Aromatase inhibitors are drugs that block the body's production of oestrogen. After surgery for stage I to III hormone receptor-positive breast cancer, they improve survival in women past the menopause compared with tamoxifen, according to a 2022 Cochrane review by Kate Roberts and colleagues at Princess Alexandra Hospital in Brisbane. The review describes the joint pain, muscle pain and stiffness the drugs cause in about half of these women, known as aromatase inhibitor-induced musculoskeletal symptoms (AIMSS), and reports that the symptoms reduce adherence to the drug in up to half of women. The acupuncture evidence rests on a small number of trials: three small sham-controlled trials, then SWOG S1200, a trial of 226 women run by the cancer research network SWOG, and a 2022 joint guideline from the Society for Integrative Oncology and ASCO that drew on it.
No Cochrane review of acupuncture
Cochrane has published two reviews on AIMSS, both by the Roberts group, and neither covers acupuncture. The 2020 review of exercise found seven trials with 400 women and rated the evidence on worst pain very low in certainty. The 2022 review of systemic therapies, drugs, supplements and other treatments taken by mouth or injection, found 17 trials with 2,034 women and rated every outcome low or very low; its methods state that it excluded acupuncture given on its own because acupuncture is not a systemic therapy, and that it excluded systematic reviews of acupuncture for AIMSS.
Systematic reviews
| Review | Searched to | Trials (women) | Compared with | Certainty |
|---|---|---|---|---|
| Chen 2017 | February 2017 | 5 (181) | Sham acupuncture, conventional treatment | Very low to low, as rated by Zhang 2024 |
| Liu 2021 | November 2019 | 7 (603) | Sham acupuncture, waiting list, conventional treatment | Very low to low, as rated by Zhang 2024 |
| Bae 2023 (network meta-analysis) | October 2021 | 17 (1,516), all treatments | Each other and inactive controls | Not stated in the abstract |
| Zhang 2024 (umbrella review) | December 2023 | 6 reviews covering 9 trials | Sham, placebo, waiting list, conventional treatment | Low for 4 of 27 outcomes, very low for 23 |
Chen and colleagues at Taipei Medical University reported lower worst-pain scores on the Brief Pain Inventory (BPI), a 0 to 10 scale, after 6 to 8 weeks of acupuncture, and 18 minor adverse events in 8 patients over 398 treatment sessions reported by one trial. Liu and colleagues at Nanjing University of Chinese Medicine pooled seven trials and reported lower BPI scores with acupuncture than with control overall. Split by control, the difference against waiting lists was significant and the difference against sham acupuncture was not, for worst pain, pain severity and the interference of pain with daily life alike. No trial reported a serious adverse event.
Bae and colleagues, an Australian and Korean group, compared every treatment tested for this pain in one network meta-analysis, a method that ranks treatments that were never tested against each other directly. Acupuncture ranked first for pain intensity and sham acupuncture second, ahead of herbal medicine, exercise, the antidepressant duloxetine, vitamin D and omega-3 fatty acids.
Zhang and colleagues at Tianjin University of Traditional Chinese Medicine assessed six earlier reviews, two published in Chinese, rather than the trials themselves. AMSTAR 2, a checklist for how a review was conducted, rated all six low or very low. Of the 27 outcomes the reviews reported, GRADE, the system that rates confidence in a result from high to very low, put 4 at low certainty and 23 at very low, mostly for inconsistency between trials, imprecision and possible publication bias. The authors concluded that acupuncture is safe and effective for this pain and that higher-quality research is needed.
Trials before S1200
- Columbia University, 2010. Crew and colleagues randomised 43 women, 38 of whom could be evaluated, to twice-weekly whole-body and ear acupuncture with points chosen for the painful joints, or to sham needling inserted shallowly away from acupuncture points, for 6 weeks. Mean worst-pain scores at 6 weeks were 3.0 with acupuncture and 5.5 with sham.
- University of Maryland, 2013. Bao and colleagues, in a trial at two centres, gave 47 evaluable women eight weekly sessions of real or sham acupuncture and found no significant difference between groups in disability or pain. The authors noted that the sham may not have been inert.
- University of Pennsylvania, 2014. Mao and colleagues randomised 67 women to electroacupuncture, in which a small current passes through the needles, to sham acupuncture with non-penetrating needles away from traditional points, or to a waiting list. Pain severity fell by 2.2 points with electroacupuncture and 0.2 on the waiting list at 8 weeks; the sham group's fall, 2.3 points, was similar to that of electroacupuncture.
The three trials disagree. Sham acupuncture sets out why shallow, off-point and non-penetrating shams each control for different things.
SWOG S1200
SWOG S1200, reported in JAMA in 2018 by Dawn Hershman and colleagues at Columbia University, enrolled 226 women with early-stage breast cancer at 11 US sites from 2012 to 2017. Each was taking an aromatase inhibitor and scored at least 3 on the BPI worst-pain item. The trial randomised them 2:1:1 to true acupuncture (110), sham acupuncture (59) or a waiting list (57). Both needling arms had 12 sessions over 6 weeks, then one a week for 6 more weeks.
The published protocol, by Heather Greenlee and colleagues, lists the true acupuncture points as a standard Traditional Chinese Medicine prescription for joint pain, which it calls Bi syndrome: TB05 (Waiguan), LI04 (Hegu), GB41 (Zulinqi), GB34 (Yanglingquan), ST41 (Jiexi) and KD03 (Taixi), which the protocol writes as SJ 5, LI 4 and KD 3. It added the National Acupuncture Detoxification Association ear protocol in one ear (Shen Men, Kidney, Liver, Upper Lung and Sympathetic), points for up to four of the most painful joints, and no electrical stimulation. The sham used shallow needling at four non-acupuncture points on the forearms and lower legs, joint-area shams, and adhesives on non-points of the ear.
At 6 weeks worst pain had fallen by 2.05 points with true acupuncture, 1.07 with sham and 0.99 on the waiting list. The adjusted differences were 0.92 points against sham and 0.96 against the waiting list, both statistically significant and both short of the 2-point difference the trial had set in advance as clinically meaningful. The authors called the improvement of uncertain clinical importance. In an analysis they added afterwards, 58% of the true acupuncture group, 33% of the sham group and 31% of the waiting list improved by at least 2 points. Mild bruising was the commonest adverse event, in 47% with true acupuncture and 25% with sham, and no event of grade 3 or higher occurred. More women given true acupuncture than sham believed they had received true acupuncture, which the authors list as a possible failure of masking.
The trial offered every participant 10 sessions of true acupuncture between weeks 24 and 52. The 52-week report, in JAMA Network Open in 2022, found worst pain 1.08 points lower with true acupuncture than with sham and 0.99 lower than on the waiting list, with 191 of the 226 women completing the trial. Between weeks 24 and 52, 13.2% of the true acupuncture group, 11.3% of the sham group and 10.6% of the waiting list reported having acupuncture.
What guidance says
The Society for Integrative Oncology's 2017 guideline on integrative therapies during and after breast cancer treatment, by Greenlee and colleagues, searched to 2015, before S1200 reported. It found five trials of acupuncture for this pain, all small and with mixed findings, and gave acupuncture a grade C, a recommendation to offer it selectively to individual patients on professional judgement and preference.
The 2022 joint guideline from the Society for Integrative Oncology and ASCO, the American Society of Clinical Oncology, on integrative medicine for pain in people with cancer, by Jun Mao and colleagues, reviewed 227 studies published from 1990 to 2021. It states that acupuncture should be recommended for aromatase inhibitor-related joint pain in adults, on an intermediate level of evidence with benefit outweighing risk, and a moderate strength of recommendation. Clinical guidelines compares how other bodies treat acupuncture, and Cancer pain covers the guideline's other recommendations.
What the reviews could not settle
The trials against a waiting list favour acupuncture; the trials against a sham disagree. Crew and S1200 found true acupuncture ahead of a shallow-needling sham, Bao found no difference, Mao found a non-penetrating sham doing as well as electroacupuncture, and the 2021 pooled sham comparison was not significant. S1200's difference from sham fell below the threshold its own designers had set. The reviews could not say whether needling at the chosen points, the sham procedures or the course of treatment accounts for the change, and the umbrella review rated nearly every pooled result very low in certainty.
The trials also do not show whether treating the pain keeps women on the drug, which S1200's authors gave as the reason for the research, since the symptoms often lead women to stop it. A 2025 scoping review of 74 reports on every kind of AIMSS treatment, by Bobo and colleagues at the Mayo Clinic, found three that measured whether women kept taking their aromatase inhibitor, and found every study short beside the 5 to 10 years for which the drugs are taken.
AcuiQ's arthralgia, joint pain and breast cancer pages list the protocols individual studies prescribed, each with its citation. Related evidence pages cover hot flushes, cancer-related fatigue and chemotherapy-induced neuropathy.