Fibromyalgia: what a 2026 network meta-analysis of 151 drug trials shows

Fibromyalgia is a chronic condition characterized by widespread musculoskeletal pain ^[1]. A 2026 BMJ Medicine network meta-analysis now measures how much drug classes deliver once certainty of evidence and clinical importance are applied ^[2].

What did the 2026 BMJ Medicine review pool?

The BMJ Medicine review, published on 5 October 2026, pooled 151 randomized trials with 32,592 adults, from searches run to 13 January 2025, so trials published since are not included ^[2]. The authors included commercially available, investigational and discontinued drugs, plus supplements such as melatonin, vitamin D, vitamin B6 and omega 3 fatty acids ^[2].

What does the evidence show by drug class?

The BMJ Medicine review reports these findings by drug class, and rated the evidence for global improvement, symptom burden and sleep as low certainty ^[2]:

• Global improvement: against placebo or usual care, sodium oxybate, selective norepinephrine reuptake inhibitors (selective NRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs) and gabapentinoids may increase patient-reported meaningful improvement or recovery by 105 to 283 per 1,000 patients ^[2].
• Symptom burden: on the Fibromyalgia Impact Questionnaire, mean differences were -8.81 points (95% CI -11.78 to -5.84) for sodium oxybate and -8.02 points (95% CI -13.40 to -2.64) for selective serotonin reuptake inhibitors, reductions in the symptom score ^[2].
• Sleep: tricyclic antidepressants, sodium oxybate and gabapentinoids may improve sleep quality ^[2].
• Pain responders: for sodium oxybate, selective NRIs, SNRIs and cyclobenzaprine, responder analyses found 111 to 202 more per 1,000 patients reaching at least a 30% pain reduction, a responder measure that differs from the pooled pain scores reported below ^[2].
• Harms: the review found that no drug increased serious adverse events, and low-certainty evidence suggested that only low-potency opioids may increase discontinuation; its authors note that short follow-up likely overestimated tolerability ^[2].

For pain, physical health and mental health, no drug had a pooled effect that met or exceeded the threshold for clinical importance ^[2]. The review authors conclude that these classes may provide typically small benefits in select outcomes ^[2].

What does the evidence say about exercise and education?

The 2017 EULAR revised recommendations rated exercise strong-for on the basis of meta-analyses ^[3]. In the EULAR graduated approach, initial management centers on patient education and nonpharmacological therapies, and the further therapies it lists were rated weak-for on the basis of meta-analyses ^[3]. The 2017 EULAR recommendations rated growth hormone, sodium oxybate, strong opioids and corticosteroids strong-against because of lack of efficacy and high risk of side effects ^[3]. These ratings predate the 2026 drug review ^[2][3]. The 2025 Chinese guideline, drawing on 57 studies, strongly recommends acupuncture and health education on moderate-quality evidence ^[4]. Both guidelines were developed outside the United States ^[3][4]. In a 2026 network meta-analysis of 21 trials (1,638 participants), searched to 15 October 2022, mind-body exercise was associated with lower pain scores than routine care (SMD -1.62, 95% CI -2.35 to -0.90) ^[5]. The same network meta-analysis associated resistance training with lower pain scores as well (SMD -1.58, 95% CI -2.67 to -0.49) ^[5]. These are standardized mean differences against routine care in a smaller network, so they are not directly comparable with the drug figures above ^[5].

Where is the evidence uncertain?

In the 2025 Health Technology Assessment network of sleep outcomes (65 studies, 8,247 participants), most studies were at high risk of bias and certainty was predominantly low to very low ^[1]. The HTA authors describe the trials as generally small, unblinded and short, with a median duration of 12 weeks ^[1].

At the point of care: what to check

• Target outcome: whether pain, sleep, function or global improvement matters most, since class benefits in the BMJ Medicine review differed by outcome ^[2].
• Label status: whether each agent discussed carries a US fibromyalgia indication, since the BMJ Medicine review pooled commercially available, investigational and discontinued drugs ^[2].
• Certainty: whether the outcome that matters to the patient is one the BMJ Medicine review rated low certainty: global improvement, symptom burden or sleep ^[2].
• Harms: whether the trial data behind an agent report discontinuation and serious adverse events over a follow-up long enough to matter, since the BMJ Medicine review's short trials likely overestimated tolerability ^[2].
• Nonpharmacological care: whether exercise and education are already in place, given the EULAR strong-for rating of exercise ^[3].