At SBM we have two basic missions, which are really just different ends of the same spectrum. We are trying to raise the ceiling of rigorous science in the practice of medicine. This is a complex topic involving factors such as statistical methods, preregistering trials, p-hacking, and various forms of publication bias. At the same time we are trying to raise the floor by addressing the worse pseudoscience that is trying (successfully, unfortunately) to infiltrate medicine. On the former task it feels like we are rowing alongside others in our field – journal editors, university departments, and statisticians. But in the latter, tackling pseudoscience, it often feels that either we are working alone, or many experts who should know better are actively working against us, and actually supporting pseudoscience in medicine. This is due in part to a frustrating double standard in which regular medicine is held to a high standard, while “alternative” medicine has a standard all its own (functionally no standard).
The latest example of this disastrous double standard is a PNAS article on acupuncture. The article is completely gullible, written as if it were a promotional piece entirely by advocates, without a hint of critical thinking or skepticism. It is also internally inconsistent, even incoherent, but does usefully give the game away for those paying attention, especially with a background in SBM. The author is Lynne Peeples, a journalist, not a scientist. This makes sense in that the piece reads like a feature, with a typical narrative structure, not like an academic review or even opinion.
The piece, after profiling a TCM practitioner, opens with a biased review of the literature on acupuncture. Peeples makes it seem as if there is a growing evidence-based that acupuncture actually works, relying heavily on reviews like one by Vickers. We have already deconstructed the Vickers review here, showing why it is biased and does not show acupuncture works. We have also written often about how negative acupuncture studies are interpreted to make it seem they are positive. She also ignored other better reviews, such as this 2020 systematic review of systematic reviews of acupuncture which found: “Evidence from SRs suggests that there are insufficient high-quality RCTs to judge the efficacy of acupuncture for chronic pain associated with various medical conditions.”
David Gorski correctly reinterpreted this from an SBM perspective – “Inconclusive literature plus no plausible physiologic mechanism for acupuncture to work equal an ineffective treatment whose effects are nonspecific placebo effects.” As I have said many times, the null hypothesis of acupuncture is that it is nothing but theatrical placebo, and there is insufficient evidence to reject the null.
Peeples later hits upon a key reason why the evidence for acupuncture is lacking, but gullibly buys that standard spin of acupuncturists:
“Because sham acupuncture can also trigger genuine analgesic and other biological pathways, however, it is not a true placebo. Both factors may help explain some earlier confusing results, according to Nielsen. She authored another paper in September that argued a failure to account for the very real effects of sham controls led to widespread underestimation of acupuncture’s efficacy, noting that funders now prefer comparisons to standard care.”
Here is what is really happening. Those “confusing results” were actually negative results, verum or true acupuncture works no better than placebo or sham acupuncture. The more rigorous the study design, the better the blinding, the smaller any apparent effect, and the best studies are negative. This is a clear signal of a null effect, and supports the conclusion that acupuncture is a theatrical placebo. Sham acupuncture has “genuine” effects, but they are all placebo effects. That makes it a great placebo, because it controls for all factors other than the acupuncture itself.
The conclusion is the killer, though – “funders now prefer comparisons to standard care”. Of course they do – because those studies are usually positive. If you are doing an unblinded pain study, it will be positive, because pain is highly subjective and amenable to placebo effects and non-specific effects. Unblinded pain studies are basically worthless for determining efficacy. The fact that acupuncture researchers have moved beyond designing more rigorous studies (because they are frustratingly negative) to doing unblinded studies is extremely telling.
Much of the article is also about possible mechanisms. It is folly, however, to speculate about how a treatment works before determining if a treatment works, but here we are. She recounts the usual nonsense – sticking a needle in the skin can release endorphins, affect immune function, and even alter brain function. I file all this under: “stuff happens”. Of course stuff happens, because you are actually sticking needles into the skin. There may even be some immediate counter-irritation type effect. But none of this explains any significant long-term effect. This is where her article gets incoherent, in the way that many pseudosciences are. They make individual arguments that do not all hold together, because they are not describing and actual underlying reality, just a made-up narrative.
She describes the clinical evidence but then cites putative mechanisms that do not explain the alleged clinical effects. She then hand waves about non-pain applications of acupuncture (asthma, fertility, addiction) and says we will have to find mechanisms for them. She buys into the false claim that sham acupuncture is not a good control but misses the implication – what the absence of superiority over sham acupuncture shows is that it does not matter where you stick the needles, or even if you stick the needles through the skin. This means, even if you think there is a worthwhile nonspecific effect here (I don’t), acupuncture points are meaningless. So then why tie a putative mechanism to specific acupuncture points? Either they matter or they don’t.
She misses the fact that these same reviews she is citing often use different points in different studies, and that acupuncturists cannot even agree on where points are or what they do. It is all tradition and personal choice – exactly as if acupuncture were a cultural belief rather than an actual reality.
She spends a lot of time touting the facia and interstitium as potential networks in the body, explaining alleged distant effects of acupuncture. But there is no evidence to support this, or that they correspond to meridians or acupuncture points. Again, she unwittingly give away the game – ” “How on earth did they understand that a point lateral in the fifth digit of the foot connected to the eye system?” asks Jeremy Pulsifer, an acupuncturist in New York City. “Modern science is finally saying, ‘We don’t know how they knew, but somehow there was an awareness that is far more sophisticated than we’ve ever given them credit for.’”
The simple answer is – they didn’t. What is more likely, that these ancient people had incredible anatomically precise (and functionally impossible) knowledge they could not possibly have had, or that this is all cultural belief? Since acupuncture points scientifically do not exist, the latter answer is preferred. She also simultaneously touts auricular acupuncture, a completely different system based entirely on ear acupoints. How could conflicting systems of acupuncture points simultaneously exist and be valid?
She fails to point out that originally acupuncture had 365 acupoints – one for each day of the year, because acupuncture was essentially astrology. It was astrology based on superstition, not some mystical hidden incredible anatomical knowledge. She also buys into the trope that acupuncture reflects ancient knowledge, when it is only a century old. The version of acupuncture that we know was retconned in the early 20th century to become part of a new Traditional Chinese Medicine, woven together from older traditions that were mutually hostile. Prior to that acupuncture used large lances and was more like bloodletting than anything else.
PNAS should be ashamed, and rightly criticized, for publishing such a blatantly biased piece promoting pseudoscience.
