Effectiveness of non-pharmacological interventions on sleep characteristics among adults with musculoskeletal pain and a comorbid sleep problem: a systematic review.

Sleep problems are common and may be associated with persistent pain. It is unclear whether non-pharmacological interventions improve sleep and pain in adults with comorbid sleep problems and musculoskeletal (MSK) pain. We conducted a systematic review on the effectiveness of non-pharmacological interventions on sleep characteristics among adults with MSK pain and comorbid sleep problems. We searched MEDLINE, EMBASE, CINAHL, Cochrane Central and PsycINFO from inception to April 2, 2021 for randomized controlled trials (RCTs), cohort, and case-control studies. Pairs of independent reviewers critically appraised and extracted data from eligible studies. We synthesized the findings qualitatively. We screened 8459 records and identified two RCTs (six articles, 467 participants). At 9 months, in adults with insomnia and osteoarthritis pain, cognitive behavioral therapy for pain and insomnia (CBT-PI) was effective at improving sleep (Insomnia Severity Index, ISI) when compared to education (OR 2.20, 95% CI 1.25, 3.90) or CBT for pain (CBT-P) (OR 3.21, 95% CI 1.22, 8.43). CBP-P vs. education was effective at increasing sleep efficiency (wrist actigraphy) in a subgroup of participants with severe pain at baseline (mean difference 5.45, 95% CI 1.56, 9.33). At 18 months, CBT-PI, CBT-P and education had similar effectiveness on sleep and pain or health outcomes. In adults with insomnia and knee osteoarthritis, CBT-I improved some sleep outcomes including sleep efficiency (diary) at 3 months (Cohen's d 0.39, 95% CI 0.24, 1.18), and self-reported sleep quality (ISI) at 6 months (Cohen's d - 0.62, 95% CI -1.01, - 0.07). The intervention was no better than placebo (behavioural desensitization) for improving other sleep outcomes related to sleep onset or pain outcomes. Short-term improvement in sleep was associated with pain reduction at 6 months (WOMAC pain subscale) (sensitivity 54.8%, specificity 81.4%). Overall, in two acceptable quality RCTs of adults with OA and comorbid insomnia, CBT-PI/I may improve some sleep outcomes in the short term, but not pain outcomes in the short or long-term. Clinically significant improvements in sleep in the short term may improve longer term pain outcomes. Further high-quality research is needed to evaluate other non-pharmacological interventions for people with comorbid sleep problems and a range of MSK conditions.