Restless legs syndrome

Restless legs syndrome (RLS) is a sensorimotor disorder: uncomfortable sensations in the legs and an urge to move them, which disturb sleep. The evidence on acupuncture for it is a small set of small trials, and several reviews published since 2008 pool the same few of them in different combinations. Two trials compared real needling or injection at acupuncture points with a sham at non-points, and they enrolled 46 people between them.

The Cochrane review

Cochrane’s 2008 review by Ye Cui and colleagues at Guang’anmen Hospital in Beijing searched to February 2007 and found 14 potentially relevant trials. Two, with 170 people, met its criteria; both were published in Chinese and both had methodological or reporting shortcomings, and they differed too much to pool. In one, acupuncture and medication did about equally well for overall remission of symptoms (risk ratio 0.97, 95% confidence interval 0.76 to 1.24). In the other, dermal needle therapy, a shallow needling of the skin, added to medication and massage, relieved the unpleasant leg sensations and the frequency of symptoms more than medication and massage alone, but not their duration. The authors concluded that the evidence was insufficient to say whether acupuncture is effective and safe for RLS. The Cochrane Library lists no version after 2008.

Later reviews

Xu and colleagues listed acupuncture among six treatments that reduced the severity of primary RLS, as distinct from RLS that accompanies kidney failure, alongside exercise, compression devices, light therapy and two kinds of brain or spinal stimulation. Harrison and colleagues found standard acupuncture more effective than control for RLS severity and some sleep outcomes, and judged that few studies existed and the quality of evidence was not high. They add that non-drug treatments may carry placebo effects that future trials should control for.

Huang and colleagues, at hospitals in Daqing and Harbin, pooled 18 studies published before October 2020, with 640 people in acupuncture groups and 447 in control groups. They rated the studies with the Newcastle-Ottawa scale, a checklist for observational cohort studies, and found their quality low to moderate. They report 4.7% of people in the acupuncture groups as unresponsive to treatment, against 32.9% in the control groups, on the “total effective rate” scales common in Chinese trials. Across the studies that used the International Restless Legs Syndrome Study Group rating scale (IRLS), scores in the acupuncture groups fell by a mean of 9.45 points from baseline, a before-and-after figure with no control group in it. The authors call for randomised trials with a sham-acupuncture arm.

Two 2025 reviews pooled a few trials under the heading of acupressure, pressure on acupuncture points by hand. In Döner and colleagues’ review of people on haemodialysis, 10 of the 12 trials came from Iran, and the acupressure subgroup was two trials: one used finger pressure, the other near-infrared light shone on acupuncture points. Gupta and colleagues put three trials in their acupressure group, the same two plus a needle acupuncture trial, and found no statistically significant difference from control (standardised mean difference −0.87, 95% confidence interval −2.21 to 0.47).

The sham-controlled trials

A 2015 trial by Pan and colleagues at Shuguang Hospital in Shanghai enrolled 38 people with idiopathic RLS, untreated and at the severe end of the IRLS (a score of 20 or more), and analysed 31. The standard group was needled at BL23 (Shenshu), GV04 (Mingmen), SP10 (Xuehai), BL57 (Chengshan), LV03 (Taichong), ST36 (Zusanli), SP06 (Sanyinjiao) and KD03 (Taixi); the comparison group was needled at 12 sites that avoided acupuncture points, to the same depth and with the same manipulation. Both groups had three sessions a week for six weeks, and the patients did not know which they had. Night-time leg activity, recorded by an ankle-worn motion sensor called an actigraph, fell in the standard group from week 2 and differed from the comparison group from week 4, as did IRLS and daytime sleepiness scores; the comparison group did not improve on any measure. Four people left because they could not tolerate the needling. No side effects were observed.

A 2022 crossover trial by Fukutome and Murashima in Japan gave eight adults an injection of saline at ST36, GB41 (Zulinqi), SP06 and BL60 (Kunlun) on one evening, and the same dose into the outer thigh, away from the main channels, a week apart and in random order. While they lay still in bed afterwards, leg discomfort scored lower after the point injections; periodic leg movements, the involuntary jerks measured on an electromyogram, did not differ. Point injection describes the method, and Sham acupuncture sets out the kinds of sham in use and why none is inert.

Trials added to usual treatment

A 2017 trial by Raissi and colleagues in Iran randomised 46 people to 10 sessions of acupuncture over four weeks plus gabapentin at 300 mg a day, or gabapentin alone, single-blind. The authors report visual analogue scale, IRLS and sleep-quality scores in favour of the acupuncture group up to eight weeks after treatment. A 2022 crossover pilot by Tsai and colleagues in Taiwan gave 23 people on haemodialysis a month of acupressure during dialysis at points including ST36, GB34 (Yanglingquan), SP06, GB39 (Xuanzhong), BL57 and LV03, and a month of observation. RLS severity was lower in the acupressure month; sleep quality was not. In the German HYDRAC pilot, published in 2025, 54 adults were randomised to self-applied acupressure, self-applied cold-water leg affusions or routine care alone for six weeks, open-label. Mean adjusted IRLS scores at six weeks were 22.9 with acupressure and 24.0 with routine care; the study was exploratory and set out to test feasibility. Acupressure covers the technique.

The guideline

The American Academy of Sleep Medicine (AASM) published its clinical practice guideline on RLS in January 2025, with an accompanying systematic review graded with GRADE, the scale that rates confidence in a result from high to very low. The AASM lists it among its current practice guidelines. It recommends gabapentin enacarbil, gabapentin, pregabalin and, where iron status is appropriate, intravenous ferric carboxymaltose (each a strong recommendation on moderate-certainty evidence), and suggests against the standard use of the dopamine agonists pramipexole, ropinirole and rotigotine. For acupuncture it made no recommendation, listing it among treatments with insufficient and inconclusive evidence. The review found one eligible trial, the 2017 Iranian trial, and analysed 33 of its participants. The IRLS difference was 2.5 points in acupuncture’s favour, with a 95% confidence interval from 10.0 points in its favour to 5.0 against; neither that nor the sleep-quality difference reached the size the task force counted as clinically significant. It rated the overall certainty very low, for imprecision and inadequate blinding.

What the reviews could not settle

Two sham-controlled trials, with 31 and eight people analysed, are the only tests of whether needling or injecting the points matters, and neither has been repeated. The trials that added acupuncture or acupressure to a drug or to routine care were open or single-blind, so they cannot separate the treatment from the extra attention it brings. Huang and colleagues pooled observational studies with randomised ones and reported response rates rather than differences against control. The randomised trials above measured outcomes for at most 12 weeks from the start of treatment. The 2008 Cochrane review has not been updated to include any of the trials above.

AcuiQ’s restless legs syndrome page lists the protocols individual studies prescribed, each with its citation. Insomnia covers the trials of acupuncture for sleep itself.