The fourth round of antibody tests in New Delhi, the results of which were made public on Wednesday during a court hearing, presents some interesting questions.
It found that 25.5% of the roughly 15,000 people tested possessed Covid-19 antibodies. The tests were carried out in the third week of October. The results are almost identical to those of the third round, which covered around 17,000 people, conducted in the first week of September, which showed that the antibodies were present in 25.1% of the people surveyed. The second round (about 15,000 were analyzed in August) showed that 29.1% of the people surveyed had antibodies. And the first round, in late June and early July, found them in 22.6% of the 21,000 people surveyed.
This is not how serological survey results should progress. Serum surveys, such as the ones being conducted in Delhi, are a measure of the prevalence of an infection (in this case, Covid-19) and therefore of immunity. Ideally, your results should show increasing prevalence, or immunity, until the level of herd immunity is reached. This is the level at which much of the population is safe because the virus cannot find enough people to infect. But there are complications in making such assumptions: On the one hand, Covid-19 antibodies don’t seem to last; on the other, the presence of antibodies is not always necessary for immunity, just as their absence is not a sign of lack of protection. In fact, the October round of the sero survey in Delhi found that 43.5% of people previously diagnosed with Covid-19 no longer possessed antibodies.
This is not entirely unexpected.
In late October, the results of an Imperial College London-led study of 365,000 people in England, conducted in three rounds between June and September, showed that the number of people with Covid-19 antibodies decreased over time. The UK study actually saw a decline in the prevalence of infection in all parts of the country, with a clear drop between the first and third rounds. The study was conducted before the second wave of infections in the UK. This may also be what the Delhi study is discovering.
However, there is some debate about the behavior of antibodies. According to a paper published in Science Immunology in early October by US researchers, including some from Massachusetts General Hospital, and based on a study of 343 patients with Covid-19, antibodies of the immunoglobulin G (IgG) variety , which usually provide long-lasting immunity, were found in patients for up to four months (the study ended then). Others, of the IgA and IgM variety, were detected 12 days after infection but did not last more than two months.
A second article, also published in the same journal around the same time, but based on a study in Canada by researchers at the University of Toronto, came to a similar conclusion: antibody levels of the IgG variety peaked between two weeks to a month after infection and then was stable for at least three months. This study covered 439 people (not all had been tested, but they all had Covid-19 symptoms).
Interestingly, as Dispatch 112 pointed out on July 23 (citing a late-June article on a prepress server), research at the Karolinska Institute and Karolinska University Hospital in Sweden showed that many infected people who were asymptomatic or only showing Mild symptoms did not test positive for antibodies although they had what the researchers called a strong T-cell response. The research was later published in Cell. T cells are the immune system’s fighting specialists (better yet, they remember viruses), recognizing and fighting infected cells. The research (of around 200 people) showed that twice as many people with Covid-19 antibodies had T-cell immunity. This could well mean that asymptomatic people or those who contract a mild infection do not generate these antibodies or see their numbers decrease little. after the infection passes, although (according to the Swedish study) they are still immune. Is that what the Delhi result shows? And what does that mean in terms of prevalence and immunity from infection?
.
