I've gotten many questions about the comments from the WHO about how people who recover from #COVID-19 and have antibodies might not be protected against subsequent infection, so this thread is to explain what that meant (yes, it's long).
Also: https://twitter.com/WHO/status/1254160937805926405?fbclid=IwAR3pP1tOH5JihuD8EMbFwDj2JcRlQFzq9P0paSja5Z3h2UkcceNKcNSOjjo

The WHO recently made some comments that were phrased a bit clumsily and then misinterpreted by everyone regarding whether or not the presence of antibodies against COVID-19 is indeed proof of immunity, and...

more importantly proof that a person cannot be reinfected a second time with COVID-19. Basically, the WHO was trying to say ...

We do not know whether or not infection ensures that a person has long-lasting immunity against SARS-CoV-2. We do not know how long whatever protection recovery confers will last, and...

we do not know whether antibodies are the appropriate correlate of protection against SARS-CoV-2 at this time. Hence, antibody tests should be taken with a grain of salt until these data can be ascertained.

Of course, that brings up more questions because most people don’t understand what “correlates of protection (CoP)” are, so let me explain. In immunology, and in particular in vaccinology, there are a number of things we can measure to determine...

whether or not an individual is protected against an infectious disease. These are known as correlates of protection. Correlates of protection are subdivided into mechanistic correlates of protection (mCoP) and nonmechanistic correlates of protection (nCoP).

Mechanistic correlates of protection refer to things we can directly measure that are themselves responsible for the protection. Nonmechanistic correlates of protection refer to variables that we can measure that are closely associated with the state of protection...

but are not themselves doing the protecting.

In the broadest possible terms, the immune system has 2 ways to deal with any given infection. The first is pretty well known, with many pop culture references: antibodies (I think I will make a separate thread for those).

The other major thing the immune system can do is kill infected cells using killer T cells. This is particularly important for dealing with viruses. The thing is, antibodies are very easy to measure.

This is, in part, why we often default to them when we check for immunity (like with titer testing).

T cell responses by contrast are much more difficult to measure. In the most basic terms, correlates of protection are evaluating either antibodies or T cell responses.

Often, the goal of determining correlates of protection is to obtain an absolute correlate, in essence, a threshold for the immune responses past which we know protection should occur.

Ultimately, the presence of memory cells is the most important mechanistic correlate of protection.
Establishing a correlate of protection however, is not at all simple.

For one thing, we need to know a great deal about the nature of the immune response to the infection or the vaccine. For instance, correlates of protection need to be defined in relation to a particular phenomenon.

The quantity of antibodies required to prevent bacteremia (the dissemination of the bacteria throughout the blood) due to pneumococcal disease is not the same as that needed to prevent otitis media.

Additionally, sometimes it is acceptable to use a different correlate of protection with a vaccine than that which is produced by the infection. The best way to evaluate immunity against measles for instance is to measure the T cell responses against the virus...

but it is far easier to just check to confirm that antibody levels against measles antigens are above 120 mIU/mL. Furthermore, it is always true that a sufficiently high challenge dose can overwhelm existing immunity.

There are also problems with simply measuring antibodies in that simply getting a number does not tell you anything about their functionality.

For instance, it isn’t guaranteed that the antibodies being detected are neutralizing- that they prevent the entry of the pathogen or toxin into the cell.

Isotype matters too. IgG is required for antibody-dependent cell-mediated cytotoxicity (the antibody binds the surface of an infected cell and recruits natural killer cells to destroy it) for instance, which is critical to the control of some viral infections (e.g. HIV).

In addition, it is not uncommon for pathogens to have multiple CoP at the same time. Live influenza vaccines for instance may induce serum IgG and mucosal IgA, and CD8+ T cell responses. Whole pertussis vaccines should induce Th1, Th17 and antibodies. Which should we measure?

The precise correlates of protection against SARS-CoV-2 have yet to be elucidated. Antibodies to the receptor binding domain of the virus should be neutralizing and sufficient to prevent infection and disease...

but we still do not know what quantity of antibodies is needed to accomplish this. We know relatively little about the T cell responses required in this process and this is further complicated by reports of cytokine storms in severe disease.

Also here is a great read on correlates of protection: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2897268/pdf/0131-10.pdf

Hope that helps.

@threadreaderapp unroll, please