An ongoing dialogue on HIV/AIDS, infectious diseases,
April 25th, 2022
Yes, Relapses After Paxlovid Happen — Now What?
Around two weeks ago, one of my long-term, very stable patients with HIV called me saying she’d just been diagnosed with COVID-19. Over 60 with hypertension, and overweight, she qualified for nirmatrelvir/r (Paxlovid) under the Emergency Use Authorization (EUA), and took it without problem.
(Certain details changed for confidentiality.)
In fact, she started to improve within 24 hours of the first dose. Complained a bit about the metallic taste, but was thrilled at this rapid recovery.
She then contacted me again a week later, saying she’d relapsed. More nasal congestion, cough, and fatigue — not as bad as when the illness started, but unmistakably a relapse. Not only that, the home antigen test, which started out as a dark line at the outset, became a barely discernible line after the treatment, and now was clearly positive again.
Her biggest concern was getting back out in the world without infecting someone. She really wasn’t that sick; she just wanted advice about when she could return to work and start socializing again.
“Avoid close contact with others until that test clears,” I said. (Which it did a few days later, and she completely recovered.) But certainly the whole thing was a head-scratcher for both of us.
Turns out these relapses do occur; this was just the first time I saw it. The investigators from Pfizer observed them in their clinical trial EPIC-HR, and reported it to the FDA. Take a look at this figure from their study, which was not included in the NEJM paper, but did appear in an FDA report:
See those lines heading upwards from day 10-14? Those are relapses. Sequencing of SARS-CoV-2 from these cases did not demonstrate resistance mutations either at baseline or at relapse that correlated with resistance. Specifically:
In summary, currently there are no clear signals of baseline or treatment-emergent NIR resistance from the preliminary analyses of clinical trial EPIC-HR. These analyses will continue to be conducted as more complete data from EPIC-HR are obtained and reported.
A detailed analysis of a similar relapsing case is available as a pre-print — viral sequence the same at 3 time points (no resistance or reinfection), respiratory multiplex PCR negative (no new pathogen), good antibody response.
Not surprisingly, as use of Paxlovid increases along with the supply, the anecdotal reports of these relapses increase in parallel. Not just in clinical practice — they’ve popped up on social media and in the press. It’s now become a standard part of my counseling to people, that this return of symptoms and test positivity might happen.
So we know that relapses happen. What don’t we know about these cases? Quite a bit, actually.
- How often does it occur? It’s tricky from the figure to get a clear picture of the numerator and denominator. And of course, that was in an unvaccinated, high-risk population. What about those who are vaccinated?
- What are the risk factors? Could it be that those with baseline high viral loads/low cycle thresholds are at greater risk of relapse? People who are severely immunocompromised? Older? With some variants more than others?
- Does antiviral treatment blunt a helpful immune response? Does the immune system need to see a certain concentration of viral antigens to provide adequate clearance? Or are those who are relapsing just the subset of people who would have had prolonged viral shedding to begin with?
- Should we assume that people who relapse become contagious again? That’s my assumption — and it seems highly likely — but it’s worth proving in a research lab that these viruses are just as replication-competent as the pre-treatment viruses.
- Does the virus develop resistance during this 5-day course? So far this hasn’t been reported in a relapsed case, at least a far as I know. Seems inevitable at some point, however, so worth looking for, again in the context of a research study.
- Should treatment courses be longer? Maybe a longer course is better, but maybe not. How about a 5- vs. 10-day blinded clinical trial, with clinical, virologic, and immunologic endpoints? (Importantly, both 5- and 10-day courses follow the rules.)
- Does this happen with other antiviral strategies? I don’t recall hearing similar patterns with remdesivir or molnupiravir, but then again viral load reductions were less robust. This could be related to the mechanism of action of viral protease inhibitors.
- Should highly vulnerable people with relapses be treated again? Somehow “do nothing” for the most at-risk people with COVID-19 (for example, those on rituximab) doesn’t seem right when we have treatment tools at our disposal. One of my colleagues pointed out to me that neither repeated courses of nirmatrelvir (something I suggested in the above-linked Boston Globe piece) or treatment with molnupiravir would be allowed under the EUAs, as they must be within 5 days of symptom onset. It’s likely also that the bebtelovimab window of treatment (7 days) would be exceeded. That leaves 3 days of IV remdesivir as the only outpatient treatment option, which is very hard to access.
Given all these unknowns, it would be enormously helpful for Pfizer to release further data on their relapsing cases. Not just how often they happened, but also how they did clinically — presumably they did well, given the overall favorable results from the study. Any further information on immune responses? And did they paradoxically have a longer time to viral clearance than the placebo arm?
Fortunately, I’m aware of several research groups who are studying these cases right now here in Boston, and undoubtedly others are doing so elsewhere. We should know more soon.
But what we do know now — and I’ll keep saying this again and again — is that this is one tricky virus, full of surprises.