Paxlovid Trial Flops for COVID Pneumonia

— Small study found no decrease in mortality among inpatients in three countries

 A photo of a blue rubber gloved hand holding a box of Paxlovid.
  • author['full_name']

    Jeremy Faust is editor-in-chief of ľֱ, an emergency medicine physician at Brigham and Women's Hospital in Boston, and a public health researcher. He is author of the Substack column Inside Medicine.

Starting point: nirmatrelvir/ritonavir (Paxlovid) used to work very well.

In 2021, nirmatrelvir/ritonavir clearly decreased hospitalization and death among COVID-19 patients who had high risks of severe disease and who had not been vaccinated or previously infected. It's a drug that saved lives -- and would have saved many more if the vaccines had not beaten it to the punch.

The key thing to know is that nirmatrelvir/ritonavir's blockbuster effect was found in people who had neither been infected with SARS-CoV-2 nor vaccinated. But by the time it actually reached the open market, the vast majority of the high-risk U.S. population had already been infected or vaccinated. That immunity provides substantial long-term protection from severe COVID-19 when future infections occur, a fact which should not be minimized.

So, we've never really had the whole story on how well this drug works during its life in the real world. Observational studies (including , and countless other studies) have helped, but we've been waiting for far more valuable trial data. In the last few months, we've gotten some news, finally.

It's good news or bad news, depending on how you see things.

It's 2024: Where Do We Stand?

Well, we already found out this spring that nirmatrelvir/ritonavir is not as useful as it once was. Pfizer finally released results from its 2021-2022 trial showing that nirmatrelvir/ritonavir had no influence on symptoms among vaccinated patients with high risks or unvaccinated standard-risk patients. That study also found that hospitalizations were not statistically different across the nirmatrelvir/ritonavir and placebo groups but the study was not specifically designed to be definitive on that question.

But, we've been awaiting the results from a randomized controlled trial from the U.K. (the ), which are expected to give key results on hospitalization and mortality for thousands of study participants.

New Trial Data: Nirmatrelvir/Ritonavir Does Not Improve Mortality in Hospitalized Patients

In the meantime, we got some other data that is important. In late May, another U.K.-based study quietly from its randomized controlled trial of nirmatrelvir/ritonavir (conducted in 2022-2023 in the U.K., Indonesia, and Nepal). In this study, patients who were being hospitalized with COVID-related pneumonia were randomized to either receive nirmatrelvir/ritonavir or not. The patients' mortality rates were compared after 28 days. There was absolutely no difference.

This figure is pretty jarring. If the study had looked at 14-day outcomes, there would have been more deaths in the nirmatrelvir/ritonavir recipients. By day 28, things were equal. In any case, not a win for nirmatrelvir/ritonavir.

I ran some math known as power calculations -- albeit informally -- that I want to share with you. As it stands, nirmatrelvir/ritonavir was 0% effective in preventing death among patients hospitalized for COVID in the new RECOVERY study.

But let's imagine that we somehow knew for sure that the drug is effective, but that due to random chance, the study was "unlucky" and found that the death rate in the drug and control groups were equal.

In this table I made for us, the left column represents the "ground truth" hypothetical. That is, if nirmatrelvir/ritonavir were really 1%, 5%, 10%, 25%, 50%, or 75% effective in decreasing mortality, how likely would the 50-50 result (or worse) that was seen in the RECOVERY trial be (right column)? Here's the approximate readout:

In other words, if nirmatrelvir/ritonavir is actually 75% effective, the odds of a study like RECOVERY finding a 50-50 death rate (or worse) in the control and drug recipients just due to chance/bad luck is zero. If nirmatrelvir/ritonavir is actually 25% effective, the odds are just 18%. There's a reasonable chance (40-48%) that nirmatrelvir/ritonavir could be 5-10% effective against death, despite the 50-50 mortality findings in this study.

Folks, that is not good news for this drug. Remember: Pfizer's initial study found an 88% decrease in hospitalizations or deaths, and literally 100% of the deaths were among placebo recipients.

While the numbers in this newer study are on the smaller side, the researchers who did the study are highly respected and used good methods. For context, the parent study () has randomized over 49,000 COVID-19 patients since 2020 to see which drugs work and which ones don't. (It was this study, for example, that found that the inexpensive steroid dexamethasone had a massive benefit for COVID patients in 2020.)

In short, this study was small, but meaningful.

Is This the End for Paxlovid?

While this is a pretty hefty blow to nirmatrelvir/ritonavir, it is not the end of the story for this drug, for two reasons. First, the RECOVERY trial results looked at patients already being hospitalized. So, we do not know if nirmatrelvir/ritonavir might still decrease hospitalizations and death in patients who are earlier in their illness. That is what Pfizer's 2021 blockbuster study found. The question is whether it remains true in 2024. The eagerly awaited PANORAMIC study will likely answer this question, as it is much larger and designed that question.

We should brace for the possibility that the PANORAMIC trial will either show nirmatrelvir/ritonavir is now completely unable to decrease hospitalizations or deaths, or that at best, it does so modestly for a narrow group of very high-risk patients. Results are anticipated this year.

Remember when I said, "It's good news or bad news, depending on how you see things." What did I mean by that? Well, the bad news is that the drug doesn't seem to work as well as it once did. The good news is the reason for this: Collectively, we have far more immunity to severe consequences of COVID than we did back in 2020-2021.

So, here's the best way I can summarize things: It isn't that nirmatrelvir/ritonavir no longer works; rather, it's that in 2024, there are relatively few people who still seem to really need it to stay alive or out of the hospital.

This piece originally appeared in .