What are the necessary conditions and criteria for a theoretical model of ME/CFS?

[MelbME]

There are a couple of groups who have sort of gone down this route. Programms such as LIINC have been setup to track people right from their Covid infection until whatever outcome they arrive at and the recent ME/CFS biobank in Germany is also setup to do something similar.

However, some of those studies have shown that even people that meet the CCC definition at 6 months or at least look very similar to ME/CFS often turn out to look rather different at a later time, as for example seen in Long-term symptom severity and clinical biomarkers in post-COVID-19/chronic fatigue syndrome:results from a prospective observational cohort. This German group surrounding Scheibenbogen is also tracking ME/CFS post EBV, but those studies are probably extremely underpowered to reveal much from an epidemiological stand-point as seen in One-Year Follow-up of Young People with ME/CFS Following Infectious Mononucleosis by Epstein-Barr Virus.

If the studies are anyways severly underpowered from an epidemiological stand-point, for example to figure out how large of a risk factor EBV infection is for ME/CFS, I do agree that it might just be sensible too just look at 1000 ME/CFS patients with a longer disease history and do a deep phenotyping study of those patients.

I do wonder though why there is so much focus on trying to look at signatures that seperate somewhere around 90% of ME/CFS patients from healthy controls and then often do so via something like a random forest classifier that just ends up looking rather random, finding marginal differences in a larger set of people without telling us anything fruitful.

I'm far more interested in a result that shows stark differences between two groups and that tells us something meaningful about pathology even if that result only applies to say 50% of ME/CFS patients. But then I wonder how large the sample set of patients really has to be. Depending on what is done it might sometimes be more important to have someone that look at a problem under a novel angle. There are certainly things wrong with the intramural study by the NIH, however what it did show us that a study expecting somewhere around 90% accuracy from a signature is destined to fail from the get go if it is really supposed to distinguish ME/CFS from HC.

We have a paper submitted to a Nature journal that you will probably be interested in. We're hoping it gets published soon. We looked at common co-morbid conditions in an ME cohort and then compared ME to cohorts of that co-morbidity to see if we could find differences that had more to do with an ME signature than the co-morbidity. We also looked for a diagnostic signature that would separate ME from these comorbid diseases. We got one that was ~75% accurate. So we think there is potential here.

 

[EndME]

We have a paper submitted to a Nature journal that you will probably be interested in. We're hoping it gets published soon. We looked at common co-morbid conditions in an ME cohort and then compared ME to cohorts of that co-morbidity to see if we could find differences that had more to do with an ME signature than the co-morbidity. We also looked for a diagnostic signature that would separate ME from these comorbid diseases. We got one that was ~75% accurate. So we think there is potential here.

Thanks a lot for sharing! I'm very much looking forward to reading it once it is published.

With AI classifiers seemingly becoming the norm, are you aware of any work going into something along the lines of "instead of trying to distinguish as many LC or ME patients from HC as possible, even if the differences in actual measurement values might be very marginal and might require combinations of things that apparently aren't pathological connected (say:cortisol+EBV), we are looking at trying to find a diagnostic signature that yield relatively consistent values in HCs and can explain pathology but who's measurement values differ drastically in LC or ME from HC, but only does so for a subset of patients".

The danger of this approach is of course that one could be identifying differences attributable to co-morbidities rather than "true" ME or LC, but it still seems worthy exploring for me, especially if one then can then look at it in a follow-up study in a similar fashion to the one you submitted to a Nature journal.

In the past, I have been rather perplexed when even the most brilliant researchers have made claims surrounding their diagnostics, for example claims surrounding "cortisol being a biomarker for LC" when they should be well aware that AI classifiers cannot distinguish signal from noise but are simply very good at finding structures independently of whether those structure represent signal or noise and have been wondering whether there shouldn't be far more focus on the explainability power of a diagnostic signature rather than purely being focusing on its differentiating capacity and gobbling up those p-values that some statistican throws at you.

 

[MelbME]

Thanks a lot for sharing! I'm very much looking forward to reading it once it is published.

With AI classifiers seemingly becoming the norm, are you aware of any work going into something along the lines of "instead of trying to distinguish as many LC or ME patients from HC as possible, even if the differences in actual measurement values might be very marginal and might require combinations of things that apparently aren't pathological connected (say:cortisol+EBV), we are looking at trying to find a diagnostic signature that yield relatively consistent values in HCs and can explain pathology but who's measurement values differ drastically in LC or ME from HC, but only does so for a subset of patients".

The danger of this approach is of course that one could be identifying differences attributable to co-morbidities rather than "true" ME or LC, but it still seems worthy exploring for me, especially if one then can then look at it in a follow-up study in a similar fashion to the one you submitted to a Nature journal.

In the past, I have been rather perplexed when even the most brilliant researchers have made claims surrounding their diagnostics, for example claims surrounding "cortisol being a biomarker for LC" when they should be well aware that AI classifiers cannot distinguish signal from noise but are simply very good at finding structures independently of whether those structure represent signal or noise and have been wondering whether there shouldn't be far more focus on the explainability power of a diagnostic signature rather than purely being focusing on its differentiating capacity and gobbling up those p-values that some statistican throws at you.

Well one does "sell" the impact of work to get published in a good paper from time to time and to help get future grants/funding, that's probably most of the reason. And maybe sometimes they're just not as experienced in the diagnostic space. In 2015 we did the first metabolomics profile of ME vs controls, decent numbers for the time. We could get 100% separation of groups using 6 metabolites. We did explore the idea of a biomarker but in that process quickly realized how far away we were (had no mechanism we were building off and hadn't validated it nor compared it to disease cohorts). I was mid-PhD but it was a good learning process to at least speak to people in the diagnostic area about how realistic the chances were.

POT (postural orthostatic tachycardia) occurs in healthy controls. POTS is when it is produces symptoms chronically and comes with an array of symptoms that aren't obviously related. So is that what you refer to as a pathology. I mean why not separate patients by comorbidities?

 

[alex3619]

But would this design be more fruitful than deeply characterising 1000 ME/CFS patients to look for pathologies and signatures to cluster?

I'm far more interested in a result that shows stark differences between two groups and that tells us something meaningful about pathology even if that result only applies to say 50% of ME/CFS patients.

A prospective study would start with healthy people and test them every time they get sick, or at least a blood sample. No waiting. At even six months the key biochemical processes may be much less apparent. This should probably be a much larger study looking at a much larger range of diseases. No point in just tracking ME or LC, especially since funding might be easier.

 

[EndME]

A prospective study would start with healthy people and test them every time they get sick, or at least a blood sample. No waiting. At even six months the key biochemical processes may be much less apparent. This should probably be a much larger study looking at a much larger range of diseases. No point in just tracking ME or LC, especially since funding might be easier.

From my understanding that is essentially what LIINC has done for LC.

It was done for EBV and MS and took somewhere around 40 years and more than 10 million participants (and yet one tiny wrong test result in that study completely changes the risk factor they had calculated). Allegedly those authors thought about ME at the same time but decided against it due to the necessary efforts and unknown results. With Covid known to cause ME/CFS as well you'd likely have to control for Covid infections as well and I think that is extremely hard.

It's just very hard to get sensible data on something that affects almost everyone (EBV & Covid) but where only a much smaller minority develop symptoms afterwards (ME/CFS), especially when you don't even know whether negative controls are actually negative controls.

It seems more sensible to me to just start off with those that are already affected. Possibly someone has a clever way around those problems but I haven't seen anyone suggest anything sensible.

 

[MelbME]

From my understanding that is essentially what LIINC has done for LC.

It was done for EBV and MS and took somewhere around 40 years and more than 10 million participants (and yet one tiny wrong test result in that study completely changes the risk factor they had calculated). Allegedly those authors thought about ME at the same time but decided against it due to the necessary efforts and unknown results. With Covid known to cause ME/CFS as well you'd likely have to control for Covid infections as well and I think that is extremely hard.

It's just very hard to get sensible data on something that affects almost everyone (EBV & Covid) but where only a much smaller minority develop symptoms afterwards (ME/CFS), especially when you don't even know whether negative controls are actually negative controls.

It seems more sensible to me to just start off with those that are already affected. Possibly someone has a clever way around those problems but I haven't seen anyone suggest anything sensible.

Well I think characterising 500-1000 ME/CFS would give a great overview of the heterogeneity and it would be a target for developing severity markers.

Obviously decode targeted 20k+ people, this wouldn't be to that level but could be a start. Need more of these large-scale data projects.

 

[Creekside]

This report caught my attention: https://www.sciencedaily.com/releases/2024/07/240726113357.htm It's about how microglia connect to neurons via tunneling nanotubes to pump waste proteins out and mitochondria in. I've thought up a model for ME based on that.

Microglia are a good candidate for playing a major role in ME, connecting immune system activation with neurological symptoms. When a glial cell connects to a neuron with a nanotube, it transfer proteins and mitochondria. That makes it likely that other material is transferred: signalling molecules, building blocks, waste molecules; held in vesicles or floating free. ME might alter the contents of microglia, and thus what gets transferred to neurons. Immune activation in the body makes that worse, so PEM might be the result of more of <whatever> being transferred into neurons, hampering their function, resulting in symptoms. If this nanotube activity depends on neural activity, that could explain why cognitive exertion results in the same PEM symptoms, but with a shorter delay.

Another complication: microglial may use these nanotubes to connect with each other, in response to stressors. Maybe ME affects how microglia connect with each other and with neurons. Maybe ME makes them connect with each other, leaving neurons with excess waste. Maybe they connect too much with neurons, preventing the microglial interconnections necessary for their proper function.

Now that these nanotubes have been discovered, models such as this can be tested. Does nanotube activity depend on neural activity? Does the rate and/or duration of connections between different cells change in ME? Measuring what gets transferred might be tricky, since these nanotubes are tiny and probably short-lived (maybe microseconds?).

Any thoughts? Assuming your tunneling nanotubes are working properly...

 

[SNT Gatchaman]

I thought that paper was worth a thread @Creekside.

 

[Tom Hill]

Something I've been meaning to work out for a while. For any theoretical model explaining the mechanisms of ME/CFS, there are necessary conditions that must be met, without which a model simply cannot account for the data. What are those? Especially with a purpose to falsify flawed models and hypotheses.

Basically it would serve as a checklist for any theoretical model, where if it fails to meet one of those conditions, it can safely be discarded. This isn't about hunches or personal hypotheses, but about making sense of the data. In science, when data contradict a model, the model must be thrown out. We now have more data than ever, thanks to Long Covid.

For example, we know that several pathogens can trigger ME/CFS, thus any model relying on a single pathogen (e.g. HIV -> AIDS) can be ruled out. Similarly, bacteria can also cause ME/CFS, and so it cannot be about a single type or family of viruses or bacteria. The idea that a pathogen never before encountered by humans is also popular, but can also be discarded either way, especially as reinfections with COVID can also cause ME/CFS, as can the very first one.

Another example, a traditional biopsychosocial model is that it's deconditioning. Well, deconditioning cannot fluctuate, and a necessary condition for a theoretical model of ME/CFS is that fluctuations are not just common, but can be very rapid, cumulative and occur even as a result of mild fitness training. Deconditioning simply cannot account for that, and thus can be dismissed as a valid hypothesis.

Similarly, people who have never heard of ME/CFS, or even of chronic illness in general, as well as people who did but did not believe in them, can develop ME/CFS, which discounts any anxiety/fear model with anticipation of possible lifelong illness affecting behavior.

Many very fit people have developed ME/CFS, which discounts any model involving prior sedentary behavior, or inability to understand or work out how to be active, having to be 'coached' into 'learning simple walking' and other things.

Even when obviously triggered by an infectious illness, severity of acute illness does not seem to predict likelihood or severity of the illness, this is overwhelmingly obvious with COVID. Also, even when not obviously triggered by an infectious illness, subsequent infectious illness often makes it worse, including mild ones. However, sometimes mild infections can make the illness better, usually temporarily. Any model has to, if not account for it, at least not contradict it.

ME/CFS affects men and women, children and adults, and so any valid theoretical model cannot require things like hormonal changes happening later in life. Similarly, there have been poor studies showing elevated childhood adversity, but people with perfectly happy childhoods have also developed it, by which this hypothesis can be safely discarded.

Cognitive exertion can be a trigger for PEM/PESE, just as much as physical exertion, and so any valid model cannot depend on physical exertion involving only muscles or locomotion.

Remission and recovery can be spontaneous, even rapid, which discounts many hypotheses that can only unfold over the long term, and vice-versa. Some remissions and recoveries unfold over many months and years, and any valid model must account for it, or at least not make those data impossible.

The heart of the scientific method is falsification. For example, any hypothesis requiring that only women can be affected can be thrown out by finding a single man. Finding false positive evidence is often easy, but isn't the way to do science. Rather, we posit things that are impossible if a model is true, and if found, then we can discard it.

Let's do some effing science!

Super.

 

[Tom Hill]

Something I've been meaning to work out for a while. For any theoretical model explaining the mechanisms of ME/CFS, there are necessary conditions that must be met, without which a model simply cannot account for the data. What are those? Especially with a purpose to falsify flawed models and hypotheses.

Basically it would serve as a checklist for any theoretical model, where if it fails to meet one of those conditions, it can safely be discarded. This isn't about hunches or personal hypotheses, but about making sense of the data. In science, when data contradict a model, the model must be thrown out. We now have more data than ever, thanks to Long Covid.

For example, we know that several pathogens can trigger ME/CFS, thus any model relying on a single pathogen (e.g. HIV -> AIDS) can be ruled out. Similarly, bacteria can also cause ME/CFS, and so it cannot be about a single type or family of viruses or bacteria. The idea that a pathogen never before encountered by humans is also popular, but can also be discarded either way, especially as reinfections with COVID can also cause ME/CFS, as can the very first one.

Another example, a traditional biopsychosocial model is that it's deconditioning. Well, deconditioning cannot fluctuate, and a necessary condition for a theoretical model of ME/CFS is that fluctuations are not just common, but can be very rapid, cumulative and occur even as a result of mild fitness training. Deconditioning simply cannot account for that, and thus can be dismissed as a valid hypothesis.

Similarly, people who have never heard of ME/CFS, or even of chronic illness in general, as well as people who did but did not believe in them, can develop ME/CFS, which discounts any anxiety/fear model with anticipation of possible lifelong illness affecting behavior.

Many very fit people have developed ME/CFS, which discounts any model involving prior sedentary behavior, or inability to understand or work out how to be active, having to be 'coached' into 'learning simple walking' and other things.

Even when obviously triggered by an infectious illness, severity of acute illness does not seem to predict likelihood or severity of the illness, this is overwhelmingly obvious with COVID. Also, even when not obviously triggered by an infectious illness, subsequent infectious illness often makes it worse, including mild ones. However, sometimes mild infections can make the illness better, usually temporarily. Any model has to, if not account for it, at least not contradict it.

ME/CFS affects men and women, children and adults, and so any valid theoretical model cannot require things like hormonal changes happening later in life. Similarly, there have been poor studies showing elevated childhood adversity, but people with perfectly happy childhoods have also developed it, by which this hypothesis can be safely discarded.

Cognitive exertion can be a trigger for PEM/PESE, just as much as physical exertion, and so any valid model cannot depend on physical exertion involving only muscles or locomotion.

Remission and recovery can be spontaneous, even rapid, which discounts many hypotheses that can only unfold over the long term, and vice-versa. Some remissions and recoveries unfold over many months and years, and any valid model must account for it, or at least not make those data impossible.

The heart of the scientific method is falsification. For example, any hypothesis requiring that only women can be affected can be thrown out by finding a single man. Finding false positive evidence is often easy, but isn't the way to do science. Rather, we posit things that are impossible if a model is true, and if found, then we can discard it.

Let's do some effing science!

This post sums up my own view about the restrictions in research space we should be focusing on, given the undeniable evidence we have of the nature of the disease, which you outline. But with much sharper knife. Grazie.

 

[Tom Hill]

Remission and recovery can be spontaneous

I think this is my only quibble in the opening post of this thread. Don’t disagree necessarily but unsure of the precise meaning. Spontaneous is not a synonym for instantaneous but it has that vibe. I’d like clarification of what a spontaneous recovery has involved, if you could offer it. I’m unfamiliar with it in the case histories I’ve come across.

 

[neophyte32]

I think this is my only quibble in the opening post of this thread. Don’t disagree necessarily but unsure of the precise meaning. Spontaneous is not a synonym for instantaneous but it has that vibe. I’d like clarification of what a spontaneous recovery has involved, if you could offer it. I’m unfamiliar with it in the case histories I’ve come across.

In any case, this absolutely never happens (and I really mean never) in severe and very severe patients who are over 30 and have been ill for several years. That much is clear.

 

[Tom Hill]

I would need to review the data more carefully and comprehensively, but at this preliminary stage, I think this theory of ME/CFS causation might be consistent with both the full range of reported ME/CFS onset experiences, and the full range of reported experiences of ME/CFS treatment and recovery.

@rvallee bringing this here as I think it’s a better place to ask this question: could you check the teleological framing in the linked post and briefly take a moment to assess if you think it perhaps could cover the breadth of cases you have such a good grasp of? Not a general assessment of the theory’s broader plausibility (it’s just a theory among a dozen here, nothing more, and there’s plenty of time for that later as we really try and pick it down to its bones on its own thread).

Also this post for additional (more sketch book) thoughts that might help clarify the theory’s possible implications:

For the latter type I am thinking dental surgery, traumatic child birth, an experience involving extreme stress - that type of thing.)

Not trying to shirk my own homework. Just seems efficient, especially when we all have exhaustion and brain fog challenges and you’ve done this thinking already.

As with all, I just want to get closer to a cure - yesterday.

 

[rvallee]

I think this is my only quibble in the opening post of this thread. Don’t disagree necessarily but unsure of the precise meaning. Spontaneous is not a synonym for instantaneous but it has that vibe. I’d like clarification of what a spontaneous recovery has involved, if you could offer it. I’m unfamiliar with it in the case histories I’ve come across.

I double-checked the meaning and it works for me:

happening or done in a natural, sudden way without any planning or outside cause

It can be gradual and fit the definition.

 

[rvallee]

@rvallee bringing this here as I think it’s a better place to ask this question: could you check the teleological framing in the linked post and briefly take a moment to assess if you think it perhaps could cover the breadth of cases you have such a good grasp of? Not a general assessment of the theory’s broader plausibility (it’s just a theory among a dozen here, nothing more, and there’s plenty of time for that later as we really try and pick it down to its bones on its own thread).

At a glance I'd say so, but what I was hoping for with this thread was really to formalize this, to go deep into the details of what is really known. It would definitely be something involving a PhD, or a full research project. There is a lot to cover, as the only value of such a checklist of conditions is dependent on it being exact, in having both no false positives and false negatives.

The kind of stuff that an AI will likely be able to do in the next year. It would have to cover a lot, and be rather conservative at first in what it includes and excludes.

 

[Tom Hill]

I double-checked the meaning and it works for me:

It can be gradual and fit the definition.

Thanks. That works for me too.

There is a lot to cover, as the only value of such a checklist of conditions is dependent on it being exact, in having both no false positives and false negatives.

Thank you for taking a look. Just want to add that I hope you are able to do this, given the value of it.

 

[rvallee]

Thank you for taking a look. Just want to add that I hope you are able to do this, given the value of it.

Oh it was just a starting point for a bigger conversation that hasn't really happened. I haven't been well enough to continue on with this, and it's really out of scope for my skills and energy anyway. A project for later, one that will require lots of different perspectives, checks and balances.

 

[Tom Hill]

checks and balances.

S4ME: Checks and Balances — sums this forum up well