Depression affects about a third of stroke survivors at any one time, according to a 2017 scientific statement from the American Heart Association, and it is linked with poorer recovery and higher mortality. Cochrane’s 2023 review of its treatment, by Allida, Hackett and colleagues, covers antidepressants, non-invasive brain stimulation and psychological therapy, and not acupuncture. The evidence on acupuncture comes from other systematic reviews, built on trials run almost entirely in China, which report lower depression scores with acupuncture on evidence their authors rate low or very low.
Does acupuncture help depression after stroke?
The reviews report that people given acupuncture, alone or added to usual care, end treatment with lower depression scores, but none rates that finding better than low certainty, and most trials compared acupuncture with drugs or with nothing rather than with a sham. A 2021 meta-analysis by Liu, Shen and colleagues pooled 17 trials with 1,402 people. Adding acupuncture to conventional treatment lowered scores on the 17-item Hamilton Depression Rating Scale (HAMD-17), a clinician-scored measure of severity, by 5.08 points against conventional treatment alone. Acupuncture on its own did not differ from antidepressants (0.43 points, with a confidence interval running from 1.61 points in acupuncture’s favour to 0.75 against), and came with fewer adverse events than the drugs (risk ratio 0.16). The authors rated the evidence low to very low because the trials disagreed with one another.
Later reviews
A 2023 network meta-analysis by Lam Ching, Zhong and colleagues compared many treatments at once, using direct and indirect comparisons across 62 trials with 5,308 people published from 2003 to 2022. Only Chinese trials met its criteria. Acupuncture alone, and acupuncture combined with repetitive transcranial magnetic stimulation (rTMS), lowered depression scores more than antidepressants, and the combination ranked first, with a 49% probability of being the most effective. Of the 17 comparisons of depression scores the authors graded, they rated 1 high, 8 moderate and 8 low in confidence. The points the trials used most were GV20 (Baihui), GV24 (Shenting) and PC06 (Neiguan).
A 2025 meta-analysis by Liang and colleagues took only electroacupuncture, in which a small current passes between needles: 22 trials with 1,640 people, all in China. Against conventional drug treatment, electroacupuncture lowered HAMD scores by 1.71 points, with high inconsistency between trials (I² 89%) and very low certainty, and made no significant difference to the Barthel Index, a score of independence in daily activities. Electroacupuncture describes the method.
How good the reviews are
Two overviews have rated the reviews themselves with AMSTAR 2, a checklist that grades how far a systematic review can be trusted. Miao and colleagues, searching to September 2024, found 38 reviews of acupuncture for depression after stroke and rated 1 moderate, 2 low and 35 very low; most had not registered a protocol, listed the studies they excluded or disclosed their funding. They reported that no serious adverse effects appeared in the reviewed studies, and that more rigorous safety evaluation is needed before conclusions about safety can be drawn. Liu, Zeng and colleagues took the 10 reviews of acupuncture added to conventional treatment, rated 9 of them very low on AMSTAR 2, and graded 20 of the 34 outcomes they reported very low on GRADE, the scale reviewers use to say how much confidence a result supports.
An evidence map by Zhou, Zhang and colleagues, searched to April 2024, counted 666 clinical studies and 34 reviews of acupuncture-related treatment for depression after stroke, the earliest from 1996. China produced 99.6% of them; Korea, Singapore and the United States one each. Most courses lasted two to four weeks, and few studies followed patients up afterwards. Where the trials come from discusses why the country of origin matters when acupuncture trials are pooled.
Trials with a sham or placebo
Three trials kept patients from knowing which treatment they had. In a 2022 single-centre trial by Cai, Shen and colleagues in Shanghai, 65 people received electroacupuncture or sham electroacupuncture with the Park sham device at GV20, Sishencong, SP06, LV03 and BL18, three times a week for four weeks. Depression scores fell more with electroacupuncture at two, four and eight weeks; two people in that group had a local haematoma.
A 2014 pilot trial by Man, Zhang and colleagues in Hong Kong gave 43 people an SSRI antidepressant and body electroacupuncture, then added either dense cranial electroacupuncture, needles on the forehead and scalp carrying current, or a non-invasive version with electrodes on the scalp. Depression scores fell further with the needled version from the first week, while the non-invasive group improved more on the Barthel Index at four weeks.
A 2015 trial by Qian and colleagues in Shanghai compared acupuncture with fluoxetine using two dummies: 68 inpatients had either acupuncture at GV26, PC06 and ST36 with a placebo pill, or fluoxetine with minimal acupuncture, daily for six weeks. The groups did not differ at six weeks or three months; the acupuncture group improved sooner, at two weeks, and reported fewer side effects at that point. Sham acupuncture explains why no sham control is inert.
What guidelines say
A 2025 review of stroke guidelines by Zhu and colleagues tabulated the traditional and complementary medicine recommendations of nine guidelines, including ones from China, Hong Kong, Australia, the United States and the World Stroke Organization. The only recommendations it lists for depression after stroke are a weak one for tai chi, yoga and qigong, from the US Veterans Affairs and Department of Defense guideline, and a weak one for Chinese herbal formulas taken with antidepressants. Its recommendations on acupuncture concern swallowing, spasticity, shoulder-hand syndrome, arm movement and daily activities. Clinical guidelines compares guideline bodies across conditions.
What the reviews could not settle
The trials behind every pooled estimate are small, short and almost all from one country, and most compared acupuncture with an antidepressant or with no added treatment, where patients and assessors could tell which group was which. The trials with a sham or placebo control had 43 to 68 participants each and followed them for three months at most. Whether acupuncture changes the course of depression after stroke beyond a few weeks, or how it compares with an antidepressant over months, is unanswered, and the 2025 overview of 38 reviews found safety too thinly evaluated for firm conclusions. Whether any treatment suits one person after a stroke is a question for the clinicians who know them.
AcuiQ’s post-stroke depression page lists the protocols individual studies prescribed, each with its citation. Depression covers the reviews of depression in general, and Stroke recovery the reviews of acupuncture for movement, swallowing and speech after stroke.