Could existing data from large-scale Covid / physical trauma studies show the biology of how ME/CFS develops from those triggers?

On another thread, @MelbME and I were discussing whether existing studies of Covid infection or severe physical trauma would allow researchers to pull out the small percentage of cases in whom this triggered ME/CFS, to study the biology of how it develops.

This would presumably involve the same sort of data-sharing/follow-up research that DecodeME allows regarding patients who gave permission for their data to be used.

I'm thinking 'omics studies, etc.

Our conversation is below.

Never sure who to tag on this stuff but @Jonathan Edwards, @Chris Ponting, @Simon M...

 

And just crossed with this:

 

If we had a bunch of patients who we knew with certainty were going to have ME/CFS triggered in a month's time and we could then follow them up, what sort of data would we want on them before, during and after the trigger? If we had infinite amounts of cash?

And are there existing big studies that have looked at those things in Covid/trauma/toxin-exposure?

 

Scheibenbogen has been running and publishing such studies both for Covid and EBV (see for instance One-year follow-up of young people with ME/CFS following infectious mononucleosis by Epstein-Barr virus and
Long-term symptom severity and clinical biomarkers in post-COVID-19/chronic fatigue syndrome: results from a prospective observational cohort. Peluso et al also have lots of data as part of LIINC, a lot of which has also been published. They seem to have more data for the acute phase (which Scheibenbogen has little of) but they have less focus on ME/CFS. Other Covid researchers also have this data for the acute phase, even for non-hospitalised patients. The question is always how much use the data is and whether it is of relevance for ME/CFS.

It also ought to be possible to do what Ascherio did for MS and EBV in the case of ME/CFS and EBV by using a similar study setup.

 

The problem is that within a short period of time you cannot know whether someone has ME/CFS or not. It might be post-viral fatigue which ought to look quite similar in the first weeks or something else that might resemble ME/CFS. So you'd really have to have a study setup which tracks patients for at least 2 years, probably longer. So what you need are extremely big sample sizes and very long-term data. If we assume that probably less than 0,5% of Covid cases end up with something we would consider to be ME/CFS, to get a cohort of just 200 ME/CFS patients, you'd have to track 40 000 Covid cases for several years. But then you don't know whether these people would have developed ME/CFS independently as well, so you'd arguably still need a control cohort. With this approach you'd probably need data of several hundred thousand participants. So you'd really require massive sample sizes. But it has been done for MS and EBV and for Covid some of these datasets exist.

For Covid however the data for acute infections has not revealed any insights to Long-Covid and there has been a tremendous amount of studies and data on this subject.

I don't think the data for hospitalised patients is a wise choise. It is too messy with things such as lung damage and PICS and the cohort is sort of the opposite of what you want to be studying (predominatly older with comorbidities and more males). But perhaps there are some choices that are smarter like a pediatric clinic or one that is somehow female dominated or in an area where there's more younger people.

 

Thanks! My question was partly aimed at trying to identify what kind of studies would be useful to look at retrospectively. For example, would you want proteomics, metabolomics, particular immune stuff, microbiome data...?

 

I actually think even just good incidence data as in the work by Ascherio on EBV would already be extremely useful. If you can establish that different infections have very different ME/CFS rates that might provide some subtle clues. Perhaps bacterial infections don't lead to ME/CFS at all or something of that sort. You could establish whether there are actually 2 age peaks etc. But that would require even more data. The rest to me seems like a possible bonus, probably leading nowhere. As far as I'm aware the other data Ascherio collected hasn't been very useful but establishing a relationship between EBV and MS seems to have been worth the effort.

Apart from proteomics etc the question arises whether if you want to capture non-biological data and how? I tend to think a BPS narrative could become very unconvincing if there was different incidence data but the people all look the same from the PS side of things.

 

There would also be the question: At what time points do you want to collect data? Are resolvent (ME/CFS-like) post-viral fatigue and ME/CFS different from the get go or is there a later time point at which they diverge?

 

There are reports of ME/CFS following surgery, though I don't know how common that is, whether it's only major surgery or certain types of surgery, and so on. But unlike Covid and accidents, surgery is a thing that we can see coming. Worth doing a prospective study, monitoring patients and taking samples before, during and after?

 

As with the Covid example and all other examples I would think that you'd need a control cohort to establish whether the rate of ME/CFS after surgery is actually higher than the rate of ME/CFS in the general population, otherwise it might be hard to establish that there is indeed an increased rate and that the ME/CFS is somehow due to surgery. Since we don't have reliable data for ME/CFS prevalence I don't think you can just pull this data from somewhere else.

Of course you could get lucky, not have a control cohort, run an experiment and see exactly which pathways lead to ME/CFS in someone that had surgery, but as a layman I find that hard to believe especially given that we've already seen that this doesn't work in Long-Covid.

The question is whether there are events that are possibly in the sweetspot. Surgeries are less common but they don't seem to have a highly reported ME/CFS rate. ME/CFS after EBV seems to be more common but almost everyone gets infected with EBV, so you need a massive control cohort to understand the rate of ME/CFS in people that never get infected with EBV.

There is the possibility that there could actually be a interesting dataset in the future (perhaps only in several decades) and I hope ME/CFS researchers can then collaborate there if that happens: EBV vaccines.

If trials with EBV vaccines were to launch in the upcoming years it would be interesting to know whether the rate of ME/CFS is higher after vaccination or not and whether these results depend on the type of vaccination, if there is any difference at all. But of course that is all hypothetical and who knows when or if such vaccines will actually launch phase 2 trials (but some have said to have a focus on ME/CFS, whatever that may mean).

 

It would depend on how likely surgery is to be a trigger point. Given how many surgeries are undertaken it is not surprising some are associated with ME/CFS onset, however if it was only say one in a million, any preemptive assessments would be prohibitively expensive for minimal return.

 

That's not what we'd be trying to establish, though - we're simply looking for a way to grab people coming down with ME/CFS as it's happening. It wouldn't matter whether the surgery had caused it. I'm basically suggesting using anecdote about surgery as a cause to see if we can bag some cases. It might only take a very few to give us some good data.

But has this before-during-after Covid thing been done, following individual patients? (I'm not at all familiar with that literature.)

 

Maybe surgery as a trigger is not as well-established as I thought it was!

 

Why not Covid vaccines?

 

Trying to identify populations at risk if ME/CFS would a brilliant way to identify differences between those who get it and those that don’t. The difficult thing is to identify populations where enough people are likely to develop the condition to make any study worth while.

There is the EBV work, but the Covid pandemic would have been an opportunity for more such studies. Early on in the pandemic people with ME were saying that given SARS viruses seem to generate a relatively high incidence of post viral ME/CFS we should be gearing up for such research, however we did not get the message across to the right people.

I wonder if we now should be planning research to start when the next pandemic emerges, though obviously we don’t have any certainty about what virus will be involved or when it will happen.

 

That would be the purpose of something like LIINC. I believe they have before-during-after data at least for a subset of patients (maybe it's more during-after data, I can't remember too well). The focus is on general Long-Covid rather than ME/CFS, but as far as I know there have been no revolutionary insights from that work. It might be that Long-Covid is too broad a definition to yield fruitful results but it might also be that ME/CFS is too rare an occurrence for the data to yield any results. As far as I am aware RECOVER also has a massive biobank repository with during-after data. I don't know if they have before data though. Might be good to find out. I think the problem is largely that nobody has any idea what any of their data is supposed to mean given how broadly they defined and tracked Long-Covid.

In 2021 and 2022 there was a lot of hype that researchers could now maybe crack post-viral syndrome such as ME/CFS because for the first time they actually had data from the onset till later and sometimes even before onset data, but it seems getting any insights at all was not quite that easy...

 

There seems to be no clear relationship between ME/CFS and Covid vaccines as far as I've seen the studies being discussed on this forum. (Of course the vaccines are excellent at preventing severe acute infections and death and as such reduce the incidence of PICS, lung damage etc and consequentially reduce the risk of Long-Covid, but it is unclear whether that bares any relevance to ME/CFS). Something similar applies to Metformin, Paxlovid etc (for Paxlovid there seems to be some preliminary evidence that it doesn’t prevent Long-Covid in non-at risk populations). The Long-Covid literature seemingly has been to broad of a mess for it to yield any insights and it doesn’t seem possible to retrospectively correct for those errors, or no one is willing to do this work. People are now sitting on some of largest datasets ever but have no idea what anything is supposed to mean or if it even means anything at all. It seems that nobody can draw any insights from any of this and that people have since moved on.

The special thing about EBV seems to be that is seems to cause ME/CFS at a higher rate (according to DecodeME) than other things. Of course that might actually not be true, but it seems to be a fruitful starting point, but first you have to figure out whether that is actually the case. ME/CFS research is so far behind that it’s hard to even to speak of triggers yet. I think the first step would have to be to actually identify those rigorously with reliable incidence rates. Another useful difference between a Covid vaccine and an EBV vaccine would hopefully be that you wouldn't have to control for reinfections which has made any work on Covid extremely hard.

 

Still is, surely? Have infection/LC rates dropped much due to vaccination?

 

The problem is that so many people have been exposed to the virus that now there is no longer a large unexposed pool of people becoming infected possibly making interpreting the differences between before and after exposure harder to interpret.

 

Not only due to vaccination (but also repeated infections, saturation effects etc), but yes the whole world seems to be working under this assumption and I think things like the ONS data suggests there is some truth to it. Certainly hospitalisations have gone down, which would likely naturally reduce LC rates (because LC is such a broad term). It also seems the initial rates that some authors reported were wrong or were rather irrelevant to ME/CFS where mixing PICS, lung damage, hosptialisation making etc made the data completely uninterpretable. What are LC rates supposed to be? Is it the 30% some authors cite or is it the 2% others cite? It is clear that with such large discrepancies the only thing that can be done with the majority of data is to discard it. GIGO.

I also wandered whether this could tell us anything about the nature of ME/CFS:
Can ME/CFS rates following LC/Covid tell us something about the immunology of ME/CFS?

But it seems to not be easy to get anything good out of this mess. For one billion dollars RECOVER yielded 0 insights precisely because just randomly collecting data doesn't seem to be a fruitful approach.