Mar 03 2010
Acupuncture for Depression
One of the basic principles of science-based medicine is that a single study rarely tells us much about any complex topic. Reliable conclusions are derived from an assessment of basic science (i.e prior probability or plausibility) and a pattern of effects across multiple clinical trials. However the mainstream media generally report each study as if it is a breakthrough or the definitive answer to the question at hand. If the many e-mails I receive asking me about such studies are representative, the general public takes a similar approach, perhaps due in part to the media coverage.
I generally do not plan to report on each study that comes out as that would be an endless and ultimately pointless exercise. But occasionally focusing on a specific study is educational, especially if that study is garnering a significant amount of media attention. And so I turn my attention this week to a recent study looking at acupuncture in major depression during pregnancy. The study concludes:
The short acupuncture protocol demonstrated symptom reduction and a response rate comparable to those observed in standard depression treatments of similar length and could be a viable treatment option for depression during pregnancy.
Plausibility
The study compared acupuncture designed specifically to treat depression, and in fact tailored to the individual patient, according to principles of Traditional Chinese Medicine (TCM). This was compared to two control groups – a control acupuncture that was not specific to depression and massage. The comparison to massage was obviously not blinded and therefore, in my opinion, of very little value as depression is highly susceptible to non-specific therapeutic effects and both interventions – acupuncture and massage – would be likely to create such non-specific effects.
The interesting aspect of this study is the comparison between treatment acupuncture (targeted for depression) and control acupuncture (not targeted for depression). The purpose of the study was to control, as much as possible, for any other variables so as to determine if the underlying TCM principles have any validity – does it matter where the needles are placed?
We can really only put this study into context if we first consider the prior probability of this claim. I would argue that there is already a large body of acupuncture research that collectively shows needle placement as a variable has no effect on clinical outcome. This one study does little to alter the balance of that evidence.
Further, from a basic science point of view, the TCM principles have essentially no plausibility. The underlying theory is that there is an undetected life force (chi) that is partly responsible for health and illness, that acupuncture needles placed in specific acupuncture points alters the flow and strength of this energy, resulting in a clinical outcome. Chi has no existence in science, however. Vitalistic philosophies such as chi were discarded over a century ago as both unnecessary and without any empirical foundation.
Any modern attempts to explain acupuncture effects with known physiological phenomena might explain non-specific needling effects, but cannot explain any differences due to needle placement, and do not provide any explanation for the location of alleged acupuncture points.
Therefore, given the extremely low prior probability of the claims of this study, nothing short of a large rigorous and replicated study would alter our assessment of validity of acupuncture as a specific intervention.
The Current Study
This new study, published in the Obstetrics and Gynecology, is not of sufficient quality to justify the conclusions of the authors. The authors did do a decent job of trying to rigorously control the comparison between the two acupuncture groups. Subjects were blinded to which group they were in, as were those evaluating the outcome. Standard depressions scales were used. They even made a reasonable attempt to blind the acupuncturists, using a novel method (to my knowledge).
They had experienced acupuncturists design a treatment and control acupuncture regimen for each subject, and then had a “junior acupuncturist” (less than two years experience) perform the treatment without being told which one they were giving.
This, in my opinion, in the crux of the methodology – were the treating acupuncturists properly blinded. The study authors took the very useful step of assessing the degree of blinding of the acupuncturists and the subjects. Unfortunately for the validity of the study, they found that the treating acupuncturists were significantly more likely to have positive expectations for the treatment group than the control group – so their blinding methods failed with respect to the treating acupuncturists. The study was therefore, at best, single blinded. Test subjects did not have any significant difference in expectations.
Because depression is so amenable to non-specific therapeutic effects, the expectations of the treating acupuncturist can plausibly have had a significant effect on the final outcomes. This is the primary weakness of the study – but there are others worth mentioning.
The author also, for some reason, did not stratify the test subject according to race, and there turned out to be significantly more African Americans in the control acupuncture group than the treatment group. Cultural beliefs can have a significant effect on responses to different kinds of placebos, particularly needles. This is therefore a potential, if unknown, confounder.
The results were also not impressive. The study used the Hamilton Rating Scale for depression:
Interpretation of Hamilton Rating Scale for Depression scores is as follows: less than 7, nondepressed; 8–13, mild depression; 14–18, moderate depression; 19–22, severe depression; more than 23, very severe depression.
At 8 weeks the control acupuncture groups has about a 9 point drop in the scale, while the treatment acupuncture group dropped 11.5 points. On this scale that is a modest clinical effect. There was also no difference in remission rates among the three groups. In addition this was a relatively small study (141 treated in total, divided among the three groups) with a 23% drop out rate.
Conclusion
Therefore we have a small and improperly blinded and randomized study showing a modest clinical effect. This does not significantly alter the low prior probability of a treatment effect from needle placement.
This study should also be considered in the context of other trials looking at acupuncture and depression. This very recent Cochrane review concluded:
We found insufficient evidence to recommend the use of acupuncture for people with depression. The results are limited by the high risk of bias in the majority of trials meeting inclusion criteria.
Specifically – there was no difference between verum acupuncture and sham acupuncture in the clinical trials reviewed.
Given the low plausibility and overall negative character of the clinical evidence, it is reasonable to conclude that no further research into acupuncture for any indication is warranted. However, acupuncture is a modality with dedicated practitioners (acupuncturists) and proponents (by contrast, for example, there is no medical specialty dedicated to a particular drug – there are no penicillinists). And therefore it is likely that further research will be conducted.
In that event, given existing research, it would be useful to conduct only highly rigorous trials, using sham and/or placebo acupuncture (where the needle or fake needle does not penetrate the skin) with adequate blinding. Such trials would need to be large with consistent replicated positive results in order to have sufficient weight to overturn the current mass of basic and clinical evidence.
![[del.icio.us]](http://images.del.icio.us/static/img/delicious.small.gif){kind=small}
Print This Post
28 Responses to “Acupuncture for Depression”
A few things to note:
1. Chi is a word often used interchangeably with ‘blood’ in both chinese and japanese. I think blood exists in science. It’s no coincidence that exercises used to balance chi are also pretty good at getting the circulation going, it’s just that in TCM the concept of Chi predated theories about a circulatory system by hundreds if not thousands of years. Lack of consolidation between eastern and western medicine is the cause of your misunderstanding.
2. Acupuncture seems to be doing better than anti-depressants. I’m surprised you failed to mention any comparison with research into the efficacy (or lack of) amongst ssris and snris.
3. Regular exercise and activity trumps any other treatment for depression in my experience, and I’ve tried them all except lithium.
John
1 – This is relevant historically as acupuncture was likely an Asian version of blood letting, unlike its more modern manifestation. But this is irrelevant to the modern use of acupuncture. For example – how would the current use, as in this study, have any plausible effect on circulation?
2 – The effectiveness of SSRIs is irrelevant to the current study or question at hand, but I do dispute that acupuncture is doing better – the review is essentially negative.
3 – “in my experience” is anecdotal.
Coming from a Korean family, I have been on the receiving end of an actual bloodletting, and my brother received acupuncture for a sore knee. No one listened to me when I said my headache would go away with a little time, or when I said that my brother’s knee is hurting because he PLAYS FOOTBALL.
Seeing additional studies over the years, there always seems to be way too much subjectivity or too many non-specific results to convince me that pricking someone with needles actually does anything at all. I understand “Eastern” medicine all too well. It’s funny how that term, when expressed in the “West”, only covers the stuff popular here…but likely using Tiger parts for aphrodesiacs or toothaches isn’t quite so appetizing.
So pregnant females like to be stimulated with needles rather than receive anti-depressants. Why not just pin each pregnant subject with a flower corsage when they come in for treatment?
Cut and paste form my comment on the cross-posted version of this over at SBM:
I didn’t read the whole study but I felt the massage group was unnecessary at best and deceptive at worst. It seems that they frequently focused on the comparison of the treatment group to the “combined control” group rather than to the control acupuncture group, which makes me think they are trying to enhance their results by dragging the control numbers down by including the totally unblinded massage group.
It seems obvious that there is very great potential for the unblinded massage group to influence the control group results up or down.
At the very least, including an unblinded “control” like massage introduces an unknown confounder. You can’t expect the placebo response for massage to be similar to control acupuncture, and you can’t expect the massage group to not have an actual therapeutic response either.
I’ve seen this before where an acupuncture study included an extra, unblinded and seemingly unneeded “control” group to their study. The last study I saw like that used the extra group to conclude that their “control” was also effective rather than their treatment was no more effective than placebo.
Just thinking about the acupuncture and it’s relation to SSRI comparison …
Receiving personal treatment, one where someone is actively engaging you both physically and mentally, is bound to affect the relative release of neurotransmitters such as serotonin. There’s a body of evidence suggesting that serotonin acts alongside other signalling molecules like dopamine for eliciting prosocial behaviors in model animals. You could probably achieve the same effect if the treatment was a full body massage twice a week for 8 weeks.
One test I’d like to see run is a comparison acupuncture, massage, and medication comparing their effects on depression. Something like starting with a sham TCM body massage; just make up a name for a known massage technique and perform it on a subject. Compare that to control group whose massage name is bland, like “full body massage” but is the same technique as the sham. You could then add in a fully double-blinded acupuncture control and treatment group like the few that have been done in the past. Finally, an SSRI group and placebo to compare the three different modes of treatment. This is just a back-of-the-napkin test that would need to be worked out better, but I think valuable data could be obtained.
Also, the effect of treatment seems small to me. Small enough that with the addition of another cohort may end up showing no significance in the end. Admittedly, I am unfamiliar with the depression testing method.
A difference in the two groups is apparent in table 2. In the specific acupuncture group there were 15.4% with a history of chronic depression, in the other two groups there were 31.2 and 28.6% respective;y.
A difference in severity that was not captured in the test they did, but which related to a history of chronic depression might explain the small differences in treatment outcome.
Steve,
1. Acupuncture has always been focused on areas of high sensitivity and easy to reach nerves, I’m not sure where you heard the bloodletting theory, but I’d be interested to know more.
2. “the review is essentially negative” – of course, ssri’s have failed to outperform placebos. The jury is still out on acupuncture, I’d say that puts it ahead, even if only in principle.
3. I’m still reaping the success of anecdotal medicine, and every case has to be prescribed based on individual experience, especially with a condition as misunderstood as depression. You can’t prescribe based on averages, you’ll end up killing someone.
It’s amazing how many people overlook this very important aspect of real life medicine – Individual case evidence trumps everything else you know.
Ahh you beat me to it! I just read over the study a bit and I noticed that too. Seems to be a bit of an oversight.
Watcher,
“Receiving personal treatment, one where someone is actively engaging you both physically and mentally, is bound to affect the relative release of neurotransmitters such as serotonin. There’s a body of evidence suggesting that serotonin acts alongside other signalling molecules like dopamine for eliciting prosocial behaviors in model animals. You could probably achieve the same effect if the treatment was a full body massage twice a week for 8 weeks.”
All of my positive experiences with alternative treatments are congruent with this theory, especially massage – it’s a no-brainer how that one can make you feel better. There’s no way to extract its active ingredient and patent it though, which keeps it out of the remit of conventional medicine.
Personal treatment experience is how you individualize a treatment to an individual – but not how you know a treatment is safe and effective in general. Personal anecdotes are misleading and it is folly to rely on them.
The literature does not show that antidepressants don’t work for depression. They clearly work for severe depression, and are controversial for mild to moderate depression (where effect sizes are by necessity smaller and are drowned out by non-specific placebo effects).
Hmm. That seems like a strawman or cop out to me. I don’t think anyone disagree’s with you when you say that one would “feel better” after treatment through massage, possibly due to brain chemistry changes. I think the problem is that acupuncture is shown to be no better than massage at treating sickness, and the data show a “no greater than placebo” effect when treatment is supposedly specific for the malady. I mean, these studies show that acupuncture (sham or control) and massage show roughly the same results with one thing in common, personal contact. If you can get the same effect from getting a cheaper rub down at the local spa rather than paying extra for an acupuncture treatment, could that be a better solution? There’s no woo/falsality involved.
Sorry it should read and the data show a “no greater than placebo” effect when treatment is supposedly specific for the malady with proper blinding and demographic distribution.
Watcher: I suppose you can get a similar experience to a spa effect if you get a massage from caring partner of some kind or another, lol…why spend even that amount of cash?
One aspect of depression, whether a symptom or cause, or both, is loss of hope. Acupuncture may offer hope to some that it will help. So for a short period of time, when hope returns, it probably does.
And then when hope retreats, we hope for more acupuncture.
True! I guess if your partner were a bad massage-giver?
Acupuncture always seems to work great for things like emotional states and non specific pain but never things like traumatic injuries, infectious disease, or cancer.
How curious.
The study produced results that the experts who ran the study determined were statistically significant and supported the conclusion they gave.
Given the actual results of the experiment, what conclusion should they publish?
What actual results would support the conclusion they gave?
@Superdave, are you saying the acupuncture only seems to be effective for conditions that can only be subjectively evaluated and that conditions with objective criteria don’t seem to benefit for it at all?
@ johnc, regarding #1, check out several excellent articles over at the Science Based Medicine web site for information on the origins of acupuncture and it supposed role in “traditional” Chinese medicine.
http://www.sciencebasedmedicine.org/?p=252
http://www.sciencebasedmedicine.org/?p=583
and the excellent length multi-part series from Kimball Atwood.
In regards to #2: Whether ssris or other antidepressants are or are not effective for any kind of depression has absolutely nothing to do with whether acupuncture is effective for depression or not; whether it’s ahead of anything else or not is also irrelevant.
johnc,
[quote]2. “the review is essentially negative” – of course, ssri’s have failed to outperform placebos. The jury is still out on acupuncture, I’d say that puts it ahead, even if only in principle.[/quote]
Would you also say that crystal therapy and prayer are “ahead” of SSRIs? It’s true that recent research has called into question the wisdom of prescribing these powerful drugs (SSRIs) for non-severe depression, but at least there is still a plausible mechanism by which they ought to work. Acupuncture essentially claims it works by undetectable magic. Since its impossible to disprove undetectable magic, you are technically correct in saying that the jury is still out on acupuncture. That being said, I doubt the jury will be coming back anytime soon.
SSRIs, on the other hand, the guidelines for their administration is already in the process of being adjusted to fit the evidence of their efficacy. I would say this puts SSRIs (and science-based medicine as a whole) ahead.
To reuse your analogy:
In the “race” to find the best medical treatment(s) for depression, SSRIs may have made a pit stop for some adjustments, but acupuncture hasn’t even shown up at the starting line.
Dammit, what is the code you all keep using to get quotation blocks?
Undetectable energy flows detected captain!! Sound the alarm.
Woowoowoowoo….
The one good thing about a poor economy is that the worried well have better things to spend their money on. A good time to hammer the woomeisters.
One dumb question:
“Obstetrics & Gynecology’s 2008 impact factor is 4.397. The journal’s ranking is the second highest impact factor out of all 61 reproductive medicine journals.” (from their website)
Why does such an article make it in such a journal?
(you get to decide what the first “such” means
)
CivilUnrest: use (blockquote) and (/blockquote), replacing the brackets with less than () symbols. If you look at the page source, you’ll see examples of it.
Steve: nothing much to add (I did try, but then realised that you’d covered almost everything I could come up with!), other than to say that it’s unsurprising that proponents of acupuncture would attempt to show efficacy in treating something like depression. It’s subjective, responsive to placebo, can be influenced by numerous confounding variables, and symptoms are often alleviated just by being proactive in dealing with it.
A journal may want to publish something controversial in order to stir up interest.
As an editorial I can understand, but not through the peer-reviewed, data submission section. It adds a legitimacy it may not deserve.
johnc,
A note regarding your first point: In Japanese, the word ‘chi’ (血) definitely means blood, but it is not the same ‘chi’ as the one referred to in Chinese medicine. That one is actually pronounced ‘ki’ (気) in Japanese, and there is no confusion between the two.
http://en.wikipedia.org/wiki/Qi
Brian
@ Superdave, are you saying that acupuncture appears to be effective only for conditions that can only be evaluated subjectively and by objective conditions do not appear to benefit from it at all?
JohnC @, with respect to # 1, check out several excellent articles on the website AltMedicine information on the origin of acupuncture and is supposed role in “traditional” Chinese medicine.
http://altmedicine.yiberkit.com/2010/04/27/acupuncture/
and duration of various parts excellent series of Kimball Atwood.
In regard to # 2: If SSRIs or other antidepressants are or are not effective for any type of depression has absolutely nothing to do with whether acupuncture is effective for depression or not, either ahead of any other thing or not is irrelevant.