Large-scale studies of the general population are needed in order to determine how wide the disease has spread, how deadly the virus really is, and whether enough of the population has become immune for social distancing restrictions to ease. Several survey results from Germany, the Netherlands, the US, and Switzerland find that anywhere from 2% to 30% of certain populations have already been infected with COVID-19. Yet these results have been heavily criticized as too preliminary, with larger-scale studies required to draw any sort of conclusions.
COVID-19 serology tests may report a high number of people falsely believed to have immunity to COVID-19 when, in fact, they do not; this is known as a false positive result. This high number of false positives has to do with something called Positive Predictive Value (PPV), which is a function of the number of people who have the disease, also known as prevalence. The way PPV works is that the more prevalent a disease is, the lower the number of false positives will be. However, if there is a relatively low number of people who have been infected, more false positive values will be detected. Even with the current global pandemic, the prevalence of COVID-19 is likely to be currently low, meaning that antibody tests run the risk of reporting a higher number of false positives.
In the UK, antibody tests must meet a minimum standard of 98% specificity and 98% sensitivity. Even with such stringent requirements, due to PPV, when screening the entire UK population a COVID-19 antibody test may result in 1–1.2 million people being wrongly identified as immune to the coronavirus. These people would then be in danger of becoming infected, infecting others, and causing a second wave in COVID-19 infections and outbreak. Due to the low prevalence of COVID-19 translating to a high number of false positives, immunity passports should only be implemented once a majority of the population is infected. Furthermore, such passports should only be issued on the strength of additional tests or medical history, such as prior positive test for SARS-CoV-2, a positive antibody result, and/or a negative PCR test, if no prior test. This would mean though that antibody screening for COVID-19 immunity would still be limited by the PCR testing capacity of a country.
Two studies of note have emerged from the preliminary antibody surveys. In the COVID-19 case cluster study (Gangelt Municipality) lead by the Institut für Virologie, on April 9, a preliminary report was published, based on testing approximately 1,000 individuals from 400 households. This study covered Heinsberg, one of the regions hardest hit by COVID-19, with 12,500 reported cases of the coronavirus infection. They reported that antibodies to the virus were found in 14% of 500 people tested. While the report indicated that the immunochemistry test used has a specificity of in excess of 99%, no data were provided on sensitivity. This, together with the small sample size and the fact that the study sampled entire households, which can lead to overestimating infections as people living together often infect one another, means that relatively few conclusions can be drawn from this report. Furthermore, while the commercial antibody test used in the Heinsberg study is reported to have more than 99% specificity, a study from Statens Serum Institut in Denmark that evaluated nine commercial SARS-CoV-2 immunoasays found the same test produced three false positives in a sample of 82 controls, indicating specificity of only 96%. This indicates that out of the 1,000 people tested in the Heinsberg survey, the test could have produced more than a dozen false positives out of the roughly 70 people the survey found to test positive for COVID-19 antibodies.
The second notable study comes out of Stanford where 3,300 people, recruited through ads on Facebook, reported fifty people tested positive for COVID-19 antibodies—approximately 1.5% of the survey participants. After adjusting the data to better reflect the county’s demographics, the researchers concluded that between 2.49% and 4.16% of Santa Clara county residents had likely been infected, suggesting that 80,000 residents had had COVID-19. This is more than fifty times the number of reported cases, and implies a low fatality rate on par with seasonal influenza. One conclusion that could be drawn from this study is that the outbreak is not medically serious, and that control measures could be relaxed. However, this study used a commercially available lateral flow rapid test rather than a fully validated laboratory test, where the sensitivity was found to be much lower than the manufacturer’s claims. Moreover, even the manufacturer’s stated sensitivity was significantly lower than the minimum performance specification set by the UK government during recent serology rapid test trials. In addition, the study carried out testing of whole households, and may have inadvertently encouraged people who believed they had previously had the virus to join the study, in the mistaken belief this would lead to a so-called immunity passport. This approach would have led to bias in the study, and an inflated prevalence value. Given these significant concerns, the study cannot not be used to build a science-led case to relax lockdown restrictions in California, let alone the US.
Based on the present evidence, the World Health Organization (WHO) recommends the use of antibody tests only in research settings. The WHO states that antibody tests should not be used for clinical decision-making until evidence supporting use for specific indications is available. Looking forward, large-scale studies using antibody tests involving tens of thousands of individuals are still needed to draw definitive conclusions and are most useful at the level of research and epidemiology.