Acupuncture for Parkinson's Disease and Dementia: Evidence

Evidence at a glance

Labels used: promising but limited, mixed, insufficient evidence.

1. Parkinson's: sleep

   Promising but limited

   Two sham-controlled trials pooled in a 2026 review (138 people) suggested better sleep scores [1]; one 2025 trial of electroacupuncture found a larger gain than sham [2]

2. Parkinson's: mood, fatigue, quality of life

   Mixed / insufficient evidence

   A 2024 review found no clear benefit for anxiety, depression or quality of life [3]; fatigue did not differ from sham [1]

3. Parkinson's: movement symptoms

   Insufficient evidence

   A review of 17 trials reported gains when acupuncture was added to usual treatment, but the designs varied too much to compare [4]

4. Mild cognitive impairment (MCI)

   Promising but limited

   Reviews report higher test scores, but the evidence was graded low or very low [5]

5. Vascular cognitive impairment

   Promising but limited

   A review of 48 trials found better scores versus usual care or medicine, with low or very low certainty [6]

6. Alzheimer's disease

   Insufficient evidence (design and quality concerns)

   A 2026 review of 21 trials reported higher scores than no treatment or placebo [7]

Parkinson's disease

The NHS says there is currently no cure for Parkinson's disease, but treatments are available to help relieve the symptoms. [8] It also says there is no clinical evidence that complementary treatments control Parkinson's symptoms, and that some should not be used instead of prescribed medicines. [8]

Research on acupuncture is mainly about non-motor symptoms:

• Sleep. A 2025 trial in eight Chinese hospitals randomly assigned 136 people with Parkinson's and insomnia to real or sham electroacupuncture for 8 weeks. It was blinded for participants and assessors, and the real group improved more. [2] A 2026 review notes that only two sham-controlled sleep trials (138 people in total) could be pooled, so the signal is real but rests on little data. [1]
• Other non-motor symptoms. A 2024 review of 13 trials (719 people) found better sleep and total Parkinson's scale scores, but no significant gain in anxiety, depression or quality of life. [3]
• Movement. A 2022 review of 17 trials reported positive effects on the movement rating scale when acupuncture was added to usual care, but the authors found great differences in design and method that made comparison difficult. [4]

A large Korean records study (23,454 people in each group) found acupuncture-related treatment users took less levodopa and had slightly lower long-term death risk. Its authors warn that these links should not be read as proof of benefit. [9] People who choose such treatment may differ in many other ways.

Mild cognitive impairment and dementia

The NHS says there is currently no cure for dementia, and advises being wary of products claiming to benefit people with dementia. [10]

• MCI. A 2020 review pooled 15 trials (1,051 people) and called for more rigorous, longer studies. [11] A 2025 review of amnestic MCI (15 trials, 908 people) found higher cognitive test scores, but graded the evidence very low or low for every outcome. [5]
• Vascular cognitive impairment. A review of 48 trials (3,778 people) found gains over medicine or usual care, with poor methodological quality and certainty rated low or very low. [6]
• Alzheimer's disease. A 2026 review of 21 trials (1,354 people) reported higher scores than no treatment or placebo and results comparable to donepezil, and said protocols need standardizing. [7] That review mixed several kinds of comparison (sham, placebo, donepezil and no treatment). [7] A "no treatment" comparison cannot separate the needle from attention and expectation.
• Sleep in dementia. A 2025 review of 10 trials (721 people) found better sleep scores, but the certainty was limited by differences between studies and weak methods. [12]

What the trial designs can and cannot show

A sham-controlled trial (for example, shallow or non-penetrating needling) can help separate the effect of a needle from the effect of attention and hope. A trial against "usual care" or "no treatment" cannot. Both are weaker when people know which treatment they received. See Sham Acupuncture and Placebo Controls for how to read these studies. Test scores that rise a point or two are also not the same as being able to manage daily life better.

Where to go next

For the wider picture see What Is Acupuncture?, Acupuncture for Insomnia and Sleep, Acupuncture for Stroke Recovery and our wellness guide to mild cognitive impairment and dementia.