Start main page content

Covid-19 Expert Advice #3: Testing for Covid-19? All you need to know about antibody tests

- The Scientists Collective

The third in a series of expert advisories on how to respond to Covid-19 which will be published from time to time by Maverick Citizen.

This series is co-written by some of South Africa’s leading medical scientists and academics, including some members of the Ministerial Advisory Committee (MAC) on Covid-19. Although not written for or on behalf of the MAC, much of whose important advice is being lost in the corridors of government, it is intended to provide practical life-saving advice based on the best emerging knowledge and evidence of Covid-19. We are learning all the time, and it is highly likely that advice will change as we learn more about the virus, so keep an eye on any new recommendations. However, the basic principles explained in this series will stand and help inform your decisions.

These are some of the common questions we hear:

“Should I ask for a Covid-19 antibody test?”

“What is the difference between an antibody test and a PCR test?”

“Should I have a rapid diagnostic test?”

“If I have antibodies, does that mean I am safe from getting Covid-19 again, and do I still need to wear a mask, socially distance and take other precautions against Covid-19?”

South Africa has seen a flurry of controversies about the possible role and place for antibody tests – also referred to as “serology” tests – in managing Covid-19. Some have argued that antibody tests are the missing weapon, while others have countered that they provide little to no real-time value or are a “blunt stick”, offering false hope. Here, we try to help you unpack what is going on, when antibody tests might be helpful to you, your doctors, scientists and policy leaders in South Africa’s response to Covid-19… and when they won’t be of help.

The bottom line:

Should I ask for a Covid-19 antibody test?”

Sorry, no, don’t use antibody tests, not yet at any rate. It becomes positive too late after getting sick to really be useful, and even the “it’s nice to know I was infected’’ is limited as the test becomes negative quite quickly. Better tests, and combinations with other tests may make them more useful with time, but some of the limitations are unlikely to be overcome ever.

They have a place in research studies, but don’t waste your money for now.

Some basic facts about antibodies:

Antibodies are proteins that are produced by the human immune system in response to an infecting pathogen, in this case, the coronavirus SARS-CoV-2, the cause of Covid-19. Their job, along with other parts of the immune system, is to identify coronavirus and get rid of it from the body. Antibody tests detect the presence of these antibodies in our blood (or less commonly, other body fluids).

The production and appearance of antibodies is the end product of a complex process involving many cells, signals and messages of our immune system that takes time to develop after coronavirus infection occurs. Therefore, when the virus first infects us and causes many of the symptoms that we now recognise in the first 7-10 days, antibodies are nowhere to be seen.

There are two main antibody types which appear and disappear at different time points:

  • IgM antibodies start to appear first, usually ~7-10 days after infection with coronavirus and disappear again after weeks to months;
  • IgG antibodies appear second, usually ~2 weeks after infection with coronavirus, and usually persist for longer, in some infections, lifelong.

This is important because some antibody tests look for both IgM and IgG antibodies in blood, and will, therefore, be positive earlier than those antibody tests that just look for IgG antibodies. So, which test and the timing of that test is important in understanding a result.

Antibodies do their job by binding to specific parts of the virus. Some of these parts are critical for the virus to cause the damage it does and therefore by binding such a specific, important part of the virus, those antibodies will neutralise the virus’ damaging effects.

We call these “neutralising antibodies” and if they are present, it tells us firstly that the person has had coronavirus infection previously (they wouldn’t have been produced if the person hadn’t been infected in the first place), and it indicates that if that person was to encounter coronavirus again, if those antibodies were still there, then a level of protection is likely. These neutralising antibodies are what we hope will be generated when we vaccinate people against Covid-19. If they are, then that vaccine will have induced “protective immunity” in that person.

Other antibodies recognise and bind to proteins of the virus which are not critical for it to do its worst and these antibodies do not generally neutralise the virus, and are therefore referred to as “non-neutralising antibodies”. Their presence merely tells us that the person was infected previously with the coronavirus, but doesn’t tell us if they will have any level of protection should re-infection occur.

Laboratory testing and vaccines © WITS UNIVERSITY

“Can the Covid-19 antibody test tell me that the symptoms I have now are caused by Covid-19?”

We have already seen that antibodies only start to appear in blood after the first week of infection, and only meaningfully from ~10-14 days. The symptoms that we get with Covid-19 – fever, cough, shortness of breath, sore throat, loss of taste and/or smell, fatigue, muscle pains etc – generally only occur during the first week to 10 days, ie, when antibodies aren’t around yet, so in this case, an antibody test is of no use to you.

The test you need in this setting is called a PCR that is able to flag the coronavirus itself in a sample taken from your nose. We know that the presence of symptoms correlates with presence of the coronavirus which are at its highest levels during the symptomatic period.

This is the test you need if you need to diagnose acute Covid-19 infection.

There is one known exception to the rule that antibody tests are no use for diagnosing Covid-19 during the acute infection. Very rarely, children may have a form of Covid-19 called “multisystem inflammatory syndrome in children” (MIS-C), during which PCR is negative but specific antibodies may be positive. We are still trying to understand why this is so.

“Can the Covid-19 antibody test tell me whether I have had Covid-19 in the past?”

Unfortunately, the answer is “maybe”. Think of it like this – for the antibody test to be positive a certain amount (level) of them needs to be present in blood. That amount depends on two main things:

  1. How strong your antibody response is: We know that the level of antibody response you make is dependent not only on you as an individual (everyone is slightly different), but also on how severely you had Covid-19. The people who have the highest levels of antibody, and therefore most likely to test positive, are those who have had the most severe illness. Therefore, people with Covid-19 who have pneumonia and are admitted to hospital will generally have higher antibody levels in blood than someone who had mild symptoms treated at home, who in turn would have higher antibody levels that someone who had no symptoms at all, ie, asymptomatic (about 30-40% of infected with coronavirus). So people who have had asymptomatic infection or very mild Covid-19 may not mount a great enough antibody response for the test to register as positive.
  2. At what time point the blood was taken for the antibody test: As we have seen above, antibodies only start to appear in blood after symptoms are getting better, so if you take the test too soon, then they may not be at a high enough level to register as positive. In addition, the information that we currently have on how long antibodies to coronavirus last in the blood only goes up to six weeks, and we think thereafter, they may wane. Hence, we can’t say with any certainty at this point whether or not the symptoms someone had over a month ago are due to Covid-19.

“But if my test result comes back positive, surely that tells me that I definitely had Covid-19 and that would help?"

Again, the somewhat disappointing answer is “no”, sorry. It is possible to have a “false positive” result, which means that the antibody that is being picked up by the test is one that looks very much like the SARS-CoV-2 but isn’t that virus. An example of such a “cross-reaction” would be to one of the four other coronaviruses that are known to be a cause of the common cold.

Hence, a positive antibody test result does not reliably prove prior SARS-CoV-2 infection.

“If I have a positive antibody test, does that mean I am protected from getting Covid-19 again?” 

Once again, the answer to this is a cautious “maybe”.

Remember our earlier discussion about neutralising and non-neutralising antibodies? If the antibody test that you take is designed to pick up neutralising antibodies (the best ones are against the “spike protein” and the “receptor-binding domain”) which we know are protective, then if you have a positive test, as long as those antibodies remain in your blood, we think that you will have good protection. However, we have also just discussed the fact that antibody levels might decline over time, so that as they wane, you may lose that protection.

However, if the antibody test picks up non-neutralising antibodies, then having a positive test tells you nothing about whether you have protection or not. 

Saying this, a further layer of complexity is brought in as we now understand that antibodies are not the only part of the immune system that could protect us against repeated coronavirus infection. Some of those cells, which form part of our complex immune system – T cells – are implicated in protection. What this means is that if you have had Covid-19 diagnosed by a PCR test when you were sick, but your antibody test later on is negative, you may still be protected. Sadly, there is still much to learn.

“If I have a positive antibody test, do I still need to wear a mask, socially distance and take other precautions against Covid-19?”


For all the reasons we’ve discussed, at the current time with the tests we have, a positive antibody test result must not be misinterpreted as proof of immunity. Do not reduce or abandon protective measures – you may not be protected yourself and it is possible (however, unlikely from our current experience) that you can become infected again, and infect others!

This is why the World Health Organisation has strongly advised against “immunity passports” or “immunity certificates”.

“So is there anything that antibody tests could be used to help us with the pandemic?”

With all the caveats around tests being falsely negative if antibody levels have dropped or if the response of our immune system was too low in the first place, antibody tests may give us useful information about the level of coronavirus infection at a population level. We call these “seroepidemiology” studies (“sero” – indicating the sample comes from blood, and “epidemiology” being the study and analysis of distribution of a disease).

Seroepidemiology studies can be undertaken across the entire country, in certain provinces or districts, or certain groups at high risk of infection, eg, hospital patients and staff, residents and staff of old age homes, prison inmates and staff etc. The information that they give us can be even more useful if we repeat the studies at future time points which gives us trends over time. Another use would be for community surveillance, eg, looking at settings where infections have occurred recently, to evaluate “hotspots” of coronavirus transmission, and to help reconstruct chains of transmission within the community or specific groups.

Antibody tests are an important tool in understanding whether the vaccines that the world (South Africa included) is developing against the Covid-19 work. After we vaccinate someone, we wait and test for an antibody response (especially the production of neutralising antibodies) over many time points. This tells us whether the vaccine induces protective immunity, ie, whether it works, and for how long and at what level those antibodies stay in the blood.

Ideally, what we want to see are high levels of neutralising antibodies that remain over time. Lifelong would be first prize!

“What’s this rapid diagnostic test (RDT) I’ve been hearing about? Is that an antibody test and is it any better than antibody tests that have to be performed in a laboratory?”

The use of the term “rapid” in this context indicates that the result is available sooner than those that have to be sent off to a laboratory, as most often, RDTs are performed at the point-of-care, ie, at your doctor’s clinic or by you in your own home.

The same problems and restrictions in interpreting RDTs for antibodies apply as for laboratory tests. RDTs do not pick up antibody levels in blood sooner or more reliably. In fact, most of the RDTs that have been produced so far seem to perform worse than laboratory-based tests, ie, the sensitivity (ability to pick up the antibodies) of the test is lower.

However, while we are on the topic of increasing the speed of tests to give us an answer, future diagnostic developments may well see faster diagnostic tests targeting the virus itself. New versions of tests similar to PCR (some are in use in other countries but very expensive and with doubtful reliability) or for viral protein (antigen). Antigen tests, (not to be confused with antibody tests!) in common with the PCR, detect coronavirus in samples from the airways so they can diagnose acute infection. They can be simple to perform and cheap but sensitivity remains an issue, which is problematic for a diagnostic test.

“What should I do while waiting for the test results?"

If you are ill with fever, cough, etc – isolate yourself!

Our expert advice is based on what is known about the topic currently and is largely aligned with national policies. We will have to watch the rapidly growing body of data to update it should significant new findings emerge.


If you are interested in learning more, we would like to refer you to two articles about Covid-19 laboratory testing in general and antibody serology in particular.

This article was first published in the Daily Maverick/Maverick Citizen.

Professor Lucille Blumberg, National Institutes of Communicable Diseases; Professor Diana Hardie, University of Cape Town/NHLS Groote Schuur; Professor Marc Mendelson, University of Cape Town; Professor Shaheen Mehtar, University of Stellenbosch / Infection Control Africa Network (ICAN); Professor Wolfgang Preiser, University of Stellenbosch / NHLS Tygerberg; Professor Wendy Stevens, University of Cape Town/NHLS Priority Programmes; Professor Francois Venter, Ezintsha, University of the Witwatersrand.