Acupuncture for headache

It is easy to ridicule a 2000-year-old treatment that can seem closer to magic than to science. Indeed, from the 1970s to around 2005, the skeptic’s point of view was understandable, because the scientific evidence to show that acupuncture worked, and why, was weak, and clinical trials were small and of poor quality.

But things have changed since then. A lot.

Thanks to the development of valid placebo controls (for example, a retractable “sham” device that looks like an acupuncture needle but does not penetrate the skin), and the publication of several large and well-designed clinical trials in the last decade, we have the start of a solid foundation for truly understanding the effectiveness of acupuncture.

How do we know if acupuncture really works for pain?

Individual large-scale clinical studies have consistently demonstrated that acupuncture provided better pain relief compared with usual care. However, most studies also showed little difference between real and sham (fake) acupuncture. In order to address this concern, a 2012 meta-analysis combined data from roughly 18,000 individual patients in 23 high-quality randomized controlled trials of acupuncture for common pain conditions. This analysis conclusively demonstrated that acupuncture is superior to sham for low back pain, headache, and osteoarthritis, and improvements seen were similar to that of other widely used non-opiate pain relievers.

And the safety profile of acupuncture is excellent, with very few adverse events when performed by a trained practitioner. Meanwhile, basic science studies of acupuncture involving animals and humans have shown other potential benefits, from lowering blood pressure to long-lasting improvements in brain function. More broadly, acupuncture research has resulted in a number of insights and advances in biomedicine, with applications beyond the field of acupuncture itself.

Is acupuncture really that good?

We understand why there may be continued skepticism about acupuncture. There has been ambiguity in the language acupuncture researchers employ to describe acupuncture treatments, and confusion surrounding the ancient concept of acupuncture points and meridians, which is central to the practice of acupuncture. Indeed, the question of whether acupuncture points actually “exist” has been largely avoided by the acupuncture research community, even though acupuncture point terminology continues to be used in research studies. So, it is fair to say that acupuncture researchers have contributed to doubts about acupuncture, and a concerted effort is needed to resolve this issue. Nevertheless, the practice of acupuncture has emerged as an important nondrug option that can help chronic pain patients avoid the use of potentially harmful medications, especially opiates with their serious risk of substance use disorder.

Finding a balanced view

A post on acupuncture last year dismissed acupuncture as a costly, ineffective, and dangerous treatment for headache. This prompted us to point out the need for a measured and balanced view of the existing evidence, particularly in comparison to other treatments. Although the responses that followed the article overwhelmingly supported acupuncture, it nevertheless remains a concern that this practice attracts this kind of attack. Acupuncture practitioners and researchers must take responsibility for addressing deficiencies in acupuncture’s knowledge base and clarifying its terminology.

That said, we need to recognize that acupuncture can be part of the solution to the immense problem of chronic pain and opiate addiction that is gripping our society. That this solution comes from an ancient practice with a theoretical foundation incompletely understood by modern science should make it even more interesting and worthy of our attention. Clinicians owe it to their patients to learn about alternative, nondrug treatments and to answer patients’ questions and concerns knowledgeably and respectfully.


Acupuncture in patients with osteoarthritis of the knee: a randomised trial. Lancet, July 2005.

Acupuncture in Patients With Chronic Low Back Pain: A Randomized Controlled Trial. JAMA Internal Medicine, February 2006.

Acupuncture in patients with tension-type headache: randomised controlled trial. BMJ, August 2005.

Acupuncture for Patients With Migraine: A Randomized Controlled Trial. JAMA, May 2005.

Acupuncture for Chronic Pain: Individual Patient Data Meta-analysis. JAMA Internal Medicine, October 2012.

Survey of Adverse Events Following Acupuncture (SAFA): a prospective study of 32,000 consultations. Acupuncture in Medicine, December 2001.

Safety of Acupuncture: Results of a Prospective Observational Study with 229,230 Patients and Introduction of a Medical Information and Consent Form. Complementary Medicine Research, April 2009.

The safety of acupuncture during pregnancy: a systematic review. Acupuncture in Medicine, June 2014.

Cost-effectiveness of adjunct non-pharmacological interventions for osteoarthritis of the knee. PLOS One, March 2017.

Paradoxes in Acupuncture Research: Strategies for Moving Forward. Evidence-Based Complementary and Alternative Medcine, 2011.

The Long-term Effect of Acupuncture for Migraine Prophylaxis: A Randomized Clinical Trial. JAMA Internal Medicine, April 2017.


  1. Bridget Chin

    As an integrative medical acupuncturist with over 12 years of practice treating patients with chronic pain and previously 7 years as a general surgeon, I greatly understand and sympathize with individuals who “do not believe in acupuncture” and colleagues who comment that “acupuncture is all placebo effect”.

    As my grandmother taught me, “You cannot teach a pond frog about the ocean, it is too big for their understanding”. Personally, it is futile to argue that the world is round with true magnetic poles to those who believe that the world is flat.

    I continue my practice because my patients clinically improve. It is fulfilling to see individuals get back to their families and their lives. Interestingly, animals and children respond positively with acupuncture. The “belief” factor is moot because it cannot be asked or answered.

    The analgesic effect of acupuncture occurs in the nervous system through binding of the mu opioid receptors, the same receptors that are bound in sham/fake acupuncture and placebo effect; hence the overlapping response.
    Cao, Xiaoding, Acupuncture & Electro-Therapeutics Research, Volume 27, Number 1, 2002, pp. 1-14(14)

    The study of the human brain and nervous system continues to advance with the use of functional MRI and better quality research on acupuncture and chronic pain. The American College of Physicians recently published their guidelines on lower back pain. ACP strongly recommended Acupuncture before taking over the counter medications for lower back pain based upon newer scientific research. Amir Qaseem, MD, PhD, MHA, et al. Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A Clinical Practice Guideline From the American College of Physicians Ann Intern Med. 2017;166(7):514-530.

    Positively, a true acupuncture effect can be identified through functional MRI. Only through multiple sessions of true acupuncture can this occur, leading to neuromodulation. Maeda Y, et al. Rewiring the primary somatosensory cortex in carpal tunnel syndrome with acupuncture(link is external). Brain: a Journal of Neurology. March 2, 2017.

    The human body is more powerful than we can imagine. Our innate ability to heal is immense and our present understanding of the nervous system is extremely microscopic. Acupuncture is one method of healing, and it is only beneficial if the individual patient responds. As caregivers with limited options of pain medications, injections and surgery, we should embrace all modalities that can benefit our patients and improve their lives instead of wasting time arguing about which one is true, false or better.

    Bridget Chin, MS, MD
    Integrative Medical Acupuncture
    Spaulding Rehabilitation Network
    Clinical Instructor, Harvard Medical School

    • bob

      This is America. If acupuncture really worked everyone and their mothers would be using/demanding it from their docs.

      Investors galore and someone would be getting rich. There would be grand acupuncture hospitals everywhere and a cottage industry around them. And then there would be an app for it. hahaha. We would not be going to seedy acupuncture shops with neon lights in a plaza next to a chinese take out.

  2. vel

    As an engineer for nearly 2 decades, including field-support for telecom systems throughout the world, I have seen attempts at recreating issues or testing real-life conditions, in the lab fail miserably. Yet, in the real world, the behavior of the system can be completely different. We spent millions of dollars and thousands of hours attempting to simulate the customer conditions in the lab, yet in many instances, we couldn’t, or we had large holes or blind-spots in our theories that needed to be re-evaluated. This sounds to me like the controlled studies of acupuncture over the years. Engineers/researchers in their labs, don’t always have the same results as the customers/patients in the real-world environment.

    I changed careers and am now acupuncturist, after 4 years of schooling and 5 years in practice. I work in private practice as well as at an orthopedic rehab clinic, and am receiving consistent referrals from orthopod MDs, PTs and OTs because of the results they see in their patients with acupuncture. With a full schedule, we are bringing in another acupuncturist to support the demand. Having a year-on-year growing practice with a steady flow of referrals from happy patients and medical providers makes one think that there’s more to it than placebo.

    As for studies…
    We need to keep the studies coming, and continue to refine them to better understand whether the placebo being used is truly inert, and to better learn the mechanisms of acupuncture. We are learning that the mechanisms and effects are multifactorial and more complex than thought.

    We also need the skeptics and doubters to continue to (civilly) question and challenge, and thus push researchers to do more and better research. We also need them to push and challenge acupuncturists to continue to push ourselves to do even better work, and to better explain our medicine in biomedical terms (based on valid research) so this wonderful medicine can expand to help even more people.

    • Bob

      One way to determine placebo effect is to apply it to subjects that are not subject to placebo. One such creature are animals and also I suspect medical sceptics and those people that have not received any significant pain or mobility benefit after trying many modalities and are still searching for more help. These types would be less susceptible to placebo as they have tried everything and failed. I would try acupuncture on injured animals so working with a Vet and determining if you can get a decent study organised may be illuminating as we can eliminate placebo effect completely here.

      • vel

        Bob — agreed. In nearly 10 years of experience observing and dispensing acupuncture … with babies and animals, it still works consistently. Studies to specifically look at that may be helpful. (There may be ones already published already.)

        Anecdotally, my first exposure to acupuncture was about 15 years ago with a dog with allergies was adopted. (With the adoption, he came with a bottle of prescription antihistamines and prednisone.) After many different treatments (allergy-shots, nutrition/dietary modifications, even an experimental immune therapy, etc..) and giving them each a year to see any changes, a neighbor recommended a vet who practiced acupuncture. After the first treatment, his itching subsided dramatically. I’d never seen anything like it before outside of steroids (but without the intense negative side-effects). Continued treatment kept him minimally reactive for years.

        When he was older, and he had severe arthritis in his hips, he would come to me nightly, and lie down by my desk to get needles. He exhibited a marked improvement in his gait and movement after the therapy. We avoided surgery until he passed.

        Re Placebo:
        I joke that my humble tiny office with it’s Ikea furniture, can’t possibly give off much of a placebo effect. If you want placebo, walk into the grand multi-story, awe-inspiring entrance halls of a major hospital. It has beautiful, rich woods, artwork and marble/granite finishes. It exudes the power of a large institution that can heal you instantly with its wealth of resources.

  3. Bob

    A lack of science studies does not mean it doesn’t work but rather says we don’t know how it works nor the key variables. So another measure is needed. The only place were it counts where something is proven to work is the real world not science experiments or studies which do not necessarily represent the real world and how effective it is in a non controlled environment where anything goes. You may well say as a sceptic that when the first antibiotic Penicillin was discovered and used on patients they where not actually cured by it even though they got better because no significant scientific studies had been done at the time to prove it. We now know many years later through many scientific studies that it did work. So did it work at the start when there were no studies done on it? If doing studies is your sole criteria in determining whether something works or not then your knowledge is as uncertain as the many studies that show effect and non effect and then you must choose which studies will support your point of view. But the real world should always be your true measure. To illustrate it even more the Germans did MRi scans of the brains of patients given prescription pain medications for back pain. One group was given sugar pills and told that they were pain medications. The MRI scans showed that the pain had decreased. The other group were given pain medication but told they were sugar pills and surprisingly the brain scans showed that they were still in pain. How is this possible given that pain meds undergo stringent testing for effectiveness above placebo with science papers published galore and yet this is what the German’s found out. If you check out the show called “In Good Shape” you can see the Professor who did the study who then went on to conclude what this all means and given that the over use of opiates have proven a disaster you can draw your own conclusions that an alternative is necessary and needed urgently and given what the Germans have shown of the effectiveness of prescription pain medication people need to keep an open mind about alternatives.

  4. Maria C.H. Lin

    Accupuncture is highly dependent on the skill of the practitioner. When It works it is wonderful. However, I do not recommend coming to China to have the treatment. The practitioner here tend to place money above all else. They only do as much as need to keep the patient returning so they can generate more income.

  5. Rani

    I have looked at all the randomized tests that are included in this post. They all show no difference between Sham and Acupuncture ! So I am not sure I understood why the conclusions of the articles are different. Did I miss something ?

    • Helene Langevin

      The conclusion of the article are based on the cited meta-analysis:
      Acupuncture for Chronic Pain: Individual Patient Data Meta-analysis. JAMA Internal Medicine, October 2012.

      Also see previous reply to R Nadeau regarding meta-analyses.

  6. MC Sherpa

    So true! The commitment of all of us healers should be to relieve patient suffering and to support methods that do that, including acupuncture, micro-nutrient supplementation, meditation, exercise, healthy food, etc. If it works, and doesn’t harm….we should support it. There are often NOT good double blind studies of folks that used micro-nutrients and healthy eating to cure their depression, for example but yet these individuals are cured. It is important that we remain open minded and receptive to modalities that may be helping, regardless of why: placebo? effects not yet measurable through scientific study? There are lots of boxes that big Pharma want us put in…but we have choice to be open minded, and should be.

  7. Mengqi Xia

    Robert, I totally agree, we should look for more cost effective ways to manage our public health. It is time for a national dialogue.

    • Md. Ali

      Finally, the money making healthcare system will not allow any cost-effective treatment to be sustainable smoothly. Many psychological treatments, costing almost nothing, previously given by least educated people in Asian villages are now given in modern hospitals/clinic at high cost.
      -Md. Ali

      • Mengqi Xia

        Hi Ali, Thank you for your comment. I believe cost effectiveness for our public health can be achieved by careful analysis of available data and negotiations with healthcare professionals, so our healthcare professionals are willing to do the jobs. Many psychological therapies in the old days although cost “nothing”, there are human time and energy (let alone facilities) involved from the side of providers. Our healthcare is a big industry which support many people’s livelihood (jobs), this industry also has a responsibility for the health of our entire society. There need to be a balance achieved in order for our health care system to work for everyone’s interests

  8. R Nadeau

    The author states, “However, most studies also showed little difference between real and sham (fake) acupuncture.” Which means it doesn’t work. So in order to rescue the argument they do a metanalysis to see if they can get the answer they want. This is typical of pseudo-medicine. If the data don’t say what you want, give the data a good massaging until you get the answer you need. Truthiness.

    • KN

      R Nadeau says: ‘The author states, “However, most studies also showed little difference between real and sham (fake) acupuncture.” Which means it doesn’t work.’

      Nadeau left out the previous sentence: “Individual large-scale clinical studies have consistently demonstrated that acupuncture provided better pain relief compared with usual care.”

      There may be some reason that there’s no measurable difference between the 2 interventions (eg: the “placebo” is not completely inert?). But one cannot (rationally) say that “it doesn’t work”, without further research into why there’s little difference between placebo and active (yet both do better than conventional care.)

      • R Nadeau

        If you look at the references for this blog post I would be concerned about the author’s opinions on what constitutes a high-quality study and her claim that a meta-analysis “conclusively” showed that acupuncture is better than placebo is suspect, in my view. The better the study design the smaller the effect of acupuncture (or indeed there is a null effect). When a positive effect is so difficult to determine it seems to me that a treatment should simply be dropped, but this is more about belief than science.

    • Helene Langevin

      The goal of a meta-analysis using individual patient data is not to “massage” data, but rather to provide a larger sample size, which increases statistical power and confidence to determine whether-or-not differences exist between groups. Meta-analyses also allow for generalization and inferences to be made beyond individual study populations and protocol-specific procedures. For this reason, most national-level policies lean heavily on pooled evidence afforded by meta-analyses. Finally, we reiterate that the effect sizes for real vs. sham acupuncture are clinically meaningful and comparable to those comparing NSAIDs to placebo pills.

      • Carol

        For an analysis of your claims by Doctors who accurately portray why your study does not show what you think it does, please see and other analysis on why what you claim is none ture at

        Acupuncture is simnply an elaborate placebo, a fiction which some people have deluded themselves into believing. It’s sad when modern medicine wastes its time on what is essentially quackery. If you want to find out if acupuncture is real, the first thing you ought to study is whether there is a human energy field, aka qi, if meridians even exist, and to be ethical this must be done well before pretending its an effective treatment for anything.

  9. Li Ping Bartlett

    Bio-medical research approaches the human being at molecular level for creation of new medicines. Historically Chinese medicine has viewed the healing power of human beings as revitalizing energy i.e. Qi, working throughout the whole body and universe to regenerate and activate its own immune system. For repeatable evidence-based research of TCM in modern scientific theory, whole body simulation is vital. Using current technology, it is difficult but with further development of computer modelling and AI technology, it may become possible.

    • Sharon

      Hello Li Ping Bartlett,
      Are you able to recommend a reputable Acupuncturist in the Greater Lansing or MetroDetroit areas? I was told that Japanese acupuncture needles are the best for treatment. Do you hold any opinion?

      Thank you,

  10. Dr. Bahia Al-Salihi

    To quote the notable late Canadian neuroscientist Dr Bruce Pomeranz who discovered that part of acupuncture analgesia is mediated by the endorphin effect:

    I think there are two ways of being a scientist or even a modern person. There’s the empirical approach, which is trial and error: Does it work? If it works, then I’ll use it. In alternative medicine you see this in spades. If chicken soup works, use it. You don’t have to have a theory about chicken soup. Then there’s the theoretical approach. To me, those are the two ways of handling yourself. If acupuncture works, then use it; it doesn’t matter whether it works through chi or endorphins.

    Modern medicine has gone down the theoretical route and alternative medicine has stayed closer to the empirical route. My favorite example to help explain the dichotomy is this: A cook will use spices – salt, pepper, cumin – and he will mix them in certain proportions and taste them. If it tastes good, he will use it next time, but there’s no theory of spices. You don’t have to know which nerves in your tongue are affected by which spice in what proportion. You do it empirically. The theoretical approach is the other one. And we could do it. We know which nerves cumin affects, we know which nerves salt affects, and we could work out an equation for which ones are the best, but we wouldn’t end up cooking for another thousand years until we figured it out.

    The Chinese were very empirical in the early days of acupuncture (2200 years ago).

    • Rajesh Verma

      Lucid description.
      At least , modern medical scientists started to experiment and testify age old practices.
      When the allopathy will start believing that homeopathy also works.
      To keep it simple- many a times; allopathic medication does not work which worked well many. Can we say that allopath is ineffective.

  11. John Marlon

    Best way to know if acupuncture works…..go spend five years in China at a medical school studying acupuncture. I can assure you at the end of 5 years of being in a better position to render an opinion.

  12. Chuck Kistler

    It’s astonishing to me that Harvard Medical School is promoting acupuncture in 2018.

    • J Keeler

      It is astonishing to me that Doctors dismiss evidence based on their notion that it cannot work. Or that the mechanisms of how it works are not understood.
      As an acupuncturist I don’t have to believe the described mechanisms for how and why acupuncture works. I use my training as a mnemonic for how to approach treating a patient. Often a patient who has tried for years and multiple types of standard medical treatment for relief.
      I don’t have to believe or understand the described mechanism of why a group of locations and depths in a body consistently relieve headaches when present, and reduce the frequency and intensity when they are not acutely presenting during treatment.
      Unfortunately acupuncture doesn’t lend itself to double blind studies. Training people for 4 to 6 years who believe they are being trained as acupuncturists, but actually are trained as sham acupuncturists, seems extreme. But how else would you really deliver blind treatment. Where the person administering treatment would not know whether they were the placebo or not?

      • Rani

        Mr. Kleer, “It is astonishing to me that Doctors dismiss evidence based on their notion that it cannot work.” – Please take a look at the referenced trials in this article. None shows a difference between sham and Acupuncture, so why are you so sure ? The Placebo effect is substantial in Acupuncture (because of the needle effect) and if Prof. Rolls theory is right, that it stems from the Ventral Tagmental Area, it explain why there is a reduction in pain (Dopamine is the neurotransmitter of the VTA).

    • Robert Haile

      Why haven’t we done large studies in the USA on the effects of fruits and vegetables, exercise, and compare them singly or in combination to antidepressants. We are hard pressed to prove that antidepressants are much more effective than placebo, especially considering side effects, let alone comparing antidepressants to lifestyle and dietary changes. Why? Big Pharma can’t make huge profits over diet and lifestyle changes. Yet isn’t it interesting that countries with single payer healthcare systems are doing these very studies in an attempt to reduce costs. For years mainstream medicine denied manipulative medicine for back pain but it works at least equally well as standard care. Ultimately, it seems that changes and efforts most patients make to help themselves to treat or even prevent pain or other health issues should be far more effective. Countries, such as Japan, who spend less than half the costs of healthcare, focus more on prevention for all, yet they have the healthiest and longest lived populations requiring far less medications. I find it hard that Harvard is not pushing harder for lest costly self maintenance health care interventions, other than financial incentives associated with pharmaceuticals and invasive interventions.

    • Penny McCrae

      It astonishes me that, in an age of chronic opioid/opiate addiction, this isn’t being researched more aggressively.

Commenting has been closed for this post.