Abstract
The value of inspiratory muscle training (IMT) in pulmonary rehabilitation in chronic obstructive pulmonary disease (COPD) is unclear. The RIMTCORE (Routine Inspiratory Muscle Training within COPD Rehabilitation) randomised controlled trial examined the effectiveness of IMT added to pulmonary rehabilitation.
In total, 611 COPD patients (Global Initiative for Chronic Obstructive Lung Disease stage II–IV) received a 3-week inpatient pulmonary rehabilitation, of which 602 patients were included in the intention-to-treat analyses. The intervention group (n=300) received highly intensive IMT and the control group (n=302) received sham IMT. The primary outcome was maximal inspiratory pressure (PImax). The secondary outcomes were 6-min walk distance, dyspnoea, quality of life and lung function. Outcomes were assessed pre- and post-pulmonary rehabilitation. ANCOVA was used.
The intervention group showed higher effects in PImax (p<0.001) and forced inspiratory volume in 1 s (p=0.013). All other outcomes in both study groups improved significantly, but without further between-group differences. Sex and pulmonary rehabilitation admission shortly after hospitalisation modified quality of life effects.
IMT as an add-on to a 3-week pulmonary rehabilitation improves inspiratory muscle strength, but does not provide additional benefits in terms of exercise capacity, quality of life or dyspnoea. A general recommendation for COPD patients to add IMT to a 3-week pulmonary rehabilitation cannot be made.
Abstract
IMT added to a 3-week PR programme improved PImax and FIV1, but not exercise capacity, dyspnoea or QoL #COPD http://ow.ly/hPGK30gfqej
Introduction
Inspiratory muscle training (IMT) aims to improve inspiratory muscle function using different training techniques that selectively load the inspiratory muscles. IMT has been widely studied in patients with chronic obstructive pulmonary disease (COPD) and several systematic reviews support the effectiveness of IMT as a stand-alone therapy [1–5]. IMT improves maximal inspiratory muscle strength (maximal inspiratory pressure (PImax)), dyspnoea, functional capacity and health-related quality of life (QoL) [1–3].
Pulmonary rehabilitation is recommended in COPD guidelines as a key element in the long-term management of COPD [6]. The most essential pulmonary rehabilitation component is physical training, especially endurance and strength training [7–9]. The effects of pulmonary rehabilitation might increase further by adding IMT, but the evidence is inconsistent. Although meta-analyses indicate that additional IMT may improve PImax, effects on other clinically relevant outcomes such as dyspnoea, QoL or functional exercise capacity remain unclear [3, 5]. Only one meta-analysis showed a nonsignificant trend regarding 6-min walk distance (6MWD) [3]. Therefore, pulmonary rehabilitation guidelines do not recommend routinely adding IMT to pulmonary rehabilitation [8].
Moreover, positive effects of additional IMT on PImax have been demonstrated, predominantly in studies that did not use a placebo control group or for IMT programmes with a longer duration (e.g. 6 months) [10–13]. In Germany, according to legal stipulations, pulmonary rehabilitation is usually provided on an inpatient basis for 3 weeks. However, a 3-week intensive inpatient pulmonary rehabilitation with daily IMT might be comparable to an 8-week pulmonary rehabilitation with three sessions per week [14]. Currently, it is unclear whether adding IMT to an intensive 3-week pulmonary rehabilitation programme leads to an improvement of PImax and other clinical outcomes.
Based on the comparison of studies with higher and lower mean values in baseline PImax (cut-off PImax <6 kPa), Gosselink et al. [3] hypothesised that IMT is more effective in patients with inspiratory muscle weakness. However, no study has examined this hypothesis on the patient level.
Most randomised controlled trials (RCTs) regarding IMT as an add-on had a small sample size. Thus, a lack of statistical power may have inhibited the detection of beneficial effects in clinical outcomes. Moreover, a lack of power precludes examining moderation effects (e.g. whether effects between groups were influenced by sex or baseline PImax). Therefore, studies with adequately powered samples are required [15].
The main hypothesis of this RCT (RIMTCORE (Routine Inspiratory Muscle Training within COPD Rehabilitation)) was that IMT routinely added to a 3-week pulmonary rehabilitation programme (intervention group) improves PImax (primary outcome) in comparison with a control group receiving pulmonary rehabilitation and sham IMT. Secondary hypotheses were that IMT (compared with sham IMT) improves lung function (e.g. forced inspiratory volume in 1 s (FIV1), forced expiratory volume in 1 s (FEV1) and vital capacity), dyspnoea, functional capacity and QoL. Furthermore, modifying effects on these outcomes were examined for sex, Global Initiative for Chronic Obstructive Lung Disease (GOLD) stage, smoking status, “pulmonary rehabilitation admission directly or up to 2 weeks after hospitalisation due to acute exacerbations of COPD” (Hosp_AECOPD) versus “pulmonary rehabilitation in stable COPD” and baseline PImax.
Methods
Study design and subjects
This RCT used a parallel group design: COPD patients who underwent a 3-week inpatient pulmonary rehabilitation in the Bad Reichenhall Clinic (Bad Reichenhall, Germany) were recruited from February 2013 to July 2014. All admitted patients with GOLD stage II–IV fulfilling inclusion and exclusion criteria (table 1) were asked to participate. Participants were randomised without stratification according to a computer-generated randomisation list (provided by the University of Würzburg, Würzburg, Germany) to receive either highly intensive strength IMT (intervention group) or sham IMT (control group). Based on results of a nonrandomised pilot study, a sample size of n=420 patients was planned (for details regarding sampling and randomisation procedures, see supplementary material).
The study was approved by the local ethics committee (Bayerische Landesärztekammer; 12107) and was registered at the German Clinical Trials Register (identifier DRKS00004609). All participants provided written informed consent. RIMTCORE was funded by Deutsche Rentenversicherung Bayern Süd (German statutory pension insurance scheme).
Interventions
All patients received an intensive pulmonary rehabilitation programme tailored to each patient's individual needs. Obligatory components (mostly 30- to 60-min sessions) included physical training (endurance training: four or five sessions per week; strength training: three sessions per week; whole-body vibration muscle training: seven sessions per week), patient education (seven or more sessions) and respiratory physiotherapy in groups (two to four sessions per week). Optional components were smoking cessation (eight sessions), mucolytic physiotherapy, saline inhalation, psychological interventions, social counselling, nutritional counselling and occupational therapy.
Inspiratory muscle training
The intervention group received high-intensity interval-based IMT routinely added to the pulmonary rehabilitation programme according to the protocol of Hill et al. [14], using a threshold training device (POWERbreathe Medic; POWERbreathe International, Southam, UK). Training intensity was individually adjusted based on the previously determined PImax. The initial training load was at least 30% of PImax and was progressively increased to at least 60%. High-intensity interval-based IMT took place 7 times per week (21 min; seven cycles of 2 min of IMT each followed by 1 min of rest). Experienced coaches supervised three sessions per week to ensure adequate training and instructed the patients regarding further training at home.
Sham IMT
The control group received sham IMT (also supervised 3 days per week). The sham training device looked identical to the verum device, but had no valve within it. Patients were instructed to breathe slowly and in a relaxed manner. Thus, the training load was <1 cmH2O, which represented no effective IMT and thus making this a placebo procedure [4].
Outcomes
Lung function, functional capacity and questionnaire data were assessed at the beginning (T0) and end (T1) of the pulmonary rehabilitation programme.
Primary outcome
The primary outcome was PImax, i.e. the maximum static inspiratory pressure a subject can generate at the mouth [16], measured using a commercially available mouth occlusion pressure device (CareFusion, Höchberg, Germany) as recommended by German guidelines [16].
Secondary outcomes
Lung function measurement
FEV1, FIV1, vital capacity, residual volume and total specific resistance were determined before and after bronchodilation by spirometry and body plethysmography (MasterLab; CareFusion) [17, 18].
6-min walk distance
6MWD measurement was carried out on a track length of 30 m according to the 2002 American Thoracic Society statement [19]. At both T0 and T1 each patient performed two tests with an interval of 1 h. The best test of each was included for analysis.
Statistical analysis
Intention-to-treat (ITT) and per-protocol analyses were performed. Results of the ITT analysis are presented unless otherwise stated. Pre–post changes within groups were estimated via the standardised response mean, with mean differences between T1 and T0 divided by the standard deviation of the difference scores. ANCOVA with T1 values as outcomes and treatment group (intervention group/control group) as predictor was used to estimate treatment effects. Adjusted mean differences (AMDs) with 95% confidence intervals and Cohen's d were calculated to quantify the between-group effects. We computed two models for each outcome. The unadjusted model only included corresponding T0 values as covariates. The adjusted model also included sex, GOLD stage, smoking status (active versus ex-/never-smoker) and baseline PImax as covariates. The residuals of all models were checked for distortions of model assumptions (e.g. normality and linearity) and robust regression analyses were computed if model assumptions did not hold. Neither adjusted/unadjusted nor linear/robust regression analyses differed in results; therefore, only results of the adjusted analyses with linear models are presented, as stated in the study protocol.
Moderator effects of sex, Hosp_AECOPD, GOLD stage, smoking status and baseline PImax were examined by (separately) including the moderator and interaction term (group membership (intervention group/control group) and moderator) in the respective model. Significant interactions with categorical moderators were further analysed by simple effect analyses.
Missing data arose for two reasons: 1) discontinuation of pulmonary rehabilitation (i.e. no data available at T1) and 2) missing data in some items (0–3%). All missing data were imputed with a multiple imputation procedure creating 10 imputed data sets. Pooled results are reported. TDI could not be imputed due to the scaling of the items and thus listwise deletion was applied (for missing data procedure details, see supplementary material). All analyses were performed using SPSS Statistics version 23 (IBM, Armonk, NY, USA) and R (R Foundation for Statistical Computing, Vienna, Austria). p<0.05 was considered significant.
Results
In total, 983 subjects with pre-diagnosed COPD were screened for eligibility (figure 1). Of these, 611 patients with confirmed COPD II–IV were randomised to the intervention group or the control group. Data of nine patients could not be analysed because no data regarding the primary outcome PImax at T0 or T1 were available, leaving data for 602 patients for the ITT analysis. During inpatient rehabilitation, 50 patients discontinued participation in the study for various reasons, leaving data for 561 patients for the per-protocol analysis. For details on planned and recruited sample size, see the supplementary material.
Baseline data
At baseline, 64.6% of the patients were male, mean±sd age was 57.8±7.4 years and mean±sd PImax was 6.71±2.29 kPa. The intervention group received a mean±sd of 19.4±4.5 IMT sessions; mean±sd training load was increased from 21±7 cmH2O (30.6% of mean PImax) at T0 to 47±14 cmH2O (68.5% of mean PImax) at T1. Table 2 presents further sample characteristics.
Comorbidities
97.5% of the patients had clinically relevant comorbidities (mean (range) 4 (1–11)). Cardiovascular (63.5%), musculoskeletal (58.1%) and metabolic disorders (54.7%) were the most frequent (table 3).
Primary and secondary outcomes
Both groups improved significantly in PImax (table 4). The intervention group improved significantly more than the control group (AMD 0.94, 95% CI 0.72–1.16). The intervention group also improved significantly more in FIV1 (AMD 0.10, 95% CI 0.02–0.19). All other secondary outcomes improved significantly in both arms, mostly with moderate to high effect sizes, but without significant differences between groups.
Effect modifiers
Sex was a significant effect modifier for CAT (p<0.001), CCQ-Total (p=0.011), CCQ-Mental (p=0.001), CCQ-Symptoms (p=0.002) and SGRQ-Total (p=0.042) (table 5). In females, the intervention group showed significantly larger effects regarding CAT and CCQ-Mental compared with the control group. In contrast, in males, the control group showed larger effects regarding CAT, CCQ-Mental, CCQ-Symptoms and SGRQ-Total than the intervention group.
Moreover, sex modified the effects for 6MWD, but only in the per-protocol analysis (p=0.038). Females in the intervention group showed higher improvements in 6MWD than in the control group, whereas no group difference was found in males (table 5).
Significant interaction effects for Hosp_AECOPD (yes/no) were found in CCQ-Total (p=0.038), CCQ-Function (p=0.021), SGRQ-Impacts (p=0.028) and SGRQ-Total (p=0.041) (table 6). In patients with pulmonary rehabilitation after Hosp_AECOPD, the control group tended to show greater improvements in these outcomes than the intervention group, whereas no differences were found for patients with stable COPD.
In addition, although not significant in the ITT analysis (p=0.242), baseline PImax modified the change in PImax in the per-protocol analysis: larger effects in favour of the intervention group were found for patients with lower baseline values in PImax (B= −0.10 (95% CI −0.20– −0.01); p=0.035). However, this effect is rather small (figure 2). Even in patients with baseline PImax ≥7 kPa, a significant increase was found at T1.
Discussion
The main results of this RCT are that adding high-intensity interval-based IMT to an intensive 3-week pulmonary rehabilitation programme led to a significant improvement in PImax and FIV1 in patients with moderate to very severe COPD compared with sham IMT, but failed to improve other clinically relevant outcomes such as dyspnoea, QoL, functional capacity and other lung function parameters. The mean increase in PImax in the intervention group versus control group of 0.94 kPa (9.59 cmH2O) corresponds to a PImax mean difference of 8.60 cmH2O, as achieved by much longer outpatient programmes [5]. Exploratory subgroup analyses showed that sex and Hosp_AECOPD might moderate the effects in QoL to a small extent, whereas baseline PImax showed no influence on any outcome.
The rationale for adding IMT to pulmonary rehabilitation is the assumption that changes in PImax may lead to changes in clinical outcomes. However, despite relevant improvements in PImax, no further improvements in clinical outcomes were found. There may be various reasons for this. First, the causal model may just be wrong and improvements in clinical PImax do not translate into clinical outcomes [25]. However, effects in clinical outcomes were shown by various studies comparing IMT alone with controls [1, 13, 14]. Second, an effect of 0.94 kPa might be too small to translate into clinically relevant improvements in the context of pulmonary rehabilitation. Hence, interventions of longer duration are necessary. For example, Magadle et al. [12] showed both larger effects in PImax as well as effects in functional capacity using an intervention lasting 6 months. Third, changes in PImax might lead to changes in clinical outcomes only if no other therapies produce changes in these outcomes at the same time. Therefore, effects of pulmonary rehabilitation on clinical outcomes might mask possible IMT effects. Fourth, subgroups of patients (e.g. females versus males) might differ in their response to IMT. We will discuss these points in the following paragraphs.
According to our study and contrary to Gosselink et al.'s [3] meta-analysis, PImax at baseline does not moderate the size of the effect in PImax. Patients with high values of PImax at T0 still benefit substantially from the intervention according to changes in PImax. Furthermore, baseline PImax did not modify the effect in any secondary outcome. Therefore, the hypothesis that (only) patients with inspiratory muscle weakness may benefit from IMT is not supported by our results. The different results of both studies may be explained by different methods. Gosselink et al. [3] examined relationships between variables of studies (i.e. mean baseline PImax and mean effect of IMT), whereas we examined relationships between patient variables. Inferences from the study level to the patient level may be misleading if patients are not drawn randomly from the same population, which is difficult to prove [26]. This reason and the large sample size of our study support the validity of our result.
The mean 6MWD improved statistically and clinically significantly in both study arms by >80 m. The AMD between groups in the ITT analysis was 1.9 m. However, according to the per-protocol analysis, IMT might improve 6MWD in females by >13 m (Cohen's d=0.21). The cause of the difference in the results between the ITT and per-protocol analyses remains unclear. Therefore, the finding of the per-protocol analysis should be treated with caution and, based on the ITT analysis, we conclude that sex has no moderator effect on 6MWD.
Similar results were observed for QoL. Both groups improved significantly with moderate to strong effect sizes, but IMT did not lead to further improvements. However, sex and Hosp_AECOPD modified QoL. Although females tended to profit from additional IMT in some QoL subscales, males and patients after AECOPD tended to have no benefit or even showed worsening effects. No previous study has reported similar effects, maybe for reasons of power. Several possible explanations for the sex effect exist. For example, inspiratory muscles in females may fatigue less quickly [27, 28]. Therefore, females may experience IMT as less stressful and might benefit more from the gain in PImax with regard to QoL. Furthermore, males may benefit more from “relaxation breathing” (control group) than females. Slow and relaxed breathing for 21 days may be comparable to relaxation training [29] and may affect psychological but not physiological parameters [30]. However, all effects are small and the clinical relevance is unclear.
In addition, we found a significant improvement in FIV1. Although several authors regard changes in FIV1 as relevant [31], the clinical significance of a difference of 100 mL remains unclear, especially as other lung function parameters did not increase.
To the best of our knowledge, this is the largest RCT evaluating add-on IMT to pulmonary rehabilitation in COPD. To enhance external validity, we did not exclude patients with comorbidities, exacerbations shortly before rehabilitation or those without inspiratory muscle weakness (“real-life study”).
Nevertheless, the study has several limitations. First, this is a single-centre study, which might limit generalisability. However, the rehabilitation programme of the Bad Reichenhall Clinic conforms to the structural requirements of German healthcare insurance providers [32] and should be comparable to other inpatient pulmonary rehabilitation programmes. Second, participants and health professionals could not be fully blinded, as is common in such studies. Thus, influences of group awareness cannot be ruled out. Third, a control group without sham training was not available. As our control setting may have had effects beyond placebo (in the sense of an active control group), the effects on QoL may have been underestimated. Fourth, it is difficult to appraise the clinical relevance of the effects of IMT in females in 6MWD and some QoL scales. According to some authors, effect sizes of 0.2 can be regarded as small, but potentially meaningful [33]. However, clear guidelines on how to interpret small effects comparisons with active control groups do not exist.
Conclusions
IMT routinely added to an intensive 3-week pulmonary rehabilitation improves PImax and FIV1. Initial PImax did not modify these effects. However, for the whole sample, IMT did not further improve functional capacity, dyspnoea or QoL. Therefore, a general recommendation to include IMT in a 3-week pulmonary rehabilitation for all COPD patients cannot be made. Concerning subgroups, IMT improved aspects of QoL, but not in males or patients after severe AECOPD.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Sample size, randomisation and missing values ERJ-02000-2017_supplement
Disclosures
Supplementary Material
H. Faller ERJ-02000-2017_Faller
M. Schuler ERJ-02000-2017_Schuler
K. Schultz ERJ-02000-2017_Schultz
T. van der Molen ERJ-02000-2017_van_der_Molen
Acknowledgements
Author contributions: K. Schultz planned and obtained funding for the study. M. Wittmann, O. Göhl, D. Stojanovic and M. Schuler contributed to the study concept and design. K. Schultz, D. Stojanovic, M. Wittmann and M. Schuler were responsible for data collection, B. Krämer, V. Huber, S. Fuchs, N. Lehbert, S. Wingart, H.J. Alma and C. de Jong contributed to data acquisition. M. Schuler performed data analysis. K. Schultz and M. Schuler drafted the first version, and were actively supported by H.J. Alma, C. de Jong, T. van der Molen, M. Wittmann and H. Faller. All authors reviewed the manuscript critically for important intellectual content and approved the final manuscript.
Footnotes
This article has supplementary material available from erj.ersjournals.com
This study is registered at the German Clinical Trials Register with identifier number DRKS00004609.
Support statement: This study was supported by Deutsche Rentenversicherung Bayern Süd. Funding information for this article has been deposited with the Crossref Funder Registry.
Conflict of interest: Disclosures can be found alongside this article at erj.ersjournals.com
- Received September 13, 2016.
- Accepted October 24, 2017.
- Copyright ©ERS 2018