Skip to main content
A Project of The Annenberg Public Policy Center

Atlas, Paul Mislead on Preexisting COVID-19 ‘Immunity’


Coronavirus task force member Dr. Scott Atlas and Sen. Rand Paul have misleadingly suggested that much of the U.S. population has immunity to the coronavirus due to previous exposure to similar viruses. But scientists say any possible protection is theoretical — and can’t be relied upon to control the pandemic.

Paul, who is a Republican from Kentucky, broached the idea of preexisting immunity to the coronavirus during a Senate hearing on Sept. 23. 

In a heated exchange with Dr. Anthony Fauci, the director of the National Institute of Allergy and Infectious Diseases, Paul argued that the reason why New York’s outbreak is now largely under control is because of community, or herd, immunity — and not because the population followed public health guidelines.

Fauci, Sept. 23: Right now, if you look at what’s going on right now, the things that are going on in New York to get their test positivity 1% or less is because they are looking at the guidelines that we have put together from the task force of the four or five things of masks, social distancing, outdoors more than indoors, avoiding crowds and washing hands.

Paul: Or they’ve developed enough community immunity that they’re no longer having the pandemic because they have enough immunity in New York City to actually stop it.

Fauci: I challenge that, senator. … You were not listening to what the director of the CDC said, that in New York, it’s about 22%. If you believe 22% is herd immunity, I believe you’re alone in that.

Paul: There’s also the preexisting immunity of those who have cross-reactivity, which is about a third of the public in many estimates … which would actually get you to about two-thirds.

Fauci: I’d like to talk to you about that also because there was a study that recently came out that preexisting immunity to coronaviruses that are common cold do not cross-react with the COVID-19.

It’s unclear which study Fauci was thinking of — NIAID’s press office did not respond to our inquiry — but Paul’s office told us he was referencing research on immune cells known as T cells from the La Jolla Institute for Immunology.

As we’ll explain, there are studies showing that some people have T cells that can recognize SARS-CoV-2, or the novel coronavirus. But there’s no reason to think those cells would contribute to herd immunity, which occurs when enough of the population is immune to prevent spread of the disease.

Scientists have generally found that people do not harbor cross-reactive antibodies that can actually prevent infection with the coronavirus and that would be meaningful for herd immunity.

Although scientists are still debating what the herd immunity threshold is, most epidemiologists estimate that between 40% to 70% of a population would need to be vaccinated or have prior exposure to the virus for an outbreak to begin to die out.

Later that day, Atlas, the newest member of the coronavirus task force and a senior fellow at Stanford University’s Hoover Institution, also brought up preexisting immunity in a press briefing at the White House.

When asked about Centers for Disease Control and Prevention Director Dr. Robert Redfield’s testimony, in which he said preliminary CDC antibody testing results showed that “more than 90% of the population remains susceptible” to COVID-19, Atlas said Redfield had “misstated” the findings.

Atlas, Sept. 23: The data on the susceptible that he was talking about was his surveillance data that showed that roughly 9% of the country has antibodies. But when you look at the CDC data state by state, much of that data is old. Some of it goes back to March or April, before many of these states had the cases. That’s point number one.

Point number two is that the immunity to the infection is not solely determined by the percent of people who have antibodies. If you look at the research — and there’s been about 24 papers at least on the immunity from T-cells — that’s a different type of immunity than antibodies. And without being boring, the reality is that — according to the papers from Sweden, Singapore and elsewhere — there is cross-immunity, highly likely from other infections, and there is also T-cell immunity. And the combination of those makes the antibodies a small fraction of the people that have immunity.

So the answer is no, it is not 90% of people that are susceptible to the infection.

A neuroradiologist who researches health policy, Atlas wasn’t trained as an immunologist, infectious disease specialist or epidemiologist. He has reportedly backed a “herd immunity” approach to the pandemic, although he denies embracing such a strategy.

Both Redfield and Fauci have expressed concerns about Atlas and the accuracy of the information he imparts to the president and the public, although Atlas maintains that he is following the science.

In a phone call on a commercial airline, Redfield was overheard saying, “Everything he says is false,” according to an NBC News report. A federal official similarly told CNN that the CDC head was worried that Atlas was providing Trump with misleading information.

When asked in a Sept. 28 CNN interview whether Atlas was misleading the president, Fauci replied, “Well, yeah, I’m concerned that sometimes things are said that are really taken either out of context or actually incorrect.”

Cross-Reactive T Cells

Paul and Atlas reference studies that suggest a good chunk of people have T cells in their bodies that “cross-react” with SARS-CoV-2, even though they’ve never been exposed to the virus before.

These studies do indeed exist — but both men inaccurately conclude that people with cross-reactive T cells will be protected from becoming infected with the virus, when that’s almost certainly not the case.

A paper published in May by researchers at the La Jolla Institute for Immunology in the journal Cell, for example, found that about half of the 11 tested blood samples from unexposed donors contained T cells that could recognize SARS-CoV-2.

All of the control volunteers had donated blood between 2015 and 2018, so it wasn’t possible for them to have developed those T cells as a result of an infection with the novel coronavirus. But all of them did have antibodies to at least one of two human common cold coronaviruses, suggesting that these were so-called memory T cells generated as a result of past cold infections.

A follow-up study published in the journal Science by many of the same scientists further supported the idea that the responding T cells were primarily memory cells, present in the body because of past human cold coronavirus infections.

Other studies have found similar results, typically finding between 20% and 50% of the samples from people who were never exposed to the coronavirus nevertheless have T cells that can recognize the pathogen.

But immunologists don’t yet know what it means for a person to have these preexisting T cells. 

Shane Crotty, a professor at the La Jolla Institute for Immunology and an author of the Cell and Science papers, explained in an Aug. 11 Twitter thread that T cells might be helpful to some people but are unlikely to contribute to herd immunity.

“We SPECULATE that it is conceivable that these T cells may potentially reduce COVID-19 disease severity, based on things we know about flu and T cells,” he said, emphasizing that there was no data to support this, as it is only a hypothesis, but it was important to keep investigating to find out.

“[E]ven if our most optimistic speculations about crossreactive T cell memory were found to be correct, it would mean that just as many people would get infected with SARS-CoV-2, but fewer would become severely ill and die from COVID-19,” he added. “T cells generally don’t completely prevent infections, they limit disease (make it shorter and/or less serious).”

Crotty also responded on Twitter to Atlas’ Sept. 23 comments at the White House, noting that “several major aspects” of what he said “are wrong.”

“There is no direct evidence that pre-existing T cell immunity affects COVID-19 infections,” he said, noting that he and other researchers “have proven that such T cells exist, but neither we nor anyone else have shown that the pre-existing T cells make COVID-19 better, worse, or no difference.”

“[W]e only SPECULATE that they may impact COVID-19 disease,” he added. “Even IF such a pre-existing T cell immunity exists, it would almost certainly not affect herd immunity.”

Previously, Crotty’s institution also batted down claims about being close to COVID-19 herd immunity. “While many people have T cells that can recognize #SARSCOV2,” the La Jolla Institute wrote in a tweet, “this does not mean they are immune or can’t pass on the disease.”

Other experts agree.

“There is evidence that people vary in the amount of T cell immunity that they have to SARS-CoV-2 and that this plausibly reflects exposure to other betacoronaviruses over the life course,” said Harvard epidemiologist Bill Hanage in a press call, referring to the sub-group of coronaviruses to which SARS-CoV-2 belongs.

“However, the nature of the immune system is such that it’s by no means obvious what this actually means,” he said. “It could lead to more severe illness, it can lead to milder illness, and so on and so forth.”

Indeed, while the preliminary evidence does not suggest that the cross-reactive T cells are harmful in this case, it is a possibility. Crotty and his La Jolla colleague Alessandro Sette explained in a scientific review that preexisting immunity can sometimes be detrimental, either by making the immune response weaker than it otherwise would have been, or by having the immune system enhance disease.

“Any appeal to existing immunity,” Hanage added, “has to handle the fact that even in places … where we’re still doing relatively large amounts of physical distancing and mask use, infection proceeds at a sort of steady pace, which does indicate that if you’d provide the opportunity for it to transmit more, it will.”

“Any reliance upon existing immunity from other betacoronaviruses, unfortunately, is wishful thinking,” he said.

Crotty and Sette also teamed up with public health scientists from Harvard to think through the ramifications of preexisting T cell immunity for the COVID-19 pandemic. Writing in Nature Reviews Immunology on Oct. 6, the group outlined several scenarios for how cross-reactive T cells might function in individuals and at a population level.

In the most plausible scenarios, the T cells help reduce severity of COVID-19, and in some cases might reduce the spread of the virus, although there is also a chance of increasing spread if the reduced symptoms made the infection harder to identify. But, as the authors write, potential for reduced transmission is already accounted for in models of disease spread since it’s based on real-world data. And again, and because this type of immunity doesn’t prevent infection, it is “not expected to ‘supplement’ herd immunity.”

The researchers did examine a scenario in which herd immunity would be affected, but explain that it is “highly unlikely” and would be immunologically “unprecedented.” In this scenario, the preexisting T cells in the upper respiratory tract would have to quickly block viral replication of the virus, eliminating all the infected cells within a day. This is the closest T cell immunity could get to the “sterilizing immunity” of neutralizing antibodies, which can intervene before any cells are infected.

All the evidence, however, indicates not only that this has never happened before in humans, but also that it doesn’t appear to be happening with COVID-19. This kind of T cell “blitz” has been reported only in animal models with other diseases and with a different type of T cell than what most people have from past coronavirus infections. 

Such a T cell response also has not been observed in past experiments with the cold coronaviruses, the authors write, “making it implausible” that cross-reactive T cells alone could prevent SARS-CoV-2 infection. Finally, the epidemiology also fails to fit with such a scenario, as outbreaks on ships report high levels of viral spread.

“In sum,” the authors write of the plausible scenarios, “the discovery of cross-reactive T cells should not substantially change projections of disease dynamics in populations or, specifically, of the proportion of the population who will become infected before transmission wanes.”

Understanding Seroprevalence

Atlas’ other claim was that Redfield was wrong when he said 90% of the American public remains susceptible to the coronavirus. While there are always caveats to the data, there is no reason to think that Redfield’s estimate, which stems from a CDC seroprevalence study, is wildly off. Such studies estimate how much of the population has developed antibodies using antibody tests.

In fact, another study published in the journal The Lancet two days after Redfield’s testimony also backs a similar figure. That study, performed by researchers at Stanford, found that 9% of people from a nationwide survey of dialysis patients had SARS-CoV-2 antibodies.

Atlas said that one reason the estimate of susceptible people was so high is that the data was old — from March or April. But other CDC data, which are derived from commercial lab surveys, indicate that about 6% of the American population has antibodies to the coronavirus as of early August. And the Stanford study also used more recent data, from July.

As the Stanford authors explain in their article, their estimate based on dialysis patients might overestimate the number of people in the general population exposed to the virus, since many dialysis patients are minorities. Or, it could underestimate the total, perhaps because the dialysis patients are less likely to be working and may be more protected than others. But it’s in line with other seroprevalence surveys, which have concluded that the vast majority of people have not been infected and are unlikely to be immune. 

“In fact, even if the seroprevalence estimates derived from the US dialysis population overestimated true seroprevalence in the overall US adult population, our data nonetheless support that fewer than 10% of the US population has seroconverted as of July, 2020, and herd immunity remains out of reach,” the authors wrote.

Although Atlas argues the percentage of susceptible people should be much lower because of T cell immunity, as we’ve already established, T cell immunity is unlikely to contribute to herd immunity and should not be counted toward the total unless new evidence arises suggesting that it does play a role.

Dr. Julie Parsonnet, an infectious disease specialist and co-author of the Stanford antibody study, told us that the overall consensus is that 9%, with or without T cells contributing, is “very unlikely” to represent herd immunity.

She emphasized that the virus can be brought under control without reaching herd immunity, but that involves using the public health strategies that have been shown to work. “They didn’t need T cells in New Zealand or Taiwan,” she said.

Former CDC chief Dr. Tom Frieden also objected to the idea that far more Americans are already immune. “Only at most 1 in 7 Americans have been infected with Covid. That means 6 in 7 are likely still vulnerable,” he said in a Sept. 30 tweet. “The US is nowhere close to herd immunity and anyone suggesting otherwise doesn’t know enough about epidemiology and public health.”

Editor’s Note: Please consider a donation to FactCheck.org. We do not accept advertising. We rely on grants and individual donations from people like you. Credit card donations may be made through our “Donate” page. If you prefer to give by check, send to: FactCheck.org, Annenberg Public Policy Center, 202 S. 36th St., Philadelphia, PA 19104.