ONE WILD WORLD — NEWSLETTER

ISSUE 057 · DEEP DIVE · MYTHS & MISINFORMATION

If you want to be fair to acupuncture, you go to Andrew Vickers and the Acupuncture Trialists' Collaboration. They did something unusual and expensive: rather than pooling published summaries, they wrote to the original trialists and obtained the raw patient-level data. Then they included only trials where the randomisation was unambiguously sound.

The updated analysis covers 39 trials and 20,827 patients with chronic headache, migraine, osteoarthritis, and back, neck or shoulder pain.

Acupuncture beat no acupuncture by close to 0.5 standard deviations. It beat sham acupuncture by close to 0.2. The effect faded by only about 15 per cent after a year. The authors' conclusion was that acupuncture has a clinically relevant effect on chronic pain that cannot be explained by placebo alone.

So the simple dismissal — it is theatre, nothing happens — does not survive contact with the best data. Something is going on.

Germany spent serious money finding out whether acupuncture should be covered by statutory health insurance. The GERAC trials were the result: patient- and observer-blinded, 340 outpatient practices, and for the back pain arm, 1,162 patients who had been in pain for an average of eight years.

Three groups. Real acupuncture following traditional Chinese medicine. Sham acupuncture, meaning shallow needling at deliberately non-acupuncture points, with no attempt to produce the deqi sensation practitioners consider essential. And conventional guideline-based care: drugs, physiotherapy, exercise.

Real acupuncture and wrong-place acupuncture performed the same. Neither was better than the other on the primary outcome.

Both of them beat the drugs and physiotherapy, by a wide margin.

The summary of the whole German research programme — ART, ARC and GERAC together — states it without hedging: these studies do not confirm the hypothesis that needling at specific points is essential to achieving satisfactory clinical effects.

A systematic review in 1998 looked at where acupuncture trials came from and what they found. The results from China, Japan and Taiwan were uniformly favourable. Not mostly favourable. Reviewers noted that no negative trial of acupuncture had been published in China at all.

Japan has since improved somewhat — roughly 20 per cent of trials in the 2000s and 30 per cent in the 2010s reported negative results. A 2023 review found Chinese studies still carried methodological problems that non-Chinese studies did not, and predictably produced larger apparent effects.

This matters for a specific reason. The difference between real and sham acupuncture in the best meta-analysis is about 0.2 standard deviations. That is a small number. Small enough that critics argue it sits comfortably inside the noise generated by publication bias, researcher expectation and the ordinary degrees of freedom in trial analysis. You do not need much of a thumb on the scale to manufacture 0.2.

Here is the deepest issue, and it is the reason this argument has run for thirty years without resolution.

Sham acupuncture is not a placebo.

A sugar pill is inert. Shallow needling is not — you are still putting metal through skin, which produces real physiological effects regardless of any meridian. The GERAC authors saw this coming and noted that both arms involved needling, so the comparison could not cleanly separate the specific from the non-specific.

It gets worse. A 2023 network meta-analysis of ten trials and 4,379 patients with chronic low back pain found that sham needling at the same points as the treatment group produced better outcomes than sham needling at different points. The authors concluded that same-point sham may not be a true placebo control at all.

Read that again. The control condition works, and it works better when you do it in the places the theory says matter. Which either rescues point specificity or demolishes the trial design, depending on who you ask — and both camps have been asking for decades.

The honest answer is that nobody has nailed it down, but the shape is reasonably clear.

Something real occurs when you put needles in people. It reliably outperforms doing nothing, and in the German data it outperformed the standard drugs-and-physio package for chronic back pain, which should trouble anyone who thinks this is pure theatre.

What has never been demonstrated is that any of it depends on the framework it is sold with. Not the meridians, not the qi, not the point locations, not the depth, not the deqi sensation. The map is not doing the work.

Which leaves a treatment that appears to be effective for exactly the wrong reasons — and a two-thousand-year-old explanation that the treatment's own best trials have quietly retired.

Vickers, A. J. et al. (2018). “Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis.” The Journal of Pain 19(5), 455–474; and Vickers et al. (2012), Archives of Internal Medicine 172(19), 1444–1453.

Haake, M. et al. (2007). “German Acupuncture Trials (GERAC) for chronic low back pain: randomized, multicenter, blinded, parallel-group trial with 3 groups.” Archives of Internal Medicine 167(17), 1892–1898.

Witt, C. M. et al. (2009). “Modellvorhaben Akupunktur — a summary of the ART, ARC and GERAC trials.” Acupuncture in Medicine.

Vickers, A., Goyal, N., Harland, R. & Rees, R. (1998). “Do certain countries produce only positive results? A systematic review of controlled trials.” Controlled Clinical Trials 19(2), 159–166.

Lee, B. et al. (2023). “Needling Point Location Used in Sham Acupuncture for Chronic Nonspecific Low Back Pain: A Systematic Review and Network Meta-Analysis.” JAMA Network Open 6(9).

Campbell, A. (2006). “Point specificity of acupuncture in the light of recent clinical and imaging studies.” Acupuncture in Medicine.

Novella, S., critique of the Vickers meta-analysis, Science-Based Medicine (2012); MacPherson, H. et al. (2014), PLOS One 9(4), e93739, on the influence of control group choice on effect size.

ONE MORE THING

You can read our MANY newsletters for free at our website:

Read our archive for free at OneWildWorld.com →