番茄社区

A in Annals of Internal Medicine suggests that one in five COVID patients who take nirmatrelvir/ritonavir (Paxlovid) may experience viral rebound, a far higher proportion than seen in prior .

In this video, study author Mark Siedner, MD, MPH, of Harvard Medical School in Boston, discusses the research and how it may inform best use of the antiviral.

The following is a transcript of his remarks:

We were encouraged to do this study because of two essentially conflicting pieces of information. When Pfizer tested Paxlovid in a clinical trial and demonstrated that it was effective at preventing hospitalization and death, they also retrospectively went back once reports of Paxlovid rebound became available and tried to estimate how common rebound was.

What they found was that it was rare; it only happened in about 2% of people in their study. And that was similar between people who were taking Paxlovid and people in the control arm.

However, as you and probably most people are aware, reports of Paxlovid rebound gained quite a bit of attention in the press and social media among us clinicians. And it seemed, at least anecdotally, that it was much more common than 2% of people taking the medicine.

And so we had a cohort set up that was essentially ideally prepared to answer a question like this, so we thought we were well positioned to do so. We tried to answer the question: how common is rebound amongst people with COVID-19 taking Paxlovid? Is it really closer to 2% or a much higher number based on what we were hearing in the press?

We began by using our health record system at Mass General Brigham, which is a large coordinated care center in Boston with a very large catchment area and patient population, and we set up automated reports to determine when people were infected with COVID-19. We went through those reports on a daily basis and notified people of our study at the time of infection -- so soon after they had been diagnosed with COVID-19. For those who consented, we enrolled them in a study and we actually went to their homes and dropped off nasal swab kits at their homes so they could actually self-collect their nasal swabs, and then we went to pick them up.

So we, multiple times a week -- about three times per week per person -- were able to collect nasal swabs from patients with acute COVID-19, bring them into our laboratories, and they were tested there thoroughly for both PCR to see the level of virus in the nose, but also in the BSL [biosafety level]-3 laboratory at the Reagan Institute we also were able to test the viral culture to see if the virus was able to replicate -- essentially, if it was still potentially contagious.

So unlike the prior studies, which I think retrospectively went back to try to find rebound, we sampled people very frequently to be able to tell when the virus came down, if it stayed down, or if it came back up and they experienced virologic rebound.

Our primary finding was that COVID-19 rebound was actually quite rare in people that weren't taking Paxlovid. It only happened in one person in the over 50 people in that group, so around 2% of the time. However, unlike in the Paxlovid clinical trial, we found that rebound occurred in over 20% of people who were taking Paxlovid. That really showed a clear difference between our study and the prior data and did, for us, really reinforce the fact that it seems like the use of Paxlovid is associated with virologic rebound.

I think it's important when people go back and look at our study and compare it to the Paxlovid clinical trial to understand the difference between our two studies. And I think people will be asking, why is your study so different? Why did you find 20% and they found 2%?

I think that the answer to that is actually relatively straightforward in that their study was not designed to this question. They were looking at what we call virologic decay -- how fast does the virus go down once you start taking therapy at day 5? At day 10? At day 14? They looked at three time points. We looked at anywhere from six to 10 time points with multiple sampling of the nose throughout that period of time. When we went back to our data set and we removed all of the data except for days 5, 10, and 14, just like they did in their study, we also found 2% of rebound and 80% of the rebound cases were missed.

So I think that their study wasn't designed to detect rebound, which is probably why they didn't find it. Our study, which was prospective with multiple days of sampling, really was designed to detect it. I believe that that's the reason we see a difference between our results and theirs.

There are only 16 rebounders in our study, so our study is relatively small. But when we took those people who took Paxlovid and looked to see what were the predictors of whether they rebounded or not, one trend we found was that people who waited a day or two between the time that they were diagnosed or when their symptoms started until they started Paxlovid, they tended to have lower chances of having rebound. Such that once people waited 2 days, we found no people who rebounded on Paxlovid -- if they had waited 2 days from the time of diagnosis or symptom onset until the time they started taking their medicine.

That is preliminary data, which I think suggests that there may be an effect here of the amount of virus that people are shedding when they start taking the treatment that may predict whether or not they rebound.

As an infectious disease clinician who still treats COVID not infrequently, I think it's really important to know that this drug has been shown to be effective in clinical trials for high-risk people. It had an 80% reduced risk of hospitalization or death. So we still use it. It's probably the best medicine we have for COVID, and it should still be used in high-risk people.

I think our data helps provide both patients and physicians information that they need to know, which is that rebound is probably more common with Paxlovid and it may occur in about 20% of people. We don't in any way think our data should contradict the clinical trial, which is a large, extremely well-designed study, I think, with very clear evidence. But we do think people should be aware of the risk of rebound and we think our data [could] potentially help lead to new studies.

For example, should we study whether there should be a gap [between] the time of diagnosis and the time of Paxlovid? I don't think our study proves that there should be. I think more information is needed to know if waiting to take Paxlovid affects whether it still prevents hospitalization and death to recommend that.

Another thing that we did demonstrate in our study was that when we took the people who took Paxlovid and we looked 5 days after they finished Paxlovid -- so 10 days after they were infected for the most part -- what we found is that all the people who rebounded at day 10 actually had a positive viral culture, but none of the people who didn't [take Paxlovid] had a positive viral culture at day 10.

That suggests to me that maybe we could advise people who are taking Paxlovid to test themselves 5 days after they're done, around the time they're leaving isolation. And if they're positive, they should probably inform their healthcare provider and potentially extend the duration of their isolation, because that means they may be rebounding and they may potentially be contagious.

Comment