Lumbar spinal stenosis

Lumbar spinal stenosis is a narrowing of the spinal canal in the lower back that presses on the nerves inside it, and the pain it causes in the buttocks and legs is called neurogenic claudication, as the 2013 Cochrane review of non-operative treatment describes it. The protocol of a 2024 trial from Beijing defines the claudication as pain, numbness, weakness or tingling in the buttocks or legs that builds while a person stands or walks and eases when they sit, lie down or bend forward. Trials of acupuncture for it mostly measure disability on the Roland-Morris Disability Questionnaire (RMDQ), 24 questions about daily activities scored from 0 to 24, and compare acupuncture with sham acupuncture, with usual care, or with other ways of giving acupuncture.

The Cochrane review

The 2013 Cochrane review by Ammendolia and colleagues searched to June 2012, accepted only randomised trials published in English, and included 21 trials with 1,851 participants. None tested acupuncture: the trials covered drugs, calcitonin, epidural injections, exercise and physical therapy, and non-operative care compared with surgery. The authors concluded that moderate- and high-quality evidence for any non-operative treatment was lacking.

The same group published an update of that review in BMJ Open in 2022, outside Cochrane, searched to July 2020. It added 23 trials, for 44 trials and 3,792 participants in all, and graded each comparison with GRADE. Two of the new trials tested acupuncture. One trial of 80 people gave low-quality evidence that acupuncture improved back and leg pain, symptoms and function against sham acupuncture, from immediately after treatment to the intermediate term. One trial of 50 people gave very low-quality evidence that acupuncture added to usual care was no better than usual care alone in the short term. The authors judged the evidence on acupuncture insufficient to draw conclusions on its effectiveness. The update kept the English-only rule, so trials published in Chinese did not enter it.

Later reviews

The 2013 review by Kim and colleagues at Pusan National University in South Korea found only trials from China, all reported in Chinese, and none compared acupuncture with sham, no treatment or usual care. In the pooled trials, acupuncture combined with other treatments or with extra stimulation improved more patients than simpler acupuncture (relative risk 1.16). The authors found the risk of bias high or uncertain and reported no conclusive evidence of effectiveness or safety.

The 2023 review by Sun and colleagues reported that acupuncture for 6 to 8 weeks reduced back pain and improved lumbar function more than sham at 6 months, on moderate-certainty evidence from 2 trials, and that at 3 months it reduced back pain on low-certainty evidence without changing leg pain. It found the evidence on other Chinese medicine therapies insufficient for a conclusion.

The 2026 network meta-analysis by Chen and colleagues compared non-surgical treatments for neurogenic claudication from stenosis with each other and with placebo. Against placebo, the review found that acupuncture lowered short-term pain by 10 points and improved function by 14 points, both on scales converted to 0 to 100 and on low-certainty evidence, and no other treatment showed a clear benefit over placebo for pain or function. No treatment, acupuncture included, improved walking distance. Its acupuncture trials gave acupuncture to 164 people in all, and the authors judged the findings inconclusive. Chen and colleagues work at Guang’anmen Hospital in Beijing, from which the two Chinese sham-controlled trials below were led.

Trials

Both Chinese trials needled the same five points on both sides of the body: BL23 (Shenshu), BL25 (Dachangshu), BL40 (Weizhong), BL57 (Chengshan) and KD03 (Taixi), which the protocol of the first trial says were chosen from expert experience and earlier research. At BL25 the needle went 40 to 70 mm deep.

The first, by Qin and colleagues in 2020, randomised 80 people over 50 to 24 sessions of acupuncture over 8 weeks or a sham with needles that did not pierce the skin. At the end of treatment RMDQ scores had fallen by 4.1 points with acupuncture and 1.5 with sham, a difference of 2.6, and leg and buttock pain was 2.9 points lower on a 0 to 10 scale. The authors wrote that acupuncture gave immediate functional recovery and pain relief, and also that the evidence was insufficient to say it offered a clinical benefit over non-insertive sham.

The second and larger, by Zhu and colleagues in 2024 in the Annals of Internal Medicine, randomised 196 people with mainly claudication symptoms at 5 hospitals in China to 18 sessions over 6 weeks. Under its protocol the sham group had needles inserted 2 to 3 mm at the same points and left unmanipulated, with patients, outcome assessors and statisticians blinded. At 6 weeks RMDQ scores had fallen from 12.6 to 8.1 with acupuncture and from 12.7 to 9.5 with sham, an adjusted difference of 1.3 points, below the 2 to 3 points the trial took as the minimal clinically important difference. A 30% improvement was 16 percentage points more common with acupuncture. Each group had 3 mild, transient adverse events. The authors wrote that acupuncture may relieve pain-specific disability, with effects that may last 24 weeks, and that the shallow-needling sham could itself have physiological effects. A 2026 secondary analysis of the 178 patients who completed treatment found a shorter current episode of symptoms associated with a larger response to acupuncture, a finding its authors called preliminary.

A 2016 pilot trial led from Pusan National University randomised 50 people to manual acupuncture, with or without electroacupuncture, for 12 to 16 sessions over 6 weeks on top of usual care, or to usual care alone. Disability on the Oswestry Disability Index did not differ at 3 months (mean difference 2.5 points, confidence interval crossing zero), and outcome assessors were not blinded. Of 524 treatment visits, 61 had an adverse event; all but one were minor, and the one severe event was judged unrelated to acupuncture.

A 2018 Japanese study by Oka and colleagues assigned 119 people with L5 nerve root pain from stenosis to acetaminophen, exercise or acupuncture by the time of their visit rather than at random, with no blinding. After 4 weeks, physical function on the Zurich Claudication Questionnaire improved significantly only in the acupuncture group, and more than with exercise.

What guidelines say

The North American Spine Society’s guideline on degenerative lumbar spinal stenosis, revised in 2011, found insufficient evidence to recommend for or against acupuncture (grade I). It based the grade on one prospective comparative study of 35 patients, which it rated as level III evidence because it was small and used no validated outcome measures, and it called for controlled trials. NASS still lists the 2011 revision among its completed guidelines, with no stenosis revision among those in progress, so it predates every trial above.

A 2021 guideline by Bussières and a multidisciplinary North American panel, published in the Journal of Pain and based on a systematic review of trials published to June 2019, says clinicians and patients may select traditional acupuncture on a trial basis for neurogenic claudication from stenosis. Every recommendation in it is conditional, and the panel rated the acupuncture evidence very low quality.

The National Institute for Health and Care Excellence (NICE) guideline NG59, on low back pain and sciatica in people over 16, was published in November 2016 and last updated in July 2026. Recommendation 1.2.8 says not to offer acupuncture for low back pain with or without sciatica, dated 2016. NG59 addresses stenosis only in recommendation 1.3.6, against epidural injections for neurogenic claudication from central canal stenosis, and makes no acupuncture recommendation for stenosis itself. Clinical guidelines compares other bodies’ positions, and Low back pain and Sciatica and lumbar disc herniation cover the neighbouring conditions.

What the reviews could not settle

The sham-controlled evidence is two trials led from one Beijing hospital, with 276 people between them, and the 2026 review that pools them comes from the same hospital. The two shams differ: one did not pierce the skin, the other needled shallowly at the real points, and the larger trial’s difference over its shallow sham fell below its own threshold for a clinically important change. Walking distance, which the 2026 review calls clinically important in stenosis, improved under none of the treatments that review compared, and its authors asked for sham-controlled trials that measure it. The reviews that excluded trials in Chinese and the one that included only Chinese trials looked at different evidence, and the Chinese-language trials in the 2013 review compared acupuncture methods with each other rather than with sham or usual care. The one trial against usual care was a pilot of 50 people.

AcuiQ’s spinal stenosis and lumbar spinal stenosis pages list the protocols individual studies prescribed, each with its citation.