No, COVID-19 vaccines don’t suppress your immune system—here's the science
A recent preprint filled with fundamental scientific flaws is fueling anti-vaccine disinformation.
This newsletter is free, but it’s able to sustain itself from support I receive from a small percentage of regular readers. If you value science-based information, consider upgrading to a paid subscription:
A recent preprint, “Immunological and Antigenic Signatures Associated with Chronic Illnesses after COVID-19 Vaccination” has fueled anti-COVID-19 vaccine rhetoric. It is being circulated by anti-vaccine advocates as proof that COVID-19 vaccines cause immunodeficiency and therefore, their claims for the past 5 years are validated.
The study doesn’t show any of this. In fact, the study can’t tell us anything about potential adverse impacts of COVID-19 vaccines.
There are also large social media accounts who, because they lack field-specific knowledge, have shared inaccurate summaries about the study and findings across widely shared posts, blogs, and videos, further contributing to this misinformation.
The inability to correctly interpret complex data in this study is a perfect example of the importance of why specialized expertise matters with complex science. In the current information ecosystem, the desire for engagement seems to supersede an ethical responsibility to defer to experts–ironically, something many of these accounts profess to their followers.
Anyway, let’s get into it.
Akiko Iwasaki—a well-known immunologist, former President of the American Association for Immunology, and senior author – shared this in her Twitter feed:
Our study revealed differences in circulating immune cells, antibodies, EBV reactivation, and spike proteins in those with PVS. We accounted for SARS-CoV-2 infection status based on anti-nucleocapsid antibodies. Larger studies are needed to validate these findings
But did this study actually do any of this? The short answer? Not at all.
This is a cross-sectional study, where two groups are compared at a single point in time to assess differences. The samples aren’t controlled, merely observed, so these types of studies are considered exploratory at best.
The two groups are blood samples from 1) people who self-reported symptoms of illness following COVID-19 vaccination and 2) people who didn’t. Here, self-reported symptoms of following vaccination are termed Post-Vaccination Syndrome (PVS), which is not a medically recognized condition.
Does this study actually identify persistent immune dysfunction after COVID-19 vaccination? No.
The big takeaways:
The study did not accurately correct for past infection. The methods used to “exclude” past infection as a variable is not accurate–the data presented actually suggest everyone has a similar history of past infection, which means the PVS symptoms reported by participants cannot be automatically attributed to vaccination.
The study didn’t actually assess T cell exhaustion. In order to actually do that, they would have needed to show markers of T cell exhaustion (TIM-3, CTLA-4, PD-1, etc) combined with reduced function (lower cytokine levels, reduced proliferation), metabolic defect, and gene expression changes. They don’t do any of this. In fact, IFN-γ and TNF-α are comparable in “PVS” samples and controls.
They did not even use a method to assess EBV reactivation. They only look at antibody levels against EBV proteins–an indication of past exposure to EBV and memory immune response. They did not assess EBV replication, which is required to show reactivation.
CD4 T cell populations aren’t actually different between groups and all are within normal ranges for healthy individuals. Effector memory cells are within the “healthy” range in both groups. There were no meaningful differences between the groups, or a divergence from normal ranges.
The authors analyzed blood samples from 42 people signed up to the LISTEN study with Yale who said they suffered from long-term effects of COVID-19 vaccination. These were compared to 22 blood samples from people also in the LISTEN study but said they did not have long-term health issues. I want to emphasize that this is an INCREDIBLY small sample size, meaning even if there were differences noted, you cannot make broad conclusions about them.
The LISTEN study (Listen to Immune, Symptom and Treatment Experiences Now) recruits people with self-reported symptoms of long COVID (post-acute sequelae of COVID-19, PASC). The groups (control versus PVS) are differentiated solely based on whether the person believe they have symptoms that are a result of vaccination.
Why do I mention this? These samples were collected between the end of 2022 and 2023 – when nearly everyone in the US had been infected with SARS-CoV-2 by September 2022, which means that it is incredibly likely that nearly all the samples came from people who had been infected with SARS-CoV-2.
Why does that matter? Well, because without correcting for the impact of the virus itself, you can’t make conclusions that 1) reported symptoms or 2) immune response differences are uniquely due to vaccination. The authors say they correct for this, but as you’ll see, this is not true.
The “uninfected controls” have similar evidence of past infection. This means there is no true control for COVID-19 itself in the context of this entire study.
The authors say they ruled out past infection by excluding samples with nucleocapsid antibodies. Nucleocapsid is a protein of the virus, SARS-CoV-2, but not a component of the vaccine. Thus, if people have antibodies against the nucleocapsid, then they must have been exposed to the virus whereas if they don’t have nucleocapsid antibodies, then they didn’t have COVID-19.
It does not automatically work that way. There are caveats to antibody testing.
You must test at the right time because it takes time for plasma cells to produce antibodies, so if samples were collected early in the course of infection, the results may yield a negative value.
COVID-19 vaccination and the memory immune response to it can reduce viral load and symptom severity. Vaccinated people also don’t necessarily produce antibodies against the nucleocapsid protein, which you can’t assume, even if they no antibodies are detected, that they weren’t infected, especially if they’re vaccinated (which everyone in this study was).
In this instance, the biggest issue is that their “uninfected” samples actually have measurable levels of nucleocapsid antibodies.
They use 2 methods: Elecsys Anti-SARS-CoV-2 immunoassay, a qualitative test that scores antibody levels based on a cut-off index (COI) to determine which are positive and negative, something this group has cited previously. But the study doesn’t report these COI levels. (it was also not in the supplementary data)
Instead, they include concentrations of anti-N antibodies in nanograms of antibody per milliliter (ng/mL), presumably from method 2, an in-house ELISA (enzyme-linked immunosorbent assay) which allows you to interpolate antibody levels based on a standard curve.
Reporting absolute levels of antibodies is fine as far as units go. Not fine when it comes to the values they report and how they normalize it. They say they separate “infected” samples based on the levels of nucleocapsid antibodies. Just take a quick look – particularly between the [-I] and [+I] groups – they overlap.
How can they say one group is “uninfected” and one is “infected” when the samples have similar levels of nucleocapsid antibodies? They can’t. The overlap of antibody levels between groups suggests this triaging by infection status was done arbitrarily.
Look at the Y-axis—it starts at 100 ng/mL.
Most assays for nucleocapsid antibodies have a detection threshold of 4 ng/mL: values above can be reliably measured and may be considered positive. Every sample has a measured level above 200 ng/mL – 50-times higher than typical detection thresholds. Even if you use a more conservative threshold of 30 ng/mL as was done here, that’s nearly 7-times higher.
These levels of nucleocapsid antibodies suggest ALL of these samples are from people with previous infection.
Here’s a new figure I made a new figure with the expanded Y-axis:
If measured antibody levels are well above detection limits including in “uninfected,” some exceeding 1,000 ng/mL, why are they calling them uninfected? They can’t.
That means their exclusion by “infection status” is false.
They also didn’t normalize these values to any negative controls such as blood samples collected before COVID-19 was circulating in order to subtract background signal (a common practice for any of these types of tests).
What does this mean – in simple terms?
They cannot say samples in their PVS group–and symptoms reported by those patients–are uniquely related to COVID-19 vaccination.
It is far more likely that symptoms – and any immune response differences – are due to previously having COVID-19.
But let’s dig into the immunology claims either way.
The study didn’t even measure T cell exhaustion.
While they state there is evidence of T cell exhaustion in PVS samples, they didn’t actually measure that.
T cell exhaustion–for non-immunologists–occurs when T cells become less responsive because of repeated stimulation. As a result, T cells are less able to mount an effector response to fend off infection, injury, or illness. T cell exhaustion is well-characterized in the context of cancers and chronic infections.
T cell exhaustion requires T cells to display a well-defined collection of hallmarks.
These include sustained expression of key T cell inhibitory molecules (surface markers): PD-1, LAG-3, TIM-3, or CTLA-4.
This must be coupled with functional impairment of T cells: reduced ability to produce key cytokines, reduced proliferation ability, & changes in genes involved in these processes (like TOX, Thymocyte selection-associated high mobility group box protein).
The study did not assess functional impairment of T cells, so they can’t say T cells were exhausted.
They state there are “significantly higher proportions of exhausted CD8 T cells,” which, for their purposes, refer to CD8 T cells (cytotoxic T cells) that express PD-1 and TIM-3. Let’s look at the data.
Now, I’m no expert (kidding–here is where I legitimately can say I am–) but these values are nearly identical—look at the median values (in the red box). The data overlap, and “significant” here refers to a p-value of 0.02, solely due to those outlier samples in the already tiny PVS group.
To me, it feels a bit like HARKing (“hypothesizing after the results are known”), a tactic used to try to manipulate data to fit a preconceived story.
There isn’t an increased proportion of T cells expressing PD-1 and TIM-3 and simply expressing PD-1 and TIM-3 doesn’t make an exhausted T cell. Are there functional impacts in T cells ability to produce cytokines?
Well, from the same figure, we can see very obviously: no.
The T cells in the PVS group are producing similar levels of cytokines as measured by TNF-α and IFN-γ. If anything, the good levels of TNF-α and IFN-γ in the PVS patient T cells contradicts the claim that there is T cell exhaustion, because these cytokines are indicative of an activated immune state, not an exhausted one.
They don’t assess function of these T cells.
Exhausted T cells have reduced cytokine production (we see the opposite here) and reduced proliferation, but no proliferation data is included.
If you want to claim T cell exhaustion, you must demonstrate T cells have reduced proliferation, express genetic markers of T cell exhaustion (like TOX), and have reduced metabolic function. You should also show the exhausted T cells were specific for the Spike protein.
None of this was conducted. They did not assess T cell exhaustion, nor did they show any differences between groups.
There were also no meaningful differences between CD4 T cells.
Their next point to support claims of “immune dysfunction among PVS patients” is higher levels of memory CD4 T cells (helper T cells) and regulatory T cells among control versus PVS samples.
CD4+ T cell level differences are minor and within normal ranges for healthy individuals across groups. Memory effector T cells account for 10-30% of all CD4 T cells and that’s exactly what we see in both groups. The other population they call out is a subtype of regulatory T cells, resting natural CD4 regulatory T cells. They are a very small fraction of CD4 T cells, accounting for ~1-3% of CD4 T cells, also what we see in both groups (note the Y-axis of that panel).
Based on these data, there is no evidence that PVS samples have impaired T cells, CD4 or otherwise.
The study also did not show EBV reactivation—or use a method to measure it.
The authors state PVS samples have evidence of Epstein-Barr Virus (EBV) reactivation, used to support their argument of immunosuppression following COVID-19 vaccination.
The data don’t show this, either.
EBV, like other herpesviruses, establishes latency after primary infection (Epstein-Barr virus causes infectious mononucleosis, colloquially mono). After recovery, EBV hangs out in B cells indefinitely. In certain circumstances–often during times of substantial immune system stress–the virus can emerge from this latency: this is reactivation.
For context, this is like cold sore or genital herpes recurrence (caused by herpes simplex viruses type 1 and 2) or developing shingles later in life following Chickenpox as a child (caused by the Varicella Zoster virus, also a Herpesvirus).
They only measure antibodies against EBV proteins, an indication of prior exposure and immune response, NOT reactivation of EBV.
To measure that, you’d need to look for evidence of viral replication by measuring increased levels of viral DNA using PCR (polymerase chain reaction).
Legitimate EBV reactivation occurs with conditions that impact immune system function: HIV, cancer treatments, organ transplantation, and…severe infections and post-acute infection syndromes (like COVID-19 and long COVID, perhaps?).
Aside from the fact there is no EBV reactivation evidence, antibody levels don’t meaningfully differ between the groups (and again, there are no appropriate negative control samples). This study does not enable you to say whether any differences are related to past infection, vaccination, or something else entirely (another reason pre-vaccine levels are needed to normalize to).
There are several more glaring issues, but for the sake of length, I’ll go quickly.
Their claim that PVS samples had higher levels of SARS-CoV-2 spike protein in blood plasma compared to control is also unsupported. The short explanation: spike protein is supposedly detected in both PVS and control samples (I say supposedly because the accuracy of the method used, a proximity assay, is also questionable).
The samples are only from one point in time, so you can’t make judgments about whether this is continued or increased production of spike protein because there aren’t prior samples to compare to. You also can’t exclude the likely possibility that spike protein is a remnant of infection, especially factoring in the discussion points above.
Dr. Dan Wilson did a great video where he tackled this section in detail here.
This study has serious methodological flaws and is being wildly overinterpreted by the authors and media outlets.
Is PVS a real phenomenon? Well, it depends on how you define PVS.
Some people have adverse reactions to vaccines – and those are real, but rare.
But we have data on those. Around 1 in 100,000 men and 1 in 1,000,000 women who receive an mRNA COVID-19 vaccine will experience myocarditis or pericarditis. Context matters here: myocarditis after COVID-19 is 10-100 times more common. Rare allergic reactions like anaphylaxis occur in around 1 in 1,000,000 individuals.
None of these are what the authors are terming PVS – their newly discovered “immunosuppression” caused by COVID-19 vaccination.
Do the data here show that people who believe they have PVS have immune dysregulation? No.
To this day, there is no evidence that COVID-19 vaccines cause negative impacts to your immune system. Now, some people have come out in defense of this by saying there is a need to study rare adverse events following vaccination.
I agree that there is a need to understand rare adverse events following vaccination.
But publishing this incredibly flawed study doesn’t contribute to that in a positive way. It actually undermines real adverse effects and fuels anti-vaccine rhetoric,
especially since it doesn’t show adverse effects from vaccination!
Look at Twitter and media headlines. If we want to study legitimate adverse effects, we need far better study designs with robust research controls, validation of functions of immune system components, and mechanistic explanations.
This fallout is made worse by several authors affiliations with pseudoscience & anti-vaccine disinformation.
There are obvious ones:
Briane Dressen, who formed the organization React19 after she has vocally claimed vaccine injury, and has made quite a name for herself since.
Danice Hertz, a gastroenterologist who has espoused anti-vaccine rhetoric even before COVID-19, and has referred patients to Simone Gold, the CEO of the anti-vaccine group America’s Frontline Doctors (also convicted for taking part in the January 6 riots).
David Putrino, a senior author, actively flirts with anti-vaccine and wellness pseudoscience.
David Putrino has made quite a name for himself by spreading disinformation about pseudoscience diagnoses like chronic Lyme. (see the American Lyme Disease Foundation for more).
Now?
Putrino is suggesting PVS–a condition not proven to exist–should be treated with Nattokinase.
Yes, this is the supplement Peter McCullough (the cardiologist who spread COVID-19 vaccine conspiracy theories and lost his medical license because of it) sells to protect people against “vaccine shedders” by “detoxing” after you’re around them. Dr. David Gorski has written about this several times.
Yes, Peter McCullough literally started a wellness company along with other anti-vaccine activists to sell disinformation and conspiracy theories – the ones that a senior author of this preprint is endorsing.
Putrino has long been dabbling with wellness disinformation. Unfortunately, this story is not a new one. The fact he has an academic appointment at a university is irrelevant (remember Mehmet Oz? Mark Hyman? Marty Makary? Jay Bhattacharya? Andrew Huberman?)
Putrino is using his authority to sell vulnerable people a story about unproven ailments with a laundry list of nebulous symptoms that he can sell you the cures for by trying to legitimize this unsupported claim of immune dysfunction following COVID-19 vaccination. That’s the only explanation I can sort out as to why this study was put online without any attempt to review the integrity of it.
Science is only as strong as the rigor behind it.
These types of flawed studies amplified by bad actors are doing irreparable harm, especially when they’re being used by those same people who are promoting “vaccine detoxes” and conspiracy theories.
COVID-19 vaccines do not cause immunosuppression, and this study does nothing to change that fact.
If we want to have legitimate discussions about vaccine safety, we need rigorous science, not preprints that feed misinformation machines and clickbait media.
Now, more than ever, we all must join in the fight for science.
Thank you for supporting evidence-based science communication. With outbreaks of preventable diseases, refusal of evidence-based medical interventions, propagation of pseudoscience by prominent public “personalities”, it’s needed now more than ever.
More science education, less disinformation.
- Andrea
ImmunoLogic is written by Dr. Andrea Love, PhD - immunologist and microbiologist. She works full-time in life sciences biotech and has had a lifelong passion for closing the science literacy gap and combating pseudoscience and health misinformation as far back as her childhood. This newsletter and her science communication on her social media pages are born from that passion. Follow on Instagram, Threads, Twitter, and Facebook, or support the newsletter by subscribing below:
The detailed dissection of the preprint is worthy of a scientific publication in and of itself. It's just painful that bad (scientific) actors need to be called out.
Why would Akiko Iwasaki do this???