Widespread antibody testing in a Californian county has revealed a much higher prevalence of coronavirus infection than official figures suggested. The findings also indicate that the virus is less deadly than current estimates of global case and death counts suggest. But some scientists have raised concerns about the accuracy of kits used in such studies because most have not been rigorously assessed to confirm they are reliable.
An analysis of the blood of some 3,300 people living in Santa Clara county in early April found that one in every 66 people had been infected with SARS-CoV-2. On the basis of that finding, the researchers estimate that between 48,000 and 82,000 of the county’s roughly 2m inhabitants were infected with the virus at that time — numbers that contrast sharply with the official case count of some 1,000 people reported in early April, according to the analysis. The work has not yet been peer reviewed.
The results are some of the first of more than a dozen ‘sero-prevalence surveys’ being carried out in cities worldwide to try to estimate populations’ true infection rates, in the absence of widespread diagnostic testing. The World Health Organisation is also running a global sero-prevalence study, known as Solidarity II.
Many surveys are using commercial antibody kits to detect antibodies against the virus in blood samples. The presence of SARS-CoV-2-specific antibodies reveals that a person had been infected for at least a week earlier, even if they have had no symptoms.
“A sero-survey gives you a snapshot in time of who is infected in your given population,” says Kanta Subbarao, a virologist at the Peter Doherty Institute for Infection and Immunity in Melbourne. This is especially important for an infection such as SARS-CoV-2, for which some people show no symptoms, or only mild ones, she says.
When combined with information about age, gender, symptoms, co-morbidities and socioeconomic status, these surveys can also help to answer questions about factors such as the role of children in spreading the infection, and the portion of cases that are asymptomatic.
“This is a really inexpensive way to get an incredible amount of information,” says Jayanta Bhattacharya, a health economist at Stanford University in California and a co-author of the study.
News of the Santa Clara analysis follows preliminary results from a similar study in Germany, released on 9 April, that tested some 500 people in a village of more than 12,000 and found that one in seven had been infected with SARS-CoV-2. The German team also looked for active infections, using diagnostic tests based on the polymerase chain reaction, and when those figures were combined with those who had antibodies, estimate that the town’s overall infection rate was 15%.
But this result might not be indicative of what’s happening across Germany, says virologist Christian Drosten, who heads the Institute of Virology at the Charité University Hospital in Berlin, because many people in the town celebrated at a carnival in February. “There was a big point source outbreak in that village,” he says.
The fact that both studies detected much higher rates of infection than official figures suggest is not surprising, says Peter Collignon, a physician and microbiologist at the Australian National University in Canberra. The virus had been spreading in the US and parts of Europe for at least a month before it was detected as spreading in the community.
But Collignon notes that the commercial antibody tests used in both studies were evaluated on using only a small number of people, which could also affect the accuracy of the survey results. Antibody kits aren’t just being used for population studies. Kits are also being marketed for testing whether individuals have had the disease. But experts warn that most tests haven’t been rigorously evaluated to ensure they are reliable.
Sero-surveys can also provide a better estimate of how deadly a virus is, using a measure known as the infection fatality rate (IFR) – the proportion of all infections, not just those confirmed through clinical testing, that result in death. An accurate IFR can improve models being used to decide public-health responses. If a disease turns out to be less deadly than previously estimated, this could reframe discussions around the measures being introduced to contain it, and their economic and social impact, says Neeraj Sood, a health economist at the University of Southern California in Los Angeles, who is leading a separate antibody study in Los Angeles and is also a co-author in the Santa Clara study.
The Santa Clara team estimated an IFR for the county of 0.1–0.2%, which would equate to about 100 deaths in 48,000-82,000 infections. As of 10 April, the county's official death count was 50 people. The study's IFR is lower than the IFR used in models by researchers at Imperial College London, which estimated an IFR for the UK on the basis of data from China to be 0.9%. In another study, the same group estimated an IFR for China of 0.66%, and a study of deaths on the Diamond Princess cruise ship estimated an IFR of 0.5%.
Figures vary in different places for several reasons, including the age distribution of the population and the extent of testing.
Fatality rate estimates have been revised down over time as more people have been tested and researchers have gained more insight into less-severe cases, as happened with swine flu in 2009, says Eran Bendavid, a population-health researcher at Stanford University who led the Santa Clara study. But scientists have concerns about the reliability of antibody tests, particularly in regards to the number of false positives they produce, which could inflate infection rate estimates.
The Santa Clara study reports using a kit purchased from Premier Biotech, based in Minneapolis, Minnesota. According to the pre-print, the manufacturer's kit performance data noted 2 false positives out of 371 true negative samples. But with that ratio of false positives to true cases, a large number of the positive cases reported in the study — 50 out of 3320 tests — could be false positives, says Marm Kilpatrick, an infectious disease researcher at the University of California Santa Cruz.
To ensure a test is sensitive enough to pick up only true SARS-CoV-2 infections, it needs to be evaluated on hundreds of positive cases of COVID-19 and thousands of negative ones, says Michael Busch, an infectious-diseases researcher and director of the Vitalant Research Institute in San Francisco, California, who is also leading a sero-prevalence survey. But most kits have not been thoroughly tested, and health agencies are particularly concerned about the accuracy of some rapid tests, says Busch.
The researchers involved in the Santa Clara study say that they assessed the sensitivity and specificity of the antibody tests in an initial 37 positive samples and 30 negative controls. The tests identified 68% of the positive samples and 100% of the negatives. An unpublished follow-up assessment in 30 positive and 88 negative controls found that the test correctly identified 28 positives and all 88 negatives, says Bendavid.
Bendavid says they adjusted for the test kit’s performance and differences in the survey population relative to the county to estimate the prevalence of SARS-CoV-2 in Santa Clara. Survey participants included a higher proportion of white, female and affluent individuals than is found in the county’s population.
Kilpatrick says another potential source of bias in the study is that participants were recruited using social media. As a result, the sample could include a disproportionally higher number of people who thought they were exposed to the virus and volunteered to get tested, he says. “The real prevalence might be half as high, a tenth as high, or it might even be the number presented in the paper – we don’t know because recruiting participants over Facebook presents an unknown bias,” he says.
Bhattacharya says the results probably undercount the prevalence in the wider population, because they miss anyone who has been infected too recently to have mounted an immune response, and exclude people in prisons, nursing homes and other institutions.
Results are expected soon from sero-prevalence surveys run by other groups around the world, including teams in China, Australia, Iceland, Italy, Germany and several others in the US.
The study confirms the widely-held belief that far more people than originally thought have been infected with the coronavirus, The Guardian quotes Arthur Reingold, an epidemiology professor at University of California – Berkeley who was not involved in the study, as saying. But, he says it doesn’t mean the shelter-in-place order will be lifted any time soon.
“The idea this would be a passport to going safely back to work and getting us up and running has two constraints: we do not know if antibodies protect you and for how long, and a very small percentage of the population even has antibodies,” he said.
Even with the adjusted rate of infection as found by the study, only 3% of the population has coronavirus – that means 97% does not. To reach herd immunity a significant portion of the population would have to be infected and recovered from coronavirus.
Abstract
Background Addressing COVID-19 is a pressing health and social concern. To date, many epidemic projections and policies addressing COVID-19 have been designed without seroprevalence data to inform epidemic parameters. We measured the seroprevalence of antibodies to SARS-CoV-2 in Santa Clara County. Methods On 4/3-4/4, 2020, we tested county residents for antibodies to SARS-CoV-2 using a lateral flow immunoassay. Participants were recruited using Facebook ads targeting a representative sample of the county by demographic and geographic characteristics. We report the prevalence of antibodies to SARS-CoV-2 in a sample of 3,330 people, adjusting for zip code, sex, and race/ethnicity. We also adjust for test performance characteristics using 3 different estimates: (i) the test manufacturer's data, (ii) a sample of 37 positive and 30 negative controls tested at Stanford, and (iii) a combination of both. Results The unadjusted prevalence of antibodies to SARS-CoV-2 in Santa Clara County was 1.5% (exact binomial 95CI 1.11-1.97%), and the population-weighted prevalence was 2.81% (95CI 2.24-3.37%). Under the three scenarios for test performance characteristics, the population prevalence of COVID-19 in Santa Clara ranged from 2.49% (95CI 1.80-3.17%) to 4.16% (2.58-5.70%). These prevalence estimates represent a range between 48,000 and 81,000 people infected in Santa Clara County by early April, 50-85-fold more than the number of confirmed cases. Conclusions The population prevalence of SARS-CoV-2 antibodies in Santa Clara County implies that the infection is much more widespread than indicated by the number of confirmed cases. Population prevalence estimates can now be used to calibrate epidemic and mortality projections.
Authors
Eran Bendavid, Bianca Mulaney, Neeraj Sood, Soleil Shah, Emilia Ling, Rebecca Bromley-Dulfano, Cara Lai, Zoe Weissberg, Rodrigo Saavedra, James Tedrow, Dona Tversky, Andrew Bogan, Thomas Kupiec, Daniel Eichner, Ribhav Gupta, John Ioannidis, Jay Bhattachaya
[link url="https://www.nature.com/articles/d41586-020-01095-0"]Nature material[/link]
[link url="https://www.theguardian.com/world/2020/apr/17/antibody-study-suggests-coronavirus-is-far-more-widespread-than-previously-thought"]The Guardian report[/link]
[link url="https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1"]medRxiv abstract (non-peer reviewed)[/link]