Background
Musculoskeletal conditions (MSCs) are common, and have important consequences for the individual and for society. MSCs are the most common cause of severe long-term pain and physical disability, and in Europe, 20 to 30
% of adults are affected at any one time [
1‐
3]. It is estimated that MSCs represent nearly 25
% of the total cost of illness in Sweden [
4], and they are one of the most common causes of health problems limiting work ability [
5]. About one in five consultations in primary care are for MSCs, and some of these patients are referred to other health professionals such as physiotherapists, occupational therapists, or chiropractors, to medical specialists such as rheumatologists, or to orthopedic surgeons. The burden that MSCs create has been recognized by the United Nations and WHO, with their endorsement of the Bone and Joint Decade 2000 to 2010 [
6]. The prevalence of many of these conditions increases markedly with age, and many patients also have some common lifestyle factors (obesity, smoking, and physical inactivity). With the increasing number of older people and the ongoing changes in lifestyle, the burden of MSCs and other non-communicable diseases is predicted to increase [
7].
MSCs form a heterogeneous group of over 200 different health problems that are linked anatomically and also linked by their association with pain and impaired physical function [
8]. They range from conditions of acute onset and short duration to lifelong disorders. For many common MSCs such as regional pain syndromes, the underlying pathogenesis is poorly understood, and it is often not possible to produce a clear-cut diagnosis. Some, such as osteoarthritis (OA), are biologically well defined but clinically less well understood, whereas others, such as rheumatoid arthritis (RA) are both biologically and clinically well defined [
9].
For decades, inactivity and bed rest were the mainstays of the management of many diseases, but there is now an increasing amount of evidence supporting the opposite view; that is, that physical activity and exercise are beneficial to health promotion and treatment. Although exercise and physical activity are closely related constructs, they have distinct meanings. The term 'physical activity' includes everyday activities that can contribute to wellbeing, whereas 'exercise' is physical activity that is planned, structured and repetitive [
10]. The focus of this overview is exercise therapy (ET), which involves the prescription of a physical-activity program that involves the client undertaking voluntary muscle contraction and/or body movement with the aim of relieving symptoms or improving function, or of improving, retaining, or slowing deterioration of health [
11].
Whereas systematic reviews of randomized trials usually summarize the evidence of one kind of intervention for a single condition, overviews of systematic reviews can summarize evidence from more than one systematic review of the same intervention for different conditions. There is, to our knowledge, only one published overview of systematic reviews that has focused on the effects of exercises for MSCs [
12]. This overview was based on reviews published up to July 2007, and did not include specific diseases such as osteoporosis and inflammatory joint diseases. Thus, the most updated evidence may not be included in this overview. The main aim of the present overview was to synthesize evidence from systematic reviews on the effects of ET on pain and physical function for eight selected MSCs. In addition, we explored the evidence for the effect of ET on disease pathogenesis, and whether particular components of exercise programs are associated with the size of the treatment effects.
Methods
Criteria for considering reviews for inclusion
For this overview, we included systematic reviews on the effects of ET for four common conditions (FM, low back pain (LBP), neck pain (NP), and shoulder pain (SP)) and four specific musculoskeletal diseases (OA, RA, ankylosing spondylitis (AS), and osteoporosis). Other related conditions and musculoskeletal malformations and traumas were not included in this overview. All types of land-based ET interventions were considered.
Search methods for the identification of reviews
The first search, of the Cochrane database of systematic reviews, was performed in March 2012. All Cochrane reviews on exercises for the aforementioned MSCs were considered for inclusion. Two authors (GS and KBH) assessed the eligibility of reviews based on the inclusion criteria presented above. Cochrane reviews with the most recent update of January 2007 or before were not included, and we replaced these reviews by searching for non-Cochrane systematic reviews in MEDLINE, EMBASE, CINAHL, AMED, and PEDro, published after this date (for the electronic search strategy for MEDLINE, see Additional file
1). Two authors (GS and KBH) then screened these records for new reviews. If several reviews fulfilled the inclusion criteria, we included only one review per combination of disease, intervention, and outcome by choosing the newest review of high quality. If two or more reviews were equal for these criteria, we chose the review with the highest number of primary studies. We also searched in the same databases for reviews with the explicit aim of investigating the effect of exercise on the pathogenesis of MSCs, with no restrictions on publication date.
Outcome measures
ET. can have clinical effects by improving dominant symptoms, and we therefore chose to primarily focus on the core symptoms of MSCs: 1) any measure of pain, and 2) physical disability or physical function.
In addition, we collected data on the effect of ET on disease pathogenesis, and whether particular components of exercise programs are associated with the magnitude of treatment effects.
We did not focus on the general health benefits or complications of exercise because small to medium randomized controlled trials (RCTs) with short-term to intermediate follow-up may be inadequate to assess other health benefits such as the cardiovascular risk profile and the incidence of complications. Other potentially relevant outcomes, such as work participation and health-related quality of life, compliance, or costs, were not evaluated.
Data collection and analysis
One reviewer (GS) extracted data on population, intervention, comparison, and outcomes (inclusion criteria) and the methodological quality of included trials. The methodological quality was assessed using the AMSTAR (A Measurement Tool to Assess Systematic Reviews) checklist [
13]. The 11 criteria shown in Table
1 were rated as 'met,' 'unclear/partly met,' or 'not met'. A second reviewer (KBH) independently verified the accuracy of the numeric results. Most reviews included several comparisons, and we included analyses that compared any exercise intervention with no or minimal exercise intervention. Because pain and function most often are continuous outcomes, data were summarized using the standardized mean difference (SMD) or weighted mean difference (WMD) with 95
% confidence interval (95
% CI) as reported in the included reviews. For dichotomous outcomes, odds ratio and 95
% CI were presented. Pooled effect estimates were presented according to the model used in the reviews. For reviews that explored whether particular components of exercise programs were associated with the magnitude of treatment effects, we extracted the coefficients from meta-regression or sub-group analyses. Results from the review investigating the evidence for the effect of ET on disease pathogenesis were reported as in the original review.
Table 1
The AMSTAR check list for methodological assessment.
1 | Was an 'a priori' design provided? |
2 | Was there duplicate study selection and data extraction? |
3 | Was a comprehensive literature search performed? |
4 | Was the status of publication (that is, 'gray' literature) used as an inclusion criterion? |
5 | Was a list of studies (included and excluded) provided? |
6 | Were the characteristics of the included studies provided? |
7 | Was the scientific quality of the included studies assessed and documented? |
8 | Was the scientific quality of the included studies used appropriately in formulating conclusions? |
9 | Were the methods used to combine the findings of studies appropriate? |
10 | Was the likelihood of publication bias assessed? |
11 | Were potential conflicts of interest included? |
Discussion
In this overview of systematic reviews on ET for muscle and bone health, we included 9 systematic reviews with a total of 224 trials and 24,059 patients. Overall, we found a substantial amount of empirical evidence supporting ET as a mainstay in the management of MSCs. However, there are differences in the number of included trials between the diagnostic groups included in the present overview. For common conditions such as chronic non-specific LBP and knee OA, there is a substantial number of trials with relatively consistent results showing that exercise has small to moderate beneficial effects on pain and function. For others, such as the inflammatory joint diseases (RA and AS), there are only one or two trials comparing exercises with non-exercise or minimal exercise interventions. When comparing the effect sizes on pain and physical functioning, there seems to be a trend towards larger effects of exercise on pain. The effect sizes were generally larger for pain than for physical function, with significant positive effects for pain in six conditions and for physical function in five conditions. This may indicate that there is no linear relationship between symptoms and functioning, and it may suggest that a relatively large reduction in symptoms is needed to improve physical function. There are also several other important components influencing function, such as environmental and personal factors.
We also intended to explore the evidence for the effect of ET on disease pathogenesis. However, we were not able to identify evidence for the influence of pathogenesis in MSCs, except for OP. Although the disease mechanisms for the common MSCs are largely unknown, there seems to be a gap in the understanding of why ET improves clinical outcome. From a clinical point of view, the most interesting question is 'which exercises for which patients?'. We therefore also explored whether particular components of exercise programs are associated with the size of the treatment effects. Because we did not include single trials comparing different exercise interventions, we could only identify indirect comparisons from the reviews. For LBP and knee OA, for which the substantial number of trials allowed explorative sub-group or meta-regression analyses, the results suggest that 'more is better'; that is, that the treatment effect increases with the number of exercise sessions. However, it is unclear whether the estimated increase is of clinical relevance.
One previous overview has addressed the effect of ET for MSCs [
12]. Dziedzic
et al. also concluded that ET is a beneficial component of the management of MSCs because it reduces pain and disability, and that there is limited evidence for the benefit of one particular approach to exercise over another. However, they also emphasized one important clinical caveat, namely that there is evidence that exercise should not be recommended for acute LBP [
12], which was also supported by an updated systematic review [
38].
There are several methodological challenges in summarizing evidence from systematic reviews only. There is a substantial number of new trials published in this field every year, and systematic reviews published some years ago may therefore not be based on the most up-to-date evidence. We therefore intended to include Cochrane reviews, because these should be regularly updated. However, this overview clearly shows that this is not the case. We excluded three out of nine eligible Cochrane reviews because they had not been updated after 2007. The decision to exclude Cochrane reviews that had not been updated during the previous five years was arbitrarily chosen and is open to debate. However, considering the substantial number of new trials that have been published on this topic in the past few years, we would suggest that including reviews that are more than 5 years old would not reflect the most up-to-date evidence. Although we performed extensive literature searches, the selection of the three non-Cochrane reviews can also be questioned. However, other systematic reviews on the effects of exercises for LBP, NP and SP also found more or less similar results. For LBP, Macedo
et al. [
39] systematically reviewed 14 RCTs on the effects of motor control exercises for persistent LBP. Their pooled estimates for the comparison between motor control exercise and minimal intervention in reducing pain and disability in both the short and long term were higher than those in the review of Ferreira
et al. [
21]. van Middelkoop
et al. [
40] provided an overview on the effects of ET in patients with chronic LBP; they included 37 RCTs that compared exercise with usual care and found that ET improved post-treatment pain intensity and disability. They also found no evidence that one particular type of ET was clearly more effective than others. In a review of nine trials on the effects of various types of exercise for prevention and cure of non-specific NP in office workers, Sihawong
et al. [
41] found strong evidence for the effects of muscle strengthening and endurance exercises in treating NP. Moderate evidence supported the use of muscle endurance exercise in reducing disability attributed to NP. Teasell
et al. [
42] reviewed the strength of evidence supporting various therapies for whiplash-associated disorder (WAD). They included 40 trials, and found that exercise and mobilization programs for acute and chronic WAD had the strongest supporting evidence, although many questions remain about the relative effects of various protocols. For SP, a recent systematic review of four RCTs examining the effects of exercise for rotator-cuff tendinopathy concluded that the available literature was supportive that exercise reduced pain and functional disability [
43]. In a systematic review from 2011, Brudvig
et al. [
44] summarized the published research evidence on the effects of therapeutic exercise and joint mobilization compared with therapeutic exercise alone in patients with shoulder dysfunction. They found no evidence for the beneficial effects of the combination of therapeutic exercise and joint mobilization versus therapeutic exercise alone for reducing pain and increasing function.
An interesting finding of the present study is that Cochrane reviews seem to be of higher methodological quality than non-Cochrane reviews, which is also consistent with reviews specifically addressing this issue [
45,
46]. Because reviews provide limited information about included trials, the conclusions from tertiary research (overview of systematic reviews) may become too general to be considered as clinically relevant. However, we suggest that findings from overviews can be valuable as a 'compass' for the clinician, and that the decision on the type, dose, and timing of intervention should be shared between the clinician and patient at each clinical encounter. Another limitation related to clinical relevance is the magnitude of the estimated effect sizes. Using a traditional and very rough definition, SMDs between 0.30 and 0.65 can be considered as small to moderate. However, whether the effects are clinically worthwhile is a complex question that should also include the patient's perceptions of risks and costs with the actual treatment. The definition of relatively broad diagnostic groups used in the present overview might also be considered as a limitation from a clinician's point of view. For the regional pain syndromes (LBP, SP, and NP) it is obvious that one size does not fit all; that is, that the effect of ET might not be equal across diagnostic sub-groups. For example, for back pain a recent randomized trial compared stratification of the management based on the patient's prognosis with non-stratified current best practice, and found that stratified care produced disease-specific and general health benefits [
47].
Although few reviews addressed potential adverse effects (AEs), exercise for people with MSCs is generally safe and well tolerated, but patients may report minor AEs of exercise such as pain and discomfort [
48]. For OP, Howe
et al. [
15] found that fractures and falls were reported as AEs in some studies, but there was no significant effect on the numbers of fractures. By contrast, results from a systematic review showed that exercise can reduce falls, fall-related fractures, and several risk factors for falls in individuals with low BMD [
49]. In general, there are few contraindications to prescription of exercise, but co-morbidities should be considered. When prescribing ET as part of the disease management, the exercise program must be individually adopted and targeted depending on the disease severity of the individual patient, and on their physical fitness and any co-morbid conditions. A number of co-morbidities may have an influence on the burden and prognosis of MSCs; in particular, inflammatory rheumatic diseases are associated with a considerably increased risk for CVD [
36].
Conclusions
In conclusion, the present overview shows that ET can decrease pain and improve physical functioning, but there were substantial differences in the amount of research evidence across the included diagnostic groups. For example, the relevant comparison for knee OA included 32 trials with more than 3,600 patients, whereas the results for RA and AS are based on one or two trials with between 50 and 150 patients. Consequently, the pooled estimates for knee OA, LBP, FM, and SP showed consistently significant effects in favor of exercise for both outcomes, whereas for NP, hip OA, RA, and AS, the effect estimates were generally smaller and not always significant. However, for the management of all MSCs included in the present overview, ET is unanimously recommended.
There are, however, important limitations when it comes to implications for clinical practice and research. Firstly, for prescription of exercise programs with optimal health benefits for the individual patient, more knowledge is needed on which particular elements and modes of ET, as well as the doses and frequency of delivery, can improve the outcomes of interest. Secondly, knowledge of the potential influence of ET on disease parthenogenesis and the long-term effects on disease progression is currently very limited. Thirdly, it is still an open question whether the magnitude of the positive effects is clinically worthwhile and whether ET is a cost-effective intervention.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors contributed to the planning of this review, and all participated in the different phases of the trial-selection process. GS and KBH extracted data, and KBH wrote the first draft of the article. All authors read and provided feedback on the draft versions of the article. All the authors have read and approved the final version. KBH is the guarantor.