Nearly a month ago, I expressed my dismay and displeasure at the infiltration fo quackademic medicine into what is arguably the premier medical journal in the world, The New England Journal of Medicine (NEJM) in the form of a highly credulous review on the use of acupuncture for low back pain that brought eternal shame on the hallowed pages of a once-great journal. As Mark Crislip put it, trust, once damaged or lost, is very hard to restore, and I definitely lost a lot of trust for the NEJM compared to what I had for it a month ago. Since then, I’ve been keeping my eyes out for other examples of quackademic medicine infiltrating various peer-reviewed journals. Unfortunately, I have not lacked for examples.
Fresh in my memory, though, remains the stench of quackademic medicine in the NEJM. Unfortunately, just as the faint odor of urine can be revived by hot and humid weather, the stench of quackademic medicine at the NEJM was restored, at least in my memory, by an article that a reader sent to me. Apparently someone was interested in just who some of the writers of the NEJM article were. At the time, when I wrote about the article, I didn’t really care too much about any of the authros other than the lead author, Dr. Brian Berman. Apparently this reader thought I should check out at least one of the other authors, which I proceeded to do. I happened to pick Dr. Helene M. Langevin of the University of Vermont. (Who knew Vermont had quackademic medicine?) One thing led to another (PubMed, specifically), and suddenly I found this gem of a quackademic medicine article by Dr. Langevin in, of all places PLoS ONE, entitled Electrical Impedance of Acupuncture Meridians: The Relevance of Subcutaneous Collagenous Bands.
Senior author? Dr. Langevin.
Funding source? What do you think? The National Center for Complementary and Alternative Medicine (NCCAM), of course!
I feel a little bit bad about having to slather a bit of the ol’ not-so-Respectful Insolence on this bit of quackademic medicine. The reason is that I know and respect Bora Zikovic, who’s the Online Community Manager at PLoS focusing on PLoS ONE. Unfortunately (or fortunately, depending on your point of view), a man’s got to do what a man’s got to do. (Or, more appropriately, a clear plastic box of blinking multicolored lights has got to do what a clear plastic box of blinking multicolored lights has got to do.) There’s also another reason to take this on, and that’s because of how PLoS ONE bills itelf.
PLoS ONE bills itself as a new model for a scientific journal. The idea is that the peer reviewers don’t evaluate the significance of the work being presented; they leave that to their readers, publishing 69% of all submissions. PLoS ONE does, however, purport to have a system of rigorous quality checks and peer review to verify that the actual science is well done, with the following requirements:
- The study presents the results of primary scientific research.
- Results reported have not been published elsewhere.
- Experiments, statistics, and other analyses are performed to a high technical standard and are described in sufficient detail.
- Conclusions are presented in an appropriate fashion and are supported by the data.
- The article is presented in an intelligible fashion and is written in standard English.
- The research meets all applicable standards for the ethics of experimentation and research integrity.
- The article adheres to appropriate reporting guidelines (e.g., CONSORT, MIAME, STROBE, EQUATOR) and community standards for data availability.
I submit that this article bit of quackademic medicine represents a rather massive failure of PLoS ONE to live up to its ideals, and I’ll explain why.
Let’s take a look at the study under consideration. I knew right away that there was a serious problem right from the very first sentence of the abstract:
The scientific basis for acupuncture meridians is unknown.
Well, I suppose that’s true enough in a very trivial fashion, you know, the same sort of fashion that the scientific basis for homeopathy is “unknown.”
And then there’s first paragraph of the introduction to the paper:
To this day, the fundamental tenets of acupuncture, the acupuncture point and meridian, remain a mystery. What are they anatomically and how do they function physiologically? Past studies have anatomically linked these Traditional Chinese anatomical structures to neurovascular bundles , , , trigger points , , , and connective tissue fascial planes . Other studies have identified functional correlates including reduced electrical impedance , ,  and enhanced migration of nuclear tracers , , , , , . However, due to multiple study design limitations – including inadequate descriptions of acupuncture point/meridian localization, small sample size, and unexplained statistical analysis – a definitive conclusion regarding the scientific basis of these structures is difficult to establish.
A better way to put it would be that there is no known physiological or anatomic structure that corresponds to acupuncture meridians, which brings up the issue of how it was supposedly possible for practitioners of traditional Chinese medicine to have identified these meridians in the first place. Science can’t tell any difference; they don’t really correspond to any nerves; and there’s no criteria by which a meridian can be distinguished from a non-meridian other than drawings on mannequins. It’s magic. It’s fairy dust. In fact, as I’ve described many times before, meridians are irrelevant to acupunture. It doesn’t matter if you stick the needles in a meridian or not, which meridian, or even if you stick the needles in. Virtually any old place will do. Given that, one wonders why Dr. Langevin even bothered to try to undertake a study. But undertake it she did, trying to use a combination of ultrasound and impedance measurements. It’s a tour de force of what Harriet Hall so famously dubbed Tooth Fairy science and Dr. RW termed quackademic medicine using the following rationale:
Given these two reported associations [allegedly between acupunture meridians and decreased electrical impedance and between decreased impedance and increased intramuscular connective tissue], we hypothesized in a previous study that intermuscular connective tissue was the anatomical basis for the reduced electrical impedance reportedly observed at acupuncture meridians. We tested this hypothesis and found that electrical impedance at Pericardium meridian-associated connective tissue was significantly reduced compared to an adjacent muscle control . The Spleen channel segment, on the other hand, showed no statistical difference. This lack of difference was attributed to unintended placement of Spleen-control needles in an adjacent intermuscular plane.
Whoa. Did Dr. Langevin actually find a physiological basis for acupuncture points, or is she measuring how much money the Tooth Fairy is leaving behind? Has she found an objective way to demonstrate that there is an anatomic difference between the tissue under meridians and the tissue elsewhere that could produce a plausible biological mechanism by which acupuncture “works”? Or is she simply measuring which sorts of teeth bring a larger haul of cash from the Tooth Fairy?
Let Harriet apply the general beat down to Tooth Fairy Science:
You could measure how much money the Tooth Fairy leaves under the pillow, whether she leaves more cash for the first or last tooth, whether the payoff is greater if you leave the tooth in a plastic baggie versus wrapped in Kleenex. You can get all kinds of good data that is reproducible and statistically significant. Yes, you have learned something. But you haven’t learned what you think you’ve learned, because you haven’t bothered to establish whether the Tooth Fairy really exists.*
In other words, carefully applied scientific methodology used to study fairy dust produces results that are still fairy dust. Prove the phenomenon exists and that acupuncture works, before you go to all these contortions to try to relate it to subcutaneous connective tissue and changes in impedance.
Sigh. Let’s see what Langevin and her team did:
To avoid confounding by an adjacent connective tissue plane and to see whether the findings from the Pericardium (PC) channel were generalizable to other body sites, we assessed the electrical impedance of skin and underlying subcutaneous connective tissue at the Large Intestine (upper arm), Liver (thigh), and Bladder meridians (calf). These sites were originally chosen because they represented a good balance of anatomical locations and meridian types (2 Yang and 1 Yin channel), and the meridians were not located close to another meridian or intermuscular tissue plane. Acupuncturists determined the location of meridian sites, and ultrasound images were obtained at each test site to record any potential structural associations with electrical impedance. There were two primary aims for this study: (1) to determine whether the electrical impedances at acupuncture meridians were significantly lower than impedances at adjacent controls, and (2) to assess whether echogenic collagen was significantly associated with electrical impedance obtained at the test sites.
Basically, Langevin’s team signed up 28 subjects (19 male, 8 female). Exclusion criteria included: age under 18 years old; pregancy; anticoagulant usage; history of a bleeding disorder; implanted ventricular defibrillator; chronic skin conditioins; chronic skin inflammation (exczema, psoriasis, for example), or a collagen disorder. They also excluded obese people with a BMI greater than 30. Meridians were identified by two different acupuncturists and in the case of any disagreements the two came to a consensus. According to the methods section, the acupunturists had an average of seven years of experience, and each represented a different acupuncture style, one Chinese, one Japanese.
Now here’s where it gets odd. The acupuncturists originally intended to insert the needles into the intermuscular connective tissue but had a hard time getting the needles in there. That’s some pretty deep insertion, which makes me wonder what relevance this even had for acupuncture, I have no idea. Neither did the acupuncturists, either, it would appear, because there’s a most telling passage in the methods:
Furthermore, the intermuscular trajectory mapped out by imaging did not consistently match the trajectory mapped out by the acupuncturists. This was particularly true for the intermuscular plane near the LV meridian where the Sartorius muscle runs obliquely along the leg and across the LV meridian path. For these reasons, we decided to focus on the skin and underlying subcutaneous connective tissue in all our subjects. In other words, impedance measurements were limited to the subcutaneous fat region above muscle.
So basically, our intrepid team of researchers decided to give up on trying to measure impedance in intramuscular collagen bands and stick to the skin and subcutaneous tissue. They even show in Figure 2 ultrasound images showing muscles with thin bands of connective tissue between them. Of course, any orthopedic surgeon could have told them that they’d have a hell of a time hitting those bands of intramuscular connective tissue. Heck, if they had asked me, I could have told them that they’d have a lot of trouble hitting those bands, even with ultrasound guidance, which, by the way, acupuncturists don’t normally have or use in deciding where to place the needles.
So, as a sloppy second, our intrepid band of acupuncturists decided to look at impedance measurements in the superficial perimuscular fascia; i.e., the band of connective tissue that surrounds each skeletal muscle, comparing impedance measurements for meridian versus control. Then, they in essence went anomaly hunting. What do I mean by that? Basically, the investigators used all sorts of statistical models looking for associations between differences in impedance and meridians and correlations between differences in impedance and ultrasound-measured tissue density. Of course one thing that must be noted is that it would be completely unremarkable to find correlations between ultrasound-measured echogenicity (the ability to reflect sound) and differences in impedance. After all, if there’s one thing ultrasound is good at, it’s measuring differences in water content, which, along with various other mechanical properties of of the tissue, determine differences in echogenicity between tissues. Consequently it would be utterly trivial to find correlations between tissue impedance and echogenicity, not at all surprising. It would also not be particularly surprising if differences were not found, because there could well be too much variation in tissue impedances to produce a statistically significant results. Either finding would not be “evidence” for acupuncture meridians.
It’s also rather remarkable that the investigators felt the need to use such sophisticated statistical methodology to look for differences in impedance between meridian and non-meridian segments and for correlations between ultrasound-measured echogenicity and differences in impedance. Let’s put it this way. if there were a clear-cut difference in impedance between the meridian and non-meridian segments tested, the investigators wouldn’t need to do all those analyses. Doing something as simple as a t-test or one-way ANOVA will usually suffice to show the difference. Even if the statistics were appropriate for the data, again, what we appear to have here is an example of the Texas sharpshooter fallacy. All those mixed model statistical methods are modeling how fast the Tooth Fairy can complete her rounds of harvesting teeth and leaving goodies under the pillows of little boys and girls.
So what did the authors find? Let’s take a look. First, they purport to have found a small difference in impedance between the Large Intestine meridian impedance and the control (345±15 Ω versus 355±15 Ω, p=0.021, at 10 kHz; 432±23 Ω versus 449±23 Ω, p=0.017 at 1 kHz). EVerything else was negative. Does this mean anything? Who knows? Does it validate the existence of the Large Intestine meridian? No, but nice try. All it says is that there might be a difference in impedance between one area on the upper arm and another area. Worse, the observers aren’t blinded at all. Without blinding, there’s the possibility of subtle bias creeping into the measurements and sampling. Would it have been so hard to have the acupuncturists place the needles and then leave the room, after which the technicians doing the measurements come into the room to do the impedance and ultrasound measurements?
Finally, the investigators undertook a multivariate analysis to see if they could identify any factors that explained the differences in impedance other than meridians. They looked at subcutaneous tissue thickness, dermal zone echogenicity, subcutaneous zone echogenicity, and perimuscular zone echogenicity. The result are the graphs in Figure 4, which are what I call star charts. About the best that all these multiple comparisons could come up with was the earth-shattering observation that percent echogenicity in the subcutaneous zone correlated with differences in impedance. Personally, I’m surprised that more measurements didn’t correlate with changes in impedance.
If there’s one thing about complementary and alternative medicine (CAM), it’s that its adherents do more stretching and twisting to try to contort data into proving that there’s something to fairy dust. This is a perfect example. It’s in essence a fishing expedition to find any differences, no matter how small or irrelevant, between acupuncture meridian points and tissue somewhere nearby. Investigators find one difference and use the sharpshooter fallacy to declare it evidence supporting the existence of acupuncture meridians. One thing I always ask myself when I see a study like this is: Why these three acupuncture meridians? Why not others? No real explanation is given. Or maybe the investigators tried other acupuncture meridians and they didn’t work. That certainly wouldn’t surprise me.
Sadly, what doesn’t surprise me either is the funding source of Dr. Langevin’s work: NCCAM. In fact, she has multiple NCCAM grants, two of which are R01s:
- 2008-2013: National Center for Complementary and Alternative Medicine RO1 (AT 001121) “Connective tissue mechanotransduction”. Principal Investigator: Helene Langevin, MD
- 2007-2011: National Center for Complementary and Alternative Medicine (RO1) Soft tissue biomechanical behavior during acupuncture in low back pain Principal Investigator: H. Langevin
Here’s the abstract from Dr. Langevin’s acupuncture R01 (5R01AT003479):
The goal of this proposal is to test the overall hypothesis that subjects with chronic or recurrent low back pain (LBP) have an abnormal soft tissue response to acupuncture needling, and that this altered response is associated with abnormal perimuscular connective tissue structure and biomechanics. We will also test whether or not altered needling responses in LBP are generalized, or localized to specific locations as predicted by traditional acupuncture theory. 80 LBP and 80 No-LBP subjects will undergo one testing session using computerized robotic acupuncture needling and ultrasound elastography. Needle torque, force and tissue displacement patterns will be compared in LBP vs. No-LBP at Meridians vs. Non-Meridians and Acupuncture Points vs. Non-Acupuncture Points in the back and leg (Aim 1). We also will evaluate soft tissue structural and biomechanical characteristics in the low back by measuring perimuscular connective tissue thickness, ultrasound signal to noise ratio, structural continuity and biomechanical parameters (stiffness and damping) (Aim 2). Accomplishing the Aims of this study will 1) advance our understanding of acupuncture mechanisms 2) provide a first step toward investigating a new dynamic pathophysiological model for LBP incorporating connective tissue and neuroplasticity and 3) provide objective outcome measures of connective tissue structure and biomechanics that can be used in future clinical trials of acupuncture and other therapies for LBP. Relevance: Despite its widespread use for the treatment low back pain (LBP), the mechanism by which acupuncture may promote healing in this condition remains largerly unknown. This lack of understanding constrains the development of improved treatments and is an obstacle to the integration of acupuncture into the mainstream management of LBP. Moreover, the mechanism of LBP itself is poorly understood, further impairing efforts to investigate treatment mechanisms. This study will provide new, objective measurements of phenomena fundamental to both acupuncture and LBP. Demonstrating that acupuncture needling responses are abnormal in LBP, and understanding why these responses are abnormal will bring us one step closer to understanding 1) why LBP occurs and 2) how acupuncture can promote healing in this common and disabling condition.
Fairy dust indeed. Your tax dollars at work, too, $361,600 combined direct and indirect costs per year, for a grand total of $1.8 million over five years. Too bad that PLoS ONE decided to provide an outlet. Perhaps its readers should head on over to the offending article and provide the sober and evidence-based scientific criticism that the peer reviewers so evidently neglected to do when the manuscript for this paper was submitted.