Open Institute for Neuro-Immune Medicine: ME/CFS: Detecting Mycotoxin Subgroup

[Medfeb]

I think with a big epidemiological study environmental signals would show up despite quite a lot of dilution. As for the GWAS proposal I am much less worried about dilution than I am about spurious associations coming uo because of internet based trawling.

Sorry for not replying on Tuesday but I wanted to follow up on this with a question.

To @Michiel Tack's post, I'm assuming we dont yet know the accuracy of the diagnostic tools when used as standalone instruments without a clinical diagnosis by a physician. But based on various other studies and reports, we know that in general clinical practice today, there's both a high rate of misdiagnosis of other diseases as CFS and underdiagnosis and misdiagnosis of actual cases of ME.

While I cant site specific studies right now, I am assuming that the genetics behind ME are not just a single gene or mutation but rather some combination of genetics that interplay with other factors and that could vary from one patient to the next. Is that a fair assumption?

If so, if we assume for the moment that 50% of the people identified as "ME/CFS" cases are actually some other condition, wont that both contribute spurious results from false positives and also make it difficult to identify the pattern in the true positives - particularly since you wont know who the true cases are.

Do we know what the diagnostic accuracy of these tools as standalone tools is? And what about this study and methodology compensates for a high rate of mischaracterization of patients in the presence of a likely complicated genetic picture?

What am I missing?

 

[Jonathan Edwards]

If so, if we assume for the moment that 50% of the people identified as "ME/CFS" cases are actually some other condition, wont that both contribute spurious results from false positives and also make it difficult to identify the pattern in the true positives - particularly since you wont know who the true cases are.

I was referring to the fact that a large epidemiological study showed no correlation between ME and zip code. (I am not sure genetics comes in to it.) It seems there were no spurious associations due to misdiagnosis because there were no associations. It is theoretically possible for all the 'non-ME' subjects to live in an exactly complementary set of zip codes from those with ME and cancel out but the probability of that is vanishingly small.

If 50% of cases do not have ME there will be 50% dilution of an environmental signal. But with a large study that is unlikely to make a clinically significant link to something like damp housing disappear. All sorts of other dilution factors operate in all such studies (not all houses in a given road are damp) but you still expect to see a statistical signal. If, say, 20% of cases of ME are due to living in a mouldy house then with big numbers the difference in incidence in a run down New York area and Tucson should show up like a sore thumb. If the cohort was diluted 50% with random other illnesses it should still show up.

I am also a bit sceptical that we can know that ME defined by a venerable group of expert physicians is any better than by a questionnaire. There was a study in the UK that indicated that the rate of diagnosis of fibromyalgia varied between specialist physicians by a factor of 100 fold. That suggests that asking physicians to do the diagnosing will give you a complete mess. A questionnaire would provide much more usable data.

 

[JemPD]

I am also a bit sceptical that we can know that ME defined by a venerable group of expert physicians is any better than by a questionnaire. There was a study in the UK that indicated that the rate of diagnosis of fibromyalgia varied between specialist physicians by a factor of 100 fold. That suggests that asking physicians to do the diagnosing will give you a complete mess. A questionnaire would provide much more usable data.

That's very interesting thank you for pointing it out.

 

[Wonko]

Surely that depends on what the questions on the questionnaire actually are? And by how they are, in practice, interpreted.

 

[Medfeb]

If 50% of cases do not have ME there will be 50% dilution of an environmental signal. But with a large study that is unlikely to make a clinically significant link to something like damp housing disappear.

I should have clarified - My question was not specific to environment issues but more general to the ability of the proposed GWAS study design to produce meaningful findings if there's a high rate of misdiagnosis and the genetic pattern is not homogenous across the true cases of ME. Its not clear to me how this study design can tolerate such high levels of misdiagnosis.

I am also a bit skeptical that we can know that ME defined by a venerable group of expert physicians is any better than by a questionnaire. There was a study in the UK that indicated that the rate of diagnosis of fibromyalgia varied between specialist physicians by a factor of 100 fold. That suggests that asking physicians to do the diagnosing will give you a complete mess. A questionnaire would provide much more usable data.

I recognize the issue with who/how an expert is defined but I dont share that broad degree of skepticism for ME experts. And any questionnaire would have to validated against some external standard and right now, all we have are experts. So we are back with @Michiel Tack's point about needing to validate the instrument.

And just realizing that these posts probably belong in a different thread. Should I move them?

 

[Snow Leopard]

A study like this needs to be done on a strict population based cohort with matched controls from the same population.

If only we had the money for that!

 

[FMMM1]

Perhaps. But depending on the study design, it could also point to a meaningless "CFS" cohort based on bad criteria and/or the use by the medical community of "CFS" as a wastebin to dump clinical cases of unexplained fatigue. That's the likely explanation for the four-fold increase in prevalence in the CDC risk factor survey

In the case of water damaged buildings, I could imagine some clustering of that factor by zipcode but could also imagine a wider and more random distribution.

I think ME is potentially a "wastebin to dump clinical cases of unexplained fatigue". E.g. MS (which also causes disabling fatigue) is caused by demyelination so you can diagnose cases of MS with a high degree of accuracy - I doubt it.

I think this issue has also surfaced in the NIH Intramural Study (Dr. Avindra Nath) approx 30%(?) of those enrolled were identified as having another (un-diagnosed) illness and promptly de-enrolled. Maybe the NIH study will help to identify common misdiagnosis.

 

[Jonathan Edwards]

If only we had the money for that!

Well we have the ME Biobank and the hope is to expand that considerably as part of the MRC study.

 

[Jonathan Edwards]

My question was not specific to environment issues but more general to the ability of the proposed GWAS study design to produce meaningful findings if there's a high rate of misdiagnosis and the genetic pattern is not homogenous across the true cases of ME. Its not clear to me how this study design can tolerate such high levels of misdiagnosis.

A GWAS study on a large cohort assumes that the genetic difference is a statistical population one - so certainly not homogeneous. The issue is whether any statistical difference will survive 50% dilution. It might not but then a slightly weaker statistical difference might not survive even with 100% pure 'ME'. My personal feeling is that any statistical difference of real interest will show up fairly robustly with a 20,000 sample and would probably survive 50% dilution.

So as an example there might be a gene allele q that is present in 20% of heathy people and 46% of PWME. With a 50% dilution that will appear as 33% in ME. That would still be likely to show up at least as a candidate for a second pass verification.

The alternative is some sort of consistent vetting by a professional - as occurs for the ME Biobank. That would be ideal but it is expensive. Doing that on 1000 ME would be good but it would still be weaker than 20,000 with only 50% purity I am pretty sure.

 

[Medfeb]

Sorry I missed this sooner, @Jonathan Edwards. Thanks for the reply.

 

[daftasabrush]

Just posting on here since I am totally unable to find this study, which is referred to as if it's published:

At the Institute for Neuro-Immune Medicine (INIM) Clinic, our clinicians discovered that some of their patients diagnosed with ME/CFS also had a mycotoxin exposure. Mycotoxins are toxic metabolites produced by molds/fungi. In a retrospective analysis, 111 ME/CFS patients' charts were reviewed in the INIM Clinic. The majority were women, aged 23-77 years old. They found 81 percent tested positive for mycotoxins, with Gliotoxin the most common (Laroche et al., 2017).​

This is still linked to from the current Research Studies page.

I can't find
- any mention of glitoxin findings in ME/CFS or chronic fatigue syndrome papers
- Who "Laroche" is - no mention on the NSU staff / research team pages, no ME/CFS or chronic fatigue syndrome publications
- Any 2016, 2017, or 2018 paper on mycotoxins in ME/CFS or chronic fatigue syndrome

Also the pilot study referred to - did it get cancelled? It says it will be recruiting SOON but the thread is from 2019, so is it still in progress, cancelled or published?

https://www.nova.edu/nim/research-s...ages/INIM-studies/19-20-PFRDG-mecfs-page.html