Should Evusheld be brought back?

An injectable drug that prevented Covid-19 in immunocompromised people was shelved by the FDA last year. It's time for another look.

Today’s Inside Medicine is a deep dive. In writing this, I spoke to officials at the FDA, trusted experts in my network, and the lead author of a new study that inspired this post. If you find this valuable, please join the community of supporters who help keep this going.

For those who already have upgraded, thank you!

Now let’s dig in…

---

Back in the day, injectable monoclonal antibodies were shown to decrease the risk of severe Covid-19. One medication called Evusheld (a combination of two monoclonal antibodies called tixagevimab and cilgavimab) was shown to prevent SARS-CoV-2 infection. The medication was expensive, but for very high-risk people— immunocompromised, vaccine non-responders—Evusheld was a lifeline, decreasing infections by over 76% for 6 months.

That was during the Alpha, Beta, and Delta variant eras. When Omicron appeared in 2022, the FDA increased the dose, an attempt to compensate for decreased stickiness between the virus and the drug. Then, in early 2023, when the Omicron XBB.1.5 subvariant appeared, the FDA effectively revoked its emergency use authorization for Evusheld altogether.

Evusheld was pulled based on lab data.

Here’s the issue: That decision was based on laboratory analyses indicating that Evusheld was “unlikely to be active” against Omicron XBB.1.5. In hindsight, I wish they’d considered clinical data from real patients. To make sure I have this right, I asked the FDA. A spokesperson shared the following with Inside Medicine readers:

“Our determination was made by a multi-disciplinary team, therefore, it would be preferable to state that FDA’s determination was based on the estimated proportion of susceptible variants circulating in the United States with susceptibility determined using in vitro neutralization assays of viruses and/or pseudoviruses, and to reference the Center for Drug Evaluation and Research Review Memorandum.”

Indeed, it wasn’t that the effectiveness of Evusheld had observably vanished in patients. It’s that scientists concluded that Evusheld probably wouldn’t work anymore because of its decreased ability to neutralize Omicron XBB.1.5 in the lab.

I have a problem with this. The clinical significance of those findings was never measured where it counts: In people.

Does Evusheld actually still work in patients?

A new study in JAMA Oncology from Italy found that in late 2022/early 2023, patients with blood cancers (like leukemia, myeloma, and lymphoma) who received Evusheld had fewer SARS-CoV-2 infections over the following 6 months.

The findings: The apparent effectiveness of Evusheld was lower compared to the original blockbuster trial. In the new study, among Evusheld recipients, the rate of infection was half that of those who didn’t receive it; not as good as the 76% reduction in the initial trial, but substantial. The results on severe disease or Covid-related hospitalization looked better; the rate of poor outcomes in Evusheld recipients was around 1/5th of those who did not get Evusheld.

A key question: Did the infections that occurred during the study happen before or during the time XBB.1.5 was circulating? To find out, I reached out to Dr. Marco Salvini, the lead author of the JAMA Oncology manuscript. Dr. Salvini shared data with us indicating that Evusheld was effective only early in the study—that is, during the time before XBB.1.5.

That implies that the FDA’s decision was possibly correct after all. Except for one thing: Patients in Italy received half the FDA’s recommended dose of Evusheld at that time. In 2022, the FDA felt that to overcome Omicron’s mutations, the original dose of Evusheld needed doubling. But over in Italy, the original dose was maintained. Nevertheless, Evusheld worked extremely well during the pre-XBB.1.5 era in the data from Italy just published. The fact that the original dose seems to have worked against Omicron variants in the fall of 2022 suggests that the FDA’s recommended double dose might have also worked against XBB.1.5 when it took over in December 2022.

A limitation: The new study from Italy was not a randomized controlled trial. It was a retrospective review; patients who received Evusheld were matched and compared to those who didn’t.

An FDA official familiar with the decision-making process made this point when defending the policy to me this week. It’s possible in observational studies that, despite all efforts to assemble a control group who did not receive the drug but who were otherwise similar to drug recipients (i.e., similar age, gender, race/ethnicity, severity of existing medical problems), there remain real unmeasured differences in the groups’ baseline characteristics. This is called “residual confounders.” It’s possible that despite looking similar on paper, patients who did not get Evusheld were sicker, and that the doctors’ decisions reflected that, even if in subconscious ways.

These were fair points. My response, however, was that the data that informed the FDA’s revoking of Evusheld’s authorization had been based on lab data alone, rather than clinical data. And if the confounders were so large, the effect shouldn’t have gone away when a new variant showed up.

That’s where the conversation ended, unfortunately.

I hope the FDA reassesses this. One on hand, newer variants may make Evusheld less effective now. This would again support the FDA’s prior decision. However, the data from Italy don’t reflect the best-case scenario because doses lower than the FDA’s advised one were used in Italy. So, I wonder how Evusheld would have performed at the FDA’s last recommended dose. Probably better.

Can Evusheld come back? Is there a double standard at work?

Based on the new clinical data, I asked FDA officials if they would consider bringing Evusheld back. Here’s their response:

“As a general principle, we remain open to considering data generated from sources other than randomized controlled trials in our decision-making.   The strengths and weaknesses of the trial design and the completeness and relevance of the data generated are amongst the important factors in determining the utility of the data.”

That sounds like a “No.”

Are other drugs receiving the same scrutiny? My epidemiologist colleague Dr. Gideon Meyerowitz-Katz pointed out to me, for example, that the FDA has allowed molnupiravir to remain on the market during the vaccine and Omicron eras, despite a massive trial showing its benefit had declined almost to nil—especially jarring compared to its original exceptional performance in the pre-vaccine days.

I’m not saying the FDA should revoke its authorization for molnupiravir. I’m saying that I find it unsatisfying that a drug with high-quality clinical data showing it barely does anything nowadays (i.e., molnupiravir) remains available, while a drug like Evusheld was pulled from the market based entirely on lab data. So, right or wrong, I think the FDA’s revocation of Evusheld’s authorization was based on premature certainty. The calls should be informed by clinical data.

What happened to the monoclonal antibodies?

Remember Regeneron? Back in 2020, that company had monoclonal antibodies against Covid-19. Trump took it when he got Covid-19. Other companies, like Eli Lilly, made similar compounds. These agents seemed to work, though they weren’t going to change the course of the pandemic. Over time, the FDA granted and then revoked authorizations for all of them, pulling them from the market, one at a time as new variants rendered them unlikely to work. Evusheld was the last one.

Why didn’t these companies alter the formulas to keep up with new variants? The problem, I learned from then-White House Covid-19 czar Dr. Ashish Jha, was that it was too expensive a gamble for the companies. If they poured a billion dollars into new versions of monoclonal antibodies only to have new variants render them useless just months later, they’d lose their shirts. To me, this felt like a good opportunity for public funding to lower the risk for these companies. (During a pandemic, mind you; I’m not in favor of freebies for these companies generally.)

In fact, a year ago in Inside Medicine, I wrote this:

As the pandemic progresses, immunocompromised people increasingly have the most at stake here. Congress should incentivize pharmaceutical companies

to make mABs that prevent infection and that can be reserved for the immunocompromised. If we end 2023 without a new mAB on the market for outpatient use, our leaders will have failed. If this happens, we should notice it and register our discontent, the same way popularity polls reflect gas prices.

2023 came and went. Our leaders failed. I’m not even sure they tried.

Now it’s 2024 and it’s at least possible that some of the shelved drugs might still work for people who really need something. Or it might be that tweaks in the formulations may be less vulnerable to mutations that decrease effectiveness than believed. That is, updating Evusheld’s recipe might lead to a drug that becomes less effective over time, but not to the point of clinical futility.

There’s hope.

The JAMA Oncology study’s findings suggest that another randomized trial of Evusheld should be conducted today. The technical term here is “equipoise.” Since the answer to the question—Does Evusheld still work?—is truly unknown, it is ethical to conduct a study where some people volunteers are randomized to get Evusheld and others placebo. (It is not ethical to do a study on a drug where you already know the answer.) It turns out that the makers of Evusheld are indeed doing a much-needed new study. However, I haven’t seen or heard about any results yet and we need an update. Come on Astra-Zeneca, makers of Evusheld, show us your latest data! Tell us whether updated Evusheld 2.0 prevents Covid-19 for immunocompromised patients.

If it does, a lot of people will be begging you to take their money.

Questions? Comments? Chime in!