Preprint Reactivated EBV, HHV6, HAdV in Sputum from ME/CFS Patients: Are autoAbs to IFN-I Impairing Antiviral Immunity?, 2025, Hannestad et al.

Full title: Reactivated EBV, HHV6, HAdV in Sputum from Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Patients: Are autoAbs to IFN-I Impairing Antiviral Immunity?

https://www.preprints.org/manuscript/202502.0185/v1

Abstract
An exhausted antiviral immune response is observed in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and long COVID. In this study, potential mechanisms behind this exhaustion were investigated.

First, the viral load of EBV, HAdV, human cytomegalovirus (HCMV), human herpesvirus 6 (HHV6), and SARS-CoV-2 was determined in sputum samples (n=29) derived from ME/CFS patients (n=13), healthy controls (n=10), elderly healthy controls (n=4), and immunosuppressed controls (n=2).

Secondly, autoAbs to type I interferon (IFN-I) in sputum were analyzed to possibly explain impaired viral immunity.

We found that ME/CFS patients released EBV at a significantly higher level compared to controls (p=0.0256).

HHV6 was present in ~50% of all participants at the same level.

HAdV was detected in two cases with immunosuppression and severe ME/CFS, respectively.

HCMV and SARS-CoV-2 were found only in immunosuppressed controls.

Notably, anti-IFN-I autoAbs in ME/CFS and controls did not differ, except in severe ME/CFS with high levels.

We conclude that ME/CFS patients, compared to controls, have a significantly higher load of EBV. IFN-I autoAbs cannot explain IFN-I dysfunction, with the possible exception of severe cases showing elevated autoAbs, also reported in severe SARS-CoV-2. We forward that additional mechanisms, such as viral evasion of IFN-I effect, may be present in ME/CFS, which demands further studies.

 

Despite there being I imagine many confounding factors and the sample size being small, the fact controls were age matched, and the diagnosis was CCC seems a good sign. I have two questions.
1) Is this a new finding? I find it hard to believe this hasn’t been tested before.
2) What does this mean. Is this evidence towards the often repeated “immune exhaustion”?

 

I believe there must have been dozens of studies looking at EBV in saliva (they did so themselvels as well: https://www.s4me.info/threads/saliv...with-me-cfs-2022-apostolou-et-al.30094/page-2). Not sure about sputum or why it would be different. Unfortunately the discussion in the paper isn't very revealing. Whilst EBV has been studied hundreds of times they simply quote the one study that had positive results, which doesn't inspire much confidence. I would guess they didn't perform any statistical corrections (indeed that appears to be the case).

The main results appear to be in table 1. Most ME/CFS patients here have a positive "PCR EBV sputum test", all elderly controls do as well and also half of the rest of controls. I think that would mean that a "PCR EBV sputum test" doesn't mean much given that its positive in so many people that are healthy. The rest of the findings aren't statistically significant.

 

If @forestglip feels up to it he might re-run the analysis. The fact that they mention a p-value in the abstract but don't mention the high likelihood of a type 1 error in the discussions isn't very confidence inspiring to me (if I understood what they did correctly). I think even with a very mild correction the results could actually still border on being significant.

 

I don't know enough to know how to test this. They don't say how they did it. The EBV values aren't normally distributed so I can't do a t-test between HD and ME/CFS, even after a log transform. Mann-whitney test gives p=0.12.

So yeah, I'm not the right person to ask.

Edit: Oh, they do say they used a Mann-Whitney (Wilcoxan rank): "Statistically significant difference was found between the HD and ME/CFS groups (p=0.0256) according to non-parametric Wilcoxon rank procedure"

So I don't know why my p value is so much higher.

Here's the data from the two groups if anyone wants to look. It's not the exact values, I pulled them from the Fig 2A image.

Spoiler: EBV (copies/mL)

y,group
0.0,HD
0.0,HD
0.0,HD
0.0,HD
0.0,ME/CFS
0.0,ME/CFS
979.4303678029924,HD
1337.734981148877,ME/CFS
1800.1933328915295,HD
3408.471279891597,HD
3459.450271974564,ME/CFS
3781.741258813013,HD
5089.098735378267,ME/CFS
5730.8692866276415,ME/CFS
5903.579410448887,ME/CFS
10850.78086635713,ME/CFS
13160.666538816113,ME/CFS
60724.72694052544,ME/CFS
271993.38971227506,ME/CFS
579933.859083633,ME/CFS
878834.6197806812,HD
878834.6197806866,HD
1182649.960969652,ME/CFS

 

Oh, nevermind, I guess I misread the plot. Or they made some mistake. Tell me that doesn't look like two overlapping green dots at the very top, making it 6 total healthy donors with EBV:


But they say only 5 healthy donors had detectable EBV:

Which matches with Fig 1 that only shows 5 had detected virus.

When I change one of the two dots I marked as really high to 0, Mann-Whitney gives me: p = 0.02556

So that's the p value without multiple test correction.

 

These seem to be the main results for EBV. Notable that the seniors had similarly elevated viral load. So perhaps it was mostly the control group that was unusual?

 

Agree it looks like more than one dot to me. I suspect it's easier to make a mistake in the text and p-value calculation than plot a wrong data point like this. So I wonder if the plot might be the correct version.

 

There are 100 % two green dots at the top.

 

So changing one data point makes the result statistically insignificant? Whatever the right version of the dataset, it's clear that these results aren't very robust.

 

I found an email address for the lead author in another paper. I'll ask him about it.

 

Good thing it’s a pre-print! That being said, I’m not confident that peer review would have caught it if there actually is an error somewhere.

 

If the plot was correct then figure 1B wouldn't be correct, so I would suspect that it indeed might be the plot.

 

I'm dopy at present but trying to understand things - what are autoAbs?

 

Autoantibodies - antibodies against a person's own tissues.

 

Thank you, @forestglip!

 

They did mention it's Mann Whitney. Just in a place that seems slightly odd so I missed it at first, in the image captions.

I replicated it with a two tailed Mann Whitney.

 

Sorry, I deleted my comment right as I realised that they did describe it (just not next to the test outcomes as I expected, so I missed it) but then your reply just came through after that. I had edited in a thing acknowledging your comment which I only saw after I commented, but then I also realised I misread the paper so I deleted it to avoid a mess.

All good!! Appreciate your dilligence

 

Version 2:
https://www.preprints.org/manuscript/202502.0185/v2

 

They didn't respond to my email and the EBV plot for HC in version 2 still shows 6 dots while the text says 5 were positive.