By David Tuller, DrPH
Last year, a team from the University of Edinburgh released a pre-print called “Replicated blood-based biomarkers for myalgic encephalomyelitis not explicable by inactivity.” At the time, I posted an interview with the lead investigator, Chris Ponting, a professor of genetics at the university.
The peer-reviewed version of the study has just been published by the journal EMBO Molecular Medicine. According to a press release from the University of Edinburgh, “The largest ever biological study of ME/CFS (myalgic encephalomyelitis/chronic fatigue syndrome) has identified consistent blood differences associated with chronic inflammation, insulin resistance, and liver disease.” The work involved collaboration between the university’s Institute of Genetics and Cancer and the Schools of Mathematics and Informatics.
The paper has received significant media coverage, including a report in The Times. A BBC article featured the following crisp summary of the implications of the research: “Prof Chris Ponting said: ‘For so long people with ME/CFS have been told it’s all in their head. It’s not. We see it in their blood.’”
The study is a complicated read for non-scientists. In short, the investigators examined levels of more than 3,000 blood-based molecular and cellular traits from a huge healthcare database repository called the UK Biobank. These biobank data were for more than 1400 people who reported having received an ME/CFS diagnosis and more than 130,000 people who did not. The analysis identified 116 traits that were significant among both female and male ME/CFS patients. Some of the findings were replicated in a separate, smaller dataset.
Furthermore, the investigators found that the results were not associated with reported levels of activity. This finding serves to undermine the long-standing theory that deconditioning is a major cause of the symptoms in ME/CFS. That theory, of course, formed the basis of the psycho-behavioral treatment approach embodied in the fraudulent PACE trial’s interventions–graded exercise therapy (GET) and a specialized form of cognitive behavior therapy (CBT).
Per the study: “Evidence of a large number of replicated and diverse blood biomarkers that differentiate between ME/CFS cases and controls should dispel any lingering perception it is caused by deconditioning and exercise intolerance.”
Although the study revealed population-level differences in the biomarkers between cases and controls, the investigators stressed that these biomarkers cannot currently be used to distinguish individuals with ME/CFS from those who don’t have it. However, they stressed the importance of expanding on their investigation, noting that the findings should help to “accelerate research into the minimum panel of blood traits required to accurately diagnose ME/CFS in real-world populations.”
The UK’s Science Media Centre, which has a long history of endorsing the discredited claims of the GET/CBT ideological brigades and whose former head compared critics of the PACE trial to Nazis, responded by posting two expert comments. One was from Alan Carson, a professor of neuropsychiatry at the University of Edinburgh, a PACE truther, and a leader in the field of functional neurological disorder (FND). (Professor Carson blocked me on then-Twitter long ago because I criticized unwarranted claims from him and others about research into FND prevalence, diagnosis, and treatment.)
In his comment, with its snide and snippy tone, Professor Carson comes across as somewhat aggrieved. That wouldn’t matter much if he were right on the facts. But he is wrong. His main complaint is that, since the study investigated more than 3,000 traits, it “is not very exciting” that 116 would appear to be significant; that would happen by chance alone, he claims. This is false. Oops!
When a study involves a large number of tests, standard practice is to use specific statistical methods to reduce the likelihood of positive results arising by chance. Professor Carson apparently did not recognize or understand that the investigators took such steps to correct for the number of tests they conducted. His comment can be dismissed outright on the grounds of cluelessness. It is disturbing that the SMC has not acknowledged this indisputable error.
Luckily, the other expert—Professor Kevin McConway, an emeritus professor of applied statistics at the Open University–offered an extensive, thoughtful, and cautiously positive review of the study. As he noted:
“I think this is an important piece of research, but it’s also important to be careful not to claim too much from its findings. There’s a lot more to do.
“The press release and the research paper both make it clear that these findings could help in finding a set of blood biomarkers that can reasonably reliably distinguish people with ME/CFS from those who do not have that condition, but that, without a lot of further work, the findings do not in themselves provide such a set of biomarkers.”
I want to preface my remarks on this study, by saying that I’m very grateful to Dr. Chris Ponting and his team, who worked on this study without compensation. I wish our ME/CFS biomedical researchers were funded properly and adequately based on objective measures like prevalance, disease burden, economic impact, need for a diagnostic test and treatment.
I have two concerns about this study from a non-scientist ME/CFS patient perspective (who happens to be obsessed with ME/CFS biomarkers).
The first concern is that 1455 ME/CFS cases used in this study from the UK biobank are based on self-report. Self-reported diagnoses are not reliable, and they have been shown not to be reliable in ME/CFS patients.
My second concern is that of the 1455 cases at the UK biobank, only 55% have PEM. Which means only 55% of the 1455 cases have ME/CFS.
Since by definition, a ME/CFS patient must present with PEM. 45% of the patients misclassified as ME/CFS patients did not have PEM, and were therefore, by definition, not ME/CFS patients. The 45% of patients within the 1455 that didn’t have PEM should have been excluded from the beginning. It would have made the number ME/CFS patients smaller, but I think it would have added to the robustness and reliability of the data analysis.
Thankfully, the researchers did do an analysis of ME/CFS patients with PEM versus patients without PEM, but it is not labelled as an analysis ME/CFS patients with PEM versus Non-MECFS patients which is what it actually is.
I think how the ME/CFS patient dataset were treated by both the UK Biobank and Dr. Ponting’s team in this analysis are a prime example of why we need a ME/CFS Biomarker Conference to lay out:
* a Biomarker protocol to agree upon which ME/CFS diagnostic criteria and patient selection criteria should be used by biobanks, and biomarker researchers to stringently enforce a uniform ME/CFS patient selection so that all blood biomarkers can be compared apples to apples.
* A biomarker roadmap to help get us from ME/CFS biomarker pilot studies, to large-scale studies like OMF BioQuest, to meetings with regulatory bodies like the FDA, CIHR, NHS et cetera.
* To showcase the dozen or so developing ME/CFS Diagnostic Biomarkers to Industry and media.
* A 2-Day CPET Biobank where blood, tissue and medical records from all ME/CFS, healthy controls, and MS patients who have done 2-Day CPET is banked for use by researchers for blood biomarkers.
From the paper:
“Among the ME/CFS cases with available Pain Questionnaire data, 55% reported long-lasting PEM-like symptoms (“Methods”). This allowed us to compare blood trait results for those with PEM versus those without PEM symptoms. For these two analyses, we used disjoint sets of controls and equal numbers of cases to match statistical power. Results from the two analyses were highly concordant (Fig. 8A). Nevertheless, whereas 26 biomarkers were significant in the PEM analysis, only 9 were significant in the non-PEM analysis. Five blood traits not previously significant (i.e., not shown in Fig. 2A) were significant only in the PEM analysis, and not in the non-PEM analysis: albumin, direct bilirubin, eosinophil percentage, haemoglobin concentration and reticulocyte count. Notably, two biomarkers of chronic inflammation—cystatin C and C-reactive protein—are significant only in the PEM analysis, and not in the non-PEM analysis. We conclude that UKB individuals with PEM have stronger ME/CFS biomarker differences than those without PEM. The full results of the analyses on the combined, male, and female cohorts can be found in Dataset EV11.
What this study suggests to me is that ME researchers need to agree on using a set of criteria for ME definition that includes PEM and doesn’t rely on self-report or medical diagnosis. It shouldn’t be acceptable in this day and age for doctors to lazily label patients with ME/CFS just because they have chronic fatigue that isn’t explained by basic tests. I would have expected FND proponents who support the positive diagnosis of FND and removal of diagnosis by exclusion in their own field to now be calling for a positive definition of ME/CFS that includes PEM for both clinical practice and research, so why don’t I hear them doing that? To my mind, it is only by employing tighter criteria that the mess that was created through decades of application of vague criteria and diagnosis by exclusion will start to be ironed out, and that needs to be done as much for those who don’t have PEM as for those who do so that other causes of chronic fatigue can also be mapped out.
(I would add that thorough exclusion of known diseases/conditions should still be a key part of the diagnostic process but should not be the basis for diagnosis of any disease/condition, physical or mental, or application of stigmatizing label such as the patient having a “functional” illness.)
For research, we should firstly define & use a definition that makes the diagnostic group very specific & distinct from normal. This definition doesn’t define patient diagnosis used in practice & care (which can be broarder). Once we target the most obvious, the most in need, then we can broaden the definition to see if the trend continues for less severe cases. Trying to reduce the noise & confounders to begin with, then later investigate subgroups that don’t follow that overly restrictive definition. The results may help to define subgroups but we have to be careful not to make them too circular reasoned. Ultimately if a treatment shows significant objective benefits for a define subgroup then the theory of biological marker can be verified.
Excellent comments here. We have clear criteria for Myalgic Encephalomyelitis in the International Consensus Criteria (ICC) which is an improvement on the Canadian Consensus Criteria (CCC). The International Consensus Primer (ICP) provides a diagnostic tool.
When the US CDC created the name CFS in the ‘80’s, knowledgeable clinicians and advocates fought to keep the name ME, unfortunately the hybrid ME/CFS with broader criteria was born. It’s nonsensical and not based on science. Strauss and the CDC held ME patients in disdain which has been well documented by Hilary Johnson in Osler’s Web and The Why. Excerpts can be read on Substack. It’s imperative to understand the history to unravel the mess we’re in.
If this is not addressed, all research based on anything less than the ICC is wasting our time and lives.
A little history lesson & addendum to my first comment.
https://virology.ws/2011/11/23/chronic-fatigue-syndrome-and-the-cdc-a-long-tangled-tale/
Thank you for outlining this. From reading your article and the comments it’s clear the tangled mess of M.E needs clearly defined pathways for further research and ongoing development of such studies.
At least it’s moving in better directions with some compassionate scientists who really do care like Chris Ponting.
I wonder if some people who do not report PEM, do so, as they are too severely ill to move etc more, to even experience a crash/flare/drop in baseline and so difficult to infact identify they are experiencing PEM as existing in such a minimal debilitated manner anyway?
I also believe that any Doctor refuting me/cfs, which all the extensive evidence is more than a little unhinged ie medically incompetent, as its blatantly obvious and has been for decades and so requires an official rebuke and enforced me/cfs training from their Professional body.
A couple of points to clarify.
1) The patients were not self-diagnosed. They reported having been diagnosed by a clinician as having some version of ME or CFS. That’s not the same as confirmed ME per acceptable criteria. But it’s not the same as self-diagnosed.
2) Only a subset of the patients were recruited at a time when they were being asked about PEM. We don’t know how many in the full patient sample would have reported or not reported having PEM.
Sandra Mayor: Thank you. Important observations.
Although not discussing ME/CFS, this review on Functional Neurological Disorder may be of interest in this discussion:
“The results suggest that FND is a neurological disorder, in view of its morphological abnormalities demonstrated in numerous brain imaging studies and considerable behavioral characteristics that are shared with canonical neurological disorders, including the many instances of clinical transition between FND and other neurological disorders. Moreover, the results indicated many instances in which acute focal cerebral hypometabolism in persons with FND receded in parallel with clinical improvement. These sources of evidence imply that FND is a neurological disorder.”
Source: Victor W. Mark. Biomarkers and Rehabilitation for Functional Neurological Disorder. J Pers Med. 2024 Sep 7;14(9):948. doi: 10.3390/jpm14090948
Full article: https://pmc.ncbi.nlm.nih.gov/articles/PMC11433361/