Blood parameters differentiate post COVID-19 condition from Myalgic Encephalomyelitis/Chronic Fatigue Syndrome and Fibromyalgia, 2025, Giménez -Orenga

Dolphin

Senior Member (Voting Rights)

Brain, Behavior, & Immunity - Health

Available online 4 July 2025, 101058
In Press, Journal Pre-proof

Blood parameters differentiate post COVID-19 condition from Myalgic Encephalomyelitis/Chronic Fatigue Syndrome and Fibromyalgia


Karen Giménez -Orenga 1, Justine Pierquin 2, Joanna Brunel 2, Benjamin Charvet 2, Eva Martín-Martínez 3, Margot Lemarinier 4, Steven Fried 4, Alexandre Lucas 4, Hervé Perron 2 5, Elisa Oltra 6
1Doctoral School. Catholic University of Valencia San Vicente Mártir
2Geneuro-Innovation, Bioparc Laënnec, Lyon, France
3National Health Service, Manises Hospital, Valencia, Spain
4Institut des Maladies Métaboliques et Cardiovasculaires, INSERM, University Toulouse III–Paul Sabatier, UMR 1297-I2MC, Toulouse, France
5GeNeuro, Geneva, Switzerland
6Department of Pathology, School of Medicine and Health Sciences, Catholic University of Valencia, C/ Quevedo 2, 46001 Valencia, Spain

Received 10 February 2025, Revised 13 June 2025, Accepted 3 July 2025, Available online 4 July 2025.



https://doi.org/10.1016/j.bbih.2025.101058


ABSTRACT​

Post-COVID-19 condition, such as Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and Fibromyalgia (FM), are characterized by fatigue, pain, shortness of breath, sleep disturbances, cognitive dysfunction and other symptoms, heavily impacting on patients daily functioning.

Moreover, over half of patients end up fulfilling ME/CFS and/or FM clinical criteria after a few months of SARS-CoV-2 infection.

Expression of the toxic human endogenous retrovirus (HERV)-W ENV protein can be induced by viral infection and HERV-W detection was correlated with acute COVID-19 severity and found significantly expressed in post-COVID-19 condition.

This study shows that HERV-W ENV may also be present in prepandemic cases of ME/CFS, FM or co-diagnosed with both clinical criteria, suggesting viral participation in these chronic diseases.

To learn whether associated antiviral mechanisms may also show differing patterns of immunological responses, we measured IgM, IgG, IgA and IgE antibody isotypes against SARS-CoV-2 spike and nucleocapsid antigens, the levels of IL-6, IL-8, IL-10, IFNγ and TNFα cytokines, the level of NfL, a neural damage biomarker, as well as some blood cell markers potentially related with fatigue.

Importantly, some of the measured variables showed a capacity to discriminate post-COVID-19 condition cases from all other participants, with 100% sensitivity and up to 71.9% specificity providing a new tool for a differential diagnosis between diseases or syndromes with so many overlapping clinical symptoms.

Interestingly, the detected markers showed moderate-to-strong correlations with patient symptoms pointing at novel therapeutic opportunities.

KEYWORDS​

HERV-W-ENV
post COVID-19 condition
long COVID-19
SARS-CoV-2
Serology
Immunoglobulins
Cytokines
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome
Fibromyalgia

 
Post-COVID-19 condition, such as Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and Fibromyalgia (FM), are characterized by fatigue, pain, shortness of breath, sleep disturbances, cognitive dysfunction and other symptoms, heavily impacting on patients daily functioning.
Defining ME/CFS as a PCC is just wrong. And no mention of PEM.
Moreover, over half of patients end up fulfilling ME/CFS and/or FM clinical criteria after a few months of SARS-CoV-2 infection.
That’s also wrong, those studies were not at all suitable to determine the prevalence of ME/CFS in PCC.
 
Defining ME/CFS as a PCC is just wrong. And no mention of PEM.
I don’t think they’re saying ME/CFS is only a sequelae of COVID. They’re just saying that the umbrella terms Long COVID or PCC includes many cases of ME/CFS.

Otherwise they wouldn’t mention non-COVID related cases of ME/CFS. Like this clause in the abstract
prepandemic cases of ME/CFS,
 
I'd summarize it as: "There's something going on with the immune systems." The question is "precisely what?" I think there's potential in figuring out what the commonalities are. Is it t-cells, or NK cells, or a specific cytokine or ratio of specific cytokines?
 
I can't find any mention of multiple testing correction whatsoever in either their legends, results, or methods.

View attachment 26916

There would be 60 pairwise comparisons just in this one panel. Am I missing something?
If you are hypothesising differences in each one, individually, do you need to? I probably wouldn't apply it in this case if I was the author, I would see it as 6 different experiments, each having 5 groups
 
If you are hypothesising differences in each one, individually, do you need to? I probably wouldn't apply it in this case if I was the author, I would see it as 6 different experiments, each having 5 groups
In this case I don't think they actually have a strong a priori justification to treat these as separate hypotheses. It's framed as an exploratory analysis for the purposes of predictive markers based on the hypothesis that there will be some differential immune perturbations between these conditions.

Given that cytokine levels can reflect immune disturbances potentially linked to the observed skewed anti-SARS-CoV-2 humoral response in some patients, we next assessed circulating levels of IL-6, IL-8, IL-10, IFNγ, and TNFα in these same cohort. In addition, NFL plasma levels were measured to evaluate potential neuronal damage and peripheral neuropathy, which are frequently reported in these patient populations (Maalmi et al., 2023).
It's not so much "I have reason to believe that interferon gamma, IL-10, IL-6, etc. etc. are each different between these conditions," it's "I have reason to believe that there will be differences in immune signaling between these groups, and those differences might be reflected in interferon gamma and/or IL-10 and/or IL-6 etc.," which is pretty much the classic case when correction is necessary. At most, they could have treated NFL as a separate hypothesis from the rest of them.

In my multi-omic analyses, I don't consider it necessary to do corrections for everything when I am trying to screen for any predictive features and determine which ones to use cumulatively as a module score, for example. And that seems to be what they're doing for the final step of the analysis. But whenever I am highlighting any of the features individually, I always do corrections because it is still a "throw things at a wall and see what sticks" scenario unless I have separate justification for each feature, in which case I would make sure to put that justification in the text with citations.

It may be more on the stringent side, but I always prefer to have significance drop out and let the visual difference in means speak for itself, rather than have an uncorrected p-value give the false impression that a weak difference on the graph is more important than it is. If my analysis relies on having significance for particular individual features to test a hypothesis, then I won't be doing that in the same part of the analysis where I'm doing an exploratory screen.
 
In this case I don't think they actually have a strong a priori justification to treat these as separate hypotheses. It's framed as an exploratory analysis for the purposes of predictive markers based on the hypothesis that there will be some differential immune perturbations between these conditions.


It's not so much "I have reason to believe that interferon gamma, IL-10, IL-6, etc. etc. are each different between these conditions," it's "I have reason to believe that there will be differences in immune signaling between these groups, and those differences might be reflected in interferon gamma and/or IL-10 and/or IL-6 etc.," which is pretty much the classic case when correction is necessary. At most, they could have treated NFL as a separate hypothesis from the rest of them.

In my multi-omic analyses, I don't consider it necessary to do corrections for everything when I am trying to screen for any predictive features and determine which ones to use cumulatively as a module score, for example. And that seems to be what they're doing for the final step of the analysis. But whenever I am highlighting any of the features individually, I always do corrections because it is still a "throw things at a wall and see what sticks" scenario unless I have separate justification for each feature, in which case I would make sure to put that justification in the text with citations.

It may be more on the stringent side, but I always prefer to have significance drop out and let the visual difference in means speak for itself, rather than have an uncorrected p-value give the false impression that a weak difference on the graph is more important than it is. If my analysis relies on having significance for particular individual features to test a hypothesis, then I won't be doing that in the same part of the analysis where I'm doing an exploratory screen.
Sure. I agree on all fronts. I was going by my first impression, having looked at the graphs, so I'm not sure what their internal framing was for the questions being asked - your take is probably more well informed than mine :)
 
I agree. A multiple comparison correction is needed. I have got so used to it being missed out that I tend to interpret raw p values in this context as needing a correction unless that is specified.

And p values are pretty irrelevant anyway if you want these tests to 'differentiate'. They don't. They overlap almost entirely.
 
Back
Top Bottom