The Office for National Statistics (ONS) have released the latest report from the COVID-19 infection survey, looking at antibody and vaccination data for the UK.
Prof Riley made a previous comment on an earlier release of the ONS COVID-19 Infection Survey looking at antibody data for the UK, this is her updated comment:
Prof Eleanor Riley, Professor of Immunology and Infectious Disease at the University of Edinburgh, said:
“I am struggling to see the value of these data, at least as currently presented. We are able to monitor the progress of the UK epidemic through the daily numbers of positive tests and hospitalisations, the ONS infection (swab) data and through remote data collection (such as through the Zoe C-19 app). The additional value of measuring antibodies would be to identify anyone who has ever been infected, the assumption being that people develop antibodies after infection (whether or not that infection was formally diagnosed at the time) and remain antibody positive for a significant period of time after they clear the virus. One would expect, therefore, that antibody positivity rates in the population would increase steadily over time as the proportion of the population that has ever been infected increases. Now that a large percentage of the adult population have been vaccinated, we would expect positivity rates to be increasing rapidly.
“However, what we saw in these data prior to vaccine roll-out was that antibody positivity rates scarcely increased over time, increasing a little in the immediate aftermath of an increase in cases but then falling again as cases fell. Now, when vaccine roll-out is in full swing, positivity rates are barely above vaccination rates, suggesting that the tests are missing a lot of people who have been infected but not vaccinated. All-in-all this suggests that the antibodies that are being measured are not very long lasting. It is not clear to me why this is the case as other studies, including a recent PHE study, suggest that antibodies persist for many months.
“It would be helpful if the ONS could publish some of the underlying data on which the antibody prevalence estimates are based. How sensitive is the assay? Is it sensitive enough for this purpose or are better assays needed? How is positivity defined? Have individuals been followed over time to see how long their antibodies persist? Presenting the data as the cumulative proportion of people in the survey who have ever tested positive would give us a much better idea of the proportion of the population who have been infected and/or effectively vaccinated to date.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This is the latest in what has now become a regular fortnightly series of ONS bulletins that estimates the percentage of people in the UK community population that would test positive for antibodies to the SARS-CoV-2 virus that can cause Covid-19. It also presents estimates of the proportion of the community population that has received at least one dose of vaccine. The data come from the ONS Infection Survey, which (among other things) tests a representative sample of the population for antibodies, and asks them about vaccination, though the vaccination estimates for England also use some administrative data on vaccinations.
“Overall, the picture in this bulletin roughly continues trends that we know about from other data, and that we’ve seen in previous bulletins. On vaccinations, unsurprisingly the pattern matches what we know from the daily announcements about the roll-out. You can see in the data how the percentage of people vaccinated rises very quickly in specific age groups as they become eligible for vaccination. That rise is now happening in people in their 50s, so that by the latest week covered in the bulletin (week ending 3 April), around 90% of that age group were vaccinated, and the percentages are very much higher than that in the older groups. The vaccination rates are lower for younger age groups, so far, but have also been increasing as people working in health and care, and people with health conditions that make them more vulnerable, are vaccinated.
“The position on antibody positivity is more complicated. For the data in this bulletin, people can test positive for antibodies either because they were previously infected with the virus, or because they have been vaccinated. So you’d expect positivity rates to increase over time as more and more people are vaccinated. The estimates do show an increase in estimated antibody positivity rates in three of the UK countries, compared to the estimates from a fortnight ago, and in round figures, about half the community population aged 16 and over would test positive for antibodies in all four countries. (The latest antibody estimates are for the week ending 28 March, a week earlier than for the vaccination estimates, because it takes longer to collect and analyse the data for antibodies.) The exception is Wales, where the rates fell slightly.
“But the slight fall in Wales perhaps gives an example of why these trends are complicated to interpret. One point is that there is quite a lot of statistical uncertainty in the antibody positivity estimates, as is always the case in estimates based on survey data, and that’s particularly true for the smaller countries such as Wales where fewer people are tested in the ONS survey. So it’s not statistically certain that the small fall in Wales truly reflects what is going on with antibodies there. But, more importantly, the overall trend in the number of people who would test positive is a complex mixture of several different trends causes by different things. People, who did not have antibodies before, will move from testing negative to testing positive, either because they have been infected or because they have been vaccinated, though antibody positivity does not develop until two to three weeks after infection or vaccination. Currently infection rates are fairly low, but they are also not known perfectly by any means (though the ONS Infection Survey does provide the best data on that, across the UK, that we have). So antibody positivity would have gone up in Wales (and elsewhere in the UK) partly because of new infections, but more because of new vaccinations. However, working in the opposite direction, someone who tested positive for antibodies in the past might no longer test positive for antibodies now. That’s because the levels of antibodies from most viruses do decrease naturally over time (a process known as ‘waning’) and can fall below the level at which they can be detected by the test. I’m no immunologist or virologist, but I understand that the rate at which this decrease occurs can vary considerably from one person to another, and can also vary according to whether the antibodies originally arose from infection or vaccination.
“ONS point out in their bulletin that one reason for declines in antibody positivity in some older age groups may be because of some people’s antibody levels, that were positive soon after their first vaccine dose, will have fallen below the level at which antibodies can be detected. There’s some evidence in the oldest groups that levels might be beginning to rise again, which may well be because more of them have had a second dose (and enough time has passed after that second dose for antibodies to increase). If that is indeed what’s going on, we’ll see such trends more clearly in future bulletins as more people get their second dose. But ONS are, rightly, very careful to point out that a lot is still unknown about how antibody levels change over time, in vaccinated people and in people who have been infected. And they emphasise that immunity from future infection depends on more than just antibody levels – other aspects of the immune system such as T cell response come into it too.
“But, overall, because of the balance between increases in antibody positivity from vaccination and infection, decreases in positivity because of natural reduction in antibody levels, and the fact that antibodies aren’t the whole story on immunity, it’s not possible to say definitely from these results either how close we might be to a level of immunity that will keep infections strongly in check. That doesn’t mean that these results aren’t important – they surely are, they just aren’t the whole story. And it’s worth mentioning that ONS are going to improve the aspects of the infection survey that test antibodies. An ONS announcement from the end of last month* says that they are increasing the numbers of people being tested for antibodies, asking them to remain in the sample for longer, and changing the actual antibody test that’s used so that it can distinguish between antibodies arising from infection with the virus and antibodies arising from vaccination. That should considerably improve our understanding of the complicated interplay of vaccination, infection, antibodies, and immunity.”
All our previous output on this subject can be seen at this weblink:
Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee. I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic. My quote above is in my capacity as an independent professional statistician.”
None others received.