Still no evidence acupuncture works

What is wrong with the reporters at the BBC? In an article today under a headline Acupuncture 'more than a placebo', they state that scientists have “proof” that acupuncture works. But the study, they way they report it, says no such thing.

Here’s what they report. Researchers used PET scans to see what was happening in the brains of 14 people during three types of treatment:

1. Blunt needles the patients knew would not pierce the skin
2. Sham (“trick”) acupuncture needles that gave the impression that the skin was being pierced although it wasn’t
3. Real acupuncture needles

They found:

Only the brain areas associated with the sensation of touch were activated when the volunteers were touched with the blunt needles.

During the trick needle treatment, an area of the brain associated with the production of natural opiates - substances that act in a non-specific way to relieve pain - were activated.

This same area was activated with the real acupuncture but, in addition, another region of the brain, the insular, was excited by the treatment.

This was a pathway known to be associated with acupuncture treatment and thought to be involved in pain modulation.

Before we go any further, a reminder of what acupuncture is. Acupuncture is the manipulation of “qi” (oddly, pronounced “chi”) by inserting needles at key points in 12 “meridians” to restore the flow of vital energy and balance yin and yang. This is supposed to restore health to the whole body. And the points are very specific: NCCAM states that the body has more than 2,000 acupuncture points; the ones you use depend on the illness you are trying to fix.

Hold on to that definition and look again at the study. It does absolutely nothing to show that acupuncture, as defined, works. All it shows is that if you stick needles in someone, a part of the brain “involved in pain modulation” is stimulated. Well whoopi-do, big surprise! I would certainly hope sticking a needle in someone would stimulate a part of the brain “involved in pain modulation”. If it didn’t you should be worried. I can think of easier ways to do this, though. For example, I would guess a good kick in the balls would also stimulate a part of the brain “involved in pain modulation”, although I doubt I would need a PET scan to prove it. This study does not even attempt to determine if anyone was cured of anything by the treatment. And remember, all that is being discussed is pain modulation: nothing about making the whole body well.

The BBC does end with:

Professor Henry McQuay, professor of pain relief at the University of Oxford and member of the Bandolier group that looks at the evidence behind different medical treatments, said: "The great bulk of the randomised controlled trials to date do not provide convincing evidence of pain relief over placebo".

(My bold.)

Precisely. So why do the BBC start by saying this is “proof” acupuncture works. It’s not even evidence it works, let alone any kind of proof.

Comments

[Mike]

Richard,

I think you're a bit off base in your analysis here. You seem to be confusing the finding of an effect related to pain "modulation" with the "sensing" of pain. Of course having a needle stuck in you will cause anyone to feel pain. But I think they're saying these needles cause the body to produce natural opiates to alleviate pain in excess of what the body might produce specifically for that needle's injury, and in other areas, such as helping with back pain or a headache.

I'm as skeptical of the existance of "qi" or accupuncture's ability to affect it as you are. But I'm not as skeptical about the possibility the body can be "tricked" into an effective pain response by needles or other means. I don't think it is manipulation of qi, just some other mechanism we don't understand yet.

Great work on the blog recently!

Mike

[Thomas Wilburn]

It's called "qi" but pronounced "chi" because it's Chinese, and the pinyin romanization system already uses "ch." Both ch and q have basically the same sound as far as the consanant goes, but one of them (I think it's the q, but my Mandarin is rusty) signals that the following vowel will be pronounced as if it had an umlaut. So "chi" and "qi" actually sound very different, and have different meanings (even before tonal differences). It's easier to explain when you can hear it.

Activation of the insular cortex also occurs with emotional content, which a needle would evoke. The release of central endorphins does not seem like much of a big deal either; endorphins do not normally rise to physiological levels that provide adequate pain relief from really painful stimuli without extreme emotional content or high levels of sympathetic arousal.

But I like Richard's idea of evaluating the insular cortex/endorphine axis using a good kick to the balls. (Probably hard to find volunteers for the study.) I bet acupuncture would fail under such conditions.

For acupuncture to work, my guess is that you need informed test subjects paired with the right experimenter (preferably, Chinese) in the right place at the right time and then sometimes it might work a little bit with a mild enough stimulus. Wait a minute; that's the placebo effect.

[Skeptico]

Comments originally posted by Mike in bold:

I think you're a bit off base in your analysis here. You seem to be confusing the finding of an effect related to pain "modulation" with the "sensing" of pain. Of course having a needle stuck in you will cause anyone to feel pain. But I think they're saying these needles cause the body to produce natural opiates to alleviate pain in excess of what the body might produce specifically for that needle's injury, and in other areas, such as helping with back pain or a headache.

I think you are confused. First, the study showed that the sham acupuncture (placebo needles) stimulated the natural opiates the same as the real needles. The article’s claim, remember, is that acupuncture is “more than a placebo”.

Second, there was no attempt to demonstrate opiates are produced “in excess of what the body might produce specifically for that needle's injury” - you are just assuming that.

Third, the effect produced by the real needles only (ie, the thing all the fuss is about), is not about opiates, it is about the insular region of the brain being excited. See this information on the parieto-insular cortex:

'Pain centres'

Second, there is growing evidence that the parieto-insular cortex is crucial for pain processing in the brain.

(Snip)

In addition, functional imaging studies in humans, and maps of electrical responses to laser-induced pain, indicate strong activation of the parieto-insular cortex during pain.

(My bold.)

So the parieto-insular cortex becomes activated during pain. This is the big discovery they are crowing about – it becomes activated when they stick needles in someone, but not when they just pretend to stick needles in. Like I said, big surprise!

I'm as skeptical of the existance of "qi" or accupuncture's ability to affect it as you are. But I'm not as skeptical about the possibility the body can be "tricked" into an effective pain response by needles or other means. I don't think it is manipulation of qi, just some other mechanism we don't understand yet.

Even if true, that’s not acupuncture.

btw thanks for the comments on the blog. Glad you like it.

[Skeptico]

Thomas:

Do I understand correctly that pinyin Romanization is the way Chinese people write their language in roman letters?

If so, how do they pronounce “chi”?

[Skeptico]

Re:

I like Richard's idea of evaluating the insular cortex/endorphine axis using a good kick to the balls.

Yes, I imagine a good kick to the balls would make you forget about your arthritis pain for a bit.

[Thomas Wilburn]

Skeptico said:

"Do I understand correctly that pinyin Romanization is the way Chinese people write their language in roman letters?

If so, how do they pronounce “chi”?"

Chi is the most natural sounding--the vowel is pronounced as in English. Qi is the "accented" version. For most intents and purposes, English-speakers can pronounce chi and qi the same, just the way it looks. Because English doesn't have tones, we butcher Mandarin anyway, so leaving the umlaut off is relatively minor. It's just good to know that although the pronunciation is almost the same, the meaning can be very different.

When actually trying to speak Mandarin, on the other hand, it's very significant, and very difficult for Americans to do correctly. I got a lot of stares from people when I was there for a month, only to be "clarified" by my professor with a word that sounded (to me) exactly the same. Talk about frustrating. It takes a lot of practice.

Pinyin is the romanization system used by the mainland Chinese, and is the official system of the PRC. Taiwan (and to some extent Hong Kong, if I remember correctly) still uses the old system, Wade-Giles. I think the new system is less confusing, personally, but neither of them is necessarily "intuitive" for non-Mandarin use. For example, the same q/ch structure also exists for j/zh and x/sh.

This isn't a big deal, really, and I wasn't taking exception with your usage. It just looked to me like there was an implied question in your parenthetical.

[Skeptico]

That's a good explanation Thomas. Thanks.

[dealga]

well done for highlighting this guff. Wrote about it myself after the same story appeared in the Observer. Fact is most people who say acupuncture works would be better off spending time with a shrink

[Castillonis]

A paper available at www.annals.org gives
some background on acupuncture and lists
numerous studies that have been published.

http://www.annals.org/cgi/reprint/136/5/374.pdf
Academia and Clinic
Ted J. Kaptchuk, OMD
"Acupuncture: Theory, Efficacy, and Practice

Kaptchuk lists post operative dental pain as
something that may be treated effectively
using acupuncture. The paper also gives
some possible acupuncture mechanisms, but
does not discuss the mechanisms in detail.
There is some constructive criticism of
the research that is cited.

1. "acupuncture analgesia may be initiated by stimulation in the muscles, of high-threshold, small-diameter nerves. ... EDITED ... Which, in turn, trigger endogenous opioids (for example endorphins and enkephalins) or stress-related hormones (for example, adrenocortiotropic hormone)" (379 Kaptchuk, OMD)

2. "... one mechanism of acupuncture may be a form of stimulation for the gene expression of neuropeptides (84,85)." (379 Kaptchuk, OMD)

3. Functional MRI acupunture point related to vision, activated by occipital lobe region ... located on lateral aspect of foot
(380 Kaptchuk, OMD)

[Grace]

Using the currently available scientific tools (100 years old at most) to investigate an ancient method of treatment (5000 years old) is never going to be enough to prove or disapprove. It's like using our human telescope to look at the universe, but how far CAN we see, and how much COULD we know???

For those skeptical about acupuncture I'd suggest they try it out themselves, especially those who are so chronically ill and nothing else works for them. Then come back here to post your comment.

Disbelieve comes from the lacking of experience and a narrow mind.

[Diathermic]

Hey Grace, Fritz Pearl said it best, “A fear of knowing is a fear of doing.”

[Skeptico]

Grace:

Re: Using the currently available scientific tools (100 years old at most) to investigate an ancient method of treatment (5000 years old) is never going to be enough to prove or disapprove.

If something has an effect it can be measured. That is all that it means to use science. (And science is more than a few hundred years old – over 2000 years old in fact: Eratosthenes calculated the circumference of the Earth in the 3rd century BC.) This comment of yours is just a lame fallacious appeal to other ways of knowing.

What tool would you use to determine if acupuncture works?

Re: It's like using our human telescope to look at the universe, but how far CAN we see, and how much COULD we know???

Yes but what tool could see further? And what tool did the ancient Chinese have that enabled them to see further than we can see now? How did they work out all this stuff that science can’t even measure now?

Re: For those skeptical about acupuncture I'd suggest they try it out themselves, especially those who are so chronically ill and nothing else works for them. Then come back here to post your comment.

A flawed method at best. Biased data selection from personal experience is not the way to arrive at what works and what doesn’t.

Re: Disbelieve comes from the lacking of experience and a narrow mind.

Another lame and fallacious argument. This time an appeal to be open minded. But are you too closed-minded to accept that maybe acupuncture doesn’t work? Because if you refuse to accept it doesn’t work, despite numerous trials that show it doesn’t, then you are the closed-minded one.

[BronzeDog]

Sorry, Grace, but I don't have god-like perception abilities. That's why an acupuncture experience wouldn't be convincing: My senses are easily tricked. I've seen it happen. I don't see how I could rule out my own mental trickery as a cause for any apparent changes if I do nothing to filter that out.

That's what double-blind placebo-controlled studies are about: They work to filter out all our human psychological tricks.

Placebo = Psychological tricks + Coincidental recovery.

Treatment = Psychological ticks + coincidental recovery + any physical effects.

Do the math: If a placebo group experiences the same amount of recovery as the treatment group, it means there's likely no physical effect. If there is a significant difference, then there likely is an effect.

[BronzeDog]

A little bit more:

Using the currently available scientific tools (100 years old at most) to investigate an ancient method of treatment (5000 years old) is never going to be enough to prove or disapprove. It's like using our human telescope to look at the universe, but how far CAN we see, and how much COULD we know???

Sounds like "How do you prove photography to a blind man?" You might want to click on that link in the "Classic Posts" section on the right. Science isn't about tools. It's about method. I fail to see how something with an observable effect couldn't be revealed in a DBCT. Perhaps someone would like to enlighten me?