Stick a needle in your eye?

Dec 15 2010 Published by under Medicine

I was talking to a Reuters reporter the other day and learned of a new acupuncture study out of China, this one about amblyopia, or "lazy eye".   Amblyopia is a common and interesting visual problem.  During development of the visual system if an eye is understimulated, the brain will not process data coming from that eye correctly, leading to a decrease in vision in the eye, despite the eye itself being structurally normal.   There are several types of amblyopia, the most common of which is called "anisometropic", but we'll use the simpler word "refractive".

Refractive amblyopia can be treated successfully, even when detected relatively late.  Standard treatment involves correcting "normal" visual problems (that is, making sure the child is using glasses if they need them) and encouraging use of the lazy eye, often by patching the good eye.  The authors of the current study reasoned that because it can be difficult and embarrassing for children to wear eye patches, and because they believe in acupuncture, a randomized controlled trial of acupuncture for amblyopia would be a good idea.

It's actually a spectacularly bad idea, but you shouldn't take my word for it.  Let's examine the study.


Let me first reassure you that the techniques used did not involve inserting needles into children's eyeballs.  But that's about all the reassurance I can give.  This study fails the initial "sniff test" for plausibility.  There is no anatomic or physiologic reason that sticking needles into a child should affect their amblyopia (although I suppose if the needles were sufficiently painful and placed only in the visual field of the lazy eye...).

This fact in and of itself essentially invalidates the study.  But let's just suppose for a moment that there was some valid reason for thinking that acupuncture might work.

Not-So Blind, Not So Random

This particular study lends itself to far too many puns, but one of the first and most obvious problems with this study is its lack of blinding, that is, every researcher and patient knows what treatment is being given.  Also, subjects (or victims?) of this experiment were recruited from eye clinics.  They were not, for example, chosen serially (at least as reported).  This can introduce bias into the experiment.  For example, parents responding to recruitment announcements might be more predisposed to "believe in" acupuncture, and therefore be more responsive to placebo effects.

All subjects of the experiment were wearing the usual corrective lenses for several months on entry to the study.  They were measured for new glasses if needed, and then randomized to either eye patching or acupuncture.  Both sets of patients continued to use corrective lenses.  The control (patching) group was instructed to use the patch for two hours a day and do near-vision activities.  This is reasonably standard, although longer periods are often used.

The test group was subjected to acupuncture an also instructed to do near-vision activity, but without the patch.

Results--Look Closely

What the researches found was that the groups had similar improvements in vision, or in study-speak, that acupuncture was non-inferior to patching.  They interpret this as meaning that acupuncture may be a nice alternative to patching for similar kids with lazy eye.   Is that what the results really say?

Well, no.  In studies of amblyopia, especially in older children as in this study, many (about a quarter) improve significantly with eyeglasses alone.  Children in both groups received this standard therapy.  Also, although two hours of patching has in some studies been found to be as good as longer periods, this is the bare minimum of patching time that is used.   This study did not include longer (and perhaps more typical) patching times.

This is one of the serious flaws in this study and any conclusions drawn from it---this study says nothing about acupuncture, because there was no adequate control group.   It might have been interesting to compare the two groups to a spectacles-only group---perhaps both acupuncture and 2-hour eye patching were equally ineffective, rather than equally effective.

But of course the fatal flaw is the one stated by the authors:

Although the treatment effect of acupuncture appears promising, the mechanism underlying its success as a treatment for amblyopia remains unclear.

The underlying mechanism is unclear because there is none.  No effect has been shown, and if it were, given the biologic implausibility, almost any explanation would be preferred to "the acupuncture did it".

While a better-designed study might be intellectually interesting, it would be ultimately useless and immediately unethical.  Acupuncture is more invasive than standard therapy and there is no reason to think it has any real advantages.  I find little comfort in the fact that the study was approved by an institutional review board and published in a major peer-reviewed journal.


Zhao, J., Lam, D., Chen, L., Wang, Y., Zheng, C., Lin, Q., Rao, S., Fan, D., Zhang, M., Leung, P., & Ritch, R. (2010). Randomized Controlled Trial of Patching vs Acupuncture for Anisometropic Amblyopia in Children Aged 7 to 12 Years Archives of Ophthalmology, 128 (12), 1510-1517 DOI: 10.1001/archophthalmol.2010.306

11 responses so far

  • WMDKitty says:

    I stopped when I hit the words "needle" and "eye" in the same sentence. I don't care what the results of the study were, needles do NOT belong in or near the eyes.

    • JJM says:

      The needle locations are depicted in the article, none is in the eye.

      This is also discussed at SBM where I commented:

      I wondered why the main outcome measure was taken at 15 weeks since they had data going to 25 weeks. According to the article (penultimate paragraph on p.5 of the PDF). “By 25 weeks, the mean BSCVA was 0.28 logMAR in the patching group and 0.22 logMAR in the acupuncture group (P=.06). The resolved rate was similar between groups (30% in the patching group vs 42.1% in the acupuncture group, P=.27).”

      Also: (“Comment” penultimate paragraph on p. 6) “By 15 weeks, the proportion of responders in both groups was similar (75.6% in the acupuncture group vs 66.7% in the patching group; P=.37), whereas the resolution rate was significantly higher in the acupuncture group (41.5%) than in the patching group (16.7%) (P=.01). However, the resolution rates were similar by 25 weeks.”

      Isn’t the longest-term result the most significant? Doesn’t it show no significant difference in the long run?

  • Kevin says:

    Question about medical ethics:

    Imagine a scenario in which a skeptical doctor does a real, well-controlled study on the effects of acupuncture. He doesn't have any expectation that it will work, which you say makes it unethical, but what if it could be used to convince people about the inefficacy (I'm not saying that last part is realistic, but just hypothetically)?

    I ask because there's a similar conundrum with the vaccine debate. Critics say there's never any double-blind studies on vaccines, but that's because no ethical doctor would ever do an experiment that deprived kids of vaccinations. If there was a chance that this sort of experiment would convince some parents that vaccines are safe, and thus over time lead to more kids getting vaccinated, could you make the case that it's ethical?

    • Chris says:

      No. Because you will always have the hard core who will move the goal posts, and actually distort the results. Especially since they start out with the false statement that there were "never any double-blind studies on vaccines." In fact there were in the 20th century, usually using children in Africa or disabled American children in institutions (like Willowbrook, where children were intentionally infected with hepatitis). After the conditions of the institutions and information about the use of children were revealed in the 1970s, several ethics rules were devised to prevent those kinds of studies. Though there may be some still being done in developing countries.

      Having had children enrolled in studies I know that there are special precautions to protect their health (and personal information). More information here.

      Also, when I have asked the proponents of the vax/unvax study how they would design the study to protect the placebo arm from measles, mumps, pertussis, Hib and any other vaccine preventable disease that seem to keep coming back, they come back with this answer: Just use the kids who parents have chosen not to vaccinate. But that presents two problems, first it would not be a random control test and second that is exactly what the several large epidemiological studies done in several countries were!

    • PalMD says:

      Oy. Ok, let's tackle these questions:
      He doesn’t have any expectation that it will work, which you say makes it unethical

      That is very different from what I actually meant. What I may not have expressed well is that subjecting humans to a study of an intervention that is prima facie wildly implausible. In this case, there is no reason to expect that futzing around with unknowable meridians would do anything, and there is a large literature on acupuncture's failure to beat "sham" acupuncture in other situations.

      I specifically find it unethical to test already proved hypotheses such as the evidence of vaccines causing autism, since there is no set of data that will ever be accepted by the goalpost-shoving antivax crowd. The benefits are so well-documented that any study like that would be unethical. The "convince me by doing it again" argument never ends.

  • daedalus2u says:

    I missed this when I first saw the abstract. These children were subjected to "5 sessions of acupuncture weekly." Does that mean they got 5 sessions weekly for 15 weeks, or 75 acupuncture sessions? Or was it 25 weeks for 125 sessions?

    That is probably a pretty good incentive to get better, so you can stop being poked with needles 5 times a week.

  • Quackwatcher says:

    The problem with relying on 'plausibility' is that almost nothing in medicine is plausible until it is found to be so. Who would have thought giving cowpox to someone could make them immune from smallpox? No one, until it was determined to be so. Who would have thought eating citrus fruit could prevent the scourge of scurvy? No one, until it was determined to be so. If you are only going to look at what your limited mind thinks is 'plausible', you will never make any interesting discoveries.

  • PalMD says:

    That is a rather silly reductionist piece of narishkeit.

    Of course there is a way to have some ideas of plausibility (and pre-test probability, but that's another story).

    If someone tells me that holding a fridge magnet over my wrist will cure herpes, I don't need to study the question to see if it's "plausible".

    The answer to your "who would've thunk it" with cowpox is lots and lots of people, including Edward Jenner, people who worked from careful observation and testing. The plausibility in that case arose from serial observations rather than a knowledge of memory B cells.

  • Quackwatcher says:

    Do you wash your hands between patients? If so, why? It was once thought implausible that it could prevent spreading of infection. Its too bad they don't teach medical history in medical school. Those who fail to learn from the mistakes of their predecessors are destined to repeat them.

  • daedalus2u says:

    Quackwatcher, you completely misunderstand what “prior plausibility” means in the context of SBM. What prior plausibility means is that there is no data inconsistent with the hypothesis. You are confusing your belief that something is plausible with the technical term of “prior plausibility”.

    Every correct idea always has high prior plausibility because there is never data that is inconsistent with correct ideas. There may be an absence of data, there may be mistaken data, but there is never reliable data that is inconsistent.

    In the context of this example, the cause of amblyopia is known, it derives from insufficient fidelity of mapping between the retina and the visual cortex that decodes the signals from the retina into meaningful visual images. The only way to “fix” amblyopia is to allow normal neurodevelopment to improve the fidelity of the mapping through neuronal remodeling. Fortunately the brain in young children still has a great deal of plasticity and self-regulates itself so as to produce the correct mapping. All that needs to be done is to improve the contrast of the visual signals (via correction of refractive errors), in some cases reduce the signal from the dominant eye via patching, and provide visual stimulation that the brain will use as data to remodel the visual cortex and other neural structures.

    For acupuncture to improve this remodeling, there needs to be physiology by which stimulation of the appropriate acupuncture points triggers the physiology of neuronal remodeling. There is no neuronal connection between the retina, the acupuncture stimulation point and the visual cortex.

    The formation of connections between the retina to the visual cortex is automatic. It derives from deep evolutionary time, every organism with vision has essentially the same process. Signals from the retina trigger neuronal remodeling such that meaningful images are decoded from the data stream. Most of the details of that process remain unknown. It is known that there is no evolutionary reason for a separate “trigger” point that could be stimulated by an acupuncture needle to evolve. Ancestor organisms hundreds of millions of years ago didn't evolve with needles sticking in them to trigger mapping of retina onto visual cortex, so such a separate trigger point could only interfere with normal neuronal remodeling.

    The neuronal remodeling needs to be local, that is local from individual cells in the retina to individual cells in the visual cortex. Each individual cell in the retina needs to form its own individual connections in the visual cortex, and those individual connections need to be modified by neuronal remodeling such that they all work together “in sync”. The “treatment” of applying appropriate visual signals is not a “simple” or unitary treatment. The visual signals stimulate the entire retina, one cell at a time, which then stimulates the entire visual cortex, one cell at a time, and this stimulation is time varying. It is the time varying and differential signaling from the retina that triggers the appropriate remodeling. These changes involve at least billions of cells and at least hundreds of billions of connections. This is well known.

    What does acupuncture do? It triggers one spot with no connection to either the retina or the visual cortex with one type of stimulation for brief periods of time. How is is plausible that such a simplistic and non-physiologically relevant treatment could improve the reliability of such a complicated process?

    Is there any animal data suggesting such a thing? Is there any known physiology that suggests such a thing? Is there any reason, other than the wishful, magical thinking of those doing the acupuncture that such a thing is plausible? No, there is not.

    Your argument that before people knew about the germ theory of disease that hand washing was an implausible intervention is disingenuous. Even in the pre-germ theory world, there was no data that suggested that hand washing would not reduce disease transmission. People may have thought it implausible because they “knew” that disease was caused by bad air, but there was no data that contradicted the idea that hand washing would reduce transmission of infections. The reason there was no data was because in fact hand washing does reduce transmission of infections. In any case, we are not in a pre-germ theory of disease world, or at least those who practice SBM are not.

    That your understanding of physiology is insufficient to tell that acupuncture for amblyopia is implausible simply tells me that you understanding of the physiology of amblyopia and of acupuncture is insufficient. You would do well to rectify that insufficiency before you go believing in things you don't understand.