Essential tremor is a movement disorder whose main sign, in the description of Shi and colleagues’ 2026 review, is an action tremor of both arms, shaking when the hands are held out or used, sometimes with tremor of the legs, head, face or voice. The evidence on acupuncture for it is one systematic review of 20 small randomised trials, all run in China, and a handful of case reports from elsewhere. None of the trials compared acupuncture with sham acupuncture, a procedure made to resemble it, so none tests whether needling the points matters. Tremor at rest in Parkinson’s disease is covered in Parkinson’s disease.
People with the condition have asked for options beyond drugs. In a 2015 online survey by Elan Louis and colleagues, sent through the International Essential Tremor Foundation’s newsletter, 11.8% of 1,418 respondents were satisfied with their care and 11.2% wanted a treatment approach other than medications and surgery.
The Cochrane review
There is none. A search of the Cochrane Database of Systematic Reviews through Europe PMC on 16 October 2026 found Cochrane reviews of four drugs for essential tremor, alprazolam, pregabalin, topiramate and zonisamide, and no review or protocol on acupuncture.
The network meta-analysis
Qingping Shi and colleagues at Heilongjiang University of Chinese Medicine in Harbin searched eight databases, four of them Chinese, to 20 October 2025, under a protocol registered on PROSPERO. They included 20 randomised trials with 1,067 people, published between 2006 and 2024, each with 26 to 80 participants and a course of 15 days to 3 months. Seven of the 20 are master’s theses. A network meta-analysis compares several treatments at once by linking them through the comparators they share, and ranks them by SUCRA, the surface under the cumulative ranking curve, a score from 0 to 100% for how likely each treatment is to rank first.
| Comparison | Trials | Acupuncture-type arm |
|---|---|---|
| Acupuncture added to propranolol, against propranolol alone | 11 | Manual, electro- or scalp acupuncture; in two trials also Chinese herbal medicine, one of them with tuina massage as well |
| Acupuncture in place of propranolol, against propranolol | 5 | Manual or scalp acupuncture, or thread embedding with or without electroacupuncture |
| One acupuncture method against another | 4 | Methods needling points on the head, or electroacupuncture at neck points with ear acupuncture, against standard acupuncture |
| Against sham acupuncture or no treatment | 0 | None |
Most trials reported response on the “total effective rate” scales common in Chinese trials, and the authors pooled it across all 20 as favouring the acupuncture groups (risk ratio 4.36, 95% confidence interval 3.14 to 6.03). On the Tremor Six Score, a bedside rating taken from a 1999 study by Louis and colleagues, 16 trials with 814 people gave scores 1.99 points lower in the acupuncture groups (mean difference −1.99, 95% CI −2.25 to −1.73), with moderate heterogeneity, the disagreement between trials beyond chance (I² 44%). Scalp acupuncture ranked first for response (SUCRA 81.5%), on one trial of 40 people, with an odds ratio against control of 12.67 and a confidence interval from 1.40 to 114.42. Manual acupuncture ranked first on the Tremor Six Score (SUCRA 76.6%).
Five trials, with 279 people, reported adverse events, and the acupuncture groups had fewer (risk ratio 0.13, 95% CI 0.07 to 0.25). The events described were nausea, abdominal discomfort, dizziness, slow heart rate and low blood pressure, and in four of the five trials the control group took propranolol, whose common side effects the authors list as slow heart rate and narrowing of the airways; one thesis reported bruising under the skin. Three trials followed people after treatment ended. One found the acupuncture group still ahead at one month; at three months, one found no difference between the groups, and in the other 13.8% of the acupuncture group relapsed against 45.5% of the control group.
The authors rated most risk-of-bias domains high or unclear. None of the 20 trials reported allocation concealment, seven did not say how they generated the random sequence, and the authors judged blinding of practitioners not feasible and of patients difficult. They describe the evidence as low to moderate in quality, report no GRADE assessment, and call the result preliminary and in need of large, well-blinded multicentre trials. They note that all 20 trials were run in China and that no standard set of points has emerged. Thread embedding and Electroacupuncture describe two of the methods.
Two of the trials
Two of the trials have English abstracts on PubMed, and both added acupuncture to propranolol, a beta blocker. In a 2010 trial by Sui Kangmin and Li Xin, 60 people were randomised to propranolol alone or with acupuncture at points including GV20 (Baihui), Sishencong, the four points around the crown (XH01A to XH01D here, EX-HN1 in the WHO standard), and KD03 (Taixi). After 30 days the total effective rate was 90.0% (27 of 30) with acupuncture added and 56.7% (17 of 30) with propranolol alone, and tremor scores were also better in the acupuncture group. In a 2006 trial by Chen Lihua and colleagues, 26 people had six weeks of propranolol with or without acupuncture; the total effective rate was 84.6% against 61.5%, with fewer adverse effects in the acupuncture group. Neither abstract describes any blinding. Reading an acupuncture trial sets out why an add-on design cannot separate the needling from the extra attention.
Case reports
Outside China the record is case reports. A 2015 report by Mir and colleagues in the United States describes a 74-year-old woman given weekly acupuncture for 10 weeks, who rated her tremor, handwriting and ability to hold a cup as better and kept the change with monthly sessions at 12 months. A 2013 Korean report by Jeong and Sun reviewed the records of three women in their 70s and 80s treated with Sa-am five-element acupuncture and pharmacopuncture, the injection of a herbal extract, here Hwangyeonhaedoktang, at CV23 (Lianquan) and CV17 (Danzhong). All three improved on a numeric rating scale, and tremor that returned improved again with the same treatment; the authors note the conclusion rests on three cases. Point injection describes the method.
Other action tremor
For tremor that is neither essential tremor nor Parkinson’s disease, AcuiQ found no trial of acupuncture in Europe PMC. Holmes tremor, a slow tremor present at rest, in posture and in action, is rare, and a 2025 scoping review by Chung and Ferraz of its drug and surgical treatment found 121 patients across 97 publications, a literature of case reports and case series. A 2025 Taiwanese case report by Huang and Chang describes a 69-year-old man with Holmes tremor after a traumatic brain injury who had acupuncture at 14 points including GV20, with Jiao’s scalp acupuncture six times a week, over 1.5 months, and had no tremor three months after the injury. A single case cannot separate treatment from recovery after injury.
What guidelines say
No guideline on essential tremor addresses acupuncture. The American Academy of Neurology’s 2011 guideline update by Theresa Zesiewicz and colleagues rates propranolol and primidone as established as effective (Level A) and grades further drugs, botulinum toxin, deep brain stimulation and thalamotomy, an operation that destroys a small target in the thalamus. Its full text does not mention acupuncture. It replaced the 2005 practice parameter, now retired, and the AAN’s guideline page records that its Guidelines Subcommittee reaffirmed it on 19 July 2025.
The International Parkinson and Movement Disorder Society’s 2019 evidence-based review by Joaquim Ferreira and colleagues set out to review pharmacological and surgical treatment only, 64 studies in all; it classed propranolol, primidone and topiramate above 200 mg a day as clinically useful, so acupuncture was outside its scope. NICE has no guideline on treating essential tremor. Its 2018 guidance on focused ultrasound thalamotomy for treatment-resistant essential tremor is under review since April 2025, and its 2006 guidance supports deep brain stimulation for tremor and dystonia outside Parkinson’s disease. Clinical guidelines lists the guidelines the evidence pages cite.
What the reviews could not settle
No trial has compared acupuncture with sham for essential tremor, so whether the points matter, or the needling at all, is untested. The trials that added acupuncture to propranolol were not blinded and cannot separate the needling from the extra attention and expectation it brings; the trials that replaced propranolol compared a procedure with a drug whose side effects patients notice. The response figure rests on “total effective rate” scales rather than a validated tremor scale, and the Tremor Six Score result comes from trials with unclear randomisation and no concealed allocation. Three trials reported follow-up after treatment, for at most three months, and they disagree. Every trial was run in China, and seven of 20 are unpublished theses. The guidelines that grade treatments for essential tremor have not assessed acupuncture.
AcuiQ’s tremor page lists the protocols individual studies prescribed, each with its citation. Sham acupuncture describes the controls a trial would need to answer the question these trials leave open.