The Data for (and Against) Coronavirus Booster Shots

Modest benefit, minimal risk

The COVID controversy du jour seems to center firmly around the booster shot.

Last week the Vaccine and Related Biological Products Advisory Committee at the FDA met to review data on the safety and efficacy of a potential third dose of the Pfizer vaccine.

I want to review the data we have available to inform decisions about dose three but first I want to clarify two things.

Number one, the boosters we are talking about are the exact same formulation as the original vaccine that have already been approved for use to prevent COVID-19. We are not talking about a modified vaccine designed specifically to generate immunity against the Delta variant. Those vaccines are in clinical trials but we have no efficacy data yet.

The second thing I want to point out is that a lot of the argument for boosters seems to center around the idea that immunity from the first two doses of mRNA vaccines is waning. There's debate about whether the observed waning reflects an immune process or is confounded by the more virulent delta variant but I want to argue that this is actually immaterial.

The utility of a booster needs to be defined, like we do for all medical interventions, on the basis of efficacy versus risk. As such the impact of waning immunity, if it exists, is only relevant insofar that it increases the chance that a booster will be proven efficacious. You can imagine a world where there has been no loss in vaccine efficacy over time, and a booster might still be worthwhile if it can increase immunity from high levels to even higher levels, which would further prevent disease or transmission.

Thus, the data presented to the FDA documenting a waning of immunity is less important to me than the data showing the effects of a booster.

To be clear, there are some worrying signs regarding the current efficacy numbers. These meta-analytic graphs tell the story better than I can.

Over time, the general trend is for vaccine efficacy to decline in terms of protection against infection, but to stay pretty strong against severe disease, hospitalization or death.

By viral variant, you see a similar change in efficacy between the Wuhan strain and Delta, with a more concerning drop in efficacy against severe disease for the gamma variant, which, fortunately, is rare at least in the US.

The theme, waning of efficacy against infection, maintenance of protection against hospitalization, appeared even in Pfizer’s own data.

Though to be fair, independent studies looking at vaccine effectiveness don’t show quite as dramatic a waning of efficacy.

Now, some would argue that as long as there is still good protection against severe disease and hospitalization, there is no problem here. I’m not one of them. Protection against mild infection clearly has public health benefits. Moreover, remember that even without waning of protection against severe disease, it is possible that boosters could increase protection from their current very high levels to even higher levels – and that’s a good thing too.

So, did they?

Pfizer’s formal booster studies looked at the neutralizing antibody titers of individuals one month after the booster dose compared to one month after dose two. This comparison, a non-inferiority test, basically asks if a booster can get you back to where you were shortly after dose 2, from an antibody perspective. Not the most compelling endpoint.

But the results were clear. Antibody titers one month after dose three were, on average, several fold higher than after dose 2.

These were antibodies directed against the original coronavirus strain, by the way, though results against Delta looked broadly similar.

In terms of hard efficacy data, that’s really it. Pfizer has successfully proven that a booster dose restores antibody levels to where they were after dose 2, or even higher, and it may be reasonable to assume that will translate into increased vaccine effectiveness, but the hard data for that link, if it exists, was not presented to the FDA.

Rather, that crucial step is provided through observational data from Israel, which jumped headfirst into boosters given internal data, now made public, that showed a decrease in efficacy against infection over time and, importantly, a decrease in efficacy against severe disease among their older age group – you can see here how the severe infection risk is higher among people vaccinated earlier.

At this point, a large proportion of the Israeli population has received a booster.

And, again based on observational data, this seems like a good move. This graph documents how infections plummeted among older individuals once boosters were introduced to that group.

Overall, the data is clear that boosters increase antibody levels, and circumstantial evidence suggests that that will translate into fewer infections. That means the utility of boosters really comes down to safety data. If adverse events are minimal, it seems like a good bet to roll them out. If not, a call for more data would be needed.

And, overall, the safety data looked very good. Despite the higher antibody titers achieved with boosters, the adverse events were, in general less frequent than those that occurred after dose 2. Light blue bars here are the rates after dose 2, compared to dark blue bars which represent dose 3.

It’s worth mentioning that individuals who had a bad reaction after dose 2 might be expected not to volunteer for dose 3, so perhaps the lower rate of adverse events here represents some selection bias, but the results are still encouraging.

The big adverse even we are all wondering about of course – myocarditis.

The data from Israel is encouraging – just 1 case of myocarditis reported with more than 3 million doses administered, but again we might be seeing the fact that those who get dose 3 probably handled dose 2 just fine.

The FDA advisory panel recommended an emergency use authorization for a booster for those above age 65 and who are at “high risk” of COVID-19, though that latter is not well-defined. At the time of this recording, I don’t know what the CDC advisory group (ACIP) will recommend.

But I think the FDA got it just right here. This data is not a slam dunk, but with the severity of COVID-19 in older adults, and the benign safety profile of a booster in that age group, the balance of efficacy and risk seems firmly in favor of authorization. 

Still, there are many key clinical questions that remain. What about those of us who received Moderna, which seems to have maintained a stronger efficacy profile than Pfizer’s vaccine? Is there a role for mixing and matching mRNA vaccines, or for boosting Johnson and Johnson with another one? What is the appropriate timing of a booster, and how dose that interact with next generation vaccines that are more specifically targeted against the Delta variant and its offshoots? What about the rest of the world? Does the marginal benefit of boosters to Americans outweigh the benefit Americans would accrue if we used those doses to vaccinate other parts of the world and reduce the emergence of new variants?

Unfortunately, without a truly robust vaccine tracking system, we may not get the answer to these questions.  The CDC estimates that more than one million Americans have already gotten boosters, and there are plenty of pharmacies and doctors offices that are administering additional doses of the various vaccines without asking too many questions about what people have had in the past. In short, pandora’s box is open, the boosters are here, and regardless of what the FDA and CDC say the provision of them are functionally in the hands of patients and doctors. Is this how the system should work? Probably not. But in this case, its hard to find a justification to deny boosters to those who really want them.

 A version of this commentary first appeared on medscape.com