilson KC, Krishnan JA, Sliwinski P, Criner GJ, Miravitlles M, Hurst JR. et al.
GRADE guidelines: 7. Rating the quality of evidence--inconsistency. J Clin
Pulmonary rehabilitation for patients with COPD during and after an
Epidemiol. 2011;64(12):1294–302.
exacerbation-related hospitalisation: back to the future? Eur Respir J. 2018;
25. Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, et al.
51(1).
GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;
64(4):401–6.
26. Bolton CE, Bevan-Smith EF, Blakey JD, Crowe P, Elkin SL, Garrod R, et al.
British Thoracic Society guideline on pulmonary rehabilitation in adults.
Thorax. 2013;68(Suppl 2):ii1–30.
27. Behnke M, Taube C, Kirsten D, Lehnigk B, Jörres RA, Magnussen H, et al.
Home-based exercise is capable of preserving hospital-based improvements
in severe chronic obstructive pulmonary disease. Respir Med. 2000;94(12):
1184–91.
28. Eaton T, Young P, Fergusson W, Moodie L, Zeng I, O'Kane F, et al. Does
early pulmonary rehabilitation reduce acute health-care utilization in COPD
patients admitted with an exacerbation? A randomized controlled study.
Respirology. 2009;14(2):230–8.
29. Kirsten DK, Taube C, Lehnigk B, Jörres RA, Magnussen H, et al. Exercise
training improves recovery in patients with COPD after an acute
exacerbation. Respir Med. 1998;92(10):1191–8.
30. Puhan MA, Spaar A, Frey M, Turk A, Brändli O, Ritscher D, et al. Early versus
late pulmonary rehabilitation in chronic obstructive pulmonary disease
patients with acute exacerbations: a randomized trial. Respiration. 2012;
83(6):499–506.
31. Daabis R, Hassan M, Zidan M. Endurance and strength training in
pulmonary rehabilitation for COPD patients. Egypt J Chest Dis Tuberc. 2017;
66(2):231–6.