Parkinson’s disease is a progressive disorder of the nervous system. Its motor symptoms are slowness of movement, stiffness, tremor at rest and, later, unsteady balance; its non-motor symptoms include poor sleep, anxiety, depression, constipation and fatigue. Trials of acupuncture have tested both. Cochrane has not published a review of acupuncture for Parkinson’s disease. The reviews in other journals divide on motor symptoms by comparison: trials that added acupuncture to medication and compared it with medication alone report benefits, and the few that compared it with sham acupuncture, a control procedure made to look like acupuncture, did not confirm them. Sham acupuncture describes the kinds of sham in use.
Reviews of motor symptoms
Most trials measure motor symptoms on the Unified Parkinson’s Disease Rating Scale (UPDRS), a clinician-scored scale on which a lower score means fewer symptoms. Certainty ratings are GRADE, or CINeMA, its counterpart for network meta-analyses, which compare several treatments at once through shared comparators.
| Review | Trials and people | Comparison | Finding |
|---|---|---|---|
| Lee and colleagues 2008, searched to January 2008 | 11 trials | Sham (3 trials, 96 people) | No significant difference on the UPDRS |
| Same review | 2 trials, 106 people | Scalp acupuncture plus drugs against drugs alone | More people improved (risk ratio 1.46) |
| Noh and colleagues 2017, searched to June 2016 | 42 trials, 2,625 people | Acupuncture plus medication against medication alone; and against sham | Better total UPDRS and several of its parts with added acupuncture; no significant difference in total UPDRS against sham |
| Mao and colleagues 2026, trials from 2002 to 2024 | 57 trials, 4,262 people | Five acupuncture-related methods plus medication against medication alone | UPDRS 0.89 points lower with manual acupuncture added (36 trials, 2,655 people); certainty moderate for that estimate, low to moderate for most comparisons |
Lee’s review judged the evidence not convincing and the trials too few, too small and of too low a quality for a firm conclusion; it noted that trials comparing acupuncture with no treatment or with drugs alone do not control for non-specific effects, such as the attention and expectation that come with any treatment. Noh’s review rated the methodological quality of the trials low and found no serious adverse events. A 2024 overview of 24 systematic reviews by Xue and colleagues rated 12 of them very low in methodological quality on AMSTAR 2, a checklist for judging how a review was conducted. Of the 105 outcomes those reviews reported, it graded 8 high, 23 moderate, 42 low and 32 very low in certainty, and called the evidence limited and inconclusive.
Sham-controlled trials of non-motor symptoms
Since 2016, five trials have compared acupuncture with sham for a single non-motor symptom.
| Trial | Symptom and setting | People and course | Primary outcome against sham |
|---|---|---|---|
| Kluger and colleagues 2016 | Fatigue; University of Colorado | 94; twice a week for 6 weeks | No difference in fatigue at 6 or 12 weeks; both groups improved |
| Fan and colleagues 2022 | Anxiety; First Affiliated Hospital of Guangzhou University of Chinese Medicine | 70; 3 times a week for 8 weeks | No difference at the end of treatment (0.22 points); 7.03 points lower on the Hamilton Anxiety Rating Scale, a clinician-scored anxiety scale, at follow-up 8 weeks later |
| Yan and colleagues 2024 | Poor sleep; same hospital | 83, of whom 78 were analysed; 3 times a week for 4 weeks | Parkinson’s Disease Sleep Scale 19.75 points higher (better) after treatment and 20.24 at week 8, 4 weeks after treatment ended |
| Liu and colleagues 2025 | Insomnia; Ningxia Medical University, Yinchuan | 60; 12 sessions over 4 weeks | Parkinson’s Disease Sleep Scale 21.4 points higher, sustained to week 16 |
| Li and colleagues 2023 | Constipation; Guangzhou University of Chinese Medicine | 78; 12 sessions over 4 weeks | Weekly complete spontaneous bowel movements rose from 3.36 to 4.62 with acupuncture and were unchanged with sham (3.10 to 3.03) |
In the fatigue trial, 63% of participants reported a noticeable improvement whichever needling they received, and the authors concluded that acupuncture helps fatigue in Parkinson’s disease mainly through non-specific or placebo effects. The anxiety and sleep trials both needled GV24 (Shenting), XH03 (Yintang, numbered GV 29 in the 2021 Chinese national standard), HT07 (Shenmen) and SP06 (Sanyinjiao), with four needles around the crown at GV21 (Qianding), GV19 (Houding) and on either side of GV20. The sleep trial added points including BL62 (Shenmai) and KD06 (Zhaohai), which the Ningxia trial also used. In each of the four Chinese trials the sham needle did not pierce the skin, and the Guangzhou and Ningxia groups each sought a patent on the guide device that held it. The sleep trial’s authors listed as limitations that the acupuncturists were not blinded, that follow-up lasted 4 weeks and that participants came from one specialist clinic. Insomnia and Constipation cover the trials in people without Parkinson’s disease.
Reviews of non-motor symptoms
A 2026 review by Cui and colleagues searched from 2015 to October 2025 for studies comparing acupuncture with sham or usual care for sleep, mood and fatigue. It found 8 studies, 4 of them randomised. It pooled the two sham-controlled sleep trials, with 138 people, and found sleep scale scores 14.52 points better with acupuncture, at moderate certainty on GRADE, with moderate inconsistency between the two. It rated the anxiety evidence preliminary, resting on one trial, and the fatigue evidence very uncertain, with no difference from sham. A 2024 review by Li and colleagues of constipation in Parkinson’s disease included 11 trials with 960 people, whose control groups received sham acupuncture or the same drugs without acupuncture. Pooling 3 trials with 297 people, it found 1.49 more complete spontaneous bowel movements a week with acupuncture alone or added to conventional treatment, a bowel movement that comes without a laxative and leaves a sense of complete emptying. The authors called for higher-quality trials.
What guidelines say
The National Institute for Health and Care Excellence’s guideline NG71, Parkinson’s disease in adults, published on 19 July 2017 and current, makes no recommendation on acupuncture. Its section on treatment without drugs covers Parkinson’s nurse specialists, physiotherapy, the Alexander Technique, occupational therapy, speech and language therapy and diet, and NICE states that it will review the guideline if new evidence is likely to change the recommendations.
The International Parkinson and Movement Disorder Society’s 2019 evidence-based review of treatments for non-motor symptoms, by Seppi and colleagues, covered trials to the end of 2016. It assessed acupuncture only for fatigue, on the strength of Kluger’s trial, which it classed as one negative, low-quality study: it found insufficient evidence on efficacy, called acupuncture investigational because both groups improved, and judged it an acceptable risk without specialised monitoring. It found no randomised trial of any treatment for anxiety that met its criteria. A Cochrane protocol published in 2026, by Li and colleagues, sets out a network meta-analysis of treatments without drugs for anxiety and depression in Parkinson’s disease that includes acupuncture among them; it is a plan for a review, not a result.
What the reviews could not settle
For motor symptoms the sham-controlled evidence remains small: Lee’s pooled sham estimate rests on 96 people, and the larger reviews that report benefit compare added acupuncture with medication alone, which cannot separate needling from attention and expectation. For non-motor symptoms, the anxiety, sleep and constipation trials in Guangzhou share a senior author, Lixing Zhuang, so only the sleep result has been repeated by an unrelated group, in Yinchuan; the one trial run outside China, on fatigue, found no difference from sham. Where the trials come from discusses what the country of a trial tells a reader. None of the five followed people beyond 16 weeks in a disease that progresses over years, and Cui’s review could not assess publication bias with so few trials. Its own table of included trials also lists each with about twice the participants the trial reports enrolling, though its pooled total of 138 matches the trials’ reports. The Cochrane review of anxiety and depression has not yet reported.
AcuiQ’s Parkinson’s disease page lists the protocols individual studies prescribed, each with its citation.