Cancer pain

Pain in people with cancer has several sources: the tumour pressing on or growing into bone, nerves and organs; procedures such as biopsies and surgery; and treatments whose side effects persist, such as nerve damage from chemotherapy or joint pain from hormone therapy. Reviews of acupuncture draw this boundary in different places, and where they draw it largely decides what they find. The Cochrane review counted only pain caused directly by the cancer and found too little evidence to judge. Later reviews that also counted pain caused by treatment, much of it joint pain from aromatase inhibitors, reported benefit against sham at moderate certainty. Joint pain from aromatase inhibitors, drugs that block oestrogen production after breast cancer, appears here only where it shapes those findings.

The Cochrane review

Carole Paley and colleagues first published the Cochrane review in 2011, with 3 trials and 204 participants, and updated it in October 2015, the latest version, with searches to July 2015. The review admitted only pain directly linked to the development of cancer, and excluded pain related to treatment, such as chemotherapy-induced neuropathic pain or pain after surgery. It included 5 randomised trials with 285 participants, which compared acupuncture with sham acupuncture, with placebo ear points, or with conventional pain medication. The authors could not pool them because the trials differed too much in cancers, methods and comparisons, and reported each one separately:

• A trial in pancreatic cancer pain reported benefit from acupuncture.
• A trial in ovarian cancer found no difference between real and sham electroacupuncture.
• A trial in late-stage cancer of unspecified type reported benefit over conventional medication.
• A trial in stomach cancer found no difference between acupuncture and conventional analgesia in the first 10 days.
• A trial of ear acupuncture for chronic neuropathic pain reported benefit over placebo.

Every trial was at high risk of bias from its small sample. The trials either reported no adverse events or did not report on them. The review concluded that there is insufficient evidence to judge whether acupuncture is effective for cancer pain in adults, and called the available evidence low quality.

The ear acupuncture trial is the one the 2011 version singled out as high quality. In that 2003 trial, Alimi and colleagues at the INSERM pain unit in Villejuif, France, randomised 90 people still in pain after at least a month of analgesics to ear needles at points where an electrical skin signal had been detected, needles at points with no signal, or seeds taped to points with no signal. At 2 months pain had fallen by 36% in the real acupuncture group and by 2% in the placebo groups. People in the seed group could see they had no needles, which the 2011 review noted may have biased the comparison in favour of the needle groups. Ear acupuncture describes the method.

Later reviews

Sham acupuncture is a control procedure made to look like acupuncture, such as needles that do not pierce the skin or shallow needling at points not used for treatment; Sham acupuncture covers the kinds in use. The reviews below kept their comparisons apart. A mean difference (MD) is the gap between groups on a 0 to 10 pain scale; a standardised mean difference (SMD) expresses the gap in units of the spread of scores, so that trials using different scales can be pooled.

He and colleagues, writing in JAMA Oncology, judged 6 of their sham-controlled trials at low risk of bias in every domain. Their pooled sham comparison was highly inconsistent between trials (I² 81%), which is why they rated it moderate rather than high. Of the 7 trials in it, 3 tested joint pain from aromatase inhibitors, including the trial by Hershman and colleagues described below; 7 of the 17 trials were run in China and 6 in the United States. The comparison with analgesics alone rests on 6 open-label trials, in which patients knew which group they were in, and the opioid figure on 2. The authors excluded short-term pain relief around surgery and concluded that acupuncture may not suit use as a stand-alone treatment for cancer pain.

Faria and colleagues in Lisbon reported larger effects against every comparator, and noted the low quality and small size of some trials and the mix of cancer types and stages. Lee and colleagues at the Korea Institute of Oriental Medicine asked whether the sham's location changes the answer. In 7 of their 8 trials the pain came from treatment rather than from the tumour. Real acupuncture beat sham needling at different points, but not sham applied at the same points as the real treatment, and the authors argued that same-point sham is not an inert placebo.

Trials

The PEACE trial, led by Mao at Memorial Sloan Kettering Cancer Center in New York and published in JAMA Oncology in 2021, randomised 360 cancer survivors with no current evidence of disease and musculoskeletal pain of at least 3 months. They received 10 weekly sessions of electroacupuncture, 10 of ear acupuncture, or usual care. In the electroacupuncture group acupuncturists chose 4 points near the pain, stimulated at 2 Hz, and at least 4 more for other symptoms. The ear protocol was the fixed five-point sequence developed by the US military, known as battlefield acupuncture. At 12 weeks average pain on the Brief Pain Inventory (BPI), a 0 to 10 scale, had fallen 1.9 points more with electroacupuncture than with usual care and 1.6 points more with ear acupuncture, and both reductions persisted to week 24. Ear acupuncture was not shown to be non-inferior to electroacupuncture at 12 weeks; 15 of 143 people stopped ear acupuncture because of adverse events, against 1 of 145 with electroacupuncture. There was no sham arm.

The IMPACT trial, led by Epstein from the same centre and published in JAMA Network Open in 2023, randomised 298 people with advanced cancer, moderate to severe musculoskeletal pain and a life expectancy of at least 6 months to weekly acupuncture or massage for 10 weeks, with monthly sessions up to 26 weeks. Worst pain on the BPI fell 2.53 points with acupuncture and 3.01 with massage, and the difference between them was not significant. With no sham or usual-care group, the trial compares two treatments and cannot separate either from the attention and expectation that come with them.

The largest sham-controlled trial in cancer treatment pain is the 2018 trial by Hershman and colleagues at 11 US centres, in 226 women with early breast cancer and joint pain from aromatase inhibitors. After 6 weeks, worst pain had fallen 0.92 points more with real acupuncture than with sham, and 0.96 points more than with a waiting list. Both were statistically significant, and both fell short of the 2-point difference the trial had set in advance as clinically meaningful.

What guidelines say

The 2022 guideline from the Society for Integrative Oncology (SIO) and the American Society of Clinical Oncology (ASCO), led by Mao, drew on 227 studies published from 1990 to 2021. For adults it states that acupuncture should be recommended for joint pain from aromatase inhibitors, and that acupuncture, reflexology or acupressure may be recommended for general cancer pain or musculoskeletal pain, each on an intermediate level of evidence with benefit outweighing risk and a moderate strength of recommendation. It found the evidence in children insufficient for any recommendation. The PEACE trial above was led by the same first author.

A 2022 guideline by the International Trustworthy Traditional Chinese Medicine Recommendations (TCM Recs) Working Group, led from Lanzhou University, published in Chinese Medicine, was developed in response to the He review and shares authors with it. Using GRADE, it made a strong recommendation for acupuncture over no treatment in moderate to severe cancer pain, a weak recommendation for acupuncture or acupressure added to analgesics to reduce pain, opioid dose and opioid side effects, and a strong recommendation for acupuncture for aromatase-inhibitor joint pain. It noted that the small trials behind it limit the strength of its recommendations.

The European Society for Medical Oncology (ESMO) guideline on cancer pain in adults, published in 2018 by a group led by Fallon, does not mention acupuncture. It covers drug treatment, radiotherapy, interventional procedures such as spinal cord stimulation, and psychological and behavioural approaches. The National Comprehensive Cancer Network’s adult cancer pain guideline sits behind a registration wall and is not reported here. Clinical guidelines compares other bodies’ positions on acupuncture.

Safety in cancer care

People having cancer treatment can have low platelet counts, which raise the risk of bleeding, or swollen limbs after lymph nodes are removed. A 2015 review of records by Cybularz and colleagues at a private cancer hospital found 98 acupuncture visits by adults with platelet counts below 100,000 per microlitre, 9 of them below 50,000, and no record of increased bruising or bleeding; the authors called for prospective trials to confirm it. In a 2018 trial in 82 women with lymphoedema after breast cancer, Bao and colleagues reported no severe adverse events from twice-weekly acupuncture, and no reduction in arm swelling against a waiting list. Bleeding and anticoagulants covers needling when blood clots poorly.

What the reviews could not settle

For pain caused by the tumour itself, the question the Cochrane review asked, the 2015 update found five small trials it could not pool. The later reviews pooled pain caused by tumours with pain caused by treatment, and in Lee’s sham-controlled set only 1 of 8 trials concerned pain from the cancer itself. The positive sham comparison in He’s review rests in large part on trials of joint pain from aromatase inhibitors, and the largest of them fell short of its own threshold for a meaningful difference. The two largest trials in cancer survivors and in advanced cancer, PEACE and IMPACT, had no sham arm, and Lee’s analysis shows that the answer against sham depends on where the sham needles go. The comparisons with analgesics come from open-label trials, and the reduction in opioid dose from 2 of them. He and colleagues found too few trials of particular kinds of cancer pain, such as nerve pain and pain from bone metastases, to say whether acupuncture behaves differently in each, and the Cochrane update found no trial that referred to bone pain at all. Related pages report trials in Chemotherapy-induced neuropathy, Pain after surgery, Chronic pain and Cancer-related fatigue.

AcuiQ’s cancer pain and cancer-related pain pages list the protocols individual studies prescribed, each with its citation.