Cardiopulmonary and metabolic responses during a 2-day CPET in [ME/CFS]: translating reduced oxygen consumption [...], Keller et al, 2024

[forestglip]

Screening phases are talked about in the actual study. For instance

They didn't give any specifics about how they determined ME/CFS status apart from what I quoted about being diagnosed and fulfilling CCC. "Had a chat with Dr. Chia" doesn't tell us they used anything more than that.

I'm not sure I understand your point on false positives and the use of the CCC in the ME group. Application of different criteria for the ME group should have no influence on your controls as false negatives don't end up in your control group afaik, the controls are not recruited by not fulfilling certain ME criteria since they are healthy and there is absolutely no evidence to suggest they have ME/CFS.

I don't mean they take people that signed up to be ME/CFS participants and if they didn't fulfill CCC, they shuffled them over to the control group. I mean people who sign up to be in the control group and don't fulfill CCC could be false negatives.

It simply doesn't seem reasonable to me, there's thousands of far more likely reasons why someone would be sedentary rather than being perfectly healthy but yet still somehow having ME/CFS.

I personally don't think it's very unlikely that someone is sedentary because of very mild or undiagnosed ME/CFS, or because the definition of ME/CFS is overly strict and people with the same pathology but different symptoms are being excluded. And I don't think it's typical for "perfectly healthy" people to be sedentary. If they used only people with a definite non-ME reason for being sedentary, that'd make me feel better, for example paralyzed in both legs from an accident and uses an electric wheelchair (although the CPET would have to be an arm exercise).

If a cancer or rheumatic arthritis study is conducted and there is little difference between healthy controls and the illness group, I don't think anybody thinks that would be an effect of everyone in the control group having cancer or rheumatic arthritis rather than the lack of differences being that one didn't observe a key pathology.

It's not likely that everyone in the control group would have cancer, but there's a chance that some members of the control group have undiagnosed or very mild cancer, throwing off the results. It's also very possible that the study wasn't looking at the right thing and the groups are perfect. But there's always a chance of the former as well - our diagnostic tests aren't perfect.

Apart from that, diagnosis in cancer is much different from diagnosis in ME. In cancer they can objectively test for cancer cells. In ME, they rely on subjective questionnaires based on arbitrary groupings of symptoms. People who fulfill IOM or even just have PEM have not been shown to have a totally different disease mechanism from people who fulfill CCC.

As with most things, the distribution of people with ME probably follows a curve of the more severe the disease, the fewer people suffer from that severity. And I don't see any evidence that there's some mechanism that happens that brings people from "perfectly healthy" to "mild ME/CFS with symptom presentation" with no people between those, with subclinical underlying pathology. There are probably very many people in this "in-between" group. If diagnosable ME is at least 0.5% of the population, there's a chance that people in the "in-between" could make up a higher proportion that that. It might not take very many of these people in the control group to make it look like the group has something similar going on.

Maybe it could indeed still be very sensible to have an additional screening round of healthy controls where those that fulfill the weakest criteria, such as Fukuda, are automatically excluded, but I don't think theres much evidence to suggest that something like that would be driving the results, or lack thereof, in this study. But maybe this is indeed a valuable idea, I wouldn't be surprised however if clinicans already automatically rule out such participants in the screening phase.

Yeah, there's a chance the screening meetings did use that. The screeners have a lot of experience in the field (at least Dr. Chia and Dr. Levine, I don't know who Dr. Moore is), so maybe they also think it's prudent to try to minimize false negatives as much as possible.

Certainly not. However, I don't think there is any evidence to suggest that they would have PEM. In my opinion someone that is a healthy control but experiences PEM is unlikely to participate in a study that purposefully puts them into PEM, especially if they are anyways sedentary, i.e. prefer to be less physically active. The reason why pwME participate in such studies is because they want to help advance ME/CFS research and as such endure the PEM, I don't think a healthy control would have a similar motivation.

I'd say being sedentary is evidence to suggest that. Not to conclusively say that they have PEM, but one (of many) potential reasons they are sedentary. I'm arguing that there are probably lots of people with very mild, hard-to-notice PEM or who haven't connected the dots about being tired two days later after an activity to realize they have PEM. And not that everyone in the control group likely has PEM, but that there's a significant chance that at least some do.

 

[forestglip]

I have to agree with ME/CFS Skeptic here that I think the risk of having false positives with a looser ME criteria is far higher (and likelier to mess with results), than having people with PEM in the control group.

I don't mean remove CCC from the inclusion criteria for the ME group. I mean add "not fulfilling CCC, IOM, NICE, Fukuda, or having any evidence of PEM" to the inclusion criteria for the control group.

And for the control group, I doubt any would fit the IOM, as it requires a 50% reduction in functioning. I don’t think people would be recruited as “healthy” controls if they had 50% reduction.

Maybe not, or maybe "healthy" to them might just mean no other diagnoses, and they didn't ask people about their level of functioning.

 

[EndME]

I don't mean they take people that signed up to be ME/CFS participants and if they didn't fulfill CCC, they shuffled them over to the control group. I mean people who sign up to be in the control group and don't fulfill CCC could be false negatives.

That is not what I'm saying either. I'm not familiar with clinical work but I don't think healthy controls are recruited on the basis of not fulfilling the CCC. I'd imagine that is neither how recruitment works nor why HCs participate in this work. They are most likely recruited on the basis of being healthy, including health records and a screening evaluation, not on the basis of not fulfilling the CCC, which is just a natural consequence.

Yeah, there's a chance the screening meetings did use that. The screeners have a lot of experience in the field (at least Dr. Chia and Dr. Levine, I don't know who Dr. Moore is), so maybe they also think it's prudent to try to minimize false negatives as much as possible.

I'd expect that. Otherwise I'd imagine they wouldn't be conducting a CPET study with healthy controls and instead just take arbitrary controls from the general population.

People who fulfill IOM or even just have PEM have not been shown to have a totally different disease mechanism from people who fulfill CCC.

Sure. But I'm not aware of any evidence that this is in any way applicable to the HCs of this study. I don't think HCs are recruited on the basis of not fulfilling the CCC.

And I don't think it's typical for "perfectly healthy" people to be sedentary.

It might depend on the definition and scale of being sedentary, but according to different studies in the US the percentage of people that are sedentary form roughly 30% of the general population (see for instance https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2996155/, https://www.fau.edu/newsdesk/articles/Physical Inactivity-AJM.php, https://www.statista.com/statistics/252063/us-sedentary-lifestyle-among-adults-by-ethnicity/). Even if the percentage is 20% that is a completely different order of magnitude than the amount of people that suffer from ME/CFS. Does that make the worries about the possibilty of false negatives driving results smaller?

 

[forestglip]

That is not what I'm saying either. I'm not familiar with clinical work but I don't think healthy controls are recruited on the basis of not fulfilling the CCC. I'd imagine that is neither how recruitment works nor why HCs participate in this work. They are most likely recruited on the basis of being healthy, including health records and a screening evaluation, not on the basis of not fulfilling the CCC.

Sure. But I'm not aware of any evidence that this is in any way applicable to the HCs of this study. I don't think HCs are recruited on the basis of not fulfilling the CCC.

I assume they initially recruit people who say they are healthy, then they make sure they don't have ME/CFS. I'd like to know what that "making sure" part is like. Whether they question them in detail to determine if they might have PEM that they haven't recognized themselves.

It might depend on the definition and scale of being sedentary, but according to different studies in the US the percentage of people that are sedentary form roughly 30% of the general population (see for instance https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2996155/, https://www.fau.edu/newsdesk/articles/Physical Inactivity-AJM.php, https://www.statista.com/statistics/252063/us-sedentary-lifestyle-among-adults-by-ethnicity/). Even if the percentage is 20% that is a completely different order of magnitude than the amount of people that suffer from ME/CFS. Does that make the worries about the possibilty of false negatives driving results smaller?

Not really, since even if it's 10 or 100 times fewer people that have undiagnosed ME vs. "healthy sedentary" people, there's still a good chance of several making it in to a group of 71.

The exact numbers in the following graph are made up, but this is an illustration showing that the more severe the disease is, the fewer people experience it, which I think is likely for ME. The red threshold line is where symptoms are prominent enough to diagnose. There could be a lot of people on the left side of the red line, greatly increasing the chances of someone appearing in the control group who has minor ME pathology.

 

[Yann04]

the more severe the disease is, the fewer people experience it, which I think is likely for ME

I would dispute that as likely being fuelled by survivorship bias in two ways:
1. People with severe ME are much less likely to go to doctors/participate in studies or to talk about their experiences and be active, because they can’t. So they will be much less seen.
2. The risk of dying from physical complications and from suicide is probably significantly higher in severe ME,

 

[forestglip]

I would dispute that as likely being fuelled by survivorship bias in two ways:
1. People with severe ME are much less likely to go to doctors/participate in studies or to talk about their experiences and be active, because they can’t. So they will be much less seen.
2. The risk of dying from physical complications and from suicide is probably significantly higher in severe ME,

I'm just basing it on what would logically make the most sense. The more extreme a pathology, the less likely someone would be to experience it, as it would take more things going wrong to a greater degree, which is less likely. It could be that's not how ME works. I don't have the energy to dig into other disease distributions, but just some examples from ChatGPT to illustrate the point:

• Infectious Diseases: Many infectious diseases, like influenza or COVID-19, show a distribution where most individuals experience mild symptoms, and a smaller proportion suffer from severe symptoms or complications.
• Chronic Diseases: Conditions like diabetes or hypertension often have many individuals with manageable or mild symptoms, while fewer individuals experience severe complications like organ damage or cardiovascular events.
• Cancer: Most people with cancer may have localized disease (which is often more treatable), while fewer have metastatic disease (which is more severe and harder to treat).

 

[Yann04]

The more extreme a pathology, the less likely someone would be to experience it, as it would take more things going wrong to a greater degree, which is less likely. It could be that's not how ME works.

Not sure that works for ME, as the worsening of the pathology seems to be caused by PEM.

Edit: Also I’m unsure if that is really a universal rule. ALS seems to be more commonly severe while mild cases of it are very rare.

 

[forestglip]

To be clear, even if a few people with sub-clinical ME got into the control group, I don't think that would make it useless. There should still be something like a "dose-response curve" where the people who were severe enough to be diagnosed should show greater reductions than the control group. It just might make it seem like healthy people also have a (smaller) reduction in CPET metrics on day 2, even if in reality, truly healthy people have no reduction, which might make it seem less useful as a biomarker.

 

[forestglip]

Not sure that works for ME, as the worsening of the pathology seems to be caused by PEM.

Well many people are mild and aren't PEMing themselves into very severe ME. They reach some kind of stable ME level. This stable level might have that distribution.

I thought I'd seen some epidemiology about mild vs. moderate vs. severe vs. very severe that showed it going down with severity. I couldn't find it now, and maybe I'm misremembering.

Edit: Also I’m unsure if that is really a universal rule. ALS seems to be more commonly severe while mild cases of it are very rare.

Yeah, there are exceptions, but I think the possibility of the decreasing distribution scenario should be considered, and would probably be more likely, unless some evidence to the contrary arises.

 

[EndME]

I assume they initially recruit people who say they are healthy, then they make sure they don't have ME/CFS. I'd like to know what that "making sure" part is like. Whether they question them in detail to determine if they might have PEM that they haven't recognized themselves.



Not really, since even if it's 10 or 100 times fewer people that have undiagnosed ME vs. "healthy sedentary" people, there's still a good chance of several making it in to a group of 71.

The exact numbers in the following graph are made up, but this is an illustration showing that the more severe the disease is, the fewer people experience it, which I think is likely for ME. The red threshold line is where symptoms are prominent enough to diagnose. There could be a lot of people on the left side of the red line, greatly increasing the chances of someone appearing in the control group who has minor ME pathology.

View attachment 22320

It is not a discussion of undiagnosed ME vs "healthy sedentary" people. It’s people that consider themselves healthy and are evaluated as such and are sedentary, but are extremely happy to participate in a clinical trial which induces activity. If I was a “healthy person” that somehow experiences subclinical ME, whatever that may be, which has made me sedentary without me realising, I certainly wouldn’t participating in a study that induces activity for the obvious reasons that my body doesn’t seem to fare well with it even if I wouldn't have realised that I experience subclinical ME.

I would imagine that completely shifts any mentioned percentages by several order of magnitudes and makes the area presented to the left red line in the graph rather irrelevant as you are looking at a significantly smaller subset of people. Apart from that the graph seems to be rather biased to the left when compared to real world data, for example the data from DecodeME (moderate ME/CFS being roughly 4-5 times more common than severe ME/CFS).

If this is anything to worry about one should be far more worried about all the false positives, which we know exist. According to some post-mortem studies in MS it has been shown that 1 in 20 pwMS in life turn out to have another disease, for instance anti-MOG associated disease. The important point is: Yet clinical trials have easily identified group differences between healthy controls and pwMS.

 

[forestglip]

If I was a “healthy person” that somehow experiences subclinical ME, whatever that may be, which has made me sedentary without me realising, I certainly wouldn’t participating in a study that induces activity for the obvious reasons that my body doesn’t seem to fare well with it even if I wouldn't have realised that I experience subclinical ME.

If I were getting paid (I don't know if they are, but probably?), were helping people with the condition, had very mild fatigue, and didn't know that two days after the CPET I would feel worse because of the CPET, I might. I was doing lots of exertion for no reason before I realized I was paying for it two days later.

I would imagine that completely shifts any mentioned percentages by several order of magnitudes and makes the area presented to the left red line in the graph rather irrelevant as you are looking at a significantly smaller subset of people. Apart from that the graph seems to be rather biased to the left when compared to real world data, for example the data from DecodeME (moderate ME/CFS being roughly 4-5 times more common than severe ME/CFS).

Yes, the graph might not be exactly fit to ME, but even so, there could be a very significant number of people between "healthy" and "very mild".

Honestly, this probably might not matter much if we can find clear differences in reductions between the groups. Would just be a more definitive finding if none in the control group had any reduction, which could be masked by false negatives.

Really, other than that I hope they try to exclude anyone who has PEM but not the other symptoms of CCC, which maybe they do, I don't think the "very very mild" false negatives are a big deal. I strayed too far from my original potential concern about the screening process. I'm not saying there's anything wrong with this study since I don't know if they did that. Just saying I hope they did.

 

[EndME]

I've asked him the following:

https://twitter.com/user/status/1809927601172992229


Could you explain why you think fitness matching is not a good idea? If many CPET measurements are collinear with VO2peak, doesn't this suggest that differences could be due to deconditioning rather than ME/CFS?

I would think that fitness matching makes comparisons more interesting because it is less likely that the differences found are due to deconditioning and more likely that they reflect ME/CFS pathology.

Conversely, abnormalities in ME/CFS patients that are also seen (same size and effect) in healthy but deconditioned controls might be less interesting because they are less likely to represent ME/CFS pathology.

So it seems really important to control for fitness level. Matching patients and controls might not be the best approach for this because it might lead to discarding the most interesting (and abnormal) ME/CFS cases for which there was no match. (Perhaps this is what you meant?)

Perhaps a better approach would be to use all data and simply control for VO2, sex and age using a regression model? Would be interested in hearing your views on this.

I am only getting more confused by Todds words. Is he really saying that PEM is defined as low VO2 and if we don't define PEM like that one is engaging in pseudoscience? Surely he can't be suggesting that...

www.x.com/sunsopeningband/status/1809932614159180160

 

[forestglip]

I'm not really following what he's saying either. But in terms of matching, shouldn't it be established that VO2 level basically equals fitness level and no other disease processes influence the correlation? Otherwise, they're not really matched for fitness level, right? Has that been established?

 

[EndME]

I'm not really following what he's saying either. But in terms of matching, shouldn't it be established that VO2 level basically equals fitness level and no other disease processes influence the correlation? Otherwise, they're not really matched, right? Has that been established?

I don't think VO2 level equals fitness (which might in itself be hard to define), otherwise athletes would just be defined by their VO2 levels plus weight. So there's probably some additional variables (for instance recovery time from VO2 max, lactate threshold, muscular pain during exercise in ME/CFS or whatever else might exist), but it's probably an extremely good indicator that is heavily correlated to fitness levels. So it would seem possible to me that disease processes can influence this correlation (someone that experiences more PEM might have different lactate levels or a worse recovery time from exertion or experience more muscular necrosis or whatever might be going on in ME/CFS).

I'm not sure that you'd want to necessarily match fitness or at least not only fitness. 2 healthy people with the same body weight not doing any different physical activity can also have a different VO2max simply because one of them grew up at altitude or because one of them has larger lungs but they might be similarly conditioned, which might be the more fruitful thing that you would want to match if you want to see that differences between ME/CFS and HCs aren't driven by deconditioning. Ideally you’d use something like step-count, or a more accurate measure, that somehow captures physical conditioning and then compare similarly conditioned ME/CFS patients and HCs to see if results are driven by deconditioning. However, this may be difficult if ME/CFS patients always have a lower step-count than HCs.

But I don't see how this would have to be a major obstacle. All you need to do is capture “deconditioning in HCs” for example taking a range of fully fit to completely sedentary HCs. If the differences in your experiment between ME/CFS patients and HCs looks equivalent to the difference of fit HCs vs deconditioned HCs and less deconditioned ME/CFS patients look more similar to less deconditioned HCs whilst the most deconditioned ME/CFS deliver the most extreme results and seem similar to HCs of similar condition, than it might seem more likely that you are studying something that might be a consequence of conditioning rather than anything else, i.e. you analyse all the various variables and then see what the outcome is. Are the experimental results better sorted by the level of "conditioning" independently of whether someone is a patient or not or does having ME/CFS created the more substantial separation?

These are just my random and completely uneducated thoughts. Ideally an exercise scientist would comment on this, unfortunately I am not able to understand what Todd is saying.

 

[forestglip]

I think we agree, but the paper seems to equate them:

Abnormal post-exertional CPET responses persisted compared to CTL matched for aerobic capacity, indicating that fitness level does not predispose to exertion intolerance in ME/CFS.

Todd said:

All “fitness matching” is doing is comparing the best-off patients with the worst-off controls.

And I don't know why he's sure that's the case, but it seems like a distinct possibility since VO2 isn't precisely tied to fitness.

 

[EndME]

So the paper matched them according to fitness which they defined via aerobic capacity, but Todd thinks this is a bad idea because fitness is not a measure of conditioning. However, at the same time he also thinks that aerobic capacity is equivalent to PEM. I guess I sort of understand his first point then but not the second point. He mentions that in a CPET all measures are highly correlated with sex and aerobic capacity. But I don't understand why that has to destroy your analysis. He seems to say that in case you're analysing your data via ANOVA that will be the case.

I'm not sure what Todd means when he says things such as

Finally, matching on collinear variables is reducing group differences because they’re correlated as a statistical artifact instead of physiological fact.

Supposedly you're matching sex, age and aerobic capacity which are correlated due to some underlying physiological phenomena related to fitness. I don't see that being a statistical artefact, whatever he may mean with that. What's the problem with any of that? You're only matching according to the aerobic capacity of the first CPET and then want to see whether these pairs matched according to "fitness" yield different results in the second CPET. Isn't that sort of what you want to do (of course preferable using condition instead of fitness as match)?

 

[DMissa]

"A recent study of post-infectious ME/CFS (N = 17) and healthy controls (N = 21) assessed a comprehensive panel of physiological, physical, cognitive, biochemical, microbiological, and immunological variables [84] (wallitt paper). Of these measures, only 8 ME/CFS and 9 controls completed a single CPET with an average VO2peak about 40% higher in the control group. Based on this small sample size and inappropriately matched control group, authors suggested that impaired ANS function in ME/CFS, evidenced by diminished HRV, abnormal tilt-related symptoms, and other abnormal orthostatic responses, led to lower metabolic energy production and work output, and may be contributed to by a reduced ‘effort preference’. Effort preference was assessed in this study using the Effort-Expenditure for Rewards Task [85], which utilized a small motor task to assess for anhedonia typically associated with major depressive disorder. The Effort-Expenditure for Rewards Task is not highly associated with measures of whole-body oxygen consumption or power output compared to conventional indices of effort (%peak HR, RER, RPE), none of which were reported in the study. Whereas the link between ANS dysfunction and impaired energy metabolism is not inconsistent with the systemic CPET data reported herein, their reasoning is misguided. It has long been known that the magnitude of cardiovascular responses to exertion is predominantly influenced by the relative level of muscle activation (number and intensity of activated muscle fibers) via feedback loop from peripheral interoceptors (e.g., Golgi tendon organs, muscle spindles, etc.) to the motor cortex then to the brainstem [85]."

pew pew pew

 

[SNT Gatchaman]

Thanks for highlighting that @DMissa. Paper gains an extra star!

 

[Dolphin]

Not much of any depth:
https://neuroimmune.cornell.edu/news/

https://twitter.com/user/status/1810718913157886098

 

[John Mac]

Cort Johnson has an article on the study

https://www.healthrising.org/blog/2024/07/10/validated-cpet-energy-chronic-fatigue-syndrome/