1. Patients launch $1.27 million crowdfunding campaign for ME/CFS gut microbiome study.
    Check out the website, Facebook and Twitter. Join in donate and spread the word!
Give ME the Money
Graham McPhee spells out some of the cold, hard facts about the dismal state of ME research and politics, and has some suggestions as to what we can do about it ...
Discuss the article on the Forums.

Invest in ME London conference 2012

Discussion in 'General ME/CFS News' started by Kate_UK, Oct 12, 2011.

  1. Ember

    Ember Senior Member

    Messages:
    1,742
    Likes:
    1,827
    To suggest that new technology makes homogeneous cohorts unnecessary today, Alex, seems to me to be less than intellectually honest. To my mind, the hypothetical nature of that suggestion needs to be stated.

    If in fact you assume that we'll go the route of becoming more “reductivistic,” then why in this context don't you accept the statements of Dr. Carruthers and Dr. Lipkin, both masters in their fields? Surely we aren't playing intellectual games here, given that lives are at stake.

    Perhaps (as you suggest) however, I'm misunderstanding your purpose.
     
  2. alex3619

    alex3619 Senior Member

    Messages:
    7,686
    Likes:
    12,520
    Logan, Queensland, Australia
    Hi Ember, I am in the process of evaluating counter-arguments that are going to be used against us. As a result I am trying to put things in context, and address issues before they even arise. That is my primary purpose, the cohort debate is only one case example of it. I have been backing, and am still backing, the adoption of strict ME definitions. What I am trying to counter is the view that this is the only way it can be done, and that it must always be done this way. At the moment it is the most sensible approach given our resource limitations. In time it may well become yesterdays science - obsolete for some purposes. So I am trying to give an understanding of how our absolute claims are not absolute but situation dependent.

    I have no influence over the researchers, hence nothing I say will matter there. What I am trying to do is change the advocacy debate to make it more robust. We will be attacked for many of the things we say, and they will win if we are not using sound argument.

    One of the things I am trying to do is construct a list of counter-arguments in my book, and then address each one separately.

    Cluster analysis of cohorts using broad definitions may be the fastest way to demonstrate validity of ME as a diagnostic entity. It could achieve, in one single study or a small number of studies, what might otherwise require dozens of studies over a decade or more. It has merit, its just not research as usual - its research taking advantage of the latest technological opportunities.

    Identifying a biochemical pathway cluster that is clearly differentiated from other clusters, specific to individual patients, gives us an opportunity to validate ME as a diagnostic entity and discredit alternatives (presuming here that ME is a valid diagnostic entity). We could in one stroke show that ME diagnosis is reliable, and that alternative diagnoses are invalid. If this fails to happen then we alternatively could show that our current diagnostic criteria are invalid, and have strong data
    to show an alternative.

    Bye, Alex
     
  3. Bob

    Bob

    Messages:
    8,844
    Likes:
    12,390
    South of England
    Oh, that's interesting... I didn't know that Lipkin was doing proteomics for his pathogen study...
    It sounds more complex and more thorough than I thought it was going to be. (I'm struggling to find any info about any of the ongoing studies.)
    I guess Lipkin could throw up some very interesting results.
    Lipkin might even be able to define sub-groups based on his findings, if he's doing a study that complex.

    Ember, I'm not disagreeing with what Lipkin said...
    Alex and I were exploring the subject theoretically, and talking about possibilities for the future.

    I'll try and explain the angle that I was coming from...

    I agree that the tightest diagnostic criteria will bring us the best results.
    And I'm in favour of using Fukuda, CCC and ICC alongside each other, or even CCC and ICC without Fukuda.
    Some of the researchers (Lipkin? Mikovits?) have also been selecting for sudden-onset patients, with viral-like onset, and some specific symptoms (I can't remember the details), and I think that's really helpful as well.
    The more of this careful selecting of the most homogeneous group of patients, the better, as far as I'm concerned.

    But we had been discussing the heterogeneous nature of 'CFS', and discussing whether we could be certain that even all 'ME' patients have exactly the same disease.
    So then we went on to talk about biomarker research, and how it could be a game changer.

    If a researcher was to design a complex enough proteomic or genetic study, with a large enough number of patients, then the huge amount of data that was amassed could be processed with IT.
    And this could lead us to a deeper understanding about subgroups, or separate cohorts.

    And it could separate patients into groups based on the data.

    (And if all the data was checked against symptoms, diagnostic criteria, family history, health history, type of onset, etc, then there could be potential for learning a lot more than we already know.)

    So, for example, if you had a single cohort of patients based on the CCC, but then when doing a genetic study, you found that this group of CCC patients could be subdived into three separate groups, based on the data, then what would this tell us?

    It could tell us that there are three different diseases, or three different triggers for the same disease, or three different varieties of the same disease, or it could be the same disease, but with three different patient responses, or it could separate the groups for reasons that we currently have no insight into at all. (I'm talking hypothetically.)

    We wouldn't know the answer straight away, but we could name each group CCC1 CCC2 and CCC3.
    And then future research could investigate the response of each group separately, to treatments.
    And each group could be investigated separately in all future research.

    This could give huge benefits...
    For example, if 80% of CCC1 patients responded to Rituximab, but only 10% of CCC2 and CCC3 responded, then this would mean that Rituximab could quickly be given to all ME patients who fitted the CCC1 cohort.
    But if we were to lump all the (hypothetically new) CCC groups together then only 33% would have responded to Rituximab (in this imaginary situation), so obviously, it would be great to get the patients into separate groups based on the data.

    Beyond that, if we were able to do a quick blood test to determine different subsets of CFS patients, then that could potentially supersede all of the current diagnostic criteria completely, even potentially making them redundant. (Hypothetically.)

    This sort of thing has actually already been going on. Jonathan Kerr suggested that he would be able to find successful therapeutic drugs based on his genetic findings. Because, if the genetics findings in a subset of his CFS patients were similar to those seen in a different illness, then the drugs for that different illness could be tested as a potential treatment for that subset of CFS patients. So in this instance, Jonathan Kerr was bi-passing some of the various diagnostic criteria, and sub-grouping CFS patietns purely based on his data. By sub-dividing CFS patients, based on sets of genetic data, appropriate treatments could be investigated for each of those cohorts, regardless of diagnostic criteria used.


    But, yes, practically speaking, for the most helpful results, and for the most homogeneous patients samples possible, then like Lipkin says, patient selection is the key.
     
    alex3619 likes this.
  4. Kate_UK

    Kate_UK Senior Member

    Messages:
    232
    Likes:
    228
  5. Enid

    Enid Senior Member

    Messages:
    3,309
    Likes:
    841
    UK
    Brilliant work Kate - thanks for posting.
     
  6. Ember

    Ember Senior Member

    Messages:
    1,742
    Likes:
    1,827
    To play devil's advocate in a climate where Dr. Carruthers' September presentation has been allowed to lie dormant for over eight months may be to risk misusing your talents, Alex. Here are Dr. Unger's comments on the debate:
    Many thanks to IiME for this month's journal articles and for including Jorgen Jelstad's “Words Matter” on their DVD!
     
  7. rlc

    rlc Senior Member

    Messages:
    822
    Likes:
    246
    Hi all, The number one problem in research in the past has been that they have forgotten what ME is, there have been endless arguments over which symptoms that should be in an ME criteria, which will never be resolved satisfactorily, because in all diseases there is always a wide variance in symptoms in patients with the same illness. Sub grouping along the lines of which patients has which symptom will only increase confusion and shows a lack of understanding of this.

    There is also been a belief that a single biomarker needs to be found, but the majority of diseases do not have a single biomarker that can confirm the diagnosis!

    In almost all disease there is considerable overlap with the symptoms of other diseases. And having a set of symptoms does not confirm a diagnosis (which is why ME definitions based on symptoms are limited in there usefulness), it should lead to the generation of a differential diagnosis list, (all the other diseases that can cause these symptoms) these then have to be ruled out one by one until there is only one diagnosis left.

    The same applies with tests results there are very few diseases that have one tests that can be run that can confirm a diagnosis, again a failed test result only generates a differential diagnosis list, so say low cortisol means the patient could have any of these 47 illnesses http://en.diagnosispro.com/differential_diagnosis-for/cortisol-lab-decreased/11102-153.html then more work has to be done to find which one it is.

    The most important question is what is ME, and the answer is ME is a illness that has a set of symptoms that can vary greatly in different patients, that isn’t one of the many other diseases that can cause these symptoms!

    And this is where previous research has fallen down, it has been based on selecting patients with the symptoms in whichever of the definitions the researchers chose, and then looking for biomarkers, the results are always contradictory because this approach guaranties mixed cohorts, the reason being that all of these definitions have a built in fault in them, they are not based on the answer to what is ME , it is a illness that has a set of symptoms that can vary greatly in different patients that isn’t one of the many other diseases that can cause these symptoms!

    These definitions have set out rules for what testing should be done that guarantee that all the other illnesses that have a set of symptoms that overlap with ME are NOT ruled out. This can be shown by comparing the testing required by the NICE guidelines found on page 140 here http://www.nice.org.uk/nicemedia/live/11824/36191/36191.pdf and the CDC guidelines http://www.cdc.gov/cfs/diagnosis/testing.html With the testing in the new IACFS/ME guidelines http://www.iacfsme.org/Portals/0/PDF/PrimerFinal3.pdf where you will see a massively increased amount of testing that is required and a far larger number of diseases that have to be ruled out. Although the IACFS/ME guidelines are not 100% complete and failed to mention things like wrong reference ranges for B12 and TSH etc, they do give a very good idea of how inadequate the testing outlined in the other criteria are. The information in the IACFS/ME guideline is not based on any new scientific breakthrough, the disease that they mention need to be ruled out have been known to cause symptoms that overlap with ME long before CFS was ever invented. If one was to be cynical one could say that previous CFS criteria have been written in such a way that not only allows ME to be portrayed as a mental illness, but require such minimal amounts of testing that it guarantees that many other people will have their true diagnosis missed and be wrongly given a CFS diagnosis, thereby increasing the numbers of people who can be said to be nuts and denied disability allowances and insurance payouts.

    The NICE guidelines in particular are quite disturbing and have include blatantly false information such as “Tests for vitamin B12 deficiency and folate levels should not be carried out unless a full blood count and mean cell volume show a macrocytosis.” and "Tests for serum ferritin in adults should not be carried out unless a full blood count and other haematological indices suggest iron deficiency”

    Macroytosis is a very late sign of B12 and folate deficiencies and serious symptoms can occur long before this happens. Disease like Hemochromatosis which effects one in every two hundred and fifty people and has symptoms that overlap with ME does not cause changes in full blood count and other haematological indices, and the disease will be missed if Ferritin isn’t checked. So in the space of two sentences they have given instructions that guarantee that tens of thousands of people will have their correct diagnoses missed!

    To correctly diagnose a patient it should go, a set of symptoms generates a differential diagnosis list, this then generates a list of tests that are to be done, they are then done and the diagnosis is found.
    The CFS definitions go a set of symptoms, generates an epic fail of a differential diagnosis list, which misses out a vast number of differential diagnoses, which then generates an epic fail of a list of tests to be done, with large amounts of important ones left out, which generates a large number of people with a large selection of un diagnosed illnesses wrongly diagnosed as CFS, which then have been finding there way into research studies, which then generates large amounts of confusing and contradictory results.

    So to get anywhere with research the first thing that has to happen is to remember what ME is, ME is a illness that has a set of symptoms that can vary greatly in different patients, that isn’t one of the many other diseases that can cause these symptoms!

    And then select patients who have these kind of symptoms and then test them for every other disease that overlaps with ME until you are left with a pure cohort! A very logical thing to do but in the past this hasn’t been happening.

    Some people are suggesting that by using modern technology a pure cohort isn’t necessarily that important, and that this modern technology will be able to find the biomarkers in the ME patients and that they can then be separated this way. However this won’t work for the simple reason that, the types of biomarkers that they are proposing to look for, things like say spinal protein signals, have never been investigated to see if they are found in all the other conditions that have the same symptoms as ME, so unless you have a pure cohort you will not know if what you are finding is biomarkers for ME or a new biomarker for any or a combination of the other diseases that get included in the mixed cohorts. Which will mean that they will have to get funding and testing will have to be done on groups of patients with each of the other diseases to see if the biomarker is found in these diseases. By which time most of the people reading will have died of old age.

    Other things that must be considered are things like the majority of ME patients due to being mainly house bound have a vitamin D deficiency, this must not only be diagnosed but treated before these patients are studied for biomarkers because vitamin D has been shown to effect both immune function and genetic expression as well as many other things, and if it isn’t treated before testing the researchers will not know if they are finding new biomarkers for vitamin D deficiency or ME.

    Thankfully though there are a lot of signs amongst comments being made by researchers that they are finally becoming aware of all these issues and that the symptoms of ME overlap with those of many diseases and that ME is a rare disease that must be separated from all the other illnesses. And we are now seeing a situation where very good researchers across many fields are communicating and sharing this information. Would be nice if the likes of the UK government were helping fund it, but I guess that is being just a little bit too hopeful.

    So it looks like we are now finally starting to move down the right road of knowing what the question is before trying to answer it! The question that should always have been asked is what is ME and the answer is, ME is a illness that has a set of symptoms that can vary greatly in different patients, that isn’t one of the many other diseases that can cause these symptoms!

    If as should always have been done all other diseases are excluded leaving only a pure cohort of ME patients. Then with the use of modern technology it is not going to take too long to be able to work out biomarkers and diagnostic scans etc that these patients fail.

    Unless we are very lucky and they find something like there is one virus found in ME patients that isn’t found in other diseases, like is the case with HIV, it is unlikely that a single test for diagnosing it will be found. But this is the case with most diseases.

    The most likely outcome is that we will end up like most other diseases, where a set of tests are used to diagnose, say immune tests, SPECT scans etc, and that there will be a reasonably small number of other diseases that can cause these failed test results, but that they can be ruled out and the diagnosis is confirmed. They will then be able to very easily study ME patients and find the cause and treatments.

    Then we will have the complete answer for what is ME, ME is a illness that has a set of symptoms that can vary greatly in different patients, that isn’t one of the many other diseases that can cause these symptoms! It is caused by such and such and these tests confirm diagnosis and it is treated with such and such

    All we can do is wait, and hope that the researchers are going in the right direction, and if anyone has any spare money and sees that any of the high quality researchers are asking for donations then give to them.

    All the best



     
  8. alex3619

    alex3619 Senior Member

    Messages:
    7,686
    Likes:
    12,520
    Logan, Queensland, Australia
    Hi rlc, I mostly agree with you, but there are some points of your argument I disagree with.

    First, ME like CFS is primarily a diagnosis of exclusion. It is not a validated disease entity, its a proposed disease entity. The primary reason for this is that Epidemic ME may be one, two or ten different entitites. We don't know if its one disease or a category of disease. It could be like cancer - a serious condition, but its a category, not a disease.

    Every diagnosis by exclusion is problematic. Psychosomatic diseases are diagnosis by exclusion - which includes hysteria - and one of the huge problem with have with ME is that this diagnostic problem lends itself to confusion with hysteria (which I regard as an erroneous diagnosis in nearly (?) all cases). A diagnosis of exclusion also lends itself to false negatives. If a person has a disease which might exclude ME, then its easy to jump to the conclusion that it does exlude ME. In many cases this may be correct, but in some cases a dual diagnosis would be necessary.

    The second issue is that you have mischaracterized the problem in using biomarkers. As the spinal proteomic study showed you can simultaneously test multiple disease groups. You could for example have MS and RA and depression cases in the study. However I do agree that you can not be certain - indeed all diagnostic tests are uncertain. Science is not about certainty. I also agree that it is unlikely that a single biomarker can do the job, and I have been calling for the testing of multiple biomarkers for some time. By combining different biomarkers you can increase both sensitivity and specificity, though there may be some trade-off between the two.

    A second issue is the failure to consider the impact of technology. As the technology advances the speed with which proteomic analysis can advance will increase vastly. This is not a slow steady development, just as it wasn't with the Human Genome Project. Its not just speed that will advance either - costs will dramatically decrease, as will resource requirements. This may be standard experimental technology in ten or twenty years (though the future is never certain).

    Is it efficient to switch to broad cohorts using current experimental methodologies and resources? No, not universally. A broad cohort has two purposes that I see - to increase the spread of candidate biomarkers and to assist validation (and refutation) of current diagnostic entities. I think the first casualty would be the Oxford definition. I suspect this would indisputably show the massive heterogeneity of this definition.

    One of the reasons I originally advanced the broad cohort argument was in relation to the PACE trial. If the PACE proponents were serious they would have gathered enough data to characterize all these patients under every diagnostic criteria and this would have been a test of the robustness of their definition versus other definitions. I think it would probably have shown the Oxford definition to be bogus however, and this is one reason why they shy away from comprehensive testing.

    I think the only way to get a pure cohort is to take every patient from one epidemic. There might be a case to take them from two geopraphically or temporally connected epidemics, but thats about it. If they come from two epidemics, or include sporadic cases, then they may not be a pure cohort.You then have the problem of finding out if patients from different epidemics have the same or a similar disorder - which can be done but then you have to use the very techniques I have been talking about.

    Just for comment on epidemics, I think there is substantive evidence that the epidemics never stopped they are just being misreported. I hope to write a blog on this. What led me to this conclusion was the repeated Q fever epidemics in the Scandinavian countries. This leads to ME or a similar condition almost automatically in roughly ten percent of cases (12%?), but its not being reported as ME, its being reported as Q fever. How many other diseases is this happening to? How about SARS for one? There is not much doubt that ME has a transmissable trigger (if not cause), and historically its been associated with polio and perhaps coxsackie virus. If we are reporting the pathogens but not the post-viral consequences, then its misreporting.

    So my conclusion is that we still do have ME epidemics. So we can get pure cohorts if we want, but we have to go to where the epidemics are.

    Bye, Alex
     
    SOC and merylg like this.
  9. Mark

    Mark Acting CEO

    Messages:
    4,532
    Likes:
    2,016
    Sofa, UK
    I don't really see the proteomics as a practical biomarker in clinical use, but it really does seem that cluster analysis of the proteomics on mixed cohorts is providing a route to understanding how those mixed cohorts should be properly subdivided. The study last year - remarkably - found radically different patterns in Lyme vs GWI vs ME/CFS, and Baraniuk's work is finding 4 distinct patterns of proteomics within patients classified as ME/CFS, which match the 4 clusters he previously identified based on analysis of symptoms and their severity. These analyses are really striking because they are looking at hundreds of variables and finding very distinct patterns which are consistent within the various subgroups defined by other means. Every population they've looked at so far has had a distinct signature and it seems reasonably likely that will continue to be the case when other conditions are similarly investigated.

    When you take that data and are able to say: This group of patients have very similar proteomics, and very similar symptoms, at that point you are starting to get a very good understanding of how the subsets should properly be defined. Even if you were to start out with a somewhat mixed cohort in some study using only ICC but including patients who have some other undiagnosed condition, the proteomics are likely to be able to tell you which patients don't really belong in that group - then you can remove those outliers and continue the study on a homogeneous group. It's early days with these results, but I think they show a lot of promise for defining the subgroups more accurately, on a biological basis, and really, such an analysis seems to me likely to be much easier to perform, more feasible, and also more reliable than the process of purely using a symptom-based definition and testing for 40 or more exclusionary conditions.

    As I say, I don't see this as necessarily producing a biomarker, but in terms of getting an understanding of the proper sub-groups and then analysing each sub-group further to see what else they have in common, how they relate on other potential biomarkers, I think it's really promising. A similar process could apply with the investigation of neuropeptides going on at Bond: the detail of that analysis, and cluster analysis on those data sets, is ultimately going to lead to a much better understanding of the disease process in those different subsets. I think it's going to take a few years, but when these various strands begin to join up we should start to see subsets of proteomics data, neuropeptide data, symptom data and other biomarker data, which gradually begin to present consistent patterns and separate out all the different conditions on a more objective basis (and yes, those subsets will probably also include some of those 'exclusionary' conditions, many of which are also unexplained syndromes).

    I think we are really talking about different things here, different types of research, and they're both valid. We all agree that for studies like Lipkin's, using the ICC and the strictest cohort definition, studying the sickest and most tightly defined subgroup, is the best means we currently have to get what is hopefully a consistent group of patients, and so that's likely to lead to more meaningful results. But I would put it like this: If you have a big barrel of mixed-up fruit, and the very question you are trying to answer is: How do you objectively distinguish which of these are oranges, then it rather begs the question to say: "I have a definition of an orange which I think is accurate so I'm only going to study those". If that definition happens to include tangerines as well, and you don't realise that, because tangerines aren't covered by your exclusions either, and they fit your criteria as well, you're still going to have a mixed cohort without knowing it. And if you want to work out and prove the exact differences between oranges, apples, pears and bananas, there is no way you can achieve that understanding by only studying the orange fruit: at some point you have to compare the different types to work out the differences between them, and I think that's the kind of broader sub-setting study that Alex, myself, and others are talking about here. To put it simply: it's impossible to investigate or prove an objective difference between ME and CFS patients by only studying ME patients.

    I do think we're talking about two very different types of research. One is where we study some aspect of ME, and there, everyone is agreed that the ICC should be used, with as much testing for exclusionary conditions as possible. But the other type of study, a study that aims to put the subsets on an objective footing, looks at those with Fukuda as well. Testing for known exclusionary conditions should certainly still take place, again nobody disputes that, but with larger mixed cohorts it is possible to study those differences between the subgroups and get a better understanding of exactly how they differ.

    One more example which was described at IiME: Dr Kogelnik described a database system at the Open Medicine Institute which already holds details of 10,000 patients, and is soon to be opened up for patients worldwide to enter their data, including data from a variety of objective tests: that's a project with enormous potential for cluster analysis to disambiguate the different subgroups, and that project clearly does not need to restrict itself to ICC definitions. It's analysis of huge datasets like these that is now more feasible than it was in the past, and I think that such analysis is one of the most promising options open to us: this can really help us to define and understand these subsets objectively, and the more data we have there, the better the chances of figuring it all out.

    It seems to me that there's this view that's sometimes expressed that seems to say that the strictly defined group of ME patients fitting ICC and who have been excluded from 40+ other conditions (through detailed and expensive testing that isn't available to most of us) form the real group of ME patients who deserve intensive study (which is absolutely fine by me), and that therefore everybody else 'just' has CFS which is actually entirely composed of people with known but undiagnosed conditions (which is not fine by me). It's a simple proposition, but I still don't see the evidence to support it. Whenever I've been presented with a list of those other conditions which 90% of CFS patients are supposed to have, undiagnosed, it strikes me that the items on those lists fall into two categories: known, discrete, serious conditions which definitely don't apply to me; and other syndromes like IBS, Sjogren's, Fibromyalgia etc which are just a list of other syndromes which frequently overlap with CFS and are no more understood, no less physical, and no more treatable than CFS is. I have no problem with supporting calls for research into cohorts of the sickest patients using the most tightly defined criteria possible - I certainly think that is the best practice and that should be the main focus - but that doesn't mean that everybody who does not fit those criteria has a well-known and understood diagnosis ready and waiting for them, neither does it mean that those other patients should not be studied, neither does it mean that many of those other patients do not have a disease which is at least closely related to ME, and neither does it mean that there's no value in studying mixed cohorts with the specific aim of objectively defining precisely what the differences are between the subgroups within those cohorts.
     
    Sing, Bob, SOC and 2 others like this.
  10. Ember

    Ember Senior Member

    Messages:
    1,742
    Likes:
    1,827
    Thanks, Mark. I'm curious to know what patient data, besides “data from a variety of objective tests,” the Open Medicine Institute is collecting. There's no indication on their website, but patients who have contributed must know.

    Since the IiME Conference, Jorgen Jelstad has posted a second blog in English. He knows how to tell a story:
    (Oops... The same article was published in this month's Journal of IiME.)
     
    SOC and Bob like this.

See more popular forum discussions.

Share This Page