Having exhausted myself for the time being on two things that irritate me a lot (namely creationist neurosurgeon Dr. Michael Egnor and the antivaccination pseudoscience being presented as "evidence" that vaccines cause autism at the Autism Omnibus), it's time for a change of pace. For all my tendency to deride certain "alternative medicine" modalities as pseudoscientific nonsense (homeopathy, anyone?), you may have noticed that I tend to take a softer line with acupuncture. No, it's not because I'm a believer. Certainly, I don't buy for a minute that somehow sticking needles in "meridians" in such a way as to "unblock the flow" of qi (life energy) does anything for serious diseases. More specifically, if acupuncture has any real effects on pain or other medical conditions, it doesn't do it by altering any sort of mystical and unmeasurable "life energy." If there is a physiologic effect of acupuncture, it must derive from human physiology, not magic, be it placebo effect, release of endorphins, or a counter-irritant effect, all of which have been postulated by various investigators as possible mechanisms for whatever clinical effects have been reported for it. Whatever the true case is, the scientific method should be able to figure it out in time.
A couple of weeks ago, I came across a rather intriguing study of acupuncture that proved irresistible for me to discuss; so I kept it in my folder of potential blog topics, the better to bring it out when the mood struck me. That day is today, because the article was formally published yesterday. It's a study of the use of acupuncture in mild to moderate hypertension, and it appeared in a solid jounal, namely Circulation, which is, as I like to joke, the house organ of the American Heart Association. The study, done by a group in Germany in collaboration with a group in China, is entitled Randomized Trial of Acupuncture to Lower Blood Pressure. Here's the abstract:
Randomized Trial of Acupuncture to Lower Blood PressureFrank A. Flachskampf, MD; Joachim Gallasch, MD; Olaf Gefeller, PhD; Junxue Gan, MD; Juntong Mao, MD; Annette B. Pfahlberg, PhD; Alois Wortmann, MD; Lutz Klinghammer, MD; Wolfgang Pflederer, MD; Werner G. Daniel, MD
From Med Klinik 2 (F.A.F., A.W., L.K., W.G.D.) and Institut für Medizininformatik, Biometrie und Epidemiologie (O.G., A.B.P.), Universitätsklinikum Erlangen, Erlangen, Germany; Klinikum Ottobeuren, Ottobeuren, Germany (J.G., W.P.); and Nanjing University of Traditional Chinese Medicine, Nanjing, People's Republic of China (J.G., J.M.).
Background-- Arterial hypertension is a prime cause of morbidity and mortality in the general population. Pharmacological treatment has limitations resulting from drug side effects, costs, and patient compliance. Thus, we investigated whether traditional Chinese medicine acupuncture is able to lower blood pressure
Methods and Results-- We randomized 160 outpatients (age, 58±8 years; 78 men) with uncomplicated arterial hypertension in a single-blind fashion to a 6-week course of active acupuncture or sham acupuncture (22 sessions of 30 minutes' duration). Seventy-eight percent were receiving antihypertensive medication, which remained unchanged. Primary outcome parameters were mean 24-hour ambulatory blood pressure levels after the treatment course and 3 and 6 months later. One hundred forty patients finished the treatment course (72 with active treatment, 68 with sham treatment). There was a significant (P
Conclusions-- Acupuncture according to traditional Chinese medicine, but not sham acupuncture, after 6 weeks of treatment significantly lowered mean 24-hour ambulatory blood pressures; the effect disappeared after cessation of acupuncture treatment.
This is actually a pretty well-designed study for the most part. It's not without problems (more on that later), but for a study of "alternative medicine" it's better than most. In essence, the investigators took patients with mild to moderate hypertension. Traditionally, diet and exercise are the first line of treatment for this condition. Unfortunately, human nature being what it is, patients tend to be pretty resistant to long term changes in diet and lifestyle. Most will start out all enthusiastic and may even stick with it for a while, but there is a strong tendency to backslide, which is one reason ultimately so many patients end up on medication (the other is that diet and exercise are not enough for a significant proportion of mildly hypertensive patients). It should be pointed out that the reason that the investigators studied patients both on and off medication was that they deemed it unlikely that they could recruit enough hypertensive patients not on any medication to have adequate statistical power to detect effects on blood pressure. Once patients were recruited, they were subjected to a randomization scheme that was meant to make both groups similar in ranges of blood pressure, antihypertensive use, age, and a number of other parameters.
The next point about this study is how the acupuncture was administered. There's a detailed description of the points used. The investigators brought in highly experienced Chinese practitioners who also had studied Western medicine (whatever that means these days) to administer the acupuncture. Patients were assigned to 1 of 4 types of hypertension according to criteria of traditional (TCM). This, of course, begs the question of how these criteria came to be, if TCM is such an "ancient" medical art. After all, a couple of thousands of years ago, the concept of blood pressure was not known, much less that having elevated blood pressure was a bad thing. Indeed, it wasn't until 1901 that Harvey Cushing introduced a version of the blood pressure cuff into clinical practice in the U.S., bringing the device, which had been invented in Europe a few years before by Scipione Riva-Rocci. In fact, this device was not initially accepted, because it was different from the previously widely used method of palpation of arteries to judge the quality of circulation. It took until 1910 for the device to catch on to the point where nearly every patient in the hospital had his or her blood pressure monitored. It was not until decades later that the longterm risks of hypertension were widely appreciated. So, why would ancient Chinese practitioners care about lowering blood pressure? Alas, the work referenced was a book from 1999 on Chinese acupuncture and was thus not accessible to me. At least the randomization was done by a Chinese physician unaware of which experimental group each subject was in.
But I digress.
Getting back to the study, based on the "type" of blood pressure each patient was judged to have, patients in the treatment group would receive what the TCM practitioner recommended. Patients in the control group would receive "sham" acupuncture, which involved sticking them with the same number of needles, just not in the "right" spots. Sham acupuncture was administered to control subjects in the same way regardless of the "type" of hypertension. Another aspect of the acupuncture given is that each session took 30 minutes, and patients started with 5 sessions a week, tapering to 3 sessions a week later in the treatment period.
In essence, the results are as described in the abstract. The difference between treatment groups was 6.4 mm Hg and 3.7 mm Hg for 24-hour systolic and diastolic blood pressures, respectively. For those receiving acupuncture, the mean 24-hour ambulatory systolic and diastolic blood pressures decreased significantly after treatment by 5.4 mm Hg, respectively. Although the number of patients not taking antihypertensive medicine was small, the pattern was similar but appeared somewhat more heterogeneous. Of note, the treatment effect disappeared after the acupuncture was stopped.
This study raises a number of interesting issues and poses some problems. First, although the decrease in blood pressure was significant and comparable to what angiotensin converting enzyme (ACE) inhibitors produce. However, I was struck at the seeming difficulty and impracticality of it all. We're talking about at least three sessions a week of a half hour each session. If patients have a tendency not to be compliant taking pills, one has to wonder how compliant patients would be with such an acupuncture regimen. Moreover, this study was done under very controlled conditions with a population free of significant disease other than arterial hypertension. People with heart failure, evidence of heart damage, or evidence of kidney disease were excluded.
One problem is that the control "treatment' is not necessarily physiologically neutral. "Sham" acupuncture, in which the needles are inserted to the same depth as acupuncture for treatment, has long been a problematic control. Indeed, as an accompanying editorial pointed out, as many as 40%-50% of patients subjected to sham acupuncture experience some degree of analgesia, and even in non-pain conditions, needling in non-acupuncture points may produce physiologic effects. Indeed, if you look at Figure 3 in the paper, you'll find something very interesting:
(Click to see a larger image at the Circulation Research website.)
I noticed two things. First, the acupuncture group had a wider variation in its pretreatment blood pressures. Second, the sham acupuncture group consistently seems to have a slightly higher blood pressure, as evidenced by a positive difference to baseline. The editorial also alluded to the possibility that sham acupuncture might actually been slightly increasing blood pressure in the control group. It was also pointed out that regression to the mean might lead one to expect that, even in the absence of treatment, blood pressures might be expected to be slightly lower at six weeks. Does that mean that that's what happened? We don't know.
Perhaps the biggest problem with this study is a problem that all studies on acupuncture suffer (and, I must admit, that nearly all surgical clinical trials suffer from also) is the impossibility of double-blinding and the control group. The patients may have been blinded to the experimental group that they were in, but the practitioners were not. Practitioners were blinded to blood pressure but not to treatment. Remember, these are TCM practitioners who really believe in acupuncture. It's quite possible that, because they believe acupuncture to be effective the practitioners administering sham acupuncture may have given off subtle signs of discomfort at administering a form of acupuncture that they believe to be useless. Also, because they were not needling points that they normally used, they may have been less sure of themselves than usual, at least in the beginning. Blood pressure is quite sensitive to emotional state; so it's not outside the pale to wonder whether patients picked up on this and if it had an effect. It is also not clear whether the assessors measuring blood pressures were fully blinded to patient treatment group.
So, what does this study tell us? Not much. Acupuncture might be mildly effective in reducing blood pressure in uncomplicated mild to moderate hypertension through an as yet unknown physiologic mechanism, but it requires considerable investment in time, with half-hour sessions done multiple times per week, making it more difficult a treatment regimen to adhere to than most antihypertensive treatment regimens. Even if this research pans out and the results of this study are replicated and extended, it doesn't appear to me that acupuncture would be all that useful in the general population, other than in a subset of highly motivated patients willing to let TCM practitioners stick needles in them for a half hour three times a week indefinitely. Personally, if I ever end up needing treatment for hypertension (and my diastolic pressure is sometimes borderline), I'd probably take medication and use a half hour three times a week to go to the gym--something I probably should be doing anyway.
Let me tell an anecdote, that has nonetheless a couple of broader implications.
A family member who suffers moderate hypertension started seeing side effects from various medications her physician had prescribed to control it. After doing some reading, she decided to try a low-sodium diet. This is not an easy thing, because most prepared foods sold in groceries and most restaurant dishes use far more salt than the target level allows. So this has required cooking most food at home, and some effort to find appropriate recipes and ingredients. Nonetheless, she has stuck to it for a year, her blood pressure is well controlled, and at this point, it seems a lifestyle change she can practice for the rest of her life.
The interesting part of this story is that her physician had never suggested this. Physicians have come to expect so little compliance with regard to "changes in diet and lifestyle" that I suspect many never order them, and of those who do, I suspect many nonetheless provide the patient little in the way of help in making those changes, and are quick to fall back to drugs at the first sign of non-compliance. Hypertension is dangerous, of course, and if it can't or won't be controlled by lifestyle changes, then it needs to be controlled by drugs. But drugs become just so easy, that we tend to overlook the drawbacks to them. (Side effects, financial costs, interactions.)
For me, the other interesting lesson to this is the state of our food industry. While the lifestyle change my family member undertook at first seemed extreme, in retrospect, she is just eating as our grandparents likely ate. There is a very real sense in which part of our food industry is simply a delivery vehicle for salt, sweeteners, and stabilized fats. This, of course, panders to our innate taste for these, while at the same time exploding the "lifestyle" diseases: hypertension, diabetes, and high blood lipids. There's not an easy solution to this. But it can't be ignored. I wonder how much disease could be prevented, and how many drugs avoided, if people more just followed the prescription: don't eat junk food. And get some exercise.
I used acupuncture to quit smoking 15 years ago! It was like flipping a switch- I went in a smoker and came out a non-smoker. Worked like a charm for me :)
It seems like we need to train some "acupuncture experts" to stick needles in the wrong places, which they will then believe are the right places.
Interesting study. I suspect that most of the BP reduction was due to NO from the placebo effect. I consider it quite plausible that there was a nocebo effect from the TCM practitioners administering a "treatment" they would think was ineffectual or perhaps even harmful, and that this raised the BP of the control group.
So how does a physiological effect occur via acupuncture? Obviously not by opening up channels for qi. However some of the meridians are associated with nerves, and those nerves are bathed in lymph, and the lymph does flow, though much more sluggishly than does the blood. I suspect that stimulation with the acupuncture needles causes the release of NO, and that this NO reacts with albumin in the lymph to form S-nitrosoalbumin, and this S-nitrosoalbumin is the actual therapeutic species which has systemic effects.
NO is one of the things that sets arterial tone and so regulates blood pressure. It is reported that acupuncture does increase local NO, however the levels they report (as NO) in the abstract (all that I have seen), cannot be actual NO concentrations. They could be nitrite. Normally NO can't be that high because of vasodilatation. Nitrite levels could easily be that high, and nitrite is commonly (but often erroneously) used as a substitute for NO. These researchers do report increases in NO due to electroacupuncture, and they did use the right technique and the concentrations of NO they report are plausible.
NO physiology does explain some of the effects observed in acupuncture, and could well be the mechanism behind it. If so, then increasing basal NO by any mechanism should be equally effective. The ideal treatment would not require inpatient treatment involving half an hour of a skilled professional to administer, and would supply additional NO (as needed) 24/7/365. I suggest that "my" bacteria (as discussed on my blog) would be more effective.
Blood pressure is a systemic effect. To affect blood pressure, there need to be systemic effects of what ever treatment is administered. Purely local effects would be ineffective.
20 yrs ago I studied under a Japanese practitioner. Worked on friends and family and cannot give much in the way of dramatic anecdotal evidence with the exception of the use of certain points for stomach cramps/ abdominal pains. For these I preferred the use of burning Chinese moxa sticks (aged mugwort)---often self-administered---on or near specific points. Drifted away from using invasive needling for the most part and found that a good overall massage or working on particular aches and pains may help the body heal itself through relaxation (I'm not referring to serious disease, here). Diet and exercise seem essential.
Here's another study that looked at acupuncture and hypertension:
http://hyper.ahajournals.org/cgi/content/abstract/48/5/838
@Blake:
It seems an easier way to do it is to use real acupuncture but tell the practitioners that they've to treat a different condition. So the sham group get treated for piles or something, and the "real" group get the "correct" treatment.
As long as there is no communication between stabber and stabbee (!) neither will know that they are treating the wrong condition or being treated for the wrong condition.
Actually, I have a different solution, which is actually a bit more workable. Remember this note from a bit earlier in the article:
Here's what we do: Have one doctor decide which type it is, then another administer the treatment. Except, for some doctors (who won't know they've been selected), give them the wrong type, randomly chosen from the other 3. This introduces some more confounding variables, but gets rid of the ones Orac mentioned. For instance, one problem this could create is what would happen if, for instance, 70% had type A, and treating for type A actually helps but treating for any other type makes it worse. Then we'd get a problem where the control is generally worse off.
When I was younger, I tried acupuncture for an annoying condition that was similar to chronic fatigue. One thing that was always part of the experience was guided relaxation, usually with music. If either or both of the groups received this treatment as well, could the relaxation alone explain the drop in blood pressure?
Incidentally, acupuncture didn't do a thing for my fatigue -- that turned out to be a hormone imbalance that was eventually diagnosed by a physician and corrected with prescription medication. However, acupuncture was the most effective thing I've ever found for the muscle-tension problems that I sometimes get in my neck and shoulders -- but only if that's where the needles were placed. Something about that mild pinch just over the area in spasm caused everything to relax very nicely.
High blood pressure was known to the "ancients" because of the symptoms (yes, there are some rather obvious indicators in TCM) and back then they knew most of their patients from birth to death.
One of the reasons that acupuncture works on lowering blood pressure is because it has a tendency to reduce excess, like plumbing, which is one of the reasons that sham AP has an effect. I would never treat a patient for HBP with just AP and not herbs that are appropriate for the diagnosis because you have to treat the underlying cause.
Ancient texts were originally translated by a Frenchman who wasn't completely familiar with the Chinese language which led to the mistranslations of meridian and qi. Deke Kendall's Dao of Chinese Medicine gives some good background on this. Looking at anatomy charts with blood, nerves and AP points layered on helps to see why quite a few of the points work the way they do. There are many AP points that do nothing because they are ceremonial. Of the 365 traditional points, only 70 or 80 are used frequently. The others are only for local use.
The intent of the practitioner is important but the soft music and lovely surroundings, aren't. Those are psychological props since AP was the original barefoot doctors medicine. Ted Kaptchuk has taught classes at Harvard about the placebo effect and is much better at putting it into perspective. He is also an entertaining and informative speaker who can make his point clear to most people.
First, the acupuncture group had a wider variation in its pretreatment blood pressures
This is actually consistent across every pretreatment measurement - see Table 2: http://circ.ahajournals.org/cgi/content/full/115/24/3121/TBL2
Anyone feel like doing Levene's test for equality of variances? I didn't see it mentioned in the study.
The University Clinic of Erlangen where the majority of the authors of this report come from is where Samuel Hahnemann took his doctoral degree!
I'm willin to believe that having needles in you that prevented you from moving much and released mild endorphins could drop blood pressure. Indeed, I suspect that's where most of the effects from acupuncture come from: forcing people to lie back and relax for a while.
If the sham acupuncture is distracting and more painful, I could see it not having the relaxant effect.
Am I the only one who looks at those wildly overlapping confidence intervals/error bars and has a red light go of in my head?
I've had acupuncture done 3 times for a problem affecting the muscles in my lower legs. It's been done by a neurologist, who also treats me with conventional medicine.
The doctor has never spoken about qi or meridians or energy imbalances, etc; I don't know if he believes in these things or not. I don't know if he places the needles in certain defined spots. The needles stay in place for ~15 minutes. They don't hurt at all; when he places them it just feels like someone is snapping a rubber band against my skin. I don't know how far below my skin the needles are placed.
What I do know is that after each treatment, I've felt incredibly relaxed, as if I've drunk a glass of wine on an empty stomach. This feeling lasts for 4-5 hours, then gradually goes away. I do feel less pain during the day that I've had the treatment, but so far that's been it. Then again, the treatments I've had have been several weeks apart, so who knows what pain relief I would have if the treatments were at least once a week.
Still, I've never had a feeling of relaxation like this from any other medical treatment I've received. For me, there definitely is a physiological response going on, though it still seems odd to me that placing very thin needles into my skin (and underneath?) would cause something like this.
It's difficult for me to believe this is simply a placebo effect. I've been given other medical treatments that I was assured were highly effective, yet they still didn't work for me. If anything, I've become sceptical of overly positive claims of effectiveness, whether they are for conventional or alternative medicine. With the acupuncture, the neurologist has made no claims one way of the other about its effects, which I have come to prefer.
Like Viscount, I wonder if heteroscedasticity affects the validity of the statistical conclusions. I'd sure like to know a lot more about the difference between the "real" acupuncture and the "sham" needling.
Thanks for this terrific analysis, Orac. I concluded as you did: not sure how this works, it does have a measurable effect, but given the choice between taking 1 pill a day or driving back and forth 3-5 times/week to an acupuncturist for 30 minute sessions that only improve the BP for a short period of time... I think I know what most Americans will choose. And good for you for choosing 30 minutes 3-5 times/week at the gym instead. Probably an even better investment.
Thank you for the posts (like this one) where you analyze experimental design in some detail, Orac. Each one is a learning experience.