Does Acupuncture Work? What the Evidence Can and Cannot Say
A plain look at what trials of acupuncture show for pain, headache and other concerns, what sham-controlled results mean, and where reviews and guidelines disagree.
Content last updated . This page has not been medically reviewed.
The honest short answer
"Does acupuncture work?" is really three questions. Is it better than doing nothing? Is it better than a pretend (sham) treatment? And is it as good as other options? The answers differ by condition, so one yes or no would mislead. This page reports what specific reviews and trials found.Why "better than sham" and "better than nothing" differ
NCCIH notes that in many studies the benefit was greater against no treatment than against sham, which suggests nonspecific effects contribute. [6] Those effects come from belief, attention and the treatment ritual, not from needling at a specific point. They are real experiences, but they are not the same as a needle effect. Sham designs have their own limits, covered on sham acupuncture and placebo controls and on the TCM page about sham needles.Guidelines also differ
The UK's NICE says not to offer acupuncture for low back pain with or without sciatica. [9] For chronic primary pain, NICE says to consider a single course of acupuncture or dry needling, within set limits of setting and clinician time. [10] The NHS says it may offer acupuncture for long-term pain, migraine-type headaches or severe hiccups. [11] These are UK positions, and other countries' bodies may differ.Frequently asked questions
Is acupuncture just a placebo?
Not entirely, according to the largest pooled analysis, which found a small advantage over sham and said placebo effects alone could not explain pain relief. [1] But the advantage over sham is modest, and other reviews question whether it is clinically meaningful. [8]
For which conditions is the evidence strongest?
Among those listed here, chronic pain conditions such as neck pain and tension-type headache have moderate-quality evidence of a modest benefit over sham. [4][5] Evidence is weaker or negative for IBS, IVF and carpal tunnel syndrome. [6]
Does it work as well as medicine?
For migraine prevention, NCCIH describes a review of nine studies finding acupuncture slightly more effective than drugs, but this was one review and comparisons vary by condition. Ask your clinician about options. [6]
Why do two reviews of the same topic reach different conclusions?
They may count different trials, use different comparison groups, or define "meaningful benefit" differently. The Cochrane back pain review used a 15-point threshold on a 100-point pain scale. [3]
Can I rely on one study showing it works for me?
One trial rarely settles a question. Reviews that pool many trials give a steadier picture, and still depend on trial quality. [1]
More details: evidence, limits and cautions
The details behind this page: cautions, limits of the evidence and where sources disagree. Open any note to read it.
How to read the evidence labels
Labels used on this page: established, promising but limited, mixed, insufficient evidence, no good evidence of benefit. Acupuncture itself is explained on What Is Acupuncture?. This page is information, not medical advice.
What the largest pooled analysis found
An update of an individual patient data meta-analysis combined raw data on 20,827 patients from 39 trials of nonspecific musculoskeletal pain, osteoarthritis, chronic headache and shoulder pain. [1] Acupuncture beat both sham and no-acupuncture control for each pain condition, with differences close to 0.5 standard deviations against no acupuncture and close to 0.2 against sham. [1] The authors concluded that pain relief could not be explained solely by placebo effects. [1] An earlier version of the analysis agreed, but noted the differences from sham were relatively modest and that factors beyond needling itself contribute. [2] More detail is on large reviews of acupuncture.Evidence by condition
| Question | What the source found | Label |
|---|---|---|
| Chronic low back pain | Cochrane review (33 studies, 8,270 people): moderate-certainty evidence of clinically important relief versus no treatment. [3] Versus sham, low-certainty evidence of about 9 points on a 0 to 100 scale, below the 15-point threshold the reviewers used for "clinically important." [3] | Mixed |
| Neck pain | Cochrane review: moderate-quality evidence of better pain relief than sham at short-term follow-up; the effect did not seem sustainable over the long term. [4] | Promising but limited |
| Tension-type headache | Cochrane review: 51% had at least a 50% drop in headache frequency with acupuncture versus 43% with sham. [5] | Promising but limited |
| Migraine prevention | NCCIH reports it may be better than sham, but the difference is small. [6] A 150-person trial in China found 20 sessions beat both sham and usual care. [7] | Promising but limited |
| Fibromyalgia | A 2019 review found better pain relief than sham, with low-to-moderate quality evidence. [6] | Promising but limited |
| Hot flashes | Better than no acupuncture, but not shown to be better than sham. [6] | Mixed |
| IBS | A 2019 review found it no more effective than sham. [6] | No good evidence of benefit (versus sham) |
| IVF success | Six studies found no difference from sham in pregnancy or live birth. [6] | No good evidence of benefit |
| Carpal tunnel syndrome | May have little or no effect compared with sham. [6] | No good evidence of benefit |
| Battlefield acupuncture | A 2021 review found no significant pain improvement against no treatment, usual care, delayed treatment or sham. [6] | No good evidence of benefit |
| Quitting smoking | The US Preventive Services Task Force made no recommendation because evidence was limited. [6] | Insufficient evidence |
Each condition has its own page: see back and neck pain, migraine and tension-type headache and fertility and IVF. One condition's result should not be carried over to another.