Acupuncture Mechanisms My favourite papers

Acupuncture research: Acupoints are fuzzy

Figure showing acupoints ST36, PC6, and LR3. Each subfigure compares the image of a small acupoint with that of its true size as derived from Molsberger et al. 2012.
Acupoints are fuzzy fields. Results of a study by Molsberger et al. 2012 showing that the position of some commonly used acupoints, when defined by trained specialists shows considerable fuzziness.

My favourite papers

I decided to start a new category today, called “My favourite papers”. In this category I will feature papers that I believe are central to the understanding of how complementary therapies, like acupuncture, moxibustion, cupping, TENS, manual therapy, and others may work. This may include specific as well as unspecific effects, since we need to understand both of them to get the best possible treatment result.

The first paper of this new category is from 2012 and was published by Albrecht Molsberger and colleagues. Those of you who are from the CAM field may know Molsberger as one of the initators of the German Acupuncture Trials (GERAC) for chronic low back pain, a large and well-designed clinical trial that led to the reimbursement of acupuncture for the treatment of low back pain by public health insurance. However, the study also spawned a heated debate whether acupuncture is merely a placebo or not.

The GERAC trial and its aftermath

In short: The GERAC results showed that the outcome of manual acupuncture at classical acupuncture points was superior to that of conventional therapy (i.e. NSAIDs, physical therapy, and exercise). The pain reduction by acupuncture was almost twice as large as by conventional therapy.

So far so good.

The trial, however, had a placebo (i.e. supposedly inactive) condition, which consisted of inserting acupuncture needles superficially at non-classical acupuncture points. To the surprise of most practitioners this placebo (or “sham”) acupuncture proved equally effective as the real thing. Critics of acupuncture usually emphasize this lack of a specific effect to argue that acupuncture is a placebo and, thus, should not be used. Despite my scientific background, I must admit that I never understood, how one could overlook the first part of the message, namely that acupuncture (no matter if true or sham) was superior to conventional therapy. Why should we withhold from patients the therapy with the best outcome perspective, even if it was a placebo? In the worst case we would trick the patient into getting better.

Possible explanations

There are several explanations for why acupuncture in this trial was not superior to sham acupuncture. Molsberger’s paper from 2012 contains one of them. Using a very straight-forward approach, he and his colleagues show that the whole concept of acupoints is questionable. In their study they asked 71 highly trained acupuncture specialists to locate a number of acupoints on the body surface. They covered the body of a volunteer with a grid drawn on transparent adhesive film. Participants then drew a point on that grid where they located each acupoint. All marks were removed before the next participant was tested.

The results showed that despite the considerable training and experience of the acupuncturists their localization of the same acupoints differed by several millimeters to even centimeters. Since the original paper contains only a schematic figure, I made some photos myself to illustrate the results on my own body using three of the most frequently used acupoints (see above).

Now what can we learn from this? Let us first listen to the authors’ conclusions. On page 1269 they state:

The concept of small acupuncture points as taught by acupuncture schools and standard textbooks is not reflected by the clinical reality.

They propose to replace the term “acupuncture point” with “acupuncture field”. When I see all the research that is being conducted in China and around the world to prove the existence of small point-like acupoints, this is a statement I could not agree more with.

The connection with the GERAC results and similar trials is drawn on page 1267:

The lack of difference between outcomes for verum and sham acupuncture in some past RCTs, where sham points were located as little as 1 cm away from verum points, could be attributable to an overlap between the point locations of verum and sham points.

This quote nicely illustrates the immense problem of so-called placebo or sham acupuncture. If one takes the results serious, the placebo conditions of many clinical trials were no placebos after all. Of course it would be no surprise then if the clinical outcome of these conditions did not differ. A sham point that lies within an acupuncture field must be considered active.

Problem solved?

Surprisingly, few people seem to be interested in these results. The original paper has only been cited five times so far. I suppose that there are still many acupuncturists that like to stick with their notion of point-like acupoints. With some exceptions, this is still the dominant way how acupuncture is being taught today. Interestingly, the early classics of Chinese medicine never speak about the size of acupoints. As I have shown in one of my own papers following the written descriptions in some of these classics makes it rather impossible to localize any acupoint with the precision implied by modern textbooks. Rather one comes to the same conclusion as Molsberger: Acupoints are fuzzy fields.

Starting from this perspective, it is much easier to motivate the hypothesis that acupuncture is mainly stimulation of peripheral nerves. This would allow us to explain its effects without the need for esoteric concepts, using only Western anatomy and physiology instead. I will elaborate in future posts how exactly this might work. For the moment it is enough to say that the finding of acupuncture points being acupuncture fields is crucial for this interpretation. No known feature of nerves can explain, why stimulating one point of the skin should have a completely different effect than stimulating a point one centimeter away. For most parts of the body we cannot even discriminate two stimuli that are less than a centimeter apart. However, if the points are in fact fields* then their centers need to separated by several centimeters. For acupoint PC6 at the wrist (see Figure above) this would make the difference between stimulating the median or the ulnar nerve, whose stimulation may have quite different effects.

Let us hope we see more unclichéd research like this in the future.

*For the neuroscientist it may be an interesting side note that the concept of acupuncture fields is very similar to that of receptive fields.