Does connective tissue have any role in the acupuncture effect?

Inspired by Chang et al Front Neurosci 2019.[1]

Photo by Clint Adair on Unsplash.

This paper was published on 20th February 2019, but it did not appear on my PubMed search until last week. The title gives away the story:

Peripheral Sensory Nerve Tissue but Not Connective Tissue Is Involved in the Action of Acupuncture.

Chang et al [1]

Well I have to say that this is completely in line with my conceptual framework. But I was intrigued by how this group came to their conclusions, and I was wary of my own potential for confirmation bias – a type of cognitive bias that involves favoring information that confirms your previously existing beliefs or biases.

Confirmation bias – a type of cognitive bias that involves favoring information that confirms your previously existing beliefs or biases


In 2010, a surprisingly young women called Nanna Goldman rather rocked my conceptual world [of acupuncture] by announcing in the journal Nature Neuroscience (with her mother and colleagues from her mother’s laboratory) that acupuncture effects were mediated by release of adenosine at the site of needling.[2] Her research also showed that the effect was only present if the needle was rotated, and there was no effect contralaterally.

Needless to say, I studied her paper very carefully, and decided to invite her to speak at a meeting so we could get to the bottom of this ground breaking research finding.

Allyson asked me, of the list of authors, “Who should we invite?” I replied, “Either the first or the last author.” Allyson came back to me some days later and said, “The first and last author are related to each other, and the first author is only 17.” I nearly fell over at this point. It turns out that Nanna was a high school student doing a holiday project in her mother’s neuroscience laboratory. Unfortunately she and her mother were not able to make it over to present their research, but I met one of the group a few years later at a conference in Stockholm.

This research was high quality, and it had opened a new area of acupuncture research that seemed to tie up nicely with the work of Helene Langevin on needle effects in connective tissue.[3–6] I was happy enough that whilst this was an interesting addition to the potential mechanisms of acupuncture, it was unlikely to explain the majority of the effects we see in clinical practice. The new aspect from the perspective of WMA (Western medical acupuncture) would be the addition of a specific supportive mechanism and therefore reason for using points near nerve trucks proximal to the symptomatic area in a peripheral pain condition. Nanna Goldman’s research had shown that adenosine released by a rotating acupuncture needle reached a concentration that mediated inhibitory effects via A1 receptors on nociceptive nerves.

So back to the current paper.

Has the potential of adenosine release from connective tissue been dispatched?

Well the paper by Chang et al certainly appears to represent high quality laboratory research. They studied the effect of acupuncture in 3 separate experimental models and then applied either bupivicaine or collagenase to the needling sites prior to acupuncture to see if the observed effects were abolished. Bupivicaine would block effects via local nerves and collagenase would destroy all the connective tissue in the area.

The experimental models were:

  1. Cocaine induced locomotor activity
  2. Immobilization-induced hypertension
  3. Mustard oil-induced visceral pain.

The research team used a specially designed robotic arm to rotate the needles in the chosen points for each experimental model, having analysed how ‘expert’ acupuncturists rotate their needles. The acronym they coined was RANT (Robotic Acupuncture Needle Twister), and the device perfectly replicated the human experts.

In all three models the effect of acupuncture was abolished by prior application of bupivicaine at the site of the RANT performance. Collagenase, on the other hand, had little or no effect, hence the conclusion in the title of the paper.

But is this conclusion fully justified?

I am intrigued to know whether or not the collagenase actually affected release of adenosine at the needle site. I would be surprised if it did, since adenosine is released by cells such as fibroblasts, not from the connective tissue structure that is destroyed by the enzyme collagenase. Ok, the lack of structure would prevent traction on fibroblasts at a distance from the needle, but would be unlikely to affect more local release.

So if adenosine was released, why was there no measured effect?

This is so obvious that it clearly blindsided the editor and reviewers of the paper, and presumably also never occurred to the researchers themselves.

If adenosine works by blockade of nociception, then the outcome measure must relate in some way to peripheral nociceptive transmission, and the acupuncture or RANT must be applied proximal to the modelled condition.

So where would you have to needle to interrupt transmission in the experimental models used?

  1. Cocaine induced locomotor activity – somewhere in the brain
  2. Immobilization-induced hypertension – somewhere in the brain
  3. Mustard oil-induced visceral pain – somewhere in splanchnic or pelvic autonomic nerves.

Where did they actually perform the needling?

  1. Cocaine induced locomotor activity – HT7 (rat forelimb)
  2. Immobilization-induced hypertension – PC6 (rat forelimb)
  3. Mustard oil-induced visceral pain – BL62-4 (rat hindlimb).

Oh my! What can I say? Was it all a complete waste of time? Not entirely of course. We know that nerves mediate the effects of acupuncture in these three models, and rotation of needles in connective tissue is less important in these circumstances. This research was not designed to test the effect of adenosine release on peripheral nerves, and used models quite different from those of Nanna Goldman et al. So we cannot dismiss connective tissue and adenosine release just yet.

At further risk of confirmation bias, I must acknowledge that I never rotate my needles in clinical practice anyway!


1         Chang S, Kwon OS, Bang SK, et al. Peripheral Sensory Nerve Tissue but Not Connective Tissue Is Involved in the Action of Acupuncture. Front Neurosci 2019;13:110. doi:10.3389/fnins.2019.00110

2         Goldman N, Chen M, Fujita T, et al. Adenosine A1 receptors mediate local anti-nociceptive effects of acupuncture. Nat Neurosci 2010;13:883–8. doi:10.1038/nn.2562

3         Langevin HM, Churchill DL, Cipolla MJ. Mechanical signaling through connective tissue: a mechanism for the therapeutic effect of acupuncture. FASEB J 2001;15:2275–82. doi:10.1096/fj.01-0015hyp

4         Langevin HM, Konofagou EE, Badger GJ, et al. Tissue displacements during acupuncture using ultrasound elastography techniques. Ultrasound Med Biol 2004;30:1173–83.

5         Langevin HM, Bouffard NA, Badger GJ, et al. Dynamic fibroblast cytoskeletal response to subcutaneous tissue stretch ex vivo and in vivo. Am J Physiol Cell Physiol 2005;288:C747–56.

6         Langevin HM, Bouffard NA, Badger GJ, et al. Subcutaneous tissue fibroblast cytoskeletal remodeling induced by acupuncture: evidence for a mechanotransduction-based mechanism. J Cell Physiol 2006;207:767–74.

Declaration of interests MC