Polycystic ovary syndrome (PCOS) is a hormonal condition marked, as the Cochrane review below describes it, by infrequent or absent periods, infertility and hirsutism, excess body hair driven by androgens. Some trials enrol only women who also have insulin resistance, in which the body needs more insulin than usual to keep blood sugar steady. Trials of acupuncture for PCOS measure one of four things: whether ovulation happens, whether a pregnancy ends in a live birth, how regular periods become, and metabolic markers such as insulin resistance. A trial can succeed on one and fail on another, so this page keeps them apart, and it keeps apart the control each trial used: sham acupuncture, a procedure made to resemble it (Sham acupuncture describes the kinds); a drug; exercise or a meeting with a therapist; or no treatment. Acupuncture during IVF is covered on its own page, IVF.
The Cochrane review
The Cochrane review by Zhang, Lim, Cheng and Lim, published in October 2025, is the fourth update of a review first published in 2011. Its searches of most databases ran to December 2024, though its search of one Chinese database, CBM, stopped in November 2015. It included 9 randomised trials with 1,606 women who ovulated rarely or not at all, one more trial than the 2019 version. Three trials compared acupuncture with sham, one low-frequency electroacupuncture with exercise or no treatment, one acupuncture with relaxation, one with clomiphene, a drug that induces ovulation, and three with Diane-35, an oral contraceptive containing an anti-androgen.
| Outcome | Trials (women) | Result | Certainty (GRADE) |
|---|---|---|---|
| Live birth | 1 (1,000) | Little or no difference (risk ratio 0.97) | Low |
| Multiple pregnancy | 1 (1,000) | Little or no difference | Low |
| Ovulations per woman during treatment | 2 (1,010) | Little or no difference | Low |
| Clinical pregnancy | 3 (1,117) | Little or no difference (risk ratio 1.07) | Low |
| Miscarriage | 1 (926) | Little or no difference | Low |
| Days between periods at 12 weeks | 1 (141) | Fewer with acupuncture | Very low |
| Adverse events | 3 (1,230) | More with acupuncture (risk ratio 1.16) | Moderate |
GRADE is the scale reviews use to say how far a result can be trusted; low means the true effect may be substantially different. The authors rated every comparison with an active treatment or with no treatment very low in certainty, and none of those trials reported live birth. Adverse events recorded in acupuncture groups included vaginal bleeding, weight gain, fatigue, dizziness, nausea and bruising under the skin. The authors concluded that they found no clear difference between acupuncture and sham on any fertility outcome or on restoring regular periods, that acupuncture is probably associated with more adverse events, and that the effectiveness of acupuncture for PCOS remains uncertain because the trials are few and report different outcomes.
Later reviews
An overview of 11 systematic reviews by Yang and colleagues at Chengdu University of Traditional Chinese Medicine, published in 2023, rated the methods of 9 of them extremely low and 2 low on AMSTAR 2, a checklist for judging how a review was done. Of the 64 results the reviews reported, the overview judged 2 high in certainty and 20 very low. The reviews reported better pregnancy and ovulation rates when acupuncture was added to drugs, and better menstrual recovery against no treatment, and the overview concluded that the efficacy and safety of acupuncture for PCOS remain uncertain.
Two network meta-analyses published in 2026 pooled trials comparing many acupuncture-related methods with drugs or placebo, most of them from China. Li and colleagues gathered 59 trials with 5,937 participants, searched to October 2023, and found lower testosterone with acupuncture than with drugs and lower scores on the Ferriman-Gallwey scale, a count of excess body hair, with electroacupuncture than with placebo; drugs ranked first for insulin resistance. Du and colleagues gathered 53 trials with 4,406 participants, searched to September 2025, and reported lower insulin resistance with acupoint injection and with acupuncture plus moxibustion. Both ranked methods by the probability of coming first, and both called the trials limited; Du’s team noted that adverse events were reported in few of them. Neither abstract gives a GRADE rating.
A 2025 review of safety by Nurwati and colleagues in Indonesia included 13 trials, searched to September 2024. No trial recorded a serious or life-threatening adverse event. Bruising at the needle sites was more frequent with acupuncture than with sham or no acupuncture, and miscarriage rates were similar to controls when low-frequency electroacupuncture was given before pregnancy.
Landmark trials
The PCOSAct trial, published in JAMA in 2017 by Wu, Stener-Victorin and colleagues, is the trial behind the Cochrane review’s live-birth result. It randomised 1,000 Chinese women with PCOS at 27 hospitals to four groups of 250 in a factorial design, which tests two treatments at once: active or control acupuncture, each combined with clomiphene or a placebo pill. Active acupuncture meant deep needling with manual and low-frequency electrical stimulation twice a week; control acupuncture meant shallow needling with no manual stimulation and a stimulator that delivered no current. Treatment lasted up to 4 ovulation cycles, and 926 women completed the trial.
- Live births were 29.4% with active acupuncture and clomiphene, 28.0% with control acupuncture and clomiphene, 13.9% with active acupuncture and placebo, and 16.8% with control acupuncture and placebo.
- Clomiphene raised the live-birth rate from 15.4% to 28.7%, a difference of 13.3 percentage points; active and control acupuncture gave 21.8% and 22.4%.
- Diarrhoea (5.0% against 1.6%) and bruising (7.4% against 1.8%) were more common with active acupuncture.
The authors concluded that the finding does not support acupuncture as an infertility treatment in these women.
A 2022 trial in Human Reproduction by Wen and colleagues at three hospitals in Guangdong randomised 342 women with PCOS and insulin resistance to acupuncture with a placebo pill, metformin with sham acupuncture, or sham acupuncture with a placebo pill, for 4 months. Insulin resistance fell by 14.7% with acupuncture, 25.0% with metformin and 8.6% with sham. The authors found acupuncture less effective than metformin and not significantly different from sham. Gastrointestinal side effects were more frequent with metformin, and bruising with acupuncture.
Smaller trials in Sweden and the United States came first. A 2011 trial in Virginia by Pastore and colleagues gave 84 women 12 sessions of true or sham acupuncture over 8 weeks and found similar ovulation rates over 5 months: 0.37 a month with true acupuncture and 0.40 with sham. In Gothenburg, Stener-Victorin’s group randomised 84 women to 16 weeks of low-frequency electroacupuncture, exercise or no treatment, in a 2011 trial led by Jedel; menstrual frequency in the electroacupuncture group rose from 0.28 to 0.69 periods a month, more than with exercise, and both active groups did better than no treatment. A 2013 trial by Johansson in the same group randomised 32 women to acupuncture or to meetings with a physical therapist for 10 to 13 weeks, and found more frequent ovulation and lower sex-steroid levels with acupuncture. Neither Swedish trial used a sham control.
Trials adding acupuncture to letrozole, the drug current guidance prefers for inducing ovulation, are newer and smaller. A 2025 trial by Ding and colleagues randomised 134 infertile women on letrozole to added acupuncture or placebo acupuncture over 3 cycles, and reported higher ovulation and pregnancy rates with acupuncture; it did not report live births in its abstract.
What guidelines say
The 2023 International Evidence-based Guideline for the assessment and management of PCOS, led by Teede and colleagues at Monash University and published the same year in four journals, updates the 2018 guideline and carries 77 evidence-based and 54 consensus recommendations. Its published recommendations do not mention acupuncture. They name letrozole the first-line drug for inducing ovulation in infertile women with PCOS who have no other infertility factor, on high-certainty evidence. On complementary therapies, two practice points apply: women taking them are encouraged to tell their healthcare professional, and information given to women should be evidence-based while respecting their preferences, complementary therapies included. The full guideline document, hosted by Monash, could not be opened for this page. Clinical guidelines compares other bodies’ positions.
What the reviews could not settle
The sham-controlled trials of acupuncture alone, PCOSAct among them, found no difference on fertility; the one sham-controlled trial that reported a difference, Ding 2025, added acupuncture to letrozole and reported no live births. The other trials that found a difference compared acupuncture with exercise, a therapist, drugs or nothing. PCOSAct is the only trial that measured live birth, so the Cochrane authors had nothing to pool it with, and all its women were Chinese. Most metabolic results come from trials in China that reviews rated low in methodological quality, and the one large sham-controlled trial of insulin resistance, Wen 2022, found no difference from sham. Where the trials come from discusses what the country of a trial tells a reader. No trial has tested acupuncture added to letrozole with live birth as its outcome. Obesity covers the trials of acupuncture for weight, and Period pain the trials in dysmenorrhoea.
AcuiQ’s polycystic ovary syndrome and anovulation pages list the protocols individual studies prescribed, each with its citation.