Cochrane Review: 'Exercise therapy for chronic fatigue syndrome', Larun et al. - New version October 2019 and new date December 2024

Exercise therapy for chronic fatigue syndrome

• Lillebeth Larun
• Kjetil G Brurberg
• Jan Odgaard-Jensen
• Jonathan R Price

Authors' declarations of interest
Version published: 02 October 2019 Version history

https://doi.org/10.1002/14651858.CD003200.pub8
____________________

I don't have the capacity to read back through this whole thread, but I've been taking a look at the review to remind myself just what it is we are asking Cochrane to withdraw. I had forgotten how many versions of their description of CFS and conclusions it includes, with the Abstract, Plain English summary, and full report.

Here they are:

Their descriptions of CFS:

Abstract:
Chronic fatigue syndrome (CFS) or myalgic encephalomyelitis (ME) is a serious disorder characterised by persistent postexertional fatigue and substantial symptoms related to cognitive, immune and autonomous dysfunction. There is no specific diagnostic test, therefore diagnostic criteria are used to diagnose CFS.

Plain English:
People with chronic fatigue syndrome have long‐lasting fatigue, joint pain, headaches, sleep problems, poor concentration and short‐term memory. These symptoms cause significant disability and distress.

Full report:
Description of the condition
Chronic fatigue syndrome (CFS) is an illness characterised by persistent, medically unexplained fatigue. Symptoms include severe, disabling fatigue, as well as musculoskeletal pain, sleep disturbance, headaches, and impaired concentration and short‐term memory (Prins 2006). Individuals experience significant disability and distress, which may be exacerbated by lack of understanding from others, including healthcare professionals.
[the paragraph continues with stuff about different diagnostic criteria]
_________________

And their overall conclusions about exercise therapy:

Abstract:
Authors' conclusions
Exercise therapy probably has a positive effect on fatigue in adults with CFS compared to usual care or passive therapies. The evidence regarding adverse effects is uncertain. Due to limited evidence it is difficult to draw conclusions about the comparative effectiveness of CBT, adaptive pacing or other interventions. All studies were conducted with outpatients diagnosed with 1994 criteria of the Centers for Disease Control and Prevention or the Oxford criteria, or both. Patients diagnosed using other criteria may experience different effects.

Plain English summary:
Key messages
People who have exercise therapy probably have less fatigue at the end of treatment than those who receive more passive therapies. We are uncertain if this improvement lasts in the long term. We are also uncertain about the risk of serious side effects from exercise therapy.

Full report
Implications for practice
Low‐ to moderate‐certainty evidence suggests that exercise therapy may contribute to alleviation of some of the symptoms of chronic fatigue syndrome (CFS), especially fatigue. Long‐term effects are in general more uncertain than short‐term effects mainly because studies did not always have long‐term follow‐up. The impact of exercise therapy on serious adverse reactions is uncertain. Due to few studies with a small number of participants it is difficult to draw conclusions about the comparative effectiveness of cognitive behavioural therapy (CBT), adaptive pacing or other interventions. This evidence is collected from outpatients diagnosed with 1994 criteria of the Centers for Disease Control and Prevention (CDC) or the Oxford criteria, or both, and people diagnosed using other criteria may experience different effects.

 

Problem being that (far as I know) all other criteria are more restrictive and less inclusive, so anybody diagnosed with those other criteria are going easily meet the 1994 CDC and especially the Oxford. In which case the disingenuous can argue GET can be safely used anyway on those patients.

 

I think the criteria aren't really relevant - the trials showed no improvement in any measures that had a notion of reliability (i.e. not subjective measures). So it doesn't matter what criteria were used.

Cochrane are embarrassing themselves by using the CFQ

 

Below is Hilda Bastian's reply to a post by Jonathan Edwards and other posts by others that criticise the 2019 Review. This is from Hilda's talkpage, here:
https://hbprojecttalk.wordpress.com...eration-hash=889c3e6b14019d5b611e1b3c1401bd91

I'm posting it here so that we can examine the points she makes about the findings of NICE and the CDC.

 

First, I note that NICE did not find the Larun et al 2019 review as useful:
https://www.nice.org.uk/guidance/ng...acological-management-of-mecfs-pdf-9265183028

1.1.3.2 Excluded studies
...
"One Cochrane review of exercise interventions (Larun 201748) pooled all exercise therapies irrespective of the type of exercise therapy, and also pooled all control arms considered ‘passive’ (including treatment as usual, relaxation and flexibility). We did not consider this methodology appropriate for decision-making. Additionally, the review did not include all critical outcomes specified in this review protocol, including cognitive function, activity levels, return to school/work, exercise performance measures and mortality."

 

The NICE Guideline committee of course was labouring under the weight of a number of people who were desperate to show that BPS methods worked. And was the analysis of the trials perfect? No, it wasn't.

But here are the results of the quality assessments GET versus standard care, as Hilda says, page 148 of the document I linked above.

SUBJECTIVE OUTCOMES
Quality of Life - (1 study, 52 weeks) LOW - mean QOL in controls 0.53; QOL in intervention 0.59

General symptom scales - (2 studies, 12-42 weeks) patient reported global impression of change positive - VERY LOW
General symptom scales - patient reported global impression of change positive versus negative/minimal change - (1 study, 12 weeks) LOW
General symptoms scales - clinical global impression of change positive versus negative/minimal change - (1 study, 134 weeks) VERY LOW

Fatigue/fatigability (CFQ) - (2 studies, 12 weeks) VERY LOW
Fatigue/fatigability (CFQ) - (1 study, 134 weeks) VERY LOW

Physical functioning SF36 Physical function - (2 studies, 12 weeks)- VERY LOW
Physical functioning SF36 Physical function - (1 study, 134 weeks)- VERY LOW

Psychological status (HADS - depression) - (2 studies, 12-52 weeks) LOW - a mean of 7.35 in controls, 6.2 in GET

Psychological status (Hospital Anxiety and Depression Scale - anxiety) Scale from: 0 to 21 - (2 studies, 12-52 weeks) LOW - a mean of 7.90 in controls, 6.86 in GET


Pain - muscle pain - (1 study, 52 weeks) - VERY LOW
Pain - joint pain - (1 study, 52 weeks) - VERY LOW

Sleep quality - (1 study, 52 weeks) - VERY LOW

Adverse events (non-serious) - (2 studies, 12-52 weeks) - VERY LOW
Adverse events (serious) - (2 studies, 12-52 weeks) - VERY LOW
Adverse events (adverse reactions) - (2 studies, 12-52 weeks) - VERY LOW

Activity levels (International Physical Activity Questionnaire high vs. low/moderate level of activity prev week) - (1 study, 12 weeks) - LOW

Return to school/work (Work and Social Adjustment Scale) Scale from: 0 to 40 - (1 study, 12 weeks) - LOW - mean in controls 25.4, mean in GET 23.5
Return to school/work (Work and social adjustment scale) 0 to 40 - (1 study, 134 weeks) - VERY LOW - mean in controls 21.1, mean in GET 20.3

Exercise performance measure (perceived exertion – Borg scale) - (2 studies, 12 weeks) - VERY LOW

OBJECTIVE OUTCOMES
Exercise performance measure (6 minute walk) - (1 study, 52 weeks) - VERY LOW
Exercise performance measure (VO2 peak/aerobic capacity) - (3 studies, 12 weeks) - VERY LOW
Exercise performance measure (Peak power) - (2 studies, 12 weeks) - VERY LOW
Exercise performance measure (Elapsed exercise test time - cycle ergometer) - (1 study, 12 weeks) - VERY LOW
Exercise performance measure (VEpeak) - (1 study, 12 weeks) - VERY LOW

 

So, I make that
1. All five of the objective outcomes rated as very low quality evidence
2. Of the 20 subjective outcomes, 14 were rated as very low quality evidence.
3. Of the 25 outcomes, 19 were rated as very low quality evidence. (Possibly I didn't get them all, I didn't double check, but they are mostly there.)
4. Any outcome that was not rated as very low quality evidence was rated as low quality evidence. 'Low quality evidence' is not something to be proud of.

5. Of all the outcomes relating to harms, all were rated as very low quality evidence. The precautionary principle suggests that we should give uncertainty about harms considerable weighting when evaluating an intervention, especially when virtually all patient organisations are expressing concern about harms.


6. Of the outcomes rated low quality evidence:
a. Quality of Life, a subjective outcome from one study. The reported benefit is small. NICE noted that there was no clinically important difference.

b. General symptom scales. One of the three outcomes for this category was rated low quality evidence, but the other two were very low quality evidence. The one that was reported as low quality was only from 1 study for 12 weeks and was self-reported. The two general symptoms scale outcomes that were reported as very low quality consisted of more studies, had a longer assessment period and/or were assessed by a clinician. NICE noted that there was no clinically important difference in general symptom scales at 134 weeks.


c. Psychological status HADS - depression.
d. Psychological status HADS - anxiety
The evidence consists of 2 studies, including one with an assessment at 12 weeks which is too short in ME/CFS as impacts of increased exertion are cumulative. The differences between controls and intervention were small. I expect that people who were able to get to sessions did feel less depressed and anxious as a result of participating in a group; there would have been a feeling of mutual support, less of a feeling of being alone. A wait list control is not an adequate control for GET. Despite that, NICE noted that there was no clinically important difference in either outcome.


e. Activity levels (subjective). This was only assessed for 12 weeks - too short. The claimed increase in activity levels was not corroborated by an increase in fitness in the objective measures. (NICE reported no clinically important differences in physical functioning SF36physical, the 6 minute walk, VO2 peak, peak power, elapsed exercise time test.)


f. Return to school/work (subjective). One study, assessed for 12 weeks. The reported benefit over controls was small. The evidence at 134 weeks was assessed as of very low quality and the reported benefit became minuscule. NICE noted that there was no clinically important difference in Return to school/work.


I think it's clear that the outcomes that were rated as having a low quality of evidence (as opposed to a very low quality of evidence) also had one or more of the following issues:

• assessed over too short a time
• very small benefit over controls, typically no clinically important benefit
• probably not arising as a result of the intervention, but rather as a result of not being on a wait-list control (e.g. meeting other patients), and not core to the illness
• benefit not corroborated by objective outcomes or other relevant outcomes

As a body of evidence for an intervention, it is not looking very convincing. Hardly any clinically important differences, and a high degree of uncertainty about any reported benefits due to low evidence quality.

 

From NICE
https://www.nice.org.uk/guidance/ng...acological-management-of-mecfs-pdf-9265183028
page 384
This has the bit about the lack of clinically important differences for many of the outcomes.

It also has the interesting news that when the NICE team had a look at the studies where they could separate out people with PEM and people without clear PEM, a benefit of GET on the general symptom scales for the PEM group remained, while there was no clinical benefit in the non-clear PEM group. Note however, that the quality of evidence for both was very low, and the assessment for the PEM group was at 12 weeks (42 weeks for the 'without clear PEM' group).

For self-reported physical functioning, when they separated out people with PEM, there was no clinically important benefit for people with PEM, but there was a clinically important benefit of GET for the people without clear PEM. Again, the quality of evidence for both was very low, and this was only at 12 weeks, which is too short a time to assess benefits.

So, what to make of that? I don't think there is evidence there of reliable differences in the impact of GET on people with Fukuda CFS regardless of whether they have PEM. There is no basis here for adding notes to the Larun et al review suggesting that GET is good for PEM or non-PEM subsets.

 

Here is the reply I posted on Hilda's talk page. A couple of my posts are not appearing on that platform, so I'm going to record them in the forum. Others might want to do the same.

Hilda,
You seem to be suggesting that the fact that NICE didn't rate all the evidence relating to GET as 'very low' somehow means that they didn't conclude in general terms that there was 'no reliable evidence of efficacy'.

I have gone through the evidence evaluation for GET (versus standard care) in NICE. While it is true that NICE did not give every outcome a rating of 'very low evidence quality', it did rate most of the outcomes in that way.

That included the evidence for harms - all three harm outcomes were rated as being of very low quality. Given the precautionary principle and the strong views of virtually every ME/CFS patient organisation in the world that there is harm, we should not dismiss the possibility of harm. Therefore, in order to recommend the treatment, we would want to see clinically important benefits for which there is good quality evidence.

But, there just aren't clinically important benefits for which there is good quality evidence.

Here are the six outcomes NICE reported as having low quality evidence ('low' is still really bad, but better than 'very low'.)

1. Quality of Life, a subjective outcome from one study. The reported benefit is small. NICE noted that there was no clinically important difference.

2. General Symptom scales. One of the three outcomes for this category was rated low quality evidence, but the other two were very low quality evidence. The one that was reported as low quality was only from 1 study for 12 weeks. The two general symptoms scale outcomes that were reported as very low quality consisted of more studies and/or had a longer assessment period. NICE noted that there was no clinically important difference in General Symptom scales at 134 weeks.

3. Psychological status HADS - depression.
4. Psychological status HADS - anxiety
The evidence consists of 2 studies, including one with an assessment at 12 weeks which is too short in ME/CFS as impacts of increased exertion are cumulative. The differences between controls and intervention were small. I expect that people who were able to get to sessions did feel less depressed and anxious as a result of participating in a group; there would have been a feeling of mutual support, less of a feeling of being alone. A wait list control is not an adequate control for GET. Despite that, NICE noted that there was no clinically important difference in either outcome.

5. Activity levels (subjective). This was only from a 12 week trial length which is too short. The claimed increase in activity levels was not corroborated by an increase in fitness in the objective measures. (NICE reported no clinically important differences in physical functioning SF36physical, the 6 minute walk, VO2 peak, peak power, elapsed exercise time test.)

6. Return to school/work (subjective). One study, from a 12 week trial length. The reported benefit over controls was small. The evidence at 134 weeks was assessed as of very low quality and the reported benefit became minuscule. NICE noted that there was no clinically important difference in Return to school/work.

Therefore, of the six outcomes with low quality evidence, all either had a benefit that was not clinically important, was assessed after too short a period of intervention and/or was not corroborated by other measures that we would expect to track in a similar way.

 

There was no clinically important difference for... ...exercise performance (6 minute walk...).

There was only one result in PACE from all objective measures that was even statistically significant (6 minute walk for the GET arm), and that was also not clinically significant, and indeed still left patients scoring down among the sickest (non-terminal) people medicine deals with.

They got nothing.