Thank you very much indeed for sharing all of this (and, indeed, for all of your work), it is enormously appreciated. I have many questions, but will wait for the preprint as I suspect many, if not all, will be answered within.
Community Symposium on the Molecular Basis of ME/CFS Sept 11 2026 (Stanford/Ron Davis)
- Thread starter Chandelier
- Start date
Incidentally, what impact would a not recognised condition with a prevalence of 1 in 1,000 have on the reference range (I believe that was the estimated figure for ME/CFS before the Covid pandemic)? After all, the test results come back as normal, but if the reference range already includes data from people with the condition, it’s a lost cause trying to spot any deviations.
Jonathan Edwards
Senior Member (Voting Rights)
I suspect none. Reference ranges tend to be worked out on a 90% basis - with 5% outliers at each end. Abnormal results from a disease at 0.1% at one end would be excluded. If there was more overlap then they would be in the 'true' normal range anyway?
Jonathan Edwards
Senior Member (Voting Rights)
Chris's comment is reassuring and makes good sense. My guess is that there are several interpretations of the 'lack of replication of DecodeME' and they may all be interesting.
The cautious position is to recognise that the Biobank cohorts may include some confounding factor due to selection bias in diagnosis or even misdiagnosis. That could include things like a weighting to people with insulin resistance because they see doctors more often or inclusion of burn-out cases that really do not fit the PEM-centred concept of ME/CFS. Even in this situation the intriguing thing is that some of this bias seems to be trackable by specific genetic tags. At worst this could help to clean up further analyses.
More interesting might be the idea that there are at least two forms of true ME/CFS. DecodeME suggested that it is the same for men and women but it looks as if the Biobank cohorts might be identifying a 'late' form in older individuals, or perhaps a more persistent form, or a form that does not involve whatever is picked out by the PEM picture.
That makes me think of narcolepsy again, which also has two age peaks and has two forms, one with a 'signature' of cataplexy and one without. We know that the genetic risks for those are different.
The cautious position is to recognise that the Biobank cohorts may include some confounding factor due to selection bias in diagnosis or even misdiagnosis. That could include things like a weighting to people with insulin resistance because they see doctors more often or inclusion of burn-out cases that really do not fit the PEM-centred concept of ME/CFS. Even in this situation the intriguing thing is that some of this bias seems to be trackable by specific genetic tags. At worst this could help to clean up further analyses.
More interesting might be the idea that there are at least two forms of true ME/CFS. DecodeME suggested that it is the same for men and women but it looks as if the Biobank cohorts might be identifying a 'late' form in older individuals, or perhaps a more persistent form, or a form that does not involve whatever is picked out by the PEM picture.
That makes me think of narcolepsy again, which also has two age peaks and has two forms, one with a 'signature' of cataplexy and one without. We know that the genetic risks for those are different.
Chris Ponting
Senior Member (Voting Rights)
Thank you @DHagen. The lesson that I take away is different. When people have been diagnosed with ME/CFS differently, by different healthcare professionals who have been mostly untrained in ME/CFS, using different criteria across decades, then the UK's (or world's) ME/CFS population will be heterogeneous. So, rather than always expecting to see replication, we expect to see this heterogeneity within and between large data sets. I'm not saying that large numbers of people with an ME/CFS diagnosis are misclassified; rather, that ME/CFS is not narrowly defined clinically, and each ME/CFS cohort will preferentially sample from a separate portion of this wider defined set.
Chandelier
Senior Member (Voting Rights)
I OCRed all the schedules and put them in the first post so that our search engine can index the participants.
I really don't understand why important info tends to be shared via images on social media...
I couldn't find any information on Paige Zuckerman on the forum.
She had a presentation slot with Brayden Yellman and mentioned FND misdiagnoses, @cstruter .
Their presentation is summarised on Bluesky.
Short excerpt:
I really don't understand why important info tends to be shared via images on social media...
I couldn't find any information on Paige Zuckerman on the forum.
She had a presentation slot with Brayden Yellman and mentioned FND misdiagnoses, @cstruter .
Their presentation is summarised on Bluesky.
Short excerpt:
Paige has had five clients diagnosed with factitious disorder, FND, or conversion disorder
Four of them ended up receiving a diagnosis within the IACC spectrum and improving with appropriate treatment and guidance for their relevant clinical diagnosis.
Eddie
Senior Member (Voting Rights)
It makes sense that DecodeMe is a more representative sample of pwME than the Biobank cohort. What I struggle to understand is how a skew in misdiagnoses could produce genes more significant than the genes that DecodeMe found. If there was a weighting towards insulin resistance in the Biobank cohort for example, does that come out as significant because insulin resistance has a much larger genetic component than ME? And if that is the case, is it possible that a very small skew of a condition with a very large genetic component could have impacted Decode?
Would such a skew overwhelm the ME genes, or is it a case of there not being enough pwME to produce statistical significance in addition to the skew genes picked up in the Biobank case?
Both your and Chris's responses are reassuring. While it would have been cool to see replication, maybe all that would have demonstrated is that the Biobank has a rigorous ME/CFS sample.
Hoopoe
Senior Member (Voting Rights)
Many doctors seem or seemed to believe that ME/CFS is just deconditioning and an unhealthy lifestyle.
Is it possible that the UK Biobank is showing us the genetic signals for the phenotype that doctors generally thought ME/CFS was, at that period of time, in the UK?
Is it possible that the UK Biobank is showing us the genetic signals for the phenotype that doctors generally thought ME/CFS was, at that period of time, in the UK?
Last edited:
Then the studies of CBT and GET would show meaningful improvement.
MinIreland
Senior Member (Voting Rights)
How can there be misdiagnosis when there are no clear criteria? How can we know who has ME/CFS and who doesn't? And for those who don't, but have similar symptoms, what do they have?
Mij
Senior Member (Voting Rights)
As someone who experienced migraines since age 11 that completely disappeared when I started menopause (after 35yrs of ME), could there be a clue there somewhere?
Could be! My understanding is that menopause very often brings significant change with regard to migraines, though the reasons are not as well understood as one might hope. For some, menopause brings an end to their migraines, for others it is the start. Migraines run in my family - I believe my maternal aunt had her migraines almost entirely cease after menopause, while for my mother, it marked a shift between the more common migraines that are primarily characterized by pain (and last for many hours or days) and "silent" migraines that feature visual and sensory disturbances but little or no pain. Thankfully, she has never had to deal with the severe aphasia that seems to afflict me.
Very glad to hear that you are free of migraines now - it seems some small mercy is still to be found!
This does seem like a potentially promising point to look for commonalities between migraine and ME.
People with e.g. depression or a known autoimmune disease or sleep deprivation might be misdiagnosed with ME/CFS and skew the data.
ME/CFS is a concept we made up to describe a collection of symptoms in a certain context.
Anyone who satisfies one of the multiple sets of criteria for ME/CFS 'has ME/CFS' under those criteria.
The hope is that the recent sets of criteria and the way they are applied are precise enough to lead to a better understanding of why people get these symptoms.
I very much share Eddie's questions.
I am unsure if what Drs. considered to be "CFS/ME" in the early 2000s in GB is consistent. If I had to guess I'd say some thought of it as entirely psychological, be it hypochondria, depression or anxiety or all at once, others possibly just thought it's the same as "chronic fatigue" and diagnosed people with all kinds of fatigue with unknown etiology while others thought of it roughly as what we would call ME today.
Of course I'd really enjoy hearing and being corrected by older people from the UK, who possibly know way better which people Drs. diagnosed with "CFS/ME" in the early 2000s.
I would disagree here, why should that mean such studies would show meaningful improvement?
Furthermore even given that all and every Dr. would see "CFS/ME" as exactly the same there's no proof that CBT+GET would help with this.
I'd even say there's evidence it DOESN'T work better than placebo even if you tried to trial people with "CFS" according to the ideas of Wessely and his friends, as in the PACE trial inclusion criteria were defined by the BPS-crew and it still failed to show an effect after substraction of all the cheating they did.
Last edited:
Maybe I don't know enough about gene expression analysis to really provide valuable input, but I'll still try and would speculate that the possibilty of a condition with a very large genetic component will surely in-/decrease with the quality of the sample.
In the UK-Biobank people were white, british and with a median age of 56, and most common comorbidity was hypertension. Also they were unhealthy but healthy enough to go out for treatment and data collection multiple times. Isn't e.g. the chance that people with these characteristics have diabetes, and therefore as it has a significant genetic component also gene expression tied to blood-sugar problems, already much higher in this sample than it would be in the average population or in healthy people with similar age and heritage?
Note that wallfish was responding to an earlier version of Hoopoe's comment, which was roughly asking if the ME/CFS cohorts back then could have included a large group of people who were just deconditioned and 'unhealthy'. Wallfish was saying this is unlikely because if deconditioning etc. were really truly someone's only problem, GET would in theory actually help. Since GET didn't help in their trials, we can conclude deconditioning is probably not an underlying driver of illness in those cohorts.
Kitty
Senior Member (Voting Rights)
I was diagnosed by a GP in 1999 after other possibilities had been ruled out. They didn't think it was psychosomatic, and were generally pretty supportive over the 30 years I was registered with the practice. One of the GPs warned me off our local ME clinic because they used dance therapy, which he didn't think was appropriate (I told him I'd studied contemporary dance for long enough that I could probably teach the class anyway!).
I've no idea about other patients, but my experience was fairly good. One GP kept wanting me to try antidepressants for a couple of years, which was inappropriate for someone not prone to depression, but I think he was genuinely trying to help.
Casterofspells
Established Member (Voting Rights)
I think this study kind of exemplifies it:
The newer and stricter Canadian consensus criteria seem capable of defining a different and more specific cohort centered around the hallmark symptom of PEM.
Meanwhile the earlier interpretations more along the lines of the of the (often used as catch all for anyone showing the symptom chronic fatigue) “CFS” label which didn’t focus on the PEM status of patients seem to describe another cohort.