Acupuncture did not improve thinking in care-home residents with dementia, but daily independence rose slightly

Dementia slowly takes away memory and other thinking skills, and with them the ability to manage daily life. Many older people with dementia end up in care homes because they can no longer live alone. No treatment cures the condition, and today's drugs can at best slow it down. Older people often take many medicines already, so there is interest in treatments that don't involve another pill.

Acupuncture, in which thin needles are placed at chosen points on the body, is one such option. Earlier studies have hinted that it might help with thinking, physical disability or sleep. A team at the School of Chinese Medicine of the University of Hong Kong tested it properly in a randomised controlled trial in care-home residents with dementia. The main result was negative: thinking did not improve. A smaller, secondary result suggests a modest gain in everyday independence.

What they did

The researchers recruited residents from 19 care homes in Hong Kong between October 2021 and August 2024. In all, 118 people took part. Their average age was 83.9, and 60.2% were women. Most had moderate thinking problems, though the type of dementia varied from person to person.

• Two groups of 59. By chance, half got acupuncture on top of their usual care and medicines; the other half simply continued usual care.
• The treatment. Experienced registered Chinese medicine practitioners gave 24 sessions, twice a week for 12 weeks, in the care homes. Each session combined needles on the forehead and scalp that carried a mild electric current with hand-turned needles elsewhere on the body.
• Main goal (primary outcome). The change in the Montreal Cognitive Assessment (MoCA), a short paper-and-pencil test of memory, attention and other thinking skills. At the start, the average score was very low, 9.6.
• Other measures. A number-repeating memory test, a scale of independence in daily activities such as eating, washing and walking (the Chinese version of the Modified Barthel Index), and scales for pain, low mood and sleep problems. Everyone was tested at the start, at 6 weeks and at 12 weeks.
• Blinding. The people scoring the tests and the statistician did not know who had acupuncture. The residents knew, as there was no fake (sham) acupuncture.

The trial shared its control group with a parallel study of acupressure, pressing on the same kind of points without needles; that study is reported separately.

What they found

Of the 118 residents, 101 (85.6%) completed the study.

• Thinking: no improvement. MoCA scores did not change significantly with acupuncture, compared either with the start or with the usual-care group, at 6 or 12 weeks. The number-repeating memory test showed no difference either.
• Daily independence: a small gain. At 12 weeks, the acupuncture group scored on average 4.12 points higher on the daily-activities scale than the usual-care group, from an average starting score of 36.3. The standardised effect size was 0.29, which the authors describe as small. No difference showed up at 6 weeks.
• Pain, mood and sleep: no improvement.
• Side effects. Acupuncture led to more side effects, but those linked to the treatment were all mild, mostly bleeding or bruising where the needles went in. Two people in the acupuncture group died, from worsening of illnesses they already had; these deaths were not counted as caused by the treatment.
• A hint in one subgroup. When the team split the residents by how far their dementia had progressed, people at the moderate stage seemed to gain 3.21 MoCA points from acupuncture compared with usual care, while those at milder or more severe stages did not. This analysis was planned but exploratory.

What to keep in mind

• The authors' own limits. The residents had dementia of different causes, which makes the group mixed. Thinking was measured with only two simple tests. Without a sham treatment, residents knew they were being treated, and their expectations could have coloured the results. Twelve weeks is short for a long-lasting, worsening disease. And although extra checks for missing data agreed with the main analysis, some bias can't be ruled out.
• The main question got a "no". Thinking was the outcome the trial was designed and sized to test, and acupuncture did not improve it.
• The independence result is secondary. The authors stress that the trial was built around thinking, not daily activities, so this finding should be read cautiously and needs its own, adequately sized trial.
• The subgroup hint is fragile. Each dementia-stage group held only 35 to 42 people, the trial was not designed to compare them, and the pattern did not hold up after correcting for the many comparisons made. The authors call it "hypothesis-generating" and say it may be random variation.
• More dropped out with needles. 13 of 59 left the acupuncture group, against 4 of 59 in the usual-care group. The authors suggest that some older people may find needling hard to accept.

Why it matters

Care homes want treatments that are safe and practical for frail residents who already take many drugs. This trial, funded by the Hong Kong government, shows that acupuncture can be delivered on site with only mild side effects, but it did not improve thinking. Its modest effect on everyday independence, which matters a great deal to residents and staff, and the possible benefit at the moderate stage are leads, not conclusions. The authors call for larger trials that sort patients by stage of dementia before anyone can say who, if anyone, benefits.