Patient-friendly summary
If you read nothing else
Bottom line
CBT for insomnia (CBT-I) is the most promising treatment for improving sleep quality in fibromyalgia; pregabalin and sodium oxybate may help modestly but carry risks, and amitriptyline, milnacipran and duloxetine show no clear sleep benefit. Medication should not be first-line for sleep in fibromyalgia.
Moderate evidencePublished
Evidence hierarchy
Study participants
Adults with fibromyalgia (ACR 1990/2010/2016, AAPT or AWMF criteria)
Full research — for clinicians and curious readers
Study Summary
This systematic review and meta-analysis pooled 47-48 randomized controlled trials (reported as 48 RCTs included; 11,094 participants in the abstract, 11,140 in the characteristics) to assess how pharmacological treatments and cognitive behavioural therapy (CBT) affect sleep quality in people with fibromyalgia. Cognitive behavioural therapy for insomnia (CBT-I) produced a statistically significant, moderate improvement in sleep quality (SMD -0.63, 95% CI -0.98 to -0.27), whereas CBT targeting pain (CBT-P) showed no significant benefit. Among medications, pregabalin (SMD -0.35) and sodium oxybate (SMD -0.72) moderately improved sleep but with uncertainty and safety concerns, while amitriptyline, milnacipran and duloxetine showed no significant benefit. The authors conclude CBT-I is a promising sleep treatment in fibromyalgia and that drug therapy should not be first-line for sleep disturbance.
Key Findings
| Finding | Detail | Impact |
|---|---|---|
| CBT for insomnia (CBT-I) significantly improved sleep quality | Meta-analysis of 7 CBT-I studies (387 patients) showed a statistically significant moderate improvement (SMD -0.63, 95% CI -0.98 to -0.27) with moderate heterogeneity (I2 = 37%) and no detected publication bias (Egger's P = 0.60). The majority of CBT trials used sleep as a primary outcome. | High |
| CBT targeting pain (CBT-P) did not improve sleep | Meta-analysis of CBT-P (4 studies, 288 patients) showed no significant effect on sleep (SMD -0.61, 95% CI -1.70 to 0.48) with substantial heterogeneity (I2 = 83%). Combined CBT-I/CBT-P did not outperform CBT-I alone, suggesting sleep interventions should target sleep directly. | High |
| Pregabalin moderately improved sleep but with uncertainty and risk | Eight studies (3834 patients, 300-600 mg over 4-26 weeks) pooled to a statistically significant moderate improvement (SMD -0.35, 95% CI -0.54 to -0.16) but with high heterogeneity (I2 = 78%). A trend toward dose-response favored 450 mg over 300 mg, with no added benefit at 600 mg. Authors note misuse/adverse-event concerns and that NICE withdrew its recommendation. | High |
| Sodium oxybate showed a significant sleep benefit but limited, uncertain evidence | Four studies (1516 patients, 4.5-6 g daily) pooled to a significant improvement (SMD -0.72, 95% CI -1.26 to -0.18) with high heterogeneity (I2 = 88%) and a wide prediction interval. One study showed increased NREM and decreased REM sleep on polysomnography. Misuse concerns limit its role. | Medium |
| Amitriptyline, milnacipran and duloxetine showed no significant sleep benefit | Amitriptyline meta-analysis (6 studies) was non-significant (SMD -0.71, 95% CI -2.05 to 0.63, I2 = 82%); milnacipran (4 studies) showed essentially no effect (SMD -0.01, 95% CI -0.07 to 0.05, I2 = 0%); a single duloxetine study (41 patients) found no benefit. Amitriptyline reliance on VAS scales (which correlate more with pain than sleep) complicates interpretation. | High |
Strengths
- Comprehensive search across five databases (PubMed, MEDLINE, Embase, Cochrane CENTRAL, CINAHL) with snowball searching of references and trial registries
- Large evidence base: 48 RCTs and roughly 11,100 participants
- Evaluated both pharmacological and CBT interventions, and is one of few reviews to compare CBT-I against CBT-P, isolating CBT-I's specific sleep benefit
- Rigorous statistical methods including random-effects models, prediction intervals, dose-response subgroup analyses, leave-one-out sensitivity and influence analyses, and Egger's test for publication bias
- Risk of bias formally assessed with the Cochrane Risk of Bias tool
Limitations
- Substantial heterogeneity for several interventions (e.g. amitriptyline I2 = 82%, pregabalin I2 = 78%, sodium oxybate I2 = 88%, CBT-P I2 = 83%), with wide prediction intervals indicating uncertainty
- Most pharmacological trials were designed around pain rather than sleep, and many used VAS scales that may reflect pain rather than true sleep quality
- CBT studies had high overall risk of bias (6 of 8 rated high risk; none low risk), with poor treatment adherence and incomplete follow-up
- Heterogeneous fibromyalgia case definitions (criteria evolved 1990-2016), intervention protocols, control arms and sleep outcome measures limit comparability
- Short median follow-up (~12 weeks for drug trials) limits conclusions about long-term effects; only a few CBT-I trials followed patients to 6 months
- Minor internal discrepancy in reported totals (47 vs 48 RCTs; 11,094 vs 11,140 participants)
Key Takeaways for Patients
What This Means for You
If poor sleep is a major problem with your fibromyalgia, a structured talking therapy called CBT for insomnia (CBT-I) had the most consistent evidence for improving sleep quality in this review.
Therapy aimed at pain coping (rather than sleep itself) did not appear to improve sleep, so sleep problems are best treated directly.
Some medications, such as pregabalin and sodium oxybate, may modestly help sleep, but the evidence is uncertain and these drugs carry safety and misuse risks, so the authors say medication should not be the first-line approach for sleep.
Commonly used fibromyalgia drugs like amitriptyline, milnacipran and duloxetine did not show a clear benefit for sleep quality in this analysis.
These conclusions come from trials of varying quality and design, so discuss the trade-offs of any treatment with your clinician.
Read the Full Paper
Access the complete peer-reviewed study from Rheumatology (Oxford)