Review Acupuncture and moxibustion for chronic fatigue syndrome: A systematic review and network meta-analysis, 2022, Fang et al

Abstract
Background: Research into acupuncture and moxibustion and their application for chronic fatigue syndrome (CFS) has been growing, but the findings have been inconsistent.

Objective: To evaluate the existing randomized clinical trials (RCTs), compare the efficacy of acupuncture, moxibustion and other traditional Chinese medicine (TCM) treatments.

Data sources: Three English-language databases (PubMed, Embase, Web of Science, and The Cochrane Library) and 4 Chinese-language biomedical databases (Chinese Biomedical Literature Database, VIP Database for Chinese Technical Periodicals, China National Knowledge Infrastructure, and Wanfang) were searched for RCTs published from database inception through August 2021.

Study selection: RCTs include acupuncture, moxibustion, traditional Chinese herbal medicine, western medicine and no control.

Data extraction and synthesis: Data were screened and extracted independently using predesigned forms. The quality of RCTs was appraised with the Cochrane Collaboration risk of bias tool. We conducted a random-effects network meta-analysis within a frequentist framework. We assessed the certainty of evidence contributing to network estimates of the main outcomes with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework.

Main outcomes and measures: The primary outcomes were the overall response rate and FS-14 scale.

Results: A total of 51 randomized controlled trials involving 3473 patients with CFS were included in this review. Forty one studies indicate low risk or unknown risk, and the GRADE scores of the combined results show low levels. Among the main indicators, traditional Chinese medicine therapies have excellent performance. However, the overall response rate is slightly different from the results obtained from the Fatigue Scale-14 total score. Moxibustion and traditional Chinese medicine (Odds ratios 48, 95% CrI 15-150) perform better in the total effective rate, while moxibustion plus acupuncture (MD 4.5, 95% CrI 3.0-5.9) is better in the FS-14 total score.

Conclusions: The effect of acupuncture and moxibustion in the treatment of CFS was significantly higher than that of other treatments. Traditional Chinese medicine should be used more widely in the treatment of CFS.

https://pubmed.ncbi.nlm.nih.gov/35945779/
https://www.ncbi.nlm.nih.gov/pmc/articles/pmid/35945779/ free access

 

I am assessing a recommendation by a doctor for acupuncture for ME/CFS, supported by this paper. I think all of the individual studies on acupuncture and moxibustion we have looked at have been flawed and unreliable, so the conclusion of this systematic review is a bit surprising and bears closer examination.

I recall the NICE ME/CFS guideline process (done slightly earlier than this paper) found no evidence to support the recommendation of acupuncture.

 

It's not starting super well.

 

No mention of PEM as a distinguishing symptom.

It does review some Chinese language papers, which is useful; it will be including papers that NICE probably did not review.

 

So, yes, Fukuda.
I'm not clear what the outcomes considered are yet, apart from the hopelessly flawed Chalder Fatigue Scale is one. FS-14 is the Chalder Fatigue Scale.

 

It's all pretty opaque with so many studies testing different things.
From Limitations, it looks as though the quality of the studies was low:

Here's the last two sentences from the Conclusion:

 

I thought I'd have a look at the one study reportedly done in Australia.

That's all of the details provided in the references. There is no journal given and when I google it, I can't find the paper.

When I look at the 51 papers listed in the supplementary materials, there is no paper with Yun as an author.

Re the Singapore study:

Again, that's all of the details, no co-authors, no journal.
Zhu, 2012 does appear in the supplementary materials list.
25 in the acupuncture arm; 35 in the Chinese herb arm
Fukuda, young people (mean age around 27), and they have only had fatigue for less than 2 years
Outcomes Fatigue Severity Scale; Hamilton Anxiety Scale; Visual Analogue Scale (of what?); Hamilton Depression Scale; overall response; adverse reaction
I can't find this paper with google.

 

Looking at the search terms, regarding the names of the disease are

No Myalgic Encephalomyelitis, no ME/CFS

There are Chinese language searches listed in the supplementary materials. It would be interesting to know more about what the searches are.

 

So, as far as I can work out so far:

The selection of participants is poor, with a Fukuda criteria meaning that the participants have chronic fatigue, but don't necessarily have ME/CFS. The participants are also typically fairly young, and have only had chronic fatigue for a relatively short time (i.e. less than a few years). We can expect high rates of improvement and even recovery in these people.

The Chalder Fatigue Scale is the main outcome, and we know that it has all sorts of problems including a ceiling on recorded worsening, and lack of clarity about comparators that tends to make 'no change' look like 'improvement'.

Some of the studies compare an active treatment with 'care as usual' or nothing at all (confusingly called 'no control'). The active treatment comes out better - which is not surprising, given the outcomes are subjective and there is no blinding.
Some of the studies compare two active treatments (e.g. acupuncture with moxibustion versus acupuncture alone). In these, treatments (other than 'Western medicine') tend to produce improvements. Moxibustion (and mixed treatments that includes moxibustion) tends to produce a surprising benefit over other treatments.

The authors have then done a network analysis, linking up the effects of the different treatments. So if acupuncture is better than no treatment in one study, and moxibustion is better than acupuncture in another study, then we can assume that moxibustion is better than no treatment. But, I think the whole edifice is built on the old 'subjective outcomes, unblinded treatments, hype the active treatment and leave the 'no treatment controls' people disgruntled, as they feel that they are missing out.

I'm just not sure why moxibustion seems to come out better than acupuncture, although there could be all sorts of things going on such as conflicts of interest, and prevailing beliefs that moxibustion is a more effective form of treatment.

What does seem clear is that traditional herbal medicines are really ineffective, with their outcomes not being very different to 'no treatment'. Which is a shame, because, as I've said before, I have liked the idea that there is a herbal medicine that is useful.

What do you think?

 

Of the 51 studies, only 11 have a control arm (rather than multiple treatments).
Actually most of the control arms were potentially blinded.
Most of the trials with plausible controls were testing acupuncture - this might be why acupuncture alone didn't come out of the review very well. Moxibustion was usually tested against real acupuncture.

Sham acupuncture - Streitbergers placebo
Zheng 2012 vs Acu
An 2014 vs Acu
Zhang 2007 vs Acu
Yang 2017 Acu
Chen 2010 Acu

Needling of non-acupuncture points
Wang 2009 vs Acu
Lu 2014 vs Acu and vs Acu +Mox
Chen 2018 Acu
Ding 2011 Acu

Usual care
Li Long 2016 vs Acu
Liang 2016 vs Acu +Mox

 

Re 'Western medicine'
There were only 4 studies with 'western medicine'
Xiao 2014 - Vitamin B1 +B6+oryzanol+paroxetine vs Moxibustion
Huang 2017 - Vitamin B1 +oryzanol vs acupuncture
Wang 2013 - Fluoxetine hydrochloride vs moxibustion
Liu Chun 2017 - Oryzanol vs moxibustion

Oryzanol - lipids extracted from rice bran, used for some medical conditions such as high cholesterol but there doesn't seem to be much evidence to support its use for anything.
Fluoxetine is Prozac

 

For example, from 4.3 of the Supplementary materials on the Chalder Fatigue Scale - total score outcome:


This is a map of the actual comparisons. You can see the substantial number of comparisons of control treatments with acupuncture, but not much else.

And here are the calculated comparisons



Those numbers described as mean differences are, I believe, differences in the Chalder Fatigue Scale score, after treatment, compared to the control group. The version reportedly used in all the studies is the 14 question scale, which is outdated (replaced by the 11 question scale). Assuming the Likert scoring was used, the total scores can range from 0 to 42. If the mean differences are the mean differences between the score of the control treatments and the active treatment (I might be misunderstanding something), then we are looking at differences of less than 5 points out of a scale going up to 42. That is very much in the realm of placebo effects.

And the rankings of the treatments seem to fit with increasing complexity (and therefore plausibility) of the treatments. Needles and burning stuff looks impressive, with lots of rituals around it, and lots of sessions and contact with the practitioner.

 

A reminder that we have this thread to talk about acupuncture generally
Acupuncture and other traditional Chinese medicine news and discussion thread

I just re-read the good article linked in the first post of that thread again. It concluded pretty much what we see here: acupuncture often shows a clinically insignificant benefit over sham treatments which is likely to be the result of problems with experimental design and implementation.

 

It is possible that some studies used the bimodal (0,1) scoring system, making the scale 0-14. I have looked hard at the paper and I can't find any clue that would tell me what scoring system the studies used. Given the studies have been undertaken over a wide range of years, I think it is very likely that studies used a range of scoring approaches (14 question and 11 question surveys; bimodal and likert scoring), and it is also possible that the review authors did not take this into account in their meta-analysis.

Even if all of the 31 studies did use a bimodal scoring system, the mean difference for acupuncture compared to the control treatment was only 2.3. Also considering the fact that the sham acupuncture techniques have been criticised as being less than perfectly blinded (the Streitbergers technique doesn't pierce the skin for example and significant proportion of participants realise this) and some studies had waitlist controls, a 2.3 point difference on a 0-14 scale is not compelling evidence.