Pelvic girdle pain is pain between the back of the iliac crest and the buttock crease, particularly near the sacroiliac joints, which can also occur at the pubic symphysis at the front of the pelvis, in the definition of the 2008 European guidelines. The same guidelines put its point prevalence in pregnant women at about 20%, and define low back pain as pain between the 12th rib and the buttock crease; the two can occur together or apart. The 2015 Cochrane review of every treatment for the two conditions found evidence from single trials that acupuncture added to usual antenatal care improved pelvic pain, and no outcome in the review reached high certainty. NICE’s antenatal care guideline makes no recommendation on acupuncture for pelvic girdle pain. This page reports what each source found.
The Cochrane review
Sarah Liddle and Victoria Pennick searched to January 2015 and included 34 randomised trials with 5,121 pregnant women: 15 trials of low back pain, 6 of pelvic pain and 13 of both. Every intervention was added to usual antenatal care. The review rates findings with GRADE, the standard scale of how far evidence can be trusted, from high to very low. Acupuncture trials differed so much in technique, comparison and outcome measure that the review could pool none of them, and it reported each trial on its own.
| Condition and comparison | Trial (women analysed) | What the review reported | Certainty |
|---|---|---|---|
| Pelvic pain: acupuncture or stabilising exercises, each added to usual care, against usual care alone | Elden 2005 (330) | Less evening pain with acupuncture than with usual care alone or with exercises | Moderate |
| Pelvic pain: acupuncture against non-penetrating sham, both added to usual care | Elden 2008 (108) | No significant difference in pain; better scores on a daily-activities scale with acupuncture | Low |
| Pelvic pain: deep against superficial needling | Lund 2006 (47) | No significant difference in evening pain | Low |
| Low back and pelvic pain: acupuncture added to usual care, against usual care, physiotherapy, sham ear acupuncture or a later start | Kvorning 2004 (72), Wedenberg 2000 (46), Wang 2009 (152), Ekdahl 2010 (32) | Each trial reported results favouring acupuncture; none could be pooled | Low |
The Wang trial tested ear acupuncture against a sham ear treatment and a waiting list, and more women improved with true ear acupuncture than with either. The Wedenberg trial compared acupuncture with physiotherapy and reported less evening pain and disability with acupuncture, but 12 of the 30 women allocated to physiotherapy were lost to follow-up, against none allocated to acupuncture. The Ekdahl trial compared acupuncture started at 20 weeks with acupuncture started at 26 weeks, and found more pain relief with the later start. Across the trials that reported harms, the review found adverse effects minor and transient, mostly in women given acupuncture: small bruises where needles went in, tiredness and passing ear tenderness. The review lists no trial of acupressure. Cochrane has published no later version, and the authors wrote that further evidence is very likely to change the estimates.
Later reviews
A 2023 review by Li and colleagues at Chengdu University of Traditional Chinese Medicine searched English and Chinese databases to January 2022 for trials of acupuncture, alone or with other treatment, in pregnancy-related low back pain. It included 12 trials with 1,302 women, against controls ranging from physiotherapy and drugs to no treatment. Pain scores fell further with acupuncture, and the “effective rate”, the share of women each trial classed as improved on a scale it set itself, was higher (odds ratio 6.92, 95% confidence interval 2.44 to 19.67), which the authors rated moderate certainty. They judged the trials of low methodological quality and reported no serious adverse events owing to acupuncture.
A 2025 network meta-analysis by Li and colleagues set acupuncture, sham acupuncture and standard care side by side, from 8 trials with 864 women searched to September 2022. Acupuncture and sham each relieved pain more than standard care, with no difference between them; on overall response and quality of life, acupuncture came out ahead of both. A second 2025 network meta-analysis, by Wang and colleagues, of 36 trials and 4,511 women across every non-drug treatment for pregnancy-related low back pain, ranked music with relaxation highest for pain and manipulation combined with acupuncture highest for physical function, and the authors called the evidence limited.
A 2022 review in BMJ Open by Yang and colleagues, of 10 trials of acupuncture for low back and pelvic pain in pregnancy, was retracted by the journal in June 2024 after an investigation found fundamental flaws in control-group selection and data extraction. Its results are not reported here.
The trials
The 2005 trial by Helen Elden and colleagues in Gothenburg randomised 386 women with pelvic girdle pain, at one hospital and 27 maternity care centres in Sweden, to six weeks of standard treatment alone, standard treatment plus acupuncture, or standard treatment plus stabilising exercises. Pain was measured morning and evening on a 0 to 100 visual analogue scale. An independent examiner also rated the severity of pelvic girdle pain before and after treatment, and found that severity fell most in the acupuncture group.
The same group then tested whether needle penetration mattered. Its 2008 double-blinded trial randomised 115 women with pain of at least 50 on the same scale to eight weeks of standard treatment plus acupuncture or plus a non-penetrating sham needle. Median pain fell from 66 to 36 with acupuncture and from 69 to 41 with sham. The authors concluded that needle penetration contributes only to a limited extent to the effects reported earlier.
In southern Sweden, a 2004 trial by Kvorning and colleagues randomised women at 24 to 37 weeks with pelvic or low back pain to acupuncture once or twice a week until delivery or recovery, or to no added treatment and no sham. At the end, 43% of acupuncture patients were less bothered by pain during activity, against 9% of controls. A 2019 Spanish trial by Vas and colleagues at 18 primary care centres had midwives without specialist acupuncture training give two sessions of ear acupuncture to 220 women at 24 to 36 weeks, in four groups of 55: true ear acupuncture, non-specific ear acupuncture, placebo ear acupuncture and standard care alone. Pain fell by 65.8% from baseline with true ear acupuncture and by 25.1% with standard care; the abstract reports the comparison with standard care.
Two later trials compared acupuncture with other active care. A 2019 French trial by Nicolian and colleagues at five maternity hospitals randomised 199 women to standard care or standard care plus five sessions with an acupuncturist midwife, open-label. Women given acupuncture had more days with pain at or below 4 out of 10 (61% against 48%), and once sick leave and lost work were counted, total costs were lower in the acupuncture group. A 2021 Swedish trial by Svahn Ekdahl and colleagues randomised 113 women with pelvic girdle pain at 12 to 28 weeks to ten acupuncture sessions from a physiotherapist or five weeks of home TENS (transcutaneous electrical nerve stimulation, a skin-surface device). The trial detected no differences between the groups, and evening pain fell within both.
Is acupuncture safe in pregnancy?
Elden and colleagues published a separate safety analysis of the 2005 trial, in which the acupuncture was strong and could begin in the second trimester. There were no serious adverse events in any group. Women given acupuncture reported 64 minor adverse events over 1,414 treatments, from 43 women, most often pain at the needle; fetal heart monitoring before, during and after 43 sessions was normal throughout. A 2015 review by David Carr found 15 controlled trials of acupuncture in pregnancy that used points traditionally forbidden there, and low back or pelvic girdle pain was the indication for 6 of them; their protocols included LI04 (Hegu) and BL60 (Kunlun). Carr found no adverse pregnancy outcomes in most of those trials, and rates in the rest no different from controls. Pregnancy and the forbidden points explains how the tradition and the trials read those points.
What guidelines say
Recommendation 1.4.15 of NICE’s antenatal care guideline (NG201), published on 19 August 2021, tells staff to consider referring women with pregnancy-related pelvic girdle pain to physiotherapy services for exercise advice, a non-rigid lumbopelvic belt or both. Its rationale says the committee found the evidence for acupuncture mixed and of poor quality, and not adequate to justify a recommendation with a substantial resource impact. The 2008 European guidelines, by Andry Vleeming and colleagues for a working group of the European Commission’s COST Action B13 on low back pain, reviewed three trials and the first Elden trial and concluded that there are indications acupuncture during pregnancy may reduce pain, but that high-quality studies are needed. Their recommended treatment for pregnant women is information, reassurance and individualised exercise.
What the reviews could not settle
Whether needling does more than a convincing sham is still open. Of the seven acupuncture trials in the Cochrane review, two had a sham arm, and they pointed in opposite directions on pain. The 2025 network meta-analysis by Li and colleagues found no difference in pain between acupuncture and sham. The Cochrane authors asked for an agreed way to classify women by their symptoms, since trials mixed low back pain, pelvic girdle pain and both, and diagnoses ranged from self-report to clinical tests. The EASE Back pilot, by Foster and colleagues in England, randomised 125 women to standard care, or standard care with true or non-penetrating acupuncture, and found a full trial feasible. Low back pain reports the evidence outside pregnancy, and Labour pain the evidence at birth.
AcuiQ’s pelvic girdle pain, symphysis pain and low back pain pages list the protocols individual studies prescribed, each with its citation.