In the past, I have criticised science journalists for not providing enough background in their reports. Both news stories and scientific papers obviously focus on new events and achievements, but they do so in the knowledge that new discoveries stand on giant shoulders. For this reason, when I cover new papers for this blog, I try to describe some of the research that led up to it, a tactic that fits with the growing cries for more context in modern journalism.

And yet, it’s perhaps churlish to expect this to be a routine part of science journalism when many scientists themselves don’t take up the practice. I bring this up in the light of a new paper, published today in Nature Neuroscience, about the controversial topic of acupuncture. I was going to do this as a straight write-up but actually the omissions in the paper are probably just as interesting than the science within it.

The gist is this: Nanna Goldman from the University of Rochester Medical Center claims to have found a biological explanation for the pain-relieving effects of acupuncture. She worked with mice that had inflamed paws, and managed to alleviate their pain by using a needle to pierce a traditional acupuncture point near the knee. This painkilling effect only happened when she rotated the needles after insertion.

This effect depended on a chemical called adenosine, which typically surges in concentration after any stress or injury. Adenosine works by docking at a protein called the adenosine A1 receptor, which has well established roles in suppressing pain and is found on neurons that transmit pain signals. Indeed, other chemicals that stimulated this protein had the same pain-relieving effects as acupuncture. Drugs that prevent the body from breaking down adenosine led to even more potent pain relief. And mice that lacked the A1 receptor altogether experienced no pain relief from the needles.

Taken on its own merits, this is a nice piece of biochemistry. But what does it really tell us about acupuncture? Does it actually validate this ancient method as a way of relieving pain? After reading the paper, you might walk away with that idea that we’re one step closer to understanding how a treatment with real medical benefits really works. It’s littered with statements like “A1 receptor activation is both necessary and sufficient for the clinical benefits of acupuncture” and “medications that interfere with A1 recep­tors or adenosine metabolism may improve the clinical benefit of acupuncture”. In the study’s press release, lead scientist Maiken Nedergaard even says, “The new findings add to the scientific heft underlying acupuncture.”

But these results have to be considered in the light of those that came before it. As mentioned above, new scientific discoveries stand on the shoulders of giants and in the case of acupuncture – one of the most well-researched of all “alternative therapies” – those shoulders are particularly large.

Many trials have demonstrated that acupuncture does have some pain-relieving effects – that is not in doubt. And as Steven Novella notes, unlike things like homeopathy or reiki, with acupuncture “something physical is actually happening… so it is therefore not impossible that a physiological response is happening”. But the big questions are whether this effect is genuine of nothing more than a placebo.

To answer that, clinical trials have used sophisticated methods, including “sham needles”, where the needle’s point retracts back into the shaft like the blade of a movie knife. It never breaks the skin, but patients can’t tell the difference from a real, penetrating needle. Last year, one such trial (which was widely misreported) found that acupuncture does help to relieve chronic back pain and outperformed “usual care”. However, it didn’t matter whether the needles actually pierce the skin, because sham needles were just as effective. Nor did it matter where the needles were placed, contrary to what acupuncturists would have us believe.

Other trials have found similar results. Going beyond individual studies and looking at all of the available evidence doesn’t much change the verdict. Last year, scientists from the Nordic Cochrane review centre did an analysis of the available evidence and after considering the 13 trials that met their stringent quality criteria, they concluded:

“A small analgesic effect of acupuncture was found, which seems to lack clinical relevance and cannot be clearly distinguished from bias… Our findings question both the traditional foundation of acupuncture, which is based on the existence of meridians and Qi sensations, and the prevailing hypothesis that acupuncture has an important effect on pain in general. If this hypothesis is wrong, and our results point to that, then acupuncture would seem to be unlikely to have an effect on pain related only to certain conditions, but further studies may examine this question.”

Goldman doesn’t really address any of these points. The introduction to her paper focuses on acupuncture’s popularity, mentioning endorsements by the WHO, the NIH and, strangely, the US Internal Revenue Service. As to the bigger question about whether it is effective, there is no sign of the recent trials or reviews that I mentioned. Instead, she briefly says that the pain-killing effects of acupuncture are “well-documented” and that “Western medicine has treated acupuncture with considerable skepticism”, citing only an editorial published in 1972.

As I’ve said, this is not an area that’s lacking in earlier research to refer to or consider. The discussion is a bit better in that it at least references one trial which showed that acupuncture has no advantage over placebo sham-needle treatments. And if anything, the results seem entirely consistent with the idea of acupuncture as an elaborate placebo.

The tissue damage inflicted by the rotating needle triggers a local flood of adenosine. If the needle is stuck in the right general area, the extra adenosine reaches the receptors on the pain-transmitting neurons and shuts down their activity. There is no need to invoke ‘qi’ flowing through ‘meridians’. Indeed, all sorts of injuries and stresses will lead to a burst of adenosine. And Goldman even says that sham needles, by stimulating but not breaking the skin, could still trigger a burst of adenosine, leading to the same pain-killing effects.

There has been so much previous work in this area that the question “How does acupuncture work?” is better replaced by “Why are acupuncture’s effects largely indistinguishable from those of sham treatments?” The new study suggests some answers but it seems unfortunate to me that Goldman didn’t include any sham-needle controls in her experiments.

Brian Berman, who has been involved in previous Cochrane reviews of acupuncture, agrees. He described the study a “very interesting” but said that “some sort of a placebo control is needed”. Edzard Ernst, former professor of complementary medicine at the University of Exeter, has written extensively about acupuncture also concurs. He told me, “It’s an interesting study but it proves nothing. We need independent replication, better controls and studies in humans.”

This is the most frustrating part of what could have been a really fascinating study. Without building on the massive amount of work that’s already been done on acupuncture, it is hard to know what to make the new and admittedly interesting results. I also wonder whether your average health journalist will know how this study fits into the bigger picture – whether it vindicates the use of acupuncture or whether it actually fits with a skeptical stance. But I suspect we won’t have to wait long to find out.

PS: The paper notes that the authors have no competing financial interests that might have affected their work. However, it is worth noting that one of the co-authors, Jurgen Schnermann, is married to one Josephine Briggs. Briggs is the director of the National Center for Complementary and Alternative Medicine, an institute that has had its fair share of controversy in the past.

Update: Yeah, the mainstream media aren’t really covering themselves in glory here. See Stuff and Nonsense for a summary. The Times probably has the best piece, in that it actually mentions previous trial data and has some great commentary from Edzard Ernst.

Reference: Nature Neuroscience http://dx.doi.org/10.1038/nn.2562

More on pain:

• Racial bias weakens our ability to feel someone else’s pain
• When pain is pleasant
• The placebo effect affects pain signalling in the spine
• Doctors repress their responses to their patients’ pain
• Thinking about money soothes sting of social rejection and physical pain
• Itch-specific neurons discovered in mice
• Pain in the eye of the beholder