Compound Interest: Antibody tests, part 1: What can antibody tests tell us?

Have you already had COVID-19? Even if you’ve had symptoms consistent with it, you may not know for certain if you didn’t have a test at the time. But newly approved antibody tests may be able to tell you if you had the infection. What exactly can these tests tell us? Part one of this two-parter on antibody tests takes a look.

Forget the tests for a minute. What are antibodies? When an infectious agent (or antigen) enters your body, your immune system detects them. Subsequently, the immune system produces antibodies, a special type of protein molecule. These destroy or neutralise the antigen, fighting off the infection.

There are different types of antibodies. The human body produces five different types (or to use the technical term, isotypes) of antibody: IgM, IgD, IgG, IgA, and IgE. In each case, ‘Ig’ stands for ‘immunoglobin’ (the type of protein that antibodies are). Each type of antibody has different characteristics and roles in the body.

All of these antibody types have two general regions. The constant region, which is identical in antibodies of the same type, and the variable region. The variable region differs for different antigens. Particular variants bind to particular antigens. Binding allows the antibodies to either destroy these antigens or mark them for destruction by other immune system cells.

Immunity to a particular disease is linked to antibodies. Once we’ve recovered from an infection, our immune cells ‘remember’ the antigen. If we get infected by the same antigen again, our immune system rapidly produces antibodies again to eradicate it. This immunity is lifelong for some diseases, but fades for others.

Antibody testing is sometimes presented as testing for immunity — though, as we’ll discuss, it can be a little more complicated than that. Antibody tests, sometimes referred to as serology tests, usually test for the presence of IgM or IgG antibodies to a certain antigen.

IgM antibodies are the first type of antibody produced in response to an infection, appearing in the blood after 5–10 days. Their production peaks around three weeks after the initial infection and then tails off. They only remain detectable by testing for 2–4 months.

IgG antibodies are the most common antibody produced in response to an infection, accounting for approximately 75% of antibodies in the blood. Their production takes longer (10–14 days) and peaks later 4–8 weeks) than IgM antibodies. However, they remain in the bloodstream for much longer. Depending on the particular antigen, they are detectable for many months or even years after infection. They’re also the primary antibody rapidly produced if we get reinfected by an antigen.

By testing for IgM and IgG antibodies, we’re trying to answer the question, “has this patient had a particular infection before?”. The answer to this question is determined from the presence or absence of these antibodies.

If an antibody test detects neither IgM nor IgG antibodies in a patient sample, it is likely that the patient has not had the infection. Because antibody production doesn’t start for several days, we can’t say for certain – they may be in the very early stages of infection. It is also a possibility that they had the infection but no longer have detectable levels of antibodies.

If an antibody test detects IgM antibodies in a patient sample, but not IgG antibodies, it is likely that the patient is in the early stages of infection. As we noted previously, IgM production starts earlier than IgG.

If an antibody test detects both IgM and IgG antibodies in a patient sample, it is likely that they have an active infection. Alternatively, they have very recently had the infection. There has been enough time since the patient’s initial infection for production of both antibody types, but IgM levels have yet to drop.

Finally, if an antibody test detects only IgG antibodies in a patient sample, it is likely that the patient has had the infection and recovered. This may also mean that the patient is immune to reinfection by the same antigen.

You might wonder why antibodies don’t guarantee immunity. If you have antibodies to a disease, you may have some degree of immunity to it. But antibody tests can’t tell us how effective a certain patient’s antibodies are at neutralising an antigen. The quantity of antibodies produced may also be important. Some antibody tests are able to determine how much of a particular antibody is in the bloodstream, but others cannot.

When it comes to SARS-CoV-2, we don’t know how long any immunity might last, even for patients who have antibodies. Better-known coronaviruses present a mixed picture. Some, like SARS, generated antibodies that gave people immunity for a few years. Immunity to others, like coronaviruses that cause the common cold, can last only a few months. Currently, there is no confirmed case of human reinfection with SARS-CoV-2, so we do not yet know for definite how long immunity lasts.

Immunity aside, how accurate are antibody tests, anyway? You might have heard terms like ‘sensitivity’ and ‘specificity’ in relation to antibody testing for coronavirus. These terms measure how often an antibody test generates a correct result. A key reason it’s taken time for accurate antibody tests to become available is because of issues with these aspects.

Sensitivity refers to how frequently the test correctly identifies antibodies in samples that contain antibodies to a particular antigen. Imagine we have ten patients, who all have antibodies to the antigen. A test with 100% sensitivity would correctly identify antibodies in samples from all these patients.

If a test had 90% sensitivity, it would mean that for every ten patients with antibodies, the test incorrectly identifies one patient as not having antibodies. This is what’s known as a false negative. So the higher a test’s sensitivity, the fewer false negatives it produces.

Specificity refers to how frequently the test correctly gives a negative result for patients who don’t have antibodies to a particular antigen. Imagine our ten patients again; this time, none of them have antibodies to the antigen. A test with 100% specificity would correctly give negative results for all these patients.

If a test had 90% specificity, it would mean that for every ten patients without antibodies, it would incorrectly identify one of them as having antibodies. This is what’s known as a false positive. It might be more problematic than a false negative, because this person may think they’ve had the infection when they haven’t. The higher a test’s sensitivity, the fewer false positives it produces.

It’s clear, then, that we want both sensitivity and specificity to be as close to 100% as possible. The recent Roche test approved in the UK claims to have a sensitivity of 100% — that is, it doesn’t produce any false negatives. The specificity of the test was 99.8% — meaning that for every 500 people tested without antibodies, only one person’s result would return a false positive.

The Roche test is one of two antibody tests to gain approval so far in the UK, and other companies’ tests have been approved in other countries. You might wonder why we need two different tests, and whether they function differently. If so, check back for part two of this antibody testing double act, which will look at how different types of antibody tests work.

Enjoy Compound Interest’s posts? Consider supporting Compound Interest on Patreon!