The proportion of people in England with coronavirus antibodies dropped by more than a quarter in the space of three months, researchers have revealed, fuelling concerns over reinfection. The findings come from the React-2 study, which is based on home finger-prick antibody test results from random participants across all 314 local authorities. The first results, based on data from 100,000 people, were released in August, revealing that about 6% of the population of England had the antibodies – protective proteins produced in response to an infection – although the team say that could be a slight underestimate. The new work – not yet peer-reviewed – extends this with more testing in two fresh cohorts, each yielding results from more than 100,000 adults. The results reveal that just 4.4% of those tested in the most recent round, between 15 and 28 September, had detectable coronavirus antibodies. “As a whole, these data suggest the possibility that decreasing population immunity will lead to an increased risk of reinfection as detectable antibodies decline in the population,” said Graham Cooke, co-author of the report and professor of infectious diseases at Imperial College London. The researchers said antibodies would be expected to decline over time, but just how long that takes, and how far levels fall, differ between viruses – factors that are important when it comes to how long protection lasts. Prof Wendy Barclay, chair of influenza virology at Imperial College London and co-author of the report, said it was not yet known what level of antibody is needed to protect someone from infection or reinfection with Covid-19, but experiments suggested it could be around the same level as the antibody test’s threshold of detection. Some have raised the possibility that other components of the immune system might continue to offer protection even if antibodies have waned – such as T-cells, which can kill infected cells, or B memory cells, which can rapidly produce new antibodies. However, the team said it is too soon to know if that is the case, or for how long that protection might last, while it is difficult to measure levels of such T-cells. “The fact that people get reinfected regularly throughout their lives with seasonal coronaviruses [that cause common colds] suggests that the immunity, whether or not it is antibody mediated and/or T-cell mediated, probably isn’t very long-lasting,” said Barclay, who added that the team suspect the body reacts to infection with the new coronavirus in a similar way. But she said the new results do not necessarily mean that immunity arising from vaccination would be short-lived. “A good vaccine may well be better than natural immunity.” Barclay added that if a second infection did occur, immunological memory might mean it would be less severe than the first. But, she noted, data is currently lacking: only a few cases of reinfection have, so far, been confirmed around the world. The team said antibodies were still good indicator of protection against reinfections, adding that the low prevalence of antibodies against the coronavirus and the rapid pace at which they had been shown to fall undermine the case for a herd immunity policy without a vaccine. “When you think that 95 out of 100 people are unlikely to be immune and therefore likely to be susceptible, then we are a long, long way from anything resembling a population-level protection against onward transmission,” said co-author Helen Ward, professor of public health at Imperial. The team said the lack of reinfections could be because it is difficult to confirm such cases – but it may also be that not enough time has passed for antibody levels to have declined to the point where most people are susceptible to reinfection. Cooke added that it remains important that the public adheres to Covid restrictions, while a vaccine is also needed. “The big picture here is that after the first wave, still the great majority of the country didn’t have evidence of protective immunity,” he said.