Background
Pulmonary arterial hypertension (PAH) is characterized by changes in mostly the distal pulmonary arteries. This causes an increase in pulmonary vascular resistance, which in turn restricts the blood flow through the pulmonary circulation [
1,
2]. In order to maintain the blood flow, the pulmonary artery pressure increase which leads to right ventricular overload, hypertrophy and dilatation [
1]. The physiologic response of the pulmonary vasculature to exercise is different from normal in individuals with PAH and they generally show a reduced exercise capacity [
3]. Advances in the understanding and management of pulmonary arterial hypertension have enabled earlier diagnosis and improved prognosis. However, despite best available therapy, symptoms of exertional dyspnea and fatigue are commonly reported and result in a reduced capacity to perform daily activities and impaired quality of life.
Exercise training has demonstrated efficacy in individuals with other respiratory and cardiovascular diseases [
4-
7]. Historically, however, exercise training has not been utilized as a form of therapy in PAH due to the perceived risk of sudden cardiac death and the theoretical possibility that exercise would lead to a worsening of pulmonary vascular hemodynamics and deterioration in right heart function. Now, with the advances in pharmacological management, determining the safety and benefits of exercise training in this population has become more relevant. The Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology and the European Respiratory Society therefore state that more data are required in order to be able to make recommendations regarding physical activity and supervised rehabilitation in patients with PAH [
1]. Recently, some studies of supervised and home-based exercise training in PAH have been published and it seems worthwhile to summarize their results in order to provide some evidence for the clinicians to rely on.
Therefore, the aim of this systematic review was to apply the meta-analytic approach in controlled trials in order to determine the effects of exercise training on functional status and exercise capacity.
Methods
Literature search
A systematic literature search was conducted in the electronic Pubmed database from its inception to December 2013 using following terms: (medical subject heading (MeSH-term)) ‘exercise’ or (MeSH term) ‘rehabilitation’ or (field term) ‘exercise training’ AND (MeSH-term) ‘pulmonary hypertension’ or (field term) ‘pulmonary arterial hypertension’; results were filtered by ‘clinical trials’ and limitations were ‘English, French, Dutch or German language’.
From this search, we only included articles specifically addressing the effects of a supervised exercise training program in patients with pulmonary hypertension. In addition, the reference lists from published original and review articles were searched manually to identify other possible eligible studies. Embase, SPORTDiscus and The Cochrane Central Register of Controlled Trials were furthermore consulted with the same search terms.
Studies were included in the meta-analysis if they were: Controlled trials 1) including a comparative control group and an intervention group; 2) with an intervention program of at least 3 weeks duration; 3) in patients groups of which at least 80% is diagnosed with PAH as defined according to the updated clinical classification of pulmonary hypertension [
1]; 4) reporting pre- and post-intervention mean and standard deviation (SD) (or standard error) of 6 minute walking distance in intervention and control groups or mean change and standard deviation (or standard error) in intervention and control groups; 5) published in a peer-reviewed journal up to December 2013. Exclusion criteria included any studies not meeting all criteria described above.
Outcomes
The primary outcome measure was change in 6 minute walking distance (6MWD). Secondary outcomes included changes in peak oxygen uptake (peak VO2), peak heart rate and World Health Organization Functional Class (WHO-FC).
A specific developed data extraction sheet was used to extract data on study source, study design, study quality, sample size, characteristics of the participants, exercise interventions, and the different outcomes in each study included in the meta-analysis. Data were extracted by the first two authors (RB and AA) independently from each other. The overall agreement rate prior to correcting discrepant items was 0.64. Disagreements were resolved by consensus.
Study quality
Study quality was assessed using an adapted PEDro-scale [
8]. This 11-item questionnaire is designed to collect data on eligibility criteria, random allocation, concealed allocation, similarity of baseline values, blinding of therapists and/or assessors, key outcomes, intention-to-treat analysis, between group differences, and point and variability measures. The quality criteria ‘blinding of participants’ and ‘blinding of therapists’ were omitted, since they were not applicable in the included intervention studies. All questions were binary. The minimum score was 0 and the maximum was 8, with a higher number reflecting a better study quality. The PEDro-scale has been reported to be valid and reliable [
9,
10]. Using Cohen’s kappa statistic, overall inter-rater agreement was 0.73. Disagreements were resolved by consensus. Trials were not excluded based on quality.
Statistical analysis
Excel 2010 and Review Manager Software (RevMan 5.1; Cochrane Collaboration, Oxford, UK) were used for statistical analyses. Descriptive data are reported as mean ± standard deviation (SD) or median and range. The mean baseline values were calculated by combining mean values from the intervention and control groups, weighted by the number of participants in the final analysis in each study group.
Effect sizes for primary and secondary outcomes were calculated by subtracting the pre-intervention value from the post-intervention value for both study groups. The net treatment effect was then obtained by subtracting the change score difference in the control group from the change score difference of the intervention group. Review Manager Software calculated the variances from the inserted pooled SDs of change scores in the intervention and control groups. However, some studies reported only the SDs or standard errors of mean at baseline and post-intervention. Therefore, change scores SDs that were missing in these studies were calculated from pre and post SD values, using the following formula: SDchange = √[(SDpre)
2 + (SDpost)
2 − 2 × corr(pre,post) × SDpre × SDpost] [
11], for which a correlation coefficient of 0.5 between the pre and post values was assumed. The primary and secondary outcomes were combined using random-effects models, a preferred approach that incorporates heterogeneity into the model [
12,
13]. Each effect size was weighted by the inverse of its variance. The results are reported as weighted means and 95% confidence intervals (CI). Two sided tests for overall effects were considered significant at p ≤ 0.05.
Statistical heterogeneity among the trials was assessed using Cochrane’s Q statistic and an alpha value for statistical significance of 0.10 indicated significant heterogeneity. In addition, we report I
2 statistics, which assesses consistency of treatment effects across trials. I
2 between 25% and 50% represents small amounts of inconsistency, whereas between 50-75% and >75% represents medium to large amounts of heterogeneity [
14]. To examine the influence of each study on the overall results, sensitivity analyses were also performed with each study deleted from the model once. Finally, funnel plots were used to assess the potential of small publication bias. Funnel plot asymmetry was evaluated by use of Begg and Egger tests, and a significant publication bias was considered if the p value was <0.10. The trim and fill computation was used to estimate the effect of publication biases on the interpretation of the results.
Discussion
This systematic review and meta-analysis is the first to focus on the effect of exercise training on physical fitness and functional capacity in patients with PAH. The analysis, in which 5 studies could be included, showed that exercise training can result in clinically relevant improvements of physical and functional capacity.
The idea that exercise training programs may potentially be beneficial for patients with PAH became only recently accepted [
1]. Exercise based rehabilitation for patients with PAH is now considered as a new adjunct treatment option in order to improve their functional capacity and quality of life. Following here, some proof of concept studies have been performed; most of them showing significant beneficial effects of exercise training on functional capacity and/or exercise capacity, but not all (Chan [
15], Martinez [
17], Fox [
19] for peak VO
2). Hence, the existing evidence is inconclusive and insufficient to result in the implementation of exercise programs in routine clinical care for patients with PAH. By pooling data from published controlled trials, we aimed to increase the statistical power and resolve the uncertainty. We could show that exercise training is effective for improving both functional and exercise capacity.
In todays practice the 6MWD is most often used as an endpoint to assess the benefit of therapies in PAH. It is a measure of functional capacity. This meta-analysis documented that an exercise training program of around 12 weeks can significantly improve the 6MWD with 72 m on average. This gain in walking distance is larger than that reported for patients with chronic obstructive pulmonary disease [
22,
23] or chronic heart failure [
24]. Whereas it remains to be determined which minimum increase in 6MWD has to be reached in patients with PAH in order to obtain a clinically relevant improvement, a change of more than 50 meters has been proposed as representative of a clinically meaningful change in most disease states [
25]. On the other hand, short-term improvements in 6MWD due to medical treatment could so far not be related to an improved longer-term prognosis [
26] even though a baseline distance less than approximately 330 m on the 6MWD has been shown to be associated with increased mortality in PAH [
27]. In line with these results, Golpe and colleagues in non-group 1 PAH patients reported that a distance less than 400 m was predictive of a significantly worse prognosis [
28]. In our study, mean baseline 6MWD was already more than 400 meter, however it is anticipated that the average observed gain in 6MWD of 72 m due to exercise training, will have shifted a part of the weaker patients towards a 6MWD larger than 400 m and thus towards a more favorable prognosis.
Overall, cardiopulmonary exercise testing is the preferred method for assessing changes associated with exercise training [
29]. Despite being less often used, cardiopulmonary exercise testing should also be considered as the gold standard measurement of exercise capacity in patients with PAH [
30]. Only three of the included studies reported exercise training related changes in oxygen uptake. Nevertheless, this was already sufficient to document a small but significant increase in peak VO
2 following exercise training. This result is also in line with several other small-scale non-controlled studies [
31] and adds further evidence to support the use of exercise training to increase peak VO
2 in this patient population. Whereas the clinical significance of these small improvements (+2.6 ml/min/kg peak VO
2) in this population remains inconclusive, Groeppenhof et al. showed that among patients with PAH, survivors have larger increases in exercise capacity compared to non-survivors [
32].
The underlying mechanisms causing these training effects might be of both central pulmonary and cardiovascular as well as peripheral origin, but need to be further clarified. In addition, whether baseline and especially training-related changes in objective exercise capacity reflect a better prognosis both in short and longer term of patients with PAH still needs to be determined [
30].
Exercise capacity is closely related to the ability to perform normal daily life activities. In turn, this closely relates to quality of life. Exercise tests can indirectly quantify shortness of breath and fatigue, two of the most common symptoms of PAH. Therefore, the change in 6MWD and peak VO
2 from baseline to the end of an exercise training program might suggest symptomatic improvement over that time-period [
33]. Although we could only show a tendency (p = 0.12) toward a functional improvement by means of the WHO-FC scale, this was only based on four studies and more studies are warranted. Moreover, the effect of exercise training on quality of life in these patients needs further research due to the importance for both clinicians and patients.
Limitations
There are several limitations to our work. Although meta-analyses are no substitute for large well-designed controlled trials, the meta-analytical technique is probably the best method to systematically review previous work. Advantages are the greater precision of the estimates and the enhanced statistical power. In order to provide the most accurate summary of the effects of exercise training for PAH, the PRISMA statement was carefully followed [
34]. A potential disadvantage of a meta-analysis is the heterogeneity of studies for instance due to variations in the exercise programs and the differences between patient populations [
35]. Although random-effects models which incorporate heterogeneity into their model were used, some caution might be warranted with the interpretation given also the small number of studies and the small sample sizes. Further, as there were only 3 randomized trials, we also included 2 non-randomized studies. The results of those may be affected by confounding bias due to the absence of random assignment.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors significantly contributed to the study and take responsibility for the integrity of the work as a whole, from inception to published article. RB conceived of the study, participated in its design and coordination, extracted and statistically analyzed the data and drafted the first version of the manuscript. AA extracted the data and revised the manuscript it critically for important intellectual content. VC participated in the study design performed and controlled the statistical analyses on the data; she provided consensus when data extracted by RB and AA was inconsistent and revised the manuscript it critically for important intellectual content. All authors read and approved the final manuscript.