Acupuncture: of the five treatments, the one with the soundest evidence — and the cheapest
Division of Pulmonary Medicine, Taipei Veterans General Hospital Hsinchu Branch · Wellness treatments
Related pagesWellness treatments overviewTranscranial magnetic stimulation
A woman of 58 had had low back pain on and off for seven or eight years. The MRI showed a mild disc protrusion, but the neurosurgeons said it was “not bad enough to need operating on”. She had taken every sort of painkiller and her stomach was starting to suffer; she wanted physiotherapy, but could not fit it around her working hours.
In the end her daughter dragged her off to have acupuncture, which she herself half believed in at best. Six weeks later she came back and said: “It is not that it does not hurt at all, but it has gone from hurting every day to hurting occasionally. I sleep better at night, and I only take a painkiller once or twice a week.”
That outcome is in fact very much in line with what the large studies show — acupuncture works for chronic pain, but it is no miracle. Its value lies in this: the effect is real, the safety is good, it is covered by the National Health Insurance (NHI), and it can let people rely less on painkillers.
How does modern medicine understand acupuncture?
Whichever theory is used to explain it, at the moment the needle goes into the skin and muscle something measurable really does happen in the body:
- Gate control theory: the needle activates the thicker sensory nerve fibres (Aδ and Aβ), which “closes the gate” in the dorsal horn of the spinal cord and suppresses the pain signals coming in along the thin fibres (C fibres). This is also why you instinctively rub a place you have knocked — the same mechanism.
- The endogenous opioid system: needling can prompt the central nervous system to release endorphins, enkephalins and dynorphins. In animal experiments, giving the opioid antagonist naloxone clearly weakens the pain-relieving effect of needling, which is important evidence in support of this mechanism.
- Local effects: the needle causes a tiny amount of tissue irritation, which locally releases adenosine, nitric oxide and other mediators, producing local pain relief and improved blood flow.
- Central modulation: functional brain imaging shows that needling changes activity in areas such as the limbic system, the hypothalamus and the periaqueductal grey matter — all of them key nodes in the modulation of pain.
How strong is the evidence?
Chronic pain: the most reliable source of evidence at present
The Acupuncture Trialists’ Collaboration brought together the raw data on individual patients from high-quality randomised controlled trials for a meta-analysis covering upwards of about 20,000 patients, in chronic low back and neck pain, osteoarthritis, chronic headache and shoulder pain. This is a methodologically rather uncommon and rather powerful approach — because what is used is the raw data, not the summary figures from each study.
The results can be looked at on two levels:
- Acupuncture is better than control groups that had no acupuncture (usual care, or a waiting list, for instance), and clearly so.
- Acupuncture is also better than “sham acupuncture” (needling at non-points, or without penetrating the skin). The difference is smaller, but it is statistically significant — and this matters, because it means the effect cannot be explained entirely as placebo.
- The effect is still broadly maintained a year later, having fallen off only slightly.
The honest conclusion is this: acupuncture genuinely does work for chronic pain, but the effect is “moderate” rather than dramatic. For many people with chronic pain, an effect that is moderate but lasting and safe is already of real value.
| Situation | State of the evidence | How to read it |
|---|---|---|
| Chronic low back pain, neck and shoulder pain | Supported by the individual-patient-data meta-analysis | The evidence is clear |
| Osteoarthritis of the knee | As above; listed in the guidelines of several countries as a non-drug treatment worth considering | The evidence is clear |
| Chronic headache, prevention of migraine | Supported by systematic reviews; can reduce the number of days with an attack | The evidence is clear |
| Nausea and vomiting caused by chemotherapy | Supported by several trials (including the PC6 Neiguan point) | Can be used as an adjunct |
| Rehabilitation after stroke | Already covered by the NHI in Taiwan; the research findings are inconsistent | Can be used as an adjunct |
| Allergic rhinitis, insomnia, menopausal symptoms | There are positive studies, but they are very heterogeneous | Can be used as an adjunct |
| Giving up smoking, losing weight | The evidence is weak | Not advisable to rely on acupuncture alone |
| Treating cancer, replacing surgery | — | No evidence |
NHI coverage: the only one of the five that is covered
Acupuncture is the only one of the five treatments in this section that is covered by the National Health Insurance (NHI), and in practice that matters a great deal — it greatly lowers the threshold, and the risk, of trying it.
The National Health Insurance Administration pays according to bands of treatment complexity. In outline the idea is:
- Ordinary acupuncture: basic needling treatment.
- Acupuncture of moderate complexity: for a defined list of conditions, needling several sites together with an adjunctive therapy.
- Acupuncture of high complexity: this applies to more complex diagnoses such as stroke, spinal cord injury, Parkinson’s disease, malignant tumours, and particular psychiatric and ophthalmic conditions. The conditions include: needling two or more anatomical sites (head, neck, trunk, limbs), combining it with one of cupping, scraping, heat therapy, electrotherapy, bloodletting or moxibustion, and a treatment time of at least 20 minutes.
The bands and the list of indications are revised from time to time, so for what is actually covered go by the assessment of the Chinese medicine practitioner and the latest announcements from the National Health Insurance Administration. One further reminder: the NHI rules provide that acupuncture and Chinese herbal medicine can both be covered at the same time, so you do not have to choose between them; if you are asked to choose, or charged extra, you can raise it with the National Health Insurance Administration.
How is a course run?
| Item | Usual arrangement |
|---|---|
| Time per session | The needles are left in for about 20–30 minutes; with the assessment and taking the needles out, about half an hour to an hour in all |
| Frequency | For an acute problem, perhaps 2–3 times a week; for a chronic one, mostly 1–2 times a week |
| Length of the course | Chronic pain usually needs weeks to months before it can be properly assessed. The courses commonly used in the research are 6–12 sessions or more |
| What it feels like | Going in, mostly a brief pricking sensation; while the needles are in, you may get the aching, full, tingling, heavy feeling known as de qi, which is normal |
| What it may be combined with | Electroacupuncture, cupping, moxibustion, scraping, tui na massage and so on |
| Afterwards | You can carry on as normal. A few people feel briefly tired or sleepy; it is best not to do strenuous exercise straight away that day |
Safety and contraindications
When carried out by a qualified practitioner of Chinese medicine using disposable sterile needles, acupuncture is distinctly safe. Serious adverse events (pneumothorax, infection, nerve injury) are very rare, and mostly have to do with needling at an inappropriate depth or with inadequate sterilisation. Even so, the following still have to be mentioned beforehand:
- A tendency to bleed, or currently taking an anticoagulant or antiplatelet drug (warfarin, a DOAC, aspirin, clopidogrel): this is not an absolute contraindication, but the practitioner needs to be told, so that deep or high-risk points can be avoided and firmer pressure applied to stop any bleeding.
- Having a pacemaker or an implantable defibrillator: the current used in electroacupuncture may interfere with the device, so it should be avoided and ordinary needling used instead.
- Infection, ulceration, a burn, severe eczema or keloid scarring of the skin in one area: that area is avoided.
- Pregnancy: some points (Hegu, Sanyinjiao and the lumbosacral region, for instance) are traditionally avoided during pregnancy, so do be sure to say how many weeks pregnant you are.
- Severe immunosuppression (during chemotherapy, with a low neutrophil count, or after an organ transplant): the risk of infection is raised, and it needs discussing with the doctor looking after you.
- A limb with lymphoedema (the arm on the affected side after breast cancer surgery, for example): needling should be avoided on that side.
- Having fainted during acupuncture before (dizziness, cold sweat, blacking out while being needled): please say so beforehand, so that you can be treated lying down and not on an empty stomach.
- An empty stomach, extreme tiredness, or after drinking: fainting is more likely, so it is better to rearrange, or to eat and rest first.
Common and minor reactions
A small bruise or a little bleeding at the needle site (pressure is all that is needed), local aching after treatment, brief tiredness or sleepiness, and occasionally mild dizziness. These usually disappear within a day.
When should you come back, or seek medical attention, straight away?
- Difficulty breathing, chest pain or a persistent cough after needling of the chest or back — although this is extremely rare, a pneumothorax still has to be ruled out.
- Redness, swelling, heat, pain or pus at a needle site, which may mean infection.
- Persistent numbness or weakness, or any abnormal sensation.
Common questions
Does acupuncture really work, or is it just a placebo?
This is the question asked most often, and the one most deserving of a serious answer. The large individual-patient-data meta-analysis shows that acupuncture is not only better than “no treatment” but also better than “sham acupuncture” — and although the difference from sham acupuncture is smaller, it is statistically significant. That means the effect cannot be explained entirely as placebo. At the same time it has to be said honestly that the effect of acupuncture is of moderate degree rather than dramatic, and that the gap from sham acupuncture really is not large, so some part of the contribution does come from non-specific factors (expectation, attention, relaxation).
Does it hurt a lot?
The needles used in acupuncture are very fine (far finer than a blood-taking needle), and going in there is mostly only a brief pricking sensation. While the needles are in you get the aching, full, tingling, heavy feeling of de qi — this is the expected sensation, not an injury. If you get a sharp stabbing pain or a feeling like an electric shock, tell the practitioner at once so that it can be adjusted.
How many sessions before I know whether it is working?
For chronic pain it usually takes six or more before even a first judgement can be made, and a full assessment often takes weeks to months. Saying “it did not work” after only one or two is usually far too early. Conversely, if you have had it for one or two months with no improvement whatsoever, you should go back and discuss with the practitioner whether to change direction, rather than carrying on indefinitely.
Can acupuncture replace painkillers or physiotherapy?
It can reduce the use of painkillers, but it is not advisable to let it replace the treatment plan as a whole. The most effective approach to chronic pain is usually a combination: exercise therapy + the medication that is needed + acupuncture + attention to sleep and mood. Acupuncture is a good part of that, but it is not the whole of it. In particular, the long-term effect of exercise and strength training in chronic low back pain is something no passive treatment can at present replace.
I am taking an anticoagulant. Can I have acupuncture?
In most cases yes, but you must be sure to say so. The practitioner will avoid deep and high-risk sites, use fewer needles, and apply firmer pressure to stop any bleeding. Please do not stop your medication on your own account in order to have acupuncture — the risk of stopping an anticoagulant is far greater than that of a bruise at a needle site.
Does the NHI cover it? Will I have to pay?
Acupuncture is covered by the NHI, in bands according to the complexity of the treatment. The NHI also provides explicitly that acupuncture and Chinese herbal medicine can both be covered at the same time, so you do not have to choose between them. Where an institution offers a special needling technique or a self-paid course, do confirm what is charged for, and how much, before treatment.
References
- Vickers AJ, Vertosick EA, Lewith G, et al; Acupuncture Trialists' Collaboration. Acupuncture for chronic pain: update of an individual patient data meta-analysis. J Pain. 2018;19(5):455-474.
- Vickers AJ, Cronin AM, Maschino AC, et al. Acupuncture for chronic pain: individual patient data meta-analysis. Arch Intern Med. 2012;172(19):1444-1453.
- Zhao ZQ. Neural mechanism underlying acupuncture analgesia. Prog Neurobiol. 2008;85(4):355-375.
- 中華民國中醫傷科醫學會. 高度複雜性針灸適應症（健保署 112/03/01 版）與中度複雜性針灸適應症. Available from: https://acot.org.tw
- 衛生福利部中央健康保險署. 中醫門診總額相關給付規定與公告. Available from: https://www.nhi.gov.tw
References follow the ICMJE (Vancouver) style. Guidelines and official documents are given with their version or year of publication; legislation and National Health Insurance (NHI) rules are given with the responsible authority and the version date. Last checked: 8 October 2026.
Service information at this hospital
What follows is the usual arrangement. The number of sessions and the time per session are the ranges commonly seen in practice, not a promise to any individual patient; the indication, the number of sessions and the cost that actually apply to you depend on assessment in the outpatient clinic and on this hospital’s latest announcements.
- Where it is done
- The Chinese medicine department, outpatient building B1
- Hours
- Weekdays 08:30–16:30 (for the actual clinic sessions, see the Chinese medicine department timetable)
- Time per session
- The needles are generally left in for 20 minutes or more; 20 minutes is also one of the conditions for the NHI’s complex acupuncture items
- Suggested number of sessions
- Chronic pain usually needs weeks to months before the effect can be properly assessed; it cannot be judged from one or two sessions.
- Cost and coverage
- Most indications are covered by the NHI (under the rules for the global budget for Chinese medicine outpatient care, with different complexity bands); anything beyond what is covered is self-paid.
- How to book
- Book the Chinese medicine outpatient clinic directly; if a relevant problem is found in the chest medicine clinic, a referral can also be made after discussion there.
- Preparing for treatment
- Do not go too long without eating before treatment, but avoid being too full either; wear loose clothing that makes it easy to expose the areas to be needled; and if you have a tendency to bleed, are taking an anticoagulant, have a pacemaker or are pregnant, please be sure to say so first.
- Contact
- Taipei Veterans General Hospital Hsinchu Branch 03-596-2134 (ask for the treatment unit)
This content is health information. It is for reference only and cannot replace a professional medical diagnosis. Whether this treatment is right for you, and how a course should be arranged, must be decided by a doctor after assessing your individual situation. All of the working behind the interactive tools on this page is done on your own device; your answers are not uploaded and are not stored. This site records only anonymous usage statistics (for example, how many people read this page to the end, how many complete a self-assessment, and which band the result falls into). It does not include what you answered, and it sets no cookies. See the privacy notice for details.
Still have questions? Add Dr. Hwa-Yen Chiu’s LINE official account @chiubreath for clinic news and updates on respiratory health.
Add on LINE