The Washington PostDemocracy Dies in Darkness

Fewer than 1 in 10 Americans show signs of past coronavirus infection, large national study finds

This suggests millions may still be vulnerable to infection, authors say

People wait in line for antibody tests outside CareWell Urgent Care in Cambridge, Mass., in August. (John Tlumacki/Boston Globe/Getty Images)
Placeholder while article actions load

Fewer than 1 in 10 Americans showed signs of past infection with the novel coronavirus as of late July, suggesting that most of the country may still be vulnerable to infection, according to one of the largest studies of its kind published Friday in the journal the Lancet.

That proportion is an estimate based on the percentage of dialysis patients whose immune systems produced coronavirus antibodies. It does not indicate exactly how many Americans may be immune to the virus, because not every infected individual develops antibodies. It is also unclear how strong a defense antibodies might confer or for how long. But, combined with similar results from studies by the Centers for Disease Control and Prevention and other institutions, it’s evident a large majority may not be protected against a disease that has already killed 200,000 Americans.

“We are still in the middle of the fight,” said Eli Rosenberg, a State University of New York at Albany epidemiologist who was not part of the study. “We’re all tired, and we’re all hoping for a vaccine. This shows us how it’s not over here, not even by a long shot.”

Researchers at Stanford University and Ascend Clinical laboratory, a company that processes lab tests for kidney dialysis patients, examined leftover blood plasma samples from a randomly selected group of 28,500 patients. Each person underwent dialysis at one of 1,300 centers in 46 states in July. Testing the plasma revealed about 8 percent had coronavirus antibodies, the molecules the immune system churns out to help fight an infection.

That’s equivalent to about 9 percent of all U.S. adults, according to Stanford University biostatistician and study co-author Maria Montez-Rath, who used patient data on region, age and sex to translate that result to the general populace. Ongoing CDC studies of leftover medical samples — not only from dialysis patients — in 10 regions found roughly similar percentages of people who had antibodies in the summer months.

This means for every case diagnosed by a nasal or saliva swab in the country, about nine more people have antibodies for the coronavirus, the study estimated. Many people remain potentially at risk. “It basically does show that a minority of the U.S. population has evidence of antibodies to SARS-CoV-2 right now,” said study co-author Shuchi Anand, a nephrologist at Stanford University.

The findings also suggest herd immunity is far off without the advent of a vaccine. Epidemiological models suggest about 70 percent of the population would need to be protected from the coronavirus to make its spread unlikely, said study co-author Julie Parsonnet, an infectious-disease epidemiologist at Stanford University, acknowledging antibody studies are “not a perfect way” to track this.

The United States began Phase 3, one of the last steps in developing a vaccine, on July 27, 2020. We take you inside the process. (Video: The Washington Post)

That’s because scientists still don’t understand coronavirus immunity well enough. “We don’t really have perfect data that antibodies give immunity,” she said. “And we don’t have perfect data that there aren’t other forms of immunity that are also important,″ which antibody tests cannot detect.

“If these numbers were taken at face value … they still would suggest that there are a lot of people” who haven’t been infected, Rosenberg said.

The study’s findings align with CDC Director Robert Redfield’s assertion to Congress this week that 90 percent of U.S. population is probably vulnerable to the virus — a statement later challenged by Scott Atlas, a pandemic adviser to President Trump. “It is not 90 percent of people that are susceptible to the infection,” Atlas said Wednesday, arguing that more people have immunity to the novel coronavirus due to protection from what’s known as “memory” T cells, the part of our immune system trained to recognize specific invaders. Some scientists theorize a number of people may have such T-cell immunity because of previous contact with other coronaviruses, such as the common cold; however, that theory has not yet been proved.

How many people are infected with the coronavirus? A major study will attempt to provide an answer.

This study’s population was more geographically diverse than those of other antibody studies in the United States and was “rigorously done,” Rosenberg said. But he said it was difficult to conclude whether dialysis patients truly represent all adults in the United States.

The study found minority groups disproportionately had coronavirus antibodies, at 16 percent for those who live in majority Black and Hispanic communities, 11 percent in majority Hispanic communities and 4 percent in majority White communities. Clinical data have suggested infections in Black and Hispanic populations are more severe, and this study indicates members of those populations are more likely to be exposed as well.

July’s snapshot also showed variation by Zip code, with “a 10-fold higher risk if you live in a very highly dense area compared to more rural settings,” Anand said.

In an accompanying editorial in the Lancet, infectious-disease doctor Barnaby Flower and public health expert Christina Atchison, both Imperial College London researchers who were not involved with the report, described the limits of studying only dialysis patients.

“Attending a health-care facility three times a week would seem like a good way to encounter SARS-CoV-2, as has been shown elsewhere,” they wrote, which would mean the report may be an overcount. But it could also be an undercount, because people who require dialysis are more likely to die of the coronavirus and, therefore, would be left out of the study.

A vaccine, or a spike in deaths: How America can build herd immunity to the coronavirus

Rosenberg, who earlier this year studied the prevalence of antibodies in 15,000 people who agreed to have their blood collected at grocery stores across the state of New York, suspects the percentages in the Lancet report may be too high. Based on dialysis patients in New York, the study found 33 percent of people had antibodies. Rosenberg said that percentage is probably an overshoot, given the results of his work. Likewise, around 18 percent of the population in the New York City metro area had antibodies in mid-July, according to a CDC estimate.

What’s valuable, Rosenberg said, is that the results are likely to be internally consistent, allowing researchers to compare regions or demographics within the dialysis patient population. “The relative picture is probably more useful,” he said.

Because it includes people from almost every state, the new report does a good job of displaying the breadth of this outbreak, Rosenberg said: The pandemic is too big a problem for any single region to handle on its own.

“This is almost a rallying cry, to me,” he said. “We need national leadership to answer this call.”