Efficacy of Nonpharmacological Interventions for Individual Features of Fibromyalgia

A Systematic Review and Meta-analysis of Randomised Controlled Trials

Burak Kundakci; Jaspreet Kaur; Siew Li Goh; Michelle Hall; Michael Doherty; Weiya Zhang; Abhishek Abhishek


Pain. 2022;163(8):1432-1445. 

In This Article


This is by far the largest SR including 167 RCTs examining the efficacy of 22 nonpharmacological interventions in fibromyalgia. It evaluated 5 patient-centred outcomes in fibromyalgia: disease-specific QoL (FIQ), pain, fatigue, sleep, and depression. The main findings are as follows: (a) several nonpharmacological interventions are effective for FIQ, pain, fatigue, sleep, and depression; (b) different interventions benefit different outcomes; and (c) exercise overall seems to improve all outcomes but the benefits vary between types of exercise for each specific outcome, for example, although all types of exercise were effective for pain apart from flexibility exercise, only aerobic and strengthening exercises were effective for sleep. These findings were also confirmed in analyses restricted to high-quality studies and in a sensitivity analysis based on the data used to compute ES, that is, change score or end point score.

Exercise is superior to usual care and has moderate-to-large ES for improving FIQ, pain, sleep, fatigue, and depression. However, studies included were prone to high RoB especially on blinding of the participant and the outcome assessor. No previous MA has assessed all exercise types for fibromyalgia. Our results are consistent with MAs in similar conditions such as osteoarthritis and chronic low back pain (CLBP).[70,83] With respect to different exercise types, all forms of exercise improved pain and depression except for flexibility exercise. Mind–body and strengthening exercises improved fatigue, whereas aerobic and strengthening exercises improved sleep. Some of our findings differ from those reported in previous SRs. For example, while we found a significant moderate ES of aerobic exercise on pain (−0.72; 95% CI −1.37 to −0.07), a Cochrane review by Busch et al. (2007) reported nonsignificant differences (0.65; 95% CI −0.09 to 1,39) for this outcome. This may be due to a larger number of studies (8 vs 5) and total sample size (512 vs 183) in the current SR.

Our study reports that education alone may not be superior to usual care for FIQ, pain, sleep, and depression but may improve fatigue. A SR by Sim and Adams[173] reported modest improvement on pain from an education program in people with fibromyalgia. However, their findings were based on a study that used a cognitive approach rather than education alone, and improvements were not maintained during follow-up. Similarly, previous reviews on CLBP and chronic musculoskeletal pain[44,111] found significant improvements. Clarke et al.[44] evaluated 2 studies (n = 122) and reported small improvements on CLBP (5/100 mm on a Visual Analogue Scale). Louw et al.[111] systematically reviewed 13 articles and reported significant findings for pain but their population group comprised different musculoskeletal conditions. Nevertheless, it is difficult to study education as a discrete component of care as it is a core requirement and an integral component of shared decision-making in any nonpharmacological intervention.

We examined 29 psychological intervention studies including CBT, mindfulness, hypnosis, acceptance, and commitment therapy that were more effective than usual care on FIQ, pain, sleep, and depression but not fatigue. Our results are broadly in line with those of previous studies.[24,69] We found that CBT was more efficacious for pain, whereas mindfulness was better than usual care for fatigue and depression. We did not find a significant effect of CBT on depression in contrast to an earlier review by Bernardy et al.[24] However, we were strict in our definition of "CBT" as an intervention; this was specified as "CBT" that was delivered by a trained professional, whereas Bernardy et al. considered as CBT, any psychological treatment based on a CBT model or framework.

Glombiewski et al.[69] reviewed 23 psychological treatment studies and reported significant improvements on fatigue. This conflicting result may reflect the type of interventions studied. Although we excluded studies including another nonpharmacological intervention (ie, exercise) in combination with psychological treatments, Glombiewski et al.'s review included multimodal programs if the psychological treatment accounted for at least 60% of treatment time. Because of the difficulty in knowing the proportion of time spent on the intervention from articles, we classified these interventions as MDT.

We defined MDT as combined nonpharmacological intervention that includes components of exercise, education, and psychological treatment. We found that MDT was superior to usual care on FIQ, pain, sleep, and depression. Further to this, it indicated larger ESs compared with exercise only, education only, and psychological treatment only. No previous MA has assessed the effect of MDT exclusively in fibromyalgia. Previous qualitative analyses,[30,93] however, suggest that MDT is effective for decreasing pain and fibromyalgia impact.

We observed differential effect of interventions on different manifestations of fibromyalgia. Aerobic, strengthening, and mind–body exercise but not flexibility exercise reduces neuroinflammation, increase endogenous opioid and serotonin release, and influence dopamine and norepinephrine levels explaining improvement in most symptoms of fibromyalgia by the first 3 exercise types.[27] Similarly, other interventions such as acupuncture and massage therapy that affect both central and peripheral pain mechanisms and have sympatholytic and inhibitory effects on the hypothalamic pituitary adrenal axis improved most symptoms.[54] Transcranial direct current stimulation improved all manifestations given the role of central wind-up in all fibromyalgia manifestations.[20] Our SR also raises the possibility that exercise and balneotherapy may be more clinically effective with combined with other interventions. There were few studies, and more such studies are needed.

This SR shows that the efficacy of nonpharmacological intervention for fibromyalgia seems to drop after 14 weeks. This suggests that patients completing a nonpharmacological treatment programme for fibromyalgia should be reviewed approximately 3 monthly and the treatments reinforced as required. We tested a wide range of nonpharmacological interventions in this study. Several interventions such as transcranial direct current stimulation, balneotherapy, mind–body exercises, and acupuncture may not be widely available across different healthcare systems. Indeed, there may be low acceptability of some of these interventions, such as psychological treatments or acupuncture, in certain cultures.

This study provides data that will help healthcare professionals and patients to select the nonpharmacological interventions, that is, most are likely to give the best results according to the patients' individual clinical features. It is a comprehensive review with no language or geographic restriction and includes all nonpharmacological interventions and their impact on disease-specific QoL (FIQ) and common symptoms in fibromyalgia—pain, fatigue, sleep, and depression.

However, the present MA is subject to several caveats. First, more than half of the studies (53%) had small sample size (<50). Second, there was a high RoB for blinding, and allocation concealment was unclear for most trials. However, because of the nature of nonpharmacological interventions, it is usually not possible to blind participants and those delivering the intervention. These studies with a small sample size and high RoB skewed the funnel plot to the left. For example, Evcik et al.[56] included 42 participants and had unclear RoB in allocation concealment and high RoB in blinding. This may have affected overall effect size and often overestimated the overall ESs.[140] Nevertheless, the sensitivity analysis for allocation concealment indicated that the findings were unlikely to be biased by study quality. In addition, it was not possible to synthesise all published evidence because of incomplete and inconsistent outcome reporting. Visual asymmetry of the funnel plots suggested that this loss of data may have biased the results. However, we attempted to impute missing data and sensitivity analyses on imputed data indicated that results are less likely to be biased because of data imputation. Moreover, most participants in the studies were predominantly middle-aged women living in a developed country. Further studies are required to confirm the generalisability of the results to all populations with fibromyalgia and in all settings. Secondary outcomes such as pain, fatigue, sleep, and depression were measured using different patient-reported outcome measures. This may have biased the ESs on these outcomes. However, the main analysis using primary outcome measure (FIQ) does not suffer from this caveat. Finally, some groups of interventions were quite heterogeneous. For example, different types of acupuncture were used in acupuncture trials including traditional acupuncture, electroacupuncture, or dry needling. Furthermore, intervention programmes and control groups are not standardised and vary considerably between studies. Substantial heterogeneity was present on subgroup analysis raising the possibility that variability in intervention such as number and duration of sessions, overall duration of treatment programme, way of delivery, expertise of the therapist, setting, and follow-up duration may have caused the heterogeneity. Researchers should describe these in detail so that their effect may be explored in future SRs.

In summary, these results suggest that several nonpharmacological interventions are effective for fibromyalgia and that different interventions improve different patient-centred outcomes. Exercise is the most beneficial for improving multiple patient-centred outcomes, but different types of exercise preferentially benefit different outcomes, supporting the possibility of individualised nonpharmacological management according to predominant symptom(s), a strategy that warrants testing in clinical trials. Future trials that address the methodological limitations highlighted above are also required, especially high-quality trials with larger sample size, longer duration of follow-up, the use of consistent terminology, and well-described intervention programmes.