Physical function and psychosocial outcomes after a 6-month self-paced aquatic exercise program for individuals with [ME/CFS], 2025, Broadbent+

[SNT Gatchaman]

Physical function and psychosocial outcomes after a 6-month self-paced aquatic exercise program for individuals with myalgic encephalomyelitis/chronic fatigue syndrome
Broadbent, Suzanne; Coetzee, Sonja; Calder, Angela; Beavers, Rosalind

PURPOSE
A randomized-controlled trial to investigate the efficacy of a 6-month self-paced aquatic exercise intervention on physical function, symptoms and psychosocial measures in individuals with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS).

METHODS
Thirty-two individuals diagnosed with ME/CFS (55.0 ± 13.9 yr) were randomized into an intervention group (INT, n = 17) or control group (CON, n = 15) for a 6-month trial of two 20-min sessions per week of self-paced aquatic movements and stretches. Pre-and post-intervention outcomes included physiological measures, 6-min walk test, hand-grip strength, Sit-to-Stand, Apley’s shoulder test, Sit–Reach test, perceived exertion, fatigue (FACIT), anxiety/depression (HADS) questionnaires, and tiredness and pain scores (VAS 0–10 scale).

RESULTS
The INT group significantly increased walk test distance (13.7%, P < 0.001), Sit-to-Stand scores (33.7%, P < 0.001) and peak expiratory pulmonary flow (12.9%, P = 0.028) post-intervention. Fatigue (29.5%, P = 0.005), depression (21.7%, P = 0.010), combined anxiety/depression scores (16.9%, P = 0.047) and resting diastolic blood pressure (4.8%, P < 0.001) also significantly improved for the INT group. Sit–Reach scores were significantly lower for the INT group compared to CON post-intervention (− 4.0 ± 10.4 vs + 4.3 ± 10.7 cm, P = 0.034). There were no adverse events or worsening of symptoms during the trial.

CONCLUSIONS
Self-paced, low–moderate-intensity aquatic exercise improved walk distance, lower limb strength, fatigue, depression and peak expiratory flow without worsening ME/CFS symptoms. This mode of low-intensity physical activity may confer mental health and physical benefits provided the activity is self-paced and within patient energy limits.

TRIAL REGISTRATION NUMBER
Australian and New Zealand Clinical Trials Registry ANZCTRN12618001683224.

Link | PDF (European Journal of Applied Physiology) [Open Access]

 

[SNT Gatchaman]

Inclusion criteria were a medical diagnosis of ME/CFS according to the Canadian Consensus Criteria (or the Fukuda Criteria if that was standard at the time of diagnosis); aged from 18 to 80 years

Exclusion criteria included […] neurological and/or inflammatory conditions that made physical activity hazardous; […] severe ME/CFS with extreme symptoms that incapacitated the individual

 

[dave30th]

So they included patients with required PEM per CCC (although optional with Fukuda) and then excluded patients with PEM?

 

[Hutan]

Recommendations for physical activity for ME/CFS have been fraught with controversy in previous years. Early stud- ies (White et al. 2011) and systematic reviews (Van Cauwen- bergh et al. 2012; Larun et al. 2017) that reported “graded exercise therapy” (GET), largely consisting of aerobic exer- cise of up to 30 min duration with fixed incremental pro- gressions, as safe and beneficial for ME/CFS, are no longer supported.

Current National Institute for Health and Care Excellence and Mayo Clinic guidelines (NICE 2021; Bate- man et al. 2021) outline best-practice recommendations for physical activity (PA) which must be individualized; under- taken within a ME/CFS management program and overseen by a trained exercise therapist; within each person’s energy limits; and flexible, allowing for regression and progression only if this is within energy limits and does not worsen the patient’s symptoms.

I suppose that is something.

And good that the trial was 6 months long.

15 controls and 17 intervention - small sample size

 

[Hutan]

Nearly 75% of total participants were female; the gender com- position of the INT group was 12 females and 5 males, and the CON group 12 females and 3 males.

Regarding physical activity (PA) levels prior to study entry, 31.25% of the cohort were doing either no PA or only essential daily activities; 59.4% were able to manage short walks; 18.6% did some gardening; 12.5% engaged in yoga and one person did a weekly Pilates session.

So, most of these participants were already doing some physical activity beyond essential daily activities.

The weight and BMI both increased in the intervention group, not significantly, but there was certainly no decrease as might have been expected if they had increased their physical activity with the two exercise sessions per week and no off-setting reduction in other activities. The mean BMI was in the overweight category both before and after the intervention. I'd be pretty disappointed if I added two sessions of physical activity to my week for 6 months and gained weight.

One physical characteristic that did change significantly was resting diastolic blood pressure. The BP changed from 129/79 to 125/71 (systolic/diastolic). I don't think that a DBP change from 79 to 71 is particularly relevant to health. And, with such a small sample size and the way blood pressure can bounce around even in healthy people, I don't think that means anything.

The other physical characteristic that did change significantly was lung peak expiratory flow (6.2 to 7.0). And fair enough, that's a good thing, although both values are pretty close to the average for a 55 year old woman, which is what the average participant was. And the p value was only 0.028, so not a very significant difference. There was no significant change in the other measures of lung function.

Resting heart rate didn't change. Resting heart rate should decrease as someone becomes fitter.

The post-between group P values are not significant for any measure in Table 1. (Annoyingly, they don't present stats on the significance of the between group change in a measure.) There is therefore little evidence that this intervention made the physical outcomes presented in Table 1 better; there's little evidence that the participants became fitter or had healthier BMIs after the 6 months.

 

[Hutan]

Onto Table 3

The only physical outcome measure on this table that had a significant post-between group P value was the sit-reach test. This is a test of touching your toes. After the 6 months, the control group was on average a bit better able to do this than the intervention group. I don't think it means anything much.

The intervention group had significant improvements in the 6-minute walk test distance and the sit to stand reps, but they still didn't look significantly different to the controls. You really can't make any claims of intervention value based on that. I mean, what is the point of having the controls (even standard care controls as are used here) if you ignore them and just look at the difference before and after the intervention?

 

[Sean]

This mode of low-intensity physical activity may confer mental health and physical benefits provided the activity is self-paced and within patient energy limits.​

Which is more-or-less what we have been saying all along. Let us figure out the best activity patterns and adaptation strategies for ourselves.

The clinician's role is supportive, not directive.

 

[Hutan]

Psychosocial measures

These outcome measures (median and SD) are shown in Figs. 1, 2, 3, 4. Individual pre-to-post responses are shown in Supplementary Figs. 9–11. The INT group significantly improved their FACIT-scored fatigue (P = 0.005, ES 0.6) pre- to post-intervention and also had reduced fatigue com- pared to the CON group at that timepoint (P = 0.003). The INT group total HADS score (combined depression and anx- iety, P = 0.047, ES 0.3) and HADS depression (P = 0.021, ES 0.3) were significantly reduced in pre- to post-intervention. Of the participants who were taking anti-depressants (40.6% of the cohort), all participants in the INT group reduced their total HADS scores compared to those in the CON group, and all but one INT participant reduced their HADS Depression component score. (Supplementary Figs. 9–11).

So, a psychosocial benefit really seems to be the best they have got from what surely must have been a rather expensive intervention. (The intervention involved time at a swimming pool twice a week with two instructors.) You've got to wonder if it might have been cheaper to have the participants just meet somewhere for a walk, or even just a chat. Perhaps the benefits had nothing much to do with the exercise and quite a lot to do with developing a support group and perhaps taking pleasure in being part of an experiment.

And, even then, the psychosocial benefits are pretty marginal. (Actually after looking at the results closely, I'd say the benefits are more like almost non-existent.)


There's the FACIT scores.
Functional Assessment of Chronic Illness Therapy – Fatigue (FACIT) (Cella et al. 2019) - higher scores means less fatigue
from left to right - INT baseline; CON baseline; INT 6 months; CON 6 months

The biggest issue is the difference between the groups in reported fatigue at the baseline. The controls had a higher level of fatigue.
After the 6 months, the controls still had about the same level of fatigue, while the intervention group had less fatigue (a change of about 5 points on a scale 0-52). But the improvement in the intervention group was about the same as the difference between the two groups at baseline. It's really nothing much and entirely within the realms of an effect of expectation/pleasing the therapists and researchers.


And there's the HADS total scores (anxiety and depression)
Same order: INT baseline; CON baseline; INT 6 months; CON 6 months
The means of all of those, pre and post, are pretty much the same. Both groups had a very small reduction in the scores. There really is nothing there.

The authors mention a reduction in the HADS scores for the intervention group:

The INT group total HADS score (combined depression and anx- iety, P = 0.047, ES 0.3) and HADS depression (P = 0.021, ES 0.3) were significantly reduced in pre- to post-intervention.

but do not acknowledge that the control group, presumably inexpensively sitting at home alone for 6 months, achieved the same sort of reduction.

I don't think it was any accident that the results were not presented in tables like the other measures. It expect it would have shown that the controls also had a significant improvement (it's probably in a supplementary table). Again, there's no point in having a controlled trial if you are going to ignore the controls in the analysis.

 

[Hutan]

This study does not prove that the people in the programme increased their physical activity or their fitness.

It provides some evidence that some people with ME/CFS can undertake a twice weekly exercise programme without coming to harm provided the approach of the instructors is cautious and the participants are able to adjust their attendance and exercise intensity depending on how they feel.

I don't think the study proves that there were physical benefits.

I don't think it proves that there were psychosocial benefits.

I think it provides some evidence that the cost of such interventions outweighs the benefits.

I think the abstract is misleading and the authors need to work on their statistical skills, specifically how to make use of data provided by a control group.

 

[Utsikt]

Participants in both INT and CON groups were asked to complete daily diaries for the duration of the study with the aim of recording fatigue and other symptoms, energy levels, sleep and other factors which might affect their health. These were to be collected post-intervention.

Within one-week post-intervention, physical and psychosocial assessments were completed at the exercise science laboratories and blood tests at the pathology centers. Diaries were collected and participants completed a study exit survey with the researchers. This qualitative data will be published separately.

It’s a shame the didn’t do objective activity monitoring. And it looks like there will be more publications.

 

[Andy]

Related previous threads
Patient experiences and the psychosocial benefits of group aquatic exercise to reduce symptoms of ME/CFS: a pilot study, 2020, Broadbent et al
Chronic fatigue syndrome participants wanted for aquatic exercise trial Published 22 February 2019

 

[Simon M]

I’ve only had the energy to skim the thread and the paper, so this might be a little off track, or repeating what’s already been said.

But I think this study is a step forward, for a couple of reasons:
– it uses several sensible objective physical measures Which we’ve been calling for for such a long time. The six minute walking test was the most useful data in the Pace trial Would be nice to see a Fitbit type measure as well, but let’s not quibble.

– As @Sean says, they’re using a self-directed method, which is a massive shift on what’s gone before, and again what we have said is more appropriate. I think it’s helpful to Tryout such an approach..

As @Hutan says, I don’t think we can conclude much from the study. My big concerns are:
– it’s a very small cohort, more of a pilot study than anything. And with a mean age of 55, much older than typical clinical trials (according to my memory, so maybe not)
– The relevantmeasure here is surely the group difference in pre-post differences. I’m sure that’s what the Pace Trial did And they’ve relied on one measure or the other, which I don’t think is Useful.. Though I don’t know how you would calculate Statistical significance on the former.

But I’d be happy to see the authors do a bigger and better version of the study

 

[Nightsong]

Small sample size (32) that is significantly lower than the anticipated sample size at trial registration (80). Participants were asked to rate immediate and 24-hour post-exercise VAS scores for tiredness and pain even though, with ME/CFS, PEM symptoms may have emerged later. No attempt at correction for multiple comparisons. Control group was care-as-usual. Fukuda was considered acceptable if it were used at time of diagnosis without evaluating whether the patients currently meet CCC or any other modern ME/CFS definition.

 

[Trish]

I have skimmed the article and looked at the data. As far as I can see there's no correction for multiple comparisons, and the ones that seemed to produce impressive p values for within group improvements were, if you look at the data, well within the range of variation up or down in the control group with no intervention, ie well within what would be expected from normal fluctuations.
Also I couldn't see any mention of assessing the participants' overall physical activity for the duration, and whether that changed. Did they, for example, replace some household duties or other forms of activity with the pool exercises and follow them with days with more rest?

For the social/depression etc measures, the control group should have been better matched, perhaps with a meet and chat session with a supportive leader provided with the same frequency and duration as the intervention group.

 

[Utsikt]

Small sample size (32) that is significantly lower than the anticipated sample size at trial registration (80).

That’s weird. It says this in the paper:

A required sample size of 34 was estimated using a G*Power (V3.1.9.7) analysis to achieve a power of 80%. With a drop-out of two participants from the CON group, the final cohort size of 32 was considered still sufficient to detect a medium effect size of 0.5 and significant differences with P < 0.050.

Also:

Participants were asked to rate immediate and 24-hour post-exercise VAS scores for tiredness and pain

Protocol:

Fatigue and tiredness will be recorded using a 0-10 Likert Scale after each session and 24- and 48-hr post-session.

 

[rvallee]

The intervention group had significant improvements in the 6-minute walk test distance and the sit to stand reps, but they still didn't look significantly different to the controls. You really can't make any claims of intervention value based on that. I mean, what is the point of having the controls (even standard care controls as are used here) if you ignore them and just look at the difference before and after the intervention?

I don't think it was any accident that the results were not presented in tables like the other measures. It expect it would have shown that the controls also had a significant improvement (it's probably in a supplementary table). Again, there's no point in having a controlled trial if you are going to ignore the controls in the analysis.

I think the abstract is misleading and the authors need to work on their statistical skills, specifically how to make use of data provided by a control group.

 

[rvallee]

It provides some evidence that some people with ME/CFS can undertake a twice weekly exercise programme without coming to harm provided the approach of the instructors is cautious and the participants are able to adjust their attendance and exercise intensity depending on how they feel.

That interpretation is mostly invalidated by scores of exercise studies that were more directive in pushing, yet despite the presence of PEM, did not report any significant worsening. It clearly can't have much to do with adaptive intensity, or something like. Even though it does, it's just that the problem is far more fundamental than this.

All of this points to a general inability to assess outcomes. If no one ever reports worsening after even intense exercise when the literal thing being studied is worsening function after mere exertion, then the entire assessment process is simply as useless as a quality control process that consists of sticking "QC passed" stickers on items without bothering to test anything.

People can literally become bedbound and lose function to the point of dying from negligent starvation, and it's not considered worsening. Simply because they don't understand the reason for the deterioration. Truth is that such data are simply uninterpretable, because the people doing the interpretation have no idea what they're dealing with. They may as well have people try to interpret radio signals in the 15th century.

 

[Hutan]

I agree to some extent @rvallee, but they delivered this intervention for 6 months. I'm not entirely sure about dropouts, but they report a 75% attendance rate. It seems pretty solid that the participants were able to sustain attendance, although it would be good to see if people were missing more sessions later in the trial.

A big issue that I and others have already noted is that we can't know if participants offset the effort of attending by reducing other exertions. The lack of much sign that the participants become physically fitter suggests that at least some probably did. If a participant replaced their yoga or pilates with aqua aerobics, there's less likelihood of a real benefit (and the study does note that some participants were doing yoga or pilates at baseline). If the participants reduced time spent on cooking good food or keeping their house clean and tidy or playing with their children, there might actually be physical and/or psychological harm.

But I think this study is a step forward, for a couple of reasons:
– it uses several sensible objective physical measures Which we’ve been calling for for such a long time. The six minute walking test was the most useful data in the Pace trial Would be nice to see a Fitbit type measure as well, but let’s not quibble.

I agree that there are good things, important good things, about this study, including the long intervention period and the understanding that exercise must be symptom contingent.

I personally don't think the six minute walking test is a good objective measure for people with mild ME/CFS, the people who are likely to choose to participate in a study like this. Unless a person slides down to a lower level of severity where walking becomes difficult, I don't think a person with mild ME/CFS on a good day is hindered when walking for 6 minutes. It therefore becomes more of a subjective outcome - how determined are you to walk fast and keep up that pace? If you come to like your instructors and want to help the research, you will probably try harder at the end pf the trial than at the beginning.

There's also a learning effect. If people haven't done a 6 minute walking test before, it can take a couple of goes to work out how best to do it. (e.g. how close to running can you go?) The fact that the 6 month results for both the intervention and control groups didn't look different suggests that at least part of the improvement in the intervention group was a learning effect. I've tried doing the test myself at home, I found a learning effect.

I don't see a call for a Fitbit measure as a quibble. I think we need to see that participants did in fact increase their physical exertion.

Further discussion about the six minute walking test has been moved to Six minute walking test

 

[jnmaciuch]

I have skimmed the article and looked at the data. As far as I can see there's no correction for multiple comparisons, and the ones that seemed to produce impressive p values for within group improvements were, if you look at the data, well within the range of variation up or down in the control group with no intervention, ie well within what would be expected from normal fluctuations.

I suspect multiple testing correction wasn't done because each outcome was treated as a separate hypothesis, and trying to reduce the rate of type I errors (false positive) would substantially increase the rate of type II errors (false negative) in such a small cohort size.

The decision to treat the outcomes as separate is, in my opinion, understandable even if it's not the most robust choice. In my statistics classes we've often discussed how much the logic of multiple testing correction actually applies to a scenario like a clinical trial measuring multiple outcomes, especially if there is expected heterogeneity within those outcomes like there would be in ME/CFS.

If I was doing the analysis I might do additional correction for the most closely related outcome measures (i.e. the three measurements related to 6MWT, or even all the physical challenge measurements). Doing a correction for all measured outcomes would probably have made it impossible to detect any change whatsoever unless it was so dramatic that the participants basically recovered from ME/CFS.

They would've made that kind of choice before doing the analysis, and I don't see other red flags indicating that they switched up their analysis choices last minute a la the PACE trial.

Fully echo the rest of your concerns, though.

 

[jnmaciuch]

The relevantmeasure here is surely the group difference in pre-post differences. I’m sure that’s what the Pace Trial did And they’ve relied on one measure or the other, which I don’t think is Useful.. Though I don’t know how you would calculate Statistical significance on the former.

Looking at the tables they actually did calculate between group differences pre and post intervention (separately for each timepoint). That would basically address what you're asking here, provided the pre intervention differences were always non-significant.

I'm guessing they didn't highlight that part of the analysis since it was overwhelmingly non-significant...

In fact, since those between group comparisons were non-significant, that kind of makes the within-group pre vs. post p-values unimpressive (dovetailing with @Trish's and @Hutan's earlier points). You'd want the improvement to be strong enough that variance in baseline levels across control and intervention does not obscure your finding. And they'd already showed in the CON vs. INT pre-intervention comparisons that the group baselines do not significantly differ.