select search filters
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

expert reaction to a preprint from the REACT-2 study looking at prevalence of antibody positivity to SARS-CoV-2

A preprint, an un-published non-peer reviewed paper from the REACT-2 programme, looked at prevalence of antibody positivity to SARS-CoV-2 in 365,000 people between 20 June and 28 September.

This Roundup accompanied an SMC Briefing.


Prof Rowland Kao, Professor of Veterinary Epidemiology and Data Science, University of Edinburgh, said:

“One of the most important considerations in our future pandemic planning is the role of immunity – are individuals who have previously been infected protected against re-infection, ultimately leading to ‘herd immunity’? This study provides evidence that the level of immune response has declined over a relatively short period (three months) indicates that such future planning cannot take for granted the beneficial effects of previous infection – importantly, should the results of this study prove robust, this implies that any strategy that relies on ‘herd immunity’ lacks credibility. While the study Is cross-sectional and does not follow individuals between time points, the size of the study and consistency of the results with other coronavirus infections are strong indicators that the measured decline is representative of the rate of decline for the population as a whole, However, while the decline is substantial, nevertheless substantial proportions of the population do retain some immune response, over four months after the peak of the epidemic. This suggests that, even with the measured decline in antibody response, the previous exposure of substantial numbers of the population should have some mitigating effects on the severity of pandemic spread over this coming winter.”


Prof Eleanor Riley, Professor of Immunology and Infectious Disease, University of Edinburgh, said:

“This study suggests that antibodies that can be detected by a simple DIY test may not persist beyond about 3 months after infection. If true, then these tests are unlikely to be useful for estimating cumulative population exposure to the virus or for individuals to use to assess their own previous exposure.

“However, it would be premature to assume that this means that immunity to SARS-CoV-2 does not last:  the study does not look at antibody concentrations, antibody function or other aspects of immunity such as T cell immunity and does not look at the trajectory of antibody levels in the same individuals over time. Nevertheless, the data do lend weight to the concern that antibodies induced by natural infection may be short-lived (as is the case for other seasonal coronaviruses).

“These data should also not be taken to infer that a vaccine would only induce short term immunity. Vaccines contain immune stimulators (adjuvants) that induce durable immune responses and the administration of multiple doses of vaccine ensures that high concentrations of antibodies (that decline only slowly over time) are achieved in the majority of vaccine recipients.”


Prof Sheila Bird, Formerly Programme Leader, MRC Biostatistics Unit, University of Cambridge, said:

“REACT-2 is an impressive England-wide study of SARS-CoV-2 antibodies (anti-spike) for randomly sampled GP-registered adults from the NHS patient list. REACT-2 has enjoyed a consistent response-rate (~ 38%) to its self-completion questionnaire and valid self-administered antibody-test rates (~30%) across three non-overlapping rounds from June to September 2020.

“Congratulations to REACT-2’s design, laboratory and analytic team and to its participants (a neighbour included!)

“REACT-2 confirms significant decline in antibody levels across 12, 18 and 24 weeks after peak infections in England but goes much further – evidences differential rates of decline at extremes of age (as immunologists anticipated) but signals no significant decline for healthcare workers (who may have been repeatedly challenged) and demonstrates a different exposure pattern for healthcare workers (earlier) versus care home workers (later). Regional differences in antibody levels (London high), also by ethnicity (non-white high), were strongly apparent.”


Dr Alexander Edwards, Associate Professor in Biomedical Technology, Reading School of Pharmacy, University of Reading, said:

“The reduction in antibody positivity appearing from the latest detailed and wide-ranging results from REACT-2 study should be taken in context of the type of rapid antibody test used. Whilst the accuracy of these rapid home tests has now been well studied, and these are extremely useful tests, what we don’t know is how closely the positive test relates to protection from reinfection, so it’s hard to understand the broader implication of this headline drop in positive % over time- termed “waning”. In general, these rapid tests have lower analytical sensitivity than lab tests and are therefore able to detect only high levels of antibody. In contrast, laboratory immunoanalysers or microplate ELISA assays (e.g. Roche and Abbott products) can detect really low levels of antibody, and provide some measure of antibody levels. In previous lab antibody test studies testing patients over time, some have found evidence of the level of antibody reducing slowly over months past infection. If you use a single cutoff value at a higher signal in these tests, this would potentially lead to the same conclusion as seen in REACT-2, i.e. fewer patients detected. When people are ill, antibody levels rise, and when you heal, antibody levels do drop naturally – this is not exactly the same as losing immunity.

“What is not clear is how quickly antibody levels would rise again if a person encounters the SARS-CoV2 virus a second time. It is possible they will still rapidly respond, and either have a milder illness, or remain protected through immune memory. So even if the rapid antibody test is no longer positive, the person may still be protected from re-infection. But we don’t know this yet- it takes time to work this out, by following large groups over many months, and this type of study is ongoing yet hard and slow.

“The large scale of this study was made possible by the use of rapid home tests, demonstrating the benefits of rapid antibody tests for population serosurveillance. A great deal of valuable information has been gathered which will take time to interpret fully. Hopefully detailed information about transmission and control measures will emerge. As we have often seen, no single test is perfect but when used together with proper methodology, we do now have a wide range of tests that we understand better, and we are using these to build more knowledge to understand the SARS-CoV2 virus, to help us combat COVID-19.”


Prof Jonathan Ball, Professor of Molecular Virology, University of Nottingham, said:

“This study confirms suspicions that antibody responses – especially in vulnerable elderly populations – decrease over time. What is less clear from this kind of study is the relationship between waning immunity and susceptibility to reinfection and the resulting severity of any subsequent infection, and this is important to know.

“Antibodies are likely to be important in protecting us from future infection and disease, but other arms of the immune system, for example cellular immunity, might also be key. Therefore, it is essential that we gain a better understanding of what protective immunity looks like and this can only be gleaned by measuring all aspects of immunity following infection and seeing how this relates to reinfection risk.”


General comment about antibody and T-cell immunity:

Prof Danny Altmann, Professor of Immunology at Imperial College London and British Society for Immunology spokesperson, said:

“In the past few months researchers around the world have gained considerably more understanding of antibody responses in COVID-19 infection. Most assays look at antibodies that bind to the Spike antigen of the virus, and in many people these antibodies wane over a period of months, though, the more severe and symptomatic the initial infection, the greater the antibody response tends to be.

“Given the fairly transient value of antibody measurements, it had been hoped that measuring the other arm of the immune response, that shown by T cells, would give a clearer picture of who has immunity. However, there remain two confounding problems in this regard. The first is that it has not yet been proved that T cell immunity (in the absence of detectable antibody) protects. The second is that measurement of T cell immunity can be muddied by the fact that many blood samples that were archived before COVID-19 emerged show a cross-reactive response against the common cold coronaviruses. This can make it hard to know if the test is picking up immunity to COVID-19 or to the common cold. As yet, nobody has been able to show if this cross-reactivity could be helpful for protection.”




“Declining prevalence of antibody positivity to SARS-CoV-2: a community study of 365,000 adults” by Ward et al was posted as a preprint at 00:01 on Tuesday 27th October.

This work is not peer-reviewed.


All our previous output on this subject can be seen at this weblink:


Declared interests

Prof Eleanor Riley: “No COI to declare”

Prof Jonathan Ball: “No CoIs”

None others received.

in this section

filter RoundUps by year

search by tag