Insomnia afflicts more than half of all patients with advanced cancer, yet few large trials have examined non-drug treatments for this population. A new multicenter exploratory randomized controlled trial from Memorial Sloan Kettering Cancer Center provides the strongest evidence to date that both acupuncture and massage therapy produce clinically meaningful improvements in insomnia for people living with advanced cancer, and that sleep and pain appear to reinforce each other in a bidirectional loop.
The study, published in the Journal of Pain and Symptom Management, drew data from a larger pragmatic trial originally designed to compare acupuncture versus massage for cancer pain. Of the 300 participants enrolled across multiple centers, 234 had clinically significant insomnia at baseline, defined as a score of 8 or higher on the Insomnia Severity Index (ISI). The researchers followed this subgroup through 10 weeks of active treatment and tracked outcomes for up to 26 weeks.
What they found
Both treatments delivered substantial and nearly identical insomnia improvements. The average ISI score dropped by 3.86 points in the acupuncture group and by 3.83 points in the massage group at week 10. Both reductions were highly statistically significant (p < 0.0001), and the between-group difference was negligible (p = 0.91), meaning neither therapy outperformed the other for sleep.
When the researchers looked at clinical response rates, the pattern held. About 35% of acupuncture participants and 33% of massage participants achieved a meaningful insomnia response, defined as either an ISI drop of 8 or more points or a final ISI score below the insomnia threshold of 8.
Pain outcomes also favored both interventions. Nearly half of acupuncture participants (48.0%) and nearly two-thirds of massage participants (62.9%) showed a 30% or greater reduction in pain severity on the Brief Pain Inventory at week 10.
Perhaps the most striking finding was the bidirectional relationship between sleep and pain. Participants whose pain responded to treatment (30% or greater reduction) experienced significantly greater insomnia improvement at week 10, with an average additional ISI drop of nearly 3 points compared to pain non-responders (delta = -2.97, p = 0.0003). The reverse was also true: participants whose insomnia responded to treatment showed significantly greater pain reduction at both week 10 and week 26 (both p < 0.0001). This suggests that improving one symptom may naturally pull the other along with it, a pattern the researchers describe as supporting symptom-cluster approaches to care.
Why it matters
Insomnia in advanced cancer is often undertreated. Medications such as benzodiazepines and Z-drugs carry risks of falls, cognitive impairment, tolerance, and dependence, and they are poorly studied in seriously ill populations. Non-pharmacologic options are recommended by clinical guidelines but remain underutilized, partly because the evidence base has been thin for patients with advanced disease.
This study, while exploratory, offers something rare: head-to-head data on two widely available integrative therapies in a sizable sample of advanced cancer patients with insomnia. The finding that both acupuncture and massage produce clinically meaningful sleep improvements, and that pain and sleep improvement travel together, gives clinicians practical options. For patients already receiving one of these therapies for pain, the sleep benefit may come as a welcome side effect. For those who decline or cannot tolerate sleep medications, acupuncture or massage may serve as an effective alternative.
The bidirectional sleep-pain finding also carries implications beyond cancer care. The relationship between poor sleep and heightened pain sensitivity is well documented in chronic pain populations, but this study demonstrates the connection in a treatment context: intervening on either symptom may produce gains in the other. This argues for treatment approaches that address symptom clusters rather than individual complaints in isolation.
Limits
Several caveats are important. The insomnia analysis was a secondary, exploratory endpoint of a trial originally powered for pain, not sleep. The study was an open-label pragmatic trial without a sham acupuncture or placebo control group, so some of the observed improvement could reflect expectation effects or natural recovery rather than specific treatment effects. The 26-week follow-up showed attenuation of some sleep-pain interaction effects, suggesting durability may vary. Additionally, the massage group showed a numerically higher pain response rate, though the study was not designed to test superiority between the two therapies, and the authors did not report a formal statistical comparison for that difference.
Bottom line
Acupuncture and massage both produce meaningful insomnia improvement in patients with advanced cancer, and sleep and pain appear to improve together. For clinicians caring for this population, these findings support offering integrative therapies as part of a symptom-management toolkit, particularly for patients who cannot or prefer not to use sleep medications.
Source
Kwag E, Fok RW, Li QS, Baser RE, Garland SN, Liou KT, Mao JJ, McConnell K. Acupuncture vs Massage for Insomnia and Pain in Advanced Cancer: Exploratory RCT. Journal of Pain and Symptom Management. 2026. doi:10.1016/j.jpainsymman.2026.07.008. PMID: 42471129.