A trifecta of wipeouts for major CAM modalities this week…

It’s been another interesting week for those who follow the literature on CAM.

To start off there is this article which is a big and well-conducted meta-analysis of Omega-3 supplementation in cardiac prevention. Even at a glance, it is apparent that the methodological niceties have been observed, since three very important comments are made up front. Firstly, the numbers are very big (over 68,000 patients in the studies analyzed). Secondly, they used 20 large studies for comparison and have included robust assessments of the heterogeneity of the studies involved. This is very important in a meta-analysis since in basic terms it accounts for the fact that different methodologies can give different answers, and different populations under study will also give variable results. Using fairly rarefied stats techniques one can compare across studies with some degree of confidence. The third piece of crucial information is to note that they have accounted for multiple comparisons. This is important because if you are performing lots of analyses, you will get some correlations as a matter of chance. The more comparisons, the more likely you are to come up with statistically significant correlations which disappear in subsequent studies. Any study which uses multiple regression techniques to get at hidden secrets in the data has to report that they have adjusted for multiple comparisons, and these authors have done just that.

So even before getting past the abstract, this looks like a credible study. It demonstrates pretty comprehensively that Omega 3 supplements aren’t effective at preventing cardiac problems like heart attacks, stroke, sudden death and arrhythmias. What makes this study more credible is that it has included both dietary and supplement studies. So whether you are getting your Omega 3’s from a capsule or from tins of tuna, it seems unlikely based on this very large and convincing study that they are doing much good.

The Cochrane Back Group has just released an updated systematic review of spinal manipulation therapy for acute low back pain (LBP). ‘Acute’ back pain by definition is pain which has been around for less than 6 weeks. The natural history of acute LBP is that 80% is better by 8 weeks, and half of the remaining patients get better over the next 8 weeks. This is a condition which has a high rate of spontaneous recovery over a couple of months no matter how agonisingly painful it is for the first couple of weeks. It is therefore not an easy condition to show a definite benefit for any proposed treatment. Unsurprisingly, there is no robustly evidence-based treatment for acute LBP apart from avoiding bed rest as much as possible. On the background of this lack of really effective treatments, manipulative therapists such as chiropractors and osteopaths have gained some acceptance while better evidence was awaited. The Cochrane acute LBP review of spinal manipulative therapy (SMT) was last updated in 2004. The new review includes 12 trials in addition to the 8 available in the previous review. The authors note that they included trials from the manual therapy, chiropractic and osteopathic literature in their search. Their conclusions deserve to be quoted in full..

“In this review, a total of 20 randomized controlled trials (RCTs) (representing 2674 participants) assessing the effects of SMT in patients with acute low-back pain were identified. Treatment was delivered by a variety of practitioners, including chiropractors, manual therapists, and osteopaths. Approximately one-third of the trials were considered to be of high methodological quality, meaning these studies provided a high level of confidence in the outcome of SMT.

Overall, we found generally low to very low quality evidence suggesting that SMT is no more effective in the treatment of patients with acute low-back pain than inert interventions, sham (or fake) SMT, or when added to another treatment such as standard medical care. SMT also appears to be no more effective than other recommended therapies. SMT appears to be safe when compared to other treatment options but other considerations include costs of care”

So that’s two strikes for two very popular and widespread CAM modalities. The third leg of the trifecta is this study in the prestigious Annals of Internal Medicine journal.

On the face of it, this looks to be a definite plus in favour of acupuncture for chronic pain conditions, but it simply isn’t what it appears. The abstract doesn’t report heterogeneity analysis, neither does it report adjusting for multiple comparisons. Both these deficiencies reduce your confidence in the strength of the conclusions. They are also just reporting pain scores, not pain-related disability or other functional goals. But let’s not quibble. Let’s suspend routine critical thinking and give them a pass on this. What are they actually reporting?

“Significant differences between true and sham acupuncture indicate that acupuncture is more than a placebo. However, these differences are relatively modest, suggesting that factors in addition to the specific effects of needling are important contributors to the therapeutic effects of acupuncture.”

The actual effect size they are reporting as ‘modest’ is actually more like ‘negligible’ when talking about chronic pain. Although it might be statistically significant, it is too small a difference to be clincially signficant. They are reporting with a straight face that 80% or so of the benefit of acupuncture that they have been able to torture out of the data is in fact not due to the specific effects of putting the needle in.

Slap me if I’m being thick here, but isn’t careful placement of the needle into the strictly defined acupoint the whole thing acupuncture is meant to be about? If this was a Big Pharma drug trial it would be wipeout that would probably precipitate a marked decline in prescriptions.

If that is the best possible result they can get by disregarding the usual standards of a meta-analysis, I think I’ll be referring my patients for a cup of tea and a chat with my mum before I send them for acupuncture, since the majority of the benefit comes from having a charismatic and sympathetic listener.

PS. I’m kidding about the referrals bit Mum, not the other bit!


My copy of Pain has just dropped through the letterbox and it contains another 2 acupuncture-related articles of relevance. This one is a test of acupuncture analgesia on human experimentally-induced pain using both sham and real acupuncture needles. It showed no useful effect in any of the usual tests of effective analgesia.

The other was a very tightly designed RCT in acute LBP, comparing usual care with placebo acupuncture (ie placing needles in the ‘wrong’ spots), sham acupuncture (using retractable needles) and real acupuncture (done according to TCM principles). Regular readers will be able to guess the result.

Yep, you guessed it.

All 3 acupuncture arms did slightly better than the ‘conventional treatment’ arm, but not any better than each other. Confirming again that acupuncture is a strongly context-based treatment where the needles are essentially a prop.