What this means is that there is a period of some months, sometimes over a year, after exposure to a pathogen when you can take a blood sample and find those specific antibodies. Lack of antibodies is not really a good indication that the person doesn't have immunity, because they might have memory B-cells in reserve, and there are also memory T-cells, which are totally different ways to fight a virus. And it's possible to have antibodies without having full protection against the pathogen. But in general, antibodies in the blood are a good indication that sometime recently that person has been exposed to the pathogen and had an immune response to it.
We are now getting to the point where we can screen populations in an attempt to determine how many of them have antibodies to the virus. This doesn't necessarily mean that they had an infection. They might have had memory B-cells to previous cold-type coronaviruses, and their immune system recognized the similar pathogen and woke up those B-cells to start making antibodies. This is the most prominent theory of why flu viruses go away in the summer. The idea is that during the winter, people are exposed to low levels of the flu virus, and awaken their immunological memory, even without an infection, to the point where the population gets herd immunity to the flu going around that winter.
In order to get herd immunity to covid-19, we need personal immunity in roughly 60-80% of the population. None of the populations studied so far have gotten close to that, but it's worth paying attention to these numbers. Aside from monitoring how close we are to herd immunity, the other consequence of knowing how many people in a population have antibodies is that it *might* allow us to get a better handle on the case fatality rate, or at least on the general "danger level" of the virus in the future. If we assume that everyone who has antibodies was actually infected, we can assume that the total cases are equal to the survivors with antibodies plus the fatalities. Divide the fatalities into the total cases, and you'd have the case fatality rate. But as I described above, I don't think that it's safe to assume that everyone who has antibodies was infected. We can't even assume that they are now totally immune, although they probably have at least partial immunity. But it does give us a reading on the overall immunological situation.
So with that introduction, here are the serology studies I know about to date:
Santa Clara County, California:
https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1
This is a serology study done in northern CA in early April, about 2 months after the first official case in this location. They advertised for volunteers on FB and screened blood samples for antibodies to the virus. Out of about 3300, 50 were positive. That's 1.5%, which is much higher than suggested by the official cases at the time. The authors noted that minority groups were underrepresented in their sample, and so corrected upwards.
My concern with this study is that they might have gotten a self-selected bias. That is, people who suspected that they had been infected might be more likely to answer the ad and go to the trouble of getting tested. While they did ask people whether or not they had been symptomatic, they didn't report anything about those answers. I suspect that the reviewers are going to call them on that. Their estimate of cases is high enough that it leads to the conclusion of a very low case fatality rate. That suggests that either the fatalities are dramatically under-counted, or their sample is over-representing cases.
So while it seems likely that the true number of cases is under-reported everywhere, due to the shortage of tests, I don't think that they are under-reported as much as these authors suggest.
Scotland
https://www.medrxiv.org/content/10.1101/2020.04.13.20060467v1In this study they are screening blood donors, which excludes symptomatic people. In mid-March, there were zero blood donations with anti-SARS2 antibodies. But in late March, there were 5 out of 500 that had neutralizing antibodies, that is, antibodies able to prevent infection, and a sixth was positive for antibodies, but not necessarily neutralizing antibodies. That's 6/500 or 1.2% of the blood-donating population.
Assuming that there is a 7-day delay in testing, the number of people who had tested positive at this time was 1600. Divide that into 5.3 million, and you get 0.03%. So far more antibodies in the population than suggested by the official tests.
I think this study is better than the one in Santa Clara, CA, because they are probably avoiding self-selection by people who really want a test. There are rumors circulating that you can get a free covid test by donating blood, but my guess is most people are not going to assume that that's the case. It's also better to test for neutralizing antibodies, rather than just any antibodies, given that we care whether or not the antibodies present will actually prevent infection.
The really interesting thing here is that the blood was all negative on March 17th, but 1.2% positive for the samples collected between March 21-23. It takes at least a few weeks to get a new antibody response going in a mild case, (see for example this paper) and two weeks prior to March 21-23, there were only a few official cases in Scotland. This suggests to me that they're getting re-activation of memory B-cells from colds, which happens more quickly, not totally new immune responses from silent infections. But we don't know that for sure.
Los Angeles
http://publichealth.lacounty.gov/phcommon/public/media/mediapubhpdetail.cfm?prid=2328This is a press release, not an academic paper. The meat of it is found in this quote:
"Based on results of the first round of testing, the research team estimates that approximately 4.1% of the county's adult population has antibody to the virus. Adjusting this estimate for statistical margin of error implies about 2.8% to 5.6% of the county's adult population has antibody to the virus- which translates to approximately 221,000 to 442,000 adults in the county who have had the infection. That estimate is 28 to 55 times higher than the 7,994 confirmed cases of COVID-19 reported to the county by the time of the study in early April. The number of COVID-related deaths in the county has now surpassed 600."
Unlike the study in Santa Clara county, they are attempting to get a random sample of the population: "Participants were recruited via a proprietary database that is representative of the county population. The database is maintained by LRW Group, a market research firm."
It's hard to say too much with so little to go on, but this at least looks encouraging. It's tempting to speculate that the reason CA has not been hit as hard as NY is that Californians passed around a cold-type coronavirus in the past few years and are now awakening their immunological memory.
Swiss military recruits
https://militaryhealth.bmj.com/content/jramc/early/2020/04/16/bmjmilitary-2020-001482.full.pdfOne recruit out of 140 in the whole company went on vacation, and four days later became symptomatic. That initial case tested positive via RT-PCR, and was isolated. 55 recruits who had been in contact (direct or indirect) with that first patient were put into quarantine. In one quarantine room, seven out of nine recruits developed symptoms. Only one of those tested positive via RT-PCR, which brings our official positives to two. The interesting thing here is that the second case was never in direct contact with the first case, only indirectly through a third person who never tested positive. The other interesting thing here is that only the two official positive cases developed antibodies. None of the others, including the six who had mild symptoms, and the person who was presumed to have carried the virus from patient 1 to patient 2, developed antibodies. It's possible that the tests (both PCR and antibodies were not sensitive enough to pick up low levels. Or it's also possible that the transmission from patient 1 to patient 2 via the third person was just through carrying the virus on his hands. But what we're not seeing here is young healthy people developing a robust immune response to the virus. It appears that they're just not very susceptible to catching it, compared to the general population.
