Antibody-Dependent Enhancement and COVID-19: Still No Evidence — Deplatform Disease

I have written at length about the prospect of antibody-dependent enhancement (ADE) in COVID-19 before here, but recently some individuals have seized on a Letter to the Editor by Yahi et al. published in the Journal of Infection which notes the presence of infection-enhancing antibodies, wherein those individuals imply (or overtly state) that ADE in COVID-19 is now confirmed and the vaccines have doomed us all. In truth, this letter and its findings don’t actually change anything from my post in December, so I’ll go through why that is and refer you to that post for a detailed explanation of ADE. To begin with, the article references a very well done study by Li et al. which noted that there was a site on the N-terminal domain of the spike protein (NTD), which upon antibody binding enhanced the ability of the virus to infect cells in vitro. These antibodies were found to be more common in individuals who had more severe cases of COVID-19. Note however that this is a retrospective design, so we cannot be sure of whether or not these antibodies are causal in the worsening of these patients’ disease. For example, in severe COVID-19 there appears to be significant epitope spreading (basically a broadening of the antibody targets on a given antigen) which could account for the appearance of these antibodies, wherein the severe disease should precede their emergence, rather than be caused by it.

Returning to the Letter to the Editor by Yahi et al., the critical limitation is that it is largely speculation and conjecture based on a modeling study with no actual in vivo evidence. The study argues that the affinity for antibodies at the infection-enhancing site is higher for Delta variants than for the ancestral strain, which is something that could readily be demonstrated with a surface plasmon resonance study, but this wasn’t done. Modeling studies are valuable and shouldn’t be disregarded but they need to be confirmed with actual data from the real world, and the authors don’t do this. The Letter also appears to conflate the Delta (B.1.617.2) variant with the Kappa (B.1.617.1) variant, which have very different properties as far as antibody neutralization (based on significant differences in the spike protein mutations each bears) and very different prevalences around the world (Kappa has nearly disappeared from much of the world and appeared to have substantially greater antibody evasion than Delta). This is especially important because the key model they present of infection enhancement through the aid of intact lipid rafts is based on the spike protein of the Kappa variant, and not the Delta variant.

When Yahi et al argue this though they really lose me:

In conclusion, ADE may occur in people receiving vaccines based on the original Wuhan strain spike sequence (either mRNA or viral vectors) and then exposed to a Delta variant.

Their data absolutely does not justify this conclusion. ADE has to be demonstrated clinically- end of story. Infection-enhancing antibodies are a common feature of numerous viral infections, and in practice, the infections in which it is actually relevant for is a very short list (basically Dengue being the main one with some weaker evidence for Zika, Ebola, and HIV). Even if antibodies are infection-enhancing, they may have effector functions in vivo that contribute to viral clearance rather than enhancing pathology, hence why modeling studies alone are inadequate to render such conclusions. For instance, in HIV, a pathogen known to benefit from ADE, antibodies were shown to be protective for functions unrelated to neutralization (antibody-dependent cellular cytotoxicity), which is independent of whether or not the antibody in question enhances infection.

I am also bothered by the fact that there is no mention of the risk of ADE among those who recovered from COVID-19 as their antibodies can decline quite rapidly and would likely have greater specificity to the variant they initially encountered than those of vaccinees, arguably representing a greater risk. In fact, antisera from vaccinees demonstrates significantly greater neutralization capacity than from recovered patients, excluding those who are critically ill.

For ADE to be meaningful, it needs to be demonstrated clinically in actual humans. The evidence to date is really not supportive. For one thing, convalescent plasma treatment was widespread earlier on in the pandemic and at a minimum, did not appear to worsen outcomes (see the original ADE post for that). For another, reinfection appears to be rare and is generally milder than the initial infection. Perhaps most importantly, it is not the vaccinated individuals that are driving the surges in hospitalizations and cases, as they represent a minority of these. Furthermore, to date, breakthrough infections remain milder and shorter-lived than infections in unvaccinated people.

I do want to point out however that I don’t believe that studies such as those by Li et al and Yahi et al lack value. Showing infection-enhancing antibodies is useful for example, as Li et al first offer the suggestion to exclude the region of the spike protein targeted by the infection-enhancing antibodies in next-generation vaccines, which is a reasonable suggestion (that of course would need to pass muster in clinical trials). However, this panic over ADE in COVID-19 is starting to get ridiculous (actually it got ridiculous quite a while ago but I digress). There really isn’t good evidence of it to date, and it would be rectifiable relatively quickly with a booster dose of vaccine. Furthermore, the specific patterns of the immune response in which enhanced disease could arise are relatively well defined and informed the choice to use technologies like mRNA and adenovirus vectors for vaccination as they avoid these types of responses.

I am formally requesting that we stop claiming evidence of ADE in COVID-19 in humans until we have actual clinical evidence of ADE in COVID-19 in humans- not in silico data, not in vitro data, not animal models.

Originally published at on August 16, 2021.



I write about vaccines here. You can find me on Twitter @enirenberg and at (where I publish the same content without a paywall)

Love podcasts or audiobooks? Learn on the go with our new app.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Edward Nirenberg

Edward Nirenberg


I write about vaccines here. You can find me on Twitter @enirenberg and at (where I publish the same content without a paywall)