A preprint study published Tuesday by researchers in England suggested that the prevalence of Covid-19 antibodies in the British population was waning, prompting scary headlines about short-lived immunity to the coronavirus. Antibodies are molecules produced by the immune system that play a key part in preventing future infections, but their role in Covid-19 isn’t totally understood, and it’s not known how long they last in the body.
The British study, called REACT (that’s short for Real-time Assessment of Community Transmission), was led by researchers at Imperial College London. It involved sending out finger-prick tests to detect antibodies to randomly selected people across the country over three rounds of testing between June and September. Like most other countries, England had its first big wave of Covid-19 cases in March and April, so the study aimed to find out whether people who got infected during that time were able to develop and sustain their levels of antibodies in the following months.
Tests on over 365,000 people across England showed that the proportion of participants who tested positive dropped from 6% to 4.4% between June and September — a drop of 26.5%. The decline in positivity rates varied across demographics, though: It was largest in people aged 75 and older and also larger in people who had a suspected, rather than confirmed, case of Covid-19.
“These data,” the authors conclude, “suggest the possibility of decreasing population immunity and increasing risk of reinfection as detectable antibodies decline in the population.”
It’s scary to think that immunity to Covid-19 may not last. A vaccine will not likely be widely available until the middle of next year, and a handful of confirmed reinfections have been documented. The REACT study prompted headlines referencing a study showing “waning antibody immunity to Covid over time” (Reuters) and “evidence of waning immunity to Covid-19” (CNN); a Twitter moment said Covid-19 antibodies “‘fall rapidly’ after infection.” But as I’ve learned over many months of covering Covid-19, alarming news headlines tend to come with many caveats. The experts I spoke to seemed to agree.
“Bottom line is that I don’t think it means what the headlines say,” Deepta Bhattacharya, PhD, an associate professor of immunobiology at the University of Arizona cancer biology program, said via email. Bhattacharya, who recently co-led a study showing Covid-19 antibodies can provide lasting immunity, explained two key points to consider: Antibodies are produced in two waves, so a dip in concentrations is expected. And the finger-prick test used in the study can only measure whether a person has antibodies past a certain threshold.
The latter is important to know for background: The finger-prick test used in the study gives a simple yes/no answer: Yes, your antibody levels are past the threshold measured by the test, or no, they’re not. These tests, explains Bhattacharya, are designed to minimize false positives (telling someone they have antibodies when they don’t) but also tend to call more false negatives (telling someone they don’t have antibodies when they actually do). “It is the lesser of two evils,” he says. “For the purposes of returning results to people and considering what they may do with the information, it is the right way to run these tests.” Essentially, outside of the purposes of this study, these tests are used by everyday people to determine whether they have immunity to Covid-19; given that there’s still a pandemic going on, it’s better for a person to mistakenly think they don’t have antibodies than mistakenly thinking that they do and relaxing their safety measures.
“For these same reasons,” Bhattacharya continues, “it is not a good approach to quantify how long antibody production lasts.” Antibody production, he explains, happens in two phases: First, there’s a sharp rise in antibody concentrations, then a sharp decline, followed by a much more stable nadir. This happens even in “super-durable” antibody responses like the one against measles, he says. The concern is that the REACT test, which looked at antibody responses four to five months after Covid-19 infection, may have measured people during the sharp decline.
Smita Iyer, PhD, an assistant professor of immunology at the University of California, Davis, explained this point further.
“The B cell response following viral infection essentially happens in two phases,” she told the Medium Coronavirus Blog via email, referring to the cells that lead the antibody response. “The first phase is swift but short-lived.” In this phase, which usually lasts two weeks but can vary, cells called plasmablasts primarily drive antibody production. These divide rapidly and are the source of the initial antibody wave, but they die off as the virus gets cleared, causing antibody levels to decline in the first few weeks after infection. “We can see anywhere from two- to ten-fold declines in [antibodies] during this phase,” Iyer explains.
Then comes the second wave, in which antibodies are generated by cells called plasma cells, which live for several decades and tend to make higher-affinity antibodies, says Iyer. (High affinity means that they’re better at sticking to the antigen — the virus, in this case.)
“In the REACT study,” explains Iyer, “the sampling period was over three months with three surveys performed cross-sectionally between late June to September. I anticipate that the majority of samples likely fell within the first phase of the B cell response.”
But even if the people who participated in the REACT study were already in the second phase, it’s possible the tests weren’t able to detect their second wave of antibodies. The second nadir, explains Bhattacharya, may be closer to the threshold for the finger-prick tests, so “one would expect a higher rate of false negatives.” He sent along this figure that he’d previously posted on Twitter, which shows that the second wave of antibodies produces a much lower concentration than the first and that the tests may not be able to detect them.
These points offer a helpful pinch of salt to take with the results of the REACT study. In addition, notes Iyer, several factors affect the generation of plasmablasts and plasma cells themselves, including viral load and the inflammatory response associated with infection.
In their preprint (this means the paper has not yet been peer-reviewed), the authors of the REACT study acknowledge that antibody levels change over time. “During any antibody response to an acute pathogen, some level of antibody waning in the months following infection is expected as short-lived plasma cells die,” they write. “Low levels of affinity-matured antibody usually continues to be produced by long-lived plasma cells, and may be sufficient to maintain levels of antibody that confer immunity.” But while antibodies to measles, influenza, and rhinovirus can be detected for years after infection, they note, “the situation for coronaviruses is less clear.”
The REACT study suggests that the antibody levels of people in England dropped past a certain threshold over the four or five months after infection, but it’s too early to draw conclusions about waning immunity from those results alone. Before doing that, says Iyer, it’s important to find out how antibody concentrations change six months to a year after infection. She also says research must address whether the presence of antibodies actually protects against reinfection and whether it matters whether those antibodies are in the blood serum or in the respiratory tract.
Other researchers are studying Covid-19 immunity in ways that address some of the concerns raised by the REACT study. The study Bhattacharya co-led, which was published in the journal Immunity earlier this month, followed individuals over time and quantified their antibody concentrations. “I think this is a better way to determine how long antibodies are produced,” he says of studies that address these issues. “Those studies point to the same conclusion — antibody production at least for now seems pretty stable.”