Should we initiate development of a new, short questionnaire to identify PEM (to aid diagnosis)?

[Jonathan Edwards]

Unless I'm reading these wrongly (and I'm reading in haste, with brainfog), we've consistently presented PEM as the essential characteristic of ME/CFS, and as having certain features. But your post above suggests that you think we might have got this wrong.

These things are not simple. I am not saying that we are entirely on the wrong track. I doubt that very much. But the point of a questionnaire would be to get a yes or no answer as to whether someone has "PEM". And the problem I see is that we have no clear consensus on what the boundaries of PEM, that would determine yes or no, are. And to go back to my Art History point, the best way to identify a specific process (like painted by Caravaggio) may have little or nothing to do with what we think of as its main characteristic (chiaroscuro).

 

[Jonathan Edwards]

to have a straightforward questionnaire to help identify whether someone has PEM.

But unless we know the boundaries of what we mean by PEM how do we do that?

 

[forestglip]

I wonder if instead of focusing on trying to create a questionnaire with optimal accuracy in separating healthy from ME/CFS, it might be better to think about making a questionnaire with virtually 100% specificity, at the cost of poor sensitivity. Meaning make a questionnaire where only people with very clear, obvious PEM test positive, but where some people who really do have PEM might not.

This would not be for clinical diagnosis, but just for research to make sure everyone in your case group really has PEM to give the best shot at getting significant findings.

Currently they try to make sure people have ME/CFS with criteria like CCC, but all the associated symptoms seem kind of arbitrary to me since we don't know how related they are to ME/CFS. If a person saying they have PEM isn't enough to be sure they have ME/CFS, then adding in things like sore lymph nodes and IBS is still sure to let in a lot of people who have unrelated conditions.

Instead, is it possible to come up with a questionnaire specifically about PEM where we can say with virtual certainty that the person actually has it if they answer a certain way?

Something like a very clear delay of at least 12 hours after exertion, and a delay that the patient has recognized as occurring at least 10 times in the past year. After the crash starts, the person feels much worse for at least 48 hours.

Edit: Corrected specificity to sensitivity and vice versa.

 

[Utsikt]

a questionnaire with virtually 100% sensitivity, at the cost of poor specificity. Meaning make a questionnaire where only people with very clear, obvious PEM test positive, but where some people who really do have PEM might not.

If the sensitivity is 100 %, it would identify all cases of PEM as PEM.

If the specificity is low, it would identify some cases on non-PEM as PEM.

So I think what you’re suggesting is a questionnaire with virtually 100 % specificity (no false positives) and lower sensitivity (some false negatives).

I’m not sure that would be possible without also utilising other diagnostic measures like excluding differential diagnoses.

 

[forestglip]

So I think what you’re suggesting is a questionnaire with virtually 100 % specificity (no false positives) and lower sensitivity (some false negatives).

Oh yes sorry mixed up my specificity and sensitivity. But yes, primarily focused on making sure there are as few false positives as possible.

I’m not sure that would be possible without also utilising other diagnostic measures like excluding differential diagnoses.

You mean excluding people who might have another condition that explains PEM? I don't really agree with that view. If they have PEM, they have PEM, whether the cause is unknown, as in most people with ME/CFS, or it's related to some other condition like a mitochondrial disorder.

The research is trying to figure out the biology behind PEM, and we don't have a good reason to believe the biology is substantially different for PEM depending on what disease it happens in.

But I suppose if one wants to specifically focus on PEM with no known cause, then exclusion criteria for various diseases could be used, which is basically also trying to exclude false positives, so same general idea as what I was saying, just even more strict.

 

[Utsikt]

The research is trying to figure out the biology behind PEM, and we don't have a good reason to believe the biology is substantially different for PEM depending on what disease it happens in.

We also don’t have a good reason to think that it’s not substantially different. We simply know far too little at this point.

 

[forestglip]

We also don’t have a good reason to think that it’s not substantially different. We simply know far too little at this point.

Sure, I think it'd also be valid to separate out people with well-characterized diseases.

 

[Utsikt]

Sure, I think it'd also be valid to separate out people with well-characterized diseases.

You’d have no way of knowing if you’ve actually drawn up the lines in a position that aligns with the underlying biology.

All of these things would be nice in theory, but I don’t see how we could make it a reality.

 

[forestglip]

You’d have no way of knowing if you’ve actually drawn up the lines in a position that aligns with the underlying biology.

I'm just suggesting that, even if imperfect, it makes more sense than the criteria being a mishmash of random symptoms that don't really even seem related. For example, one requirement for CCC:

6. At Least One Symptom from Two of the Following Categories:

a. Autonomic Manifestations: orthostatic intolerance–neurally mediated hypotenstion (NMH), postural orthostatic tachycardia syndrome (POTS), delayed postural hypotension; light-headedness; extreme pallor; nausea and irritable bowel syndrome; urinary frequency and bladder dysfunction; palpitations with or without cardiac arrhythmias; exertional dyspnea.

b. Neuroendocrine Manifestations: loss of thermostatic stability–subnormal body temperature and marked diurnal fluctuation, sweating episodes, recurrent feelings of feverishness and cold extremities; intolerance of extremes of heat and cold; marked weight change–anorexia or abnormal appetite; loss of adaptability and worsening of symptoms with stress.

c. Immune Manifestations: tender lymph nodes, recurrent sore throat, recurrent flu-like symptoms, general malaise, new sensitivities to food, medications and/or chemicals.

So you could fulfill this specific category (which to be fair is only one category of the criteria) if you have heart palpitations and weight change, and you could also fulfill it if you have nausea/IBS and tender lymph nodes. These seem like quite distinct symptom groupings to me, and I'm not sure how much it's adding in making sure that people who don't have PEM are not included.

All of these things would be nice in theory, but I don’t see how we could make it a reality.

Just as I suggested: Instead of trying to define some amorphous grouping of various symptoms from all over the body, the inclusion criteria for studies of PEM or ME/CFS would be something like:

Something like a very clear delay of at least 12 hours after exertion, and a delay that the patient has recognized as occurring at least 10 times in the past year. After the crash starts, the person feels much worse for at least 48 hours.

Something where anyone on this forum would say, yep that person definitely has PEM, not chronic fatigue, muscle weakness, depression, etc. And maybe also exclude people with diagnosed diseases that are known to have some kind PEM-like presentation to make it even more homogenous.

 

[Utsikt]

@forestglip we’re still running into the issue of justifying why this approach would be better for determining who should be included in studies or not, because we don’t know what we’re trying to optimise for, because we don’t know how the relevant parts of the underlying biology affects how the symptoms present.

 

[Kitty]

@forestglip we’re still running into the issue of justifying why this approach would be better for determining who should be included in studies or not

We're also not separating research studies and initial diagnosis by GPs. They're very different.

DecodeME seems to have shown that in-depth questionnaires can be useful, though I guess a genetics study is a very particular case. It has a way of showing whether or not a cohort has things in common, which isn't necessarily true in other study types.

I'm more concerned about initial diagnosis, as it's mostly done by non-specialists. They're either trained to recognise one model of ME/CFS but don't have the medical knowledge to spot signs of something else, or they're trained, experienced physicians who don't know much about ME/CFS.

We're relying for diagnosis on recognition of a pattern of worsening after exertion, but there's a lot of focus on how it presents. Maybe we should look at the pattern instead. It ought to be distinctive enough; it might exist in other diseases, but they often have pathology that would eventually differentiate them from ME/CFS. Apart from this pattern we call PEM, the other distinctive thing is a complete absence of detectable pathology. An unusual presentation of OI could potentially be a third leg.

 

[EndME]

I can see that others have suggested that a better questionnaire for diagnosis will only have extremely limited use and that seems very reasonable to me. Especially when doctors obviously don’t care about questionnaires for things they don’t even believe in. Only a different kind of education would probably change much.

I do think there might be some limited use for a better questionnaire for some big data studies especially if things could be made part of EHR records. I think we’ve seen researchers trying their best there but many big data studies struggling to actually know whether the cohort they end up with would be consistent with what members described here as having ME/CFS and a better questionnaire might be a bit helpful there. But I think there are also some very strong limits. If I remember correctly there were studies by RECOVER using current PEM questionnaires suggesting something like 20% of the healthy population had PEM. I doubt the big problem here is the questionnaire. You’re never gonna get anywhere, when the people conducting the work have fundamentally no idea what they’re doing. A different questionnaire will not make those situations any better. We very frequently discuss Long-Covid studies on S4ME. The most discussed Long-Covid study has been the one by Rob Wüst. I don’t think that had anything to do with how PEM was screened in said study but much more that many people had the feeling or even some confidence that what the authors were studying something that somewhat reassembled ME/CFS based on much more nuanced things than questionnaire data which made us think the study could be relevant (all patients being screened thoroughly for other problems, the population being non-hospitalised after acute Covid and younger, there being a female predominance, everyone going from working full-time to not working anymore, interviews of patients describe something that seemed familar etc). I think it may be useful to ask some of those researchers we find to be smart what kind of better questionnaire data would really be useful to them.

This whole effort would be somewhat large, involving patients, clinicians and people with expertise in questionnaire design. I think something that is of similar difficulty conceptually but that I think would lead to a much better understanding of PEM and in my opinion could be much more valuable would simply be to do a long-term activity and symptom tracking studying using a medical grade activity tracking device (recording some features like step count, heart rate, body temperature etc) combined with some patient questionnaire answering (I suppose you’d want this medical device to not feedback any data to patients so the purpose would be very different to what some other studies are currently doing). These devices are pretty cheap nowadays, I don’t think that the cost of such a study would be substantially higher than what is being proposed here.

 

[Evergreen]

There are a lot of things get repeated so many times that everyone believes them, but may not actually be true. I'd like to know what's actually true about PEM.

Regarding whether we actually have evidence to suggest that delayed onset is the key feature of PEM in ME/CFS, I read this sentence in Wormgoor & Rodenburg 2023, discussed here:

The delayed onset and the broad constellation of symptom deteriorations distinguish ME/CFS from other diseases with severe fatigue or deconditioning (6, 16–21).

I've been looking at the references given for this. So far I don't see anything in 6, 16, 17, 18 that supports this claim. Note I have to skim rather than read in full, so I may have missed some details. Please tell me if I have and I will amend this post.

Reference 19 is Hodges et al. 2017, a 2-day CPET study. This is at least a study including both ME and MS participants, and did find that MS patients had a higher workload at RER on day 2 CPET, whereas CFS patients had a lower workload on day 2 CPET. Symptoms were not assessed, though, so we don't know whether patients would have reported PEM at the time, or been aware that their functioning was a bit lower. We also don't know when any symptoms started - during CPET 1? Immediately after? A few hours after? A day after? I don't think reference 19 supports the claim either.

20 is Klebek et al. 2020 (Jason's group), a questionnaire-based study. People with "ME and CFS" did report more DSQ-defined PEM, but people with post-polio syndrome reported plenty of it too. Timing of onset of PEM was not explored. If anything, this just demonstrates that either people's concerns about the Jason team's definition of PEM are justified, or people's belief that PEM is unique to ME/CFS or uncommon in other diseases is not justified. So I don't think reference 20 supports Wormgoor et al's claim.




Reference 21 is van Campen et al. 2021, a 2-day CPET study comparing males with ME/CFS with males with idiopathic chronic fatigue. As in the study by Hodges above, those with idiopathic chronic fatigue improved their CPET performance on day 2, whereas those with ME/CFS did the opposite. But again, symptoms weren't assessed, and nor was timing of onset of symptoms. We don't know if patients were reporting PEM from CPET 1. All we know is that their functioning on a maximal exercise test was lower than the day before.

These references do not provide any support for the idea that delayed onset of post-exertional symptoms distinguishes ME/CFS from other diseases. The 2-day CPET studies comparing ME/CFS with other diseases demonstrate that people with ME/CFS function worse on a maximal exercise test on day 2 while people with other diseases and healthy people function better. That is a very interesting finding, but it does not tell us anything about whether onset of PEM is immediate or delayed.

I think we can collect better data that will allow us to make accurate claims about PEM. If delayed onset really does distinguish ME/CFS from other diseases, then great. But if what's actually happening is that people with other fatiguing illness and ME/CFS both have increased symptoms and reduced functioning for 1-12 hours post-exertion, but people with other diseases recover thereafter while ME/CFS don't, or get worse symptoms, then it's a duration issue. And if people with ME/CFS don't reliably report increased symptoms even though we can see their functioning is deteriorating, we need to find a way to detect that without subjecting them to maximal exercise tests.

So I think we need exploratory studies.

 

[Kitty]

So I think we need exploratory studies.

I keep harping on about this, but we also need to know what any questionnaire or study or info sheet is for.

Diagnosis of ME/CFS?
Identification of trial cohorts?
Identification of PEM as an individual feature of ME/CFS (or any other condition) in order to study it?

Each might need a different approach.

When it comes to diagnosis, the patient community has placed a lot of stress on PEM as the cardinal feature of ME/CFS. Might researchers and sympathetic physicians have begun to think it's the only important feature of ME/CFS? Or the only identifiable one?

If they have, it's not surprising problems have arisen.

 

[Evergreen]

I keep harping on about this, but we also need to know what any questionnaire or study or info sheet is for.

Diagnosis of ME/CFS?
Identification of trial cohorts?
Identification of PEM as an individual feature of ME/CFS (or any other condition) in order to study it?

Each might need a different approach.

When it comes to diagnosis, the patient community has placed a lot of stress on PEM as the cardinal feature of ME/CFS. Might researchers and sympathetic physicians have begun to think it's the only important feature of ME/CFS? Or the only identifiable one?

If they have, it's not surprising problems have arisen.

What I'm advocating is studies to find out what post-exertional phenomena look lke in ME/CFS vs other diseases so that we might be able to describe PEM in a more accurate way, or define what we mean by PEM in ME/CFS.

So I'm going back a good few steps from the idea this thread started with, because (a) I don't think a questionnaire will aid diagnosis and (b) I don't think we have the information we need to even educate others about PEM, let alone draft a useful questionnaire. I do think understanding post-exertional phenomena in ME/CFS and other diseases better would move us forward.

So the kind of study I'm advocating is for...knowledge.

When we have that knowledge, we will have a better idea of how to use it.

 

[Mij]

The Post Exertional Malaise symptoms on that chart don't reflect my experience much. I think if they want t learn more about PEM vs controls they should focus on the delayed aspect and how the distinctive delayed part begins. I feel completely fine DURING exercise, but 11-13hrs later at a specific point in my day the distinctive symptoms start to manifest and takes it course. Same exact distinctive onset, course and recovery period for the last 28 years.

 

[Evergreen]

The Post Exertional Malaise symptoms on that chart don't reflect my experience much. I think if they want t learn more about PEM vs controls they should focus on the delayed aspect and how the distinctive delayed part begins. I feel completely fine DURING exercise, but 11-13hrs later at a specific point in my day the distinctive symptoms start to manifest and takes it course. Same exact distinctive onset, course and recovery period for the last 28 years.

Yeah, the DSQ results reported in table 4 of Kleber et al. (above) include during and after and "after" is not defined, so people could be describing different things but they will look like they might be the same.

Against that, Hanson's group did a 2-day CPET study where symptoms rise immediately and peak at about 2 days for both controls and pwME (Moore et al. 2023). If I'm reading the figure below correctly, it suggests that the difference is how much our symptoms rise and how much longer they stay up. The controls' "PEM" symptoms might be highest at 2 days but they're probably still at a pretty negligible level, and not interfering with daily activities. Whereas by 2 days, the pwME's symptoms are soaring and functioning must be significantly affected.

 

[Kitty]

we might be able to describe PEM in a more accurate way, or define what we mean by PEM in ME/CFS

What I meant is, why are we doing that? Are we sure it's useful? Or necessary, or possible?

We can't make progress until we understand what the question is, why we're asking it, and what (if anything) the answer's going to tell us.

There's a bit of a Catch-22 with PEM, in that it's hard for a doctor to identify it if the patient doesn't already know they have it. If they haven't yet spotted the pattern, their answers are likely to be confusing no matter how good the questions are.

So to play devil's advocate, there's an argument a doctor could simply ask about PEM. People who do have it can say so; people who aren't yet sure might not be able to give clear enough answers for the doctor to reach a conclusion; and people who don't have it but are convinced they do will give the 'right' answers anyway.

That might be oversimplified, but I think it's one of the problems with placing too much emphasis on PEM for diagnosis.

 

[Utsikt]

There's a bit of a Catch-22 with PEM, in that it's hard for a doctor to identify it if the patient doesn't already know they have it. If they haven't yet spotted the pattern, their answers are likely to be confusing no matter how good the questions are.

@Bivox (Sommerfelt) is able to identify PEM in children that don’t know anything about it. By asking the right open-ended questions when seeing them and excluding the known alternate explanations.

I think we’d get much further by educating doctors about PEM and ME/CFS in general so they would be able to recognise the concept, than to give them a questionnaire they can use on the patients.

 

[Evergreen]

What I meant is, why are we doing that? Are we sure it's useful? Or necessary, or possible?

Gotcha. I think it's necessary because relying on patients' impressions and clinicians' impressions is not enough - too variable, too vulnerable to bias - and because of what has gone on in the last few years:

• Awareness of ME/CFS has risen because of long covid
• PEM started being talked about in very particular ways e.g. "If you've got PEM, you shouldn't exercise/you need to be careful with exercise."
• People with other diseases see some definitions of PEM and (rightly) say, "That happens to me too", because the definitions often do little more than describe pathological fatigue.
• People with ME/CFS argue that no, people with other diseases don't have real PEM, because real PEM is delayed/renders you bedbound/whatever.
• Researchers have responded by taking PEM into account, using existing definitions and the DSQ-PEM, which arguably wasn't ready for the task.
• People can say things like "Ha! Look! People with PEM do better in rehab than people without PEM." (Entirely possible because of regression to the mean.)

There are big potential consequences to the above e.g. people avoiding activity more than they need to or when they don't need to at all, antagonism between people with ME/CFS and other diseases, health professionals being even more irritated by us than they are already, incorrect conclusions being reached about the advisability of exercise for pwME (or indeed, pwPEM) etc.

So if we can get data that tell us what the-particular-pathological-response-to-exertion-in-ME/CFS is, we can start spreading accurate information, and do better studies.