Background

Exercise training is commonly recommended for individuals with fibromyalgia. This review is one of a series of reviews about exercise training for fibromyalgia that will replace the review titled “Exercise for treating fibromyalgia syndrome”, which was first published in 2002.

Objectives

To evaluate the benefits and harms of mixed exercise training protocols that include two or more types of exercise (aerobic, resistance, flexibility) for adults with fibromyalgia against control (treatment as usual, wait list control), non exercise (e.g. biofeedback), or other exercise (e.g. mixed versus flexibility) interventions.
Specific comparisons involving mixed exercise versus other exercises (e.g. resistance, aquatic, aerobic, flexibility, and whole body vibration exercises) were not assessed.

Search methods

We searched the Cochrane Library, MEDLINE, Embase, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), Thesis and Dissertations Abstracts, the Allied and Complementary Medicine Database (AMED), the Physiotherapy Evidence Databese (PEDro), Current Controlled Trials (to 2013), WHO ICTRP, and ClinicalTrials.gov up to December 2017, unrestricted by language, to identify all potentially relevant trials.

Selection criteria

We included randomised controlled trials (RCTs) in adults with a diagnosis of fibromyalgia that compared mixed exercise interventions with other or no exercise interventions. Major outcomes were health‐related quality of life (HRQL), pain, stiffness, fatigue, physical function, withdrawals, and adverse events.

Data collection and analysis

Two review authors independently selected trials for inclusion, extracted data, and assessed risk of bias and the quality of evidence for major outcomes using the GRADE approach.

Main results

We included 29 RCTs (2088 participants; 98% female; average age 51 years) that compared mixed exercise interventions (including at least two of the following: aerobic or cardiorespiratory, resistance or muscle strengthening exercise, and flexibility exercise) versus control (e.g. wait list), non‐exercise (e.g. biofeedback), and other exercise interventions. Design flaws across studies led to selection, performance, detection, and selective reporting biases. We prioritised the findings of mixed exercise compared to control and present them fully here.

Twenty‐one trials (1253 participants) provided moderate‐quality evidence for all major outcomes but stiffness (low quality). With the exception of withdrawals and adverse events, major outcome measures were self‐reported and expressed on a 0 to 100 scale (lower values are best, negative mean differences (MDs) indicate improvement; we used a clinically important difference between groups of 15% relative difference). Results for mixed exercise versus control show that mean HRQL was 56 and 49 in the control and exercise groups, respectively (13 studies; 610 participants) with absolute improvement of 7% (3% better to 11% better) and relative improvement of 12% (6% better to 18% better). Mean pain was 58.6 and 53 in the control and exercise groups, respectively (15 studies; 832 participants) with absolute improvement of 5% (1% better to 9% better) and relative improvement of 9% (3% better to 15% better). Mean fatigue was 72 and 59 points in the control and exercise groups, respectively (1 study; 493 participants) with absolute improvement of 13% (8% better to 18% better) and relative improvement of 18% (11% better to 24% better). Mean stiffness was 68 and 61 in the control and exercise groups, respectively (5 studies; 261 participants) with absolute improvement of 7% (1% better to 12% better) and relative improvement of 9% (1% better to 17% better). Mean physical function was 49 and 38 in the control and exercise groups, respectively (9 studies; 477 participants) with absolute improvement of 11% (7% better to 15% better) and relative improvement of 22% (14% better to 30% better). Pooled analysis resulted in a moderate‐quality risk ratio for all‐cause withdrawals with similar rates across groups (11 per 100 and 12 per 100 in the control and intervention groups, respectively) (19 studies; 1065 participants; risk ratio (RR) 1.02, 95% confidence interval (CI) 0.69 to 1.51) with an absolute change of 1% (3% fewer to 5% more) and a relative change of 11% (28% fewer to 47% more). Across all 21 studies, no injuries or other adverse events were reported; however some participants experienced increased fibromyalgia symptoms (pain, soreness, or tiredness) during or after exercise. However due to low event rates, we are uncertain of the precise risks with exercise. Mixed exercise may improve HRQL and physical function and may decrease pain and fatigue; all‐cause withdrawal was similar across groups, and mixed exercises may slightly reduce stiffness. For fatigue, physical function, HRQL, and stiffness, we cannot rule in or out a clinically relevant change, as the confidence intervals include both clinically important and unimportant effects.

We found very low‐quality evidence on long‐term effects. In eight trials, HRQL, fatigue, and physical function improvement persisted at 6 to 52 or more weeks post intervention but improvements in stiffness and pain did not persist. Withdrawals and adverse events were not measured.

It is uncertain whether mixed versus other non‐exercise or other exercise interventions improve HRQL and physical function or decrease symptoms because the quality of evidence was very low. The interventions were heterogeneous, and results were often based on small single studies. Adverse events with these interventions were not measured, and thus uncertainty surrounds the risk of adverse events.

Authors’ conclusions

Compared to control, moderate‐quality evidence indicates that mixed exercise probably improves HRQL, physical function, and fatigue, but this improvement may be small and clinically unimportant for some participants; physical function shows improvement in all participants. Withdrawal was similar across groups. Low‐quality evidence suggests that mixed exercise may slightly improve stiffness. Very low‐quality evidence indicates that we are ‘uncertain’ whether the long‐term effects of mixed exercise are maintained for all outcomes; all‐cause withdrawals and adverse events were not measured. Compared to other exercise or non‐exercise interventions, we are uncertain about the effects of mixed exercise because we found only very low‐quality evidence obtained from small, very heterogeneous trials. Although mixed exercise appears to be well tolerated (similar withdrawal rates across groups), evidence on adverse events is scarce, so we are uncertain about its safety. We downgraded the evidence from these trials due to imprecision (small trials), selection bias (e.g. allocation), blinding of participants and care providers or outcome assessors, and selective reporting.