Review Effectiveness and tolerance of exercise interventions for long COVID: a systematic review of randomised controlled trials, 2025, McDowell et al

Abstract
Objectives To examine the effectiveness of exercise interventions to improve long COVID symptoms and the tolerance of exercise interventions among people with long COVID.

Design Systematic review.

Data sources Medline via EBSCOhost, Embase via OVID and CENTRAL via the Cochrane Library up to 28 February 2023.

Eligibility criteria for selecting studies Inclusion criteria were: (1) participants with long COVID, as defined by study authors; (2) random assignment to either an exercise intervention or a comparison group and (3) a quantitative measure of at least 1 of the 12 core long COVID outcomes. Exclusion criteria were: (1) signs or symptoms not reasonably attributable to prior SARS-CoV-2 infection; (2) pre-exposure or postexposure prophylaxis for COVID-19 or the prevention of long COVID symptoms and (3) interventions where the primary exercise component is breathing or respiratory muscle training.

Data extraction and synthesis Two reviewers independently extracted data, and studies were narratively synthesised.

Results Eight studies were included. Follow-up periods ranged from 2 to 28 weeks (mean=8.5 weeks). Sample sizes ranged from 39 to 119 (mean=56). All studies were in adults (mean age=49.9 years) and both sexes (mean female proportion=53.9%). Four studies were at low risk of bias, two were unclear and two were high. The evidence suggests that exercise interventions lead to short-term improvements in dyspnoea, fatigue, physical function and the physical domain of quality of life among people with long COVID. Of the five studies that reported adverse events, rates were low and, when reported, mild. Of the seven studies that reported sufficient relevant information, 1 of 252 participants who received exercise discontinued the intervention due to tolerance-related issues.

Conclusion Available evidence suggests that exercise interventions may be beneficial and tolerable among some people with long COVID. However, the evidence base consists of a limited number of studies with small sample sizes and short follow-up periods.

https://bmjopen.bmj.com/content/15/3/e082441

https://bmjopen.bmj.com/content/bmjopen/15/3/e082441.full.pdf

 

‘When you don’t account for all of the bias, some of the studies had a low risk of bias.’

How did this get published?!

————

Risk of bias
The study risk of bias is presented in table 2. Briefly, when excluding consideration of participant blinding, four studies had a low overall risk of bias.36 38 39 41 Two studies had an unclear overall risk of bias,35 37 and two studies had a high overall risk of bias as primary outcomes reported in the studies did not align with the trial registrations.34 40

 

The BMJ link doesn’t work, only the pdf.

The PubMed link works, but the link to BMJ is also broken there.

https://pubmed.ncbi.nlm.nih.gov/40122540/

Edit: the BMJ link works now.

 

They should probably have mentioned in the abstract that none of the studies assessed PEM.

————

Postexertional malaise
No studies in the current review directly assessed outcomes related to postexertional symptoms which sometimes occur among people with long COVID.4 7 PEM is characterised by an increase in pain, fatigue, cognitive problems and other associated symptoms that persist for an unusually long period of time following mental and/or physical exertion.14 47 It is a hallmark feature of ME/CFS—another chronic multisymptom illness with substantial symptom overlap with long COVID.13 Even minimal exertion can significantly worsen the entire symptom complex of ME/CFS,14 and exercise challenges, frequently used in ME/CFS research, often lead to an increase in symptoms such as fatigue, sore throat, headaches, muscle pain, joint pain, confusion and decrease in energy.15–19 This typically lasts for days, but patients have reported that it can last for years or that it has never resolved.48 49 Although no studies in the current review directly assessed postexertional symptoms, drop-out rates and rates of adverse events were low among those that reported relevant information, suggesting that exercise interventions may be tolerable among some people with long COVID.

 

As participants cannot be blinded to the treatment allocation in exercise trials, this domain is inherently at high risk among all exercise interventions. As this review was limited to exercise trials, this domain was not considered in the overall risk of bias assessment, so as to capture between-study differences across the other domains and to differentiate between studies at higher and lower risk of bias. Studies were given an overall risk of bias assessment of low (ie, all domains were low risk), unclear (ie, all domains were low or unclear risk) or high (ie, at least one domain was high risk).

————

If you can’t determine the risk of bias based on the available info, surely the default should be ‘high’?

 

These are the categories of bias that were assessed:

1. Randomisation sequence generation
2. Allocation concealment
3. Selective reporting
4. Other
5. Blinding of participants
6. Blinding of assessor
7. Incomplete outcome data

All studies got a score of low risk of bias in the category ‘other’. Going by what you’re saying, the two studies you mentioned should have had a high risk of bias in this category.

 

This is the protocol for the review:
https://www.hiqa.ie/sites/default/files/2023-03/Long-COVID-Interventions_Protocol.pdf

 

The used the original The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials
https://www.bmj.com/content/343/bmj.d5928

 

From Cochrane:

3. Assess the risk of bias in trial results, not the quality of reporting or methodological problems that are not directly related to risk of bias
The quality of reporting, such as whether details were described or not, affects the ability of systematic review authors and users of medical research to assess the risk of bias but is not directly related to the risk of bias. Similarly, some aspects of trial conduct, such as obtaining ethical approval or calculating sample size, are not directly related to the risk of bias. Conversely, results of a trial that used the best possible methods may still be at risk of bias.

For example, blinding may not be feasible in many non-drug trials, and it would not be reasonable to consider the trial as low quality because of the absence of blinding. Nonetheless, many types of outcome may be influenced by participants’ knowledge of the intervention received, and so the trial results for such outcomes may be considered to be at risk of bias because of the absence of blinding, despite this being impossible to achieve

4. Assessments of risk of bias require judgment
Assessment of whether a particular aspect of trial conduct renders its results at risk of bias requires both knowledge of the trial methods and a judgment about whether those methods are likely to have led to a risk of bias. We decided that the basis for bias assessments should be made explicit, by recording the aspects of the trial methods on which the judgment was based and then the judgment itself.

————

I know protocols are flawed, but they didn’t even follow the protocol in regards to bias from not being able to blind the participants.

 

More from Cochrane (my bolding in the middle):

Assessments of risk of bias and synthesis of results
Summary assessments of the risk of bias for an outcome within each trial should inform the meta-analysis. The two preferable analytical strategies are to restrict the primary meta-analysis to studies at low risk of bias or to present meta-analyses stratified according to risk of bias. The choice between these strategies should be based on the context of the particular review and the balance between the potential for bias and the loss of precision when studies at high or unclear risk of bias are excluded. Meta-regression can be used to compare results from studies at high and low risk of bias, but such comparisons lack power, 15 and lack of a significant difference should not be interpreted as implying the absence of bias.

————

It seems to me like they should have only included results from the trials with a low risk of bias or be very clear about the levels of risk of bias in the studies.

They do not mention the risk of bias at all during the discussion of the results.

The abstract makes it look like there were four studies with low bias, which is clearly not the case.

The conclusion in the abstract does not mention the bias at all.

 

The lead author, McDowell, is a senior HTA analyst at HIQA.

The Health Information and Quality Authority (HIQA) is an independent authority established to drive high-quality and safe care for people using our health and social care services in Ireland.

Health technology assessment (HTA) is a multidisciplinary research process that collects and summarises information about a health technology.

So these are the people that should know better.

https://www.hiqa.ie/areas-we-work/health-technology-assessment

Edit: it seems to resemble the Larun situation in an Irish context instead of Norway.

 

There were obviously many more than 8 trials of exercise for LC that should meet the very broad criteria they chose, so this is a very selective analysis with poor methodology of a very selective set of trials with very poor methodologies.

I seriously don't understand how this EBM stuff is taken seriously. It's all arbitrary, rules are suggestions at best, no one seems to care that this is completely unreliable and should never be used for anything more important than choosing breakfast.

It's propaganda. This stuff exists to manufacture consent for ideas responding to institutional wants. It has nothing at all to do with scientific research, academic studies or even professional assessment of anything.

Also, following their nonsense, if you ignore the fact that they ignore a major source of bias purely out of convenience, then they don't do that. Just ignore the paragraph where they explain that they do that, and they're not doing that. Screw this, let's everyone just sign contracts and then afterward just choose which parts you want to apply and which don't. Nothing matters anyway.

Gotta laugh at the throwaway line they put out in the notice about the CEO stepping down that she cemented Cochrane's supremacy as the masters of clinical knowledge. What a sick joke this industry is.

 

Could you give an example?