A few weeks ago, a colleague of mine got Covid-19. She was scheduled to board a cross-country flight the following week, with important meetings in two cities on her schedule. She really wanted to make it.
But the infection hit her hard, and on day two of her symptoms, she asked her doctor if she could get a prescription for Paxlovid. It’s currently the most effective antiviral medication on the market for treating and reducing the severity of outpatient Covid infections.
Her doctor said no — she wasn’t in the typical group for which the drug was recommended, she was told.
What followed, frankly, sucked: She was so sick, and her recovery took so long, that she ultimately delayed the trip by a week, causing significant disruption to her schedule and her life. (She’s fine now.)
Hearing her story, I wondered if withholding Paxlovid had really been the right call.
Yes, the drug is technically Food and Drug Administration (FDA)-approved for treating infection only in people at high risk for being hospitalized or dying as a result of Covid-19 infection, that is, those over 65 or with certain chronic medical conditions. And my colleague is not in that category. She’s in her 30s, with no risk factors for severe infection. And while the medication has interactions with lots of other drugs, which makes prescribing it a little tricky, my colleague doesn’t take any other medications.
But doctors prescribe medications off-label (that is, for indications other than the ones they’re approved for) all the time. I sometimes did so when I practiced as a physician if there was good evidence to suggest a drug’s benefit in a certain patient outweighed its risk.
I myself am low-risk for severe Covid-19 but recently took Paxlovid to shorten the length of my illness and keep some important plans — and I’m hearing about lots of peers who’ve either done or prescribed the same.
There’s reason to believe that Paxlovid could have big benefits for healthier people: A small body of literature suggests that the drug’s effects may make it a good strategy for shortening the course of illness, reducing transmissibility, and lowering the risk of long Covid.
Although there’s no shortage of Paxlovid in the US, the drug has been underused among the high-risk populations who would derive the greatest good from its use. But we also might be underusing it in the lower-risk people in whom the drug hasn’t been as rigorously studied — but who could have other benefits from its use.
Paxlovid is not a perfect drug — the usual five-day course isn’t enough to completely eradicate infection in some people who take it, occasionally leading to “rebound” Covid. And people taking certain medications or with certain kidney conditions might be unable to take the drug at all.
Still, if taking the medication enables low-risk Covid-infected people to feel lousy for fewer days, keep plans that really matter to them, and transmit fewer infections onward, shouldn’t prescribers be prescribing it more liberally? And do they really need to wait for a study to tell them to do so?
Understanding Paxlovid’s potential to benefit healthy people means rethinking what we want it to accomplish
A big reason why many doctors are hesitant to prescribe Paxlovid to lower-risk people has to do with how the drug was studied.
When experts try to understand whether any particular drug is effective, a key part of their question is to better define what “effective” really means to them. Are they trying to use the drug to make people feel a little better? Or to save lives?
It’s easier for scientists to determine a drug’s effectiveness if they define it very narrowly upfront. But on the flip side, defining effectiveness too narrowly makes it hard to tell if the drug has other benefits they’re not looking for.
When it comes to Paxlovid, the drug was initially studied to answer the question, “Can this drug save lives and reduce hospitalizations in the people at highest risk from Covid-19 infection?” We now know the answer is a resounding “yes.”
But Paxlovid’s initial studies can’t answer all of our questions about the drug’s ability to achieve more modest goals in younger, healthier people. Scientists weren’t looking for those specific outcomes when they designed early Paxlovid studies — often called endpoints — so those studies can’t definitively tell them whether the drug can produce those outcomes.
Still, it’s reasonable to wonder how the drug could help other people and whether it’s worth prescribing more broadly. And there is some data that speaks to this.
We know what Paxlovid does well in the highest-risk people — but less about what it does well in those at lower risk
Let’s first talk about what we know with some certainty: In people at highest risk for a bad outcome related to Covid-19 infection, Paxlovid saves lives.
Historically, having physical risk factors (including older age, chronic diseases, and immunosuppression) and being unvaccinated have put people at higher risk for the worst Covid-19 outcomes.
It’s on the basis of this study that the FDA gave Paxlovid full approval earlier this year. (The agency first approved Paxlovid under an emergency use authorization in December 2021 based on preliminary data.)
Less clear are Paxlovid’s benefits in lower-risk people.
“There are four things” we should think about as potential endpoints for Paxlovid treatment in young, healthy, vaccinated people, said Peter Chin-Hong, an infectious disease doctor at the University of California San Francisco. Those include:
- Shortening the duration of illness
- Reducing onward transmission
- Reducing long Covid risk
- Preventing infection altogether through pre-exposure prophylaxis
Most of those endpoints haven’t been as rigorously studied in lower-risk people. Nevertheless, “there is enough data there to show me that it’s not harmful, and it’s potentially beneficial to some individuals,” Chin-Hong said.
Preliminary data from another trial (EPIC-SR; the trial was ended early and results have not yet been published in a peer-reviewed journal) suggests the drug was less useful in people whose risk fell somewhere in between the highest and the lowest for severe Covid.
Let’s call these people “medium-risk” — it’s a group that includes people who are unvaccinated, or who’ve been vaccinated but have other risk factors.
In this trial, the drug wasn’t particularly powerful at lowering the number of days people were sick: The group who got Paxlovid had symptoms for around 12 days (a median) — only one day fewer than the non-Paxlovid group’s 13 days of illness. The drug also more meagerly reduced the risk of hospitalization or death, from 1.6 percent in the non-Paxlovid group to 0.8 percent in the Paxlovid group.
Another study (EPIC-PEP) tried to examine Chin-Hong’s last endpoint: Paxlovid’s ability to prevent infection if a person takes it after exposure but before symptoms start. In medium-risk people, it didn’t achieve that result.
In all of these studies, the drug was broadly safe. It had minimal side effects — the most common one, a nasty metallic taste in the mouth, was tolerable for most people.
Looking at this data, you could conclude that Paxlovid has clear benefits in high-risk people, and lower-risk people have less to gain from taking it.
But less isn’t nothing! Even if the drug shortens the duration of symptoms one day, “that one day is very meaningful to a lot of people,” Chin-Hong said.
With other infections, that’s reason enough to justify antiviral treatment. For example, the flu drug oseltamivir can reduce symptoms by about a day in healthy people. And that small gain is generally thought to be enough to justify the prescription.
“This concept does have legs,” Chin-Hong said, referring to the prospect of using Paxlovid to shorten disease course in young, healthy people. It just doesn’t have a broad evidence base.
Data from several small studies is mixed, but suggest Paxlovid could do some good in lower-risk people
So there’s some data on the “medium risk” people. Is there any data on the lowest risk?
Not directly: There’s no large, high-quality clinical trial out there to answer the question, “What are the benefits of Paxlovid treatment in young, healthy, vaccinated people?”
But a few studies can give us some clues.
One small study enrolled 36 vaccinated people aged 5 to 66 with the goal of understanding rebound Covid-19 infections. About one-third of the people in the study got Paxlovid treatment; some were healthy at baseline, while others weren’t.
The study wasn’t designed to understand how long each group took to test negative after infection. Still, the data it gathered showed people who took Paxlovid took only four days to test negative, compared to seven days for people who didn’t. That suggests Paxlovid reduced the number of days they were infectious.
Other studies of higher-risk people also suggest that in the noses of people treated with Paxlovid for certain omicron infections, viral populations waned several days sooner than they did in people who weren’t treated with Paxlovid.
Biologically, it makes sense that these findings would translate to benefit among low-risk people on two important endpoints, said Chin-Hong — specifically, shorter disease courses and lower likelihood of onward transmission.
There is also data suggesting Paxlovid could reduce the risk of long Covid after infection. Several studies are currently underway to evaluate the drug for preventing long Covid symptoms across a range of risk groups.
Bottom line: Health care workers should be prescribing more Paxlovid
Even though it’s biologically plausible that Paxlovid could help low-risk people in a variety of ways, it’s unclear how willing health care providers are to prescribe the drug to these populations. (As it is, not enough providers are on board with prescribing Paxlovid to the highest-risk people who should clearly be receiving it.)
Prescribers are often loath to prescribe unfamiliar medications without clear guidelines and clear supportive evidence — especially medications like Paxlovid, which can potentially interact with a lot of other medications. Suggesting this kind of off-label use “might be heresy with some of my [infectious disease] colleagues,” said Chin-Hong.
When data is weak or mixed, doctors may in good faith withhold certain drugs from certain patients. The point here is that in the same situation, they can also in good faith do the opposite and prescribe it.
At least one expert has expressed good reason to think both ways about the issue: Paul Sax, clinical director of infectious diseases at the Brigham and Women’s Hospital in Boston and author of a popular blog hosted by the New England Journal of Medicine. Last year, he wrote that given sufficient supply, “I’d certainly recommend [Paxlovid] even for low-risk symptomatic people.” He noted that among a young, healthy subgroup in the EPIC-SR trial, there were 70 percent fewer hospitalizations in those who took Paxlovid.
But this June, Sax co-authored a retrospective study suggesting Paxlovid neither reduces hospitalizations nor symptoms for young, healthy, vaccinated people. In an email to Vox, he said these days, he’s not recommending the drug for mild cases.
There are a lot of takes on who the right patients are to get Paxlovid, and where providers practice may have as much to do with their prescribing habits as anything else. For example, last December, California’s state public health officer recommended providers consider all Covid-positive patients for Paxlovid treatment, whatever their baseline risk.
Although a Paxlovid study on younger, healthier people is unlikely, they can still get the medication (and in any case, other options are on the horizon)
There isn’t a high-quality study on the horizon aimed at measuring how well Paxlovid achieves low-stakes endpoints in lower-risk populations, said Chin-Hong. Pfizer, the drug’s manufacturer, is unlikely to ever do that trial, he said; “the juice isn’t worth the squeeze. And it’s FDA-approved already, so people can use it off-label.”
So for the foreseeable future, doctors and providers will have to make these decisions for themselves.
If he’d been seeing my colleague, Chin-Hong would have prescribed her Paxlovid, he said: It would have given her a chance to feel better sooner, and reduced her risk of transmission — both good outcomes, and at very low risk to her.
He noted this conversation may soon be moot: The oral antiviral drug ensitrelvir, already approved in Japan, shortens illness regardless of a person’s risk status — and lacks the drug interactions that complicate Paxlovid prescriptions. It’s been fast-tracked for FDA approval.
But in the meantime, he says, prescribers should feel liberated to prescribe Paxlovid off-label even to young, healthy, vaccinated patients when there’s no risk of drug interaction, he said. “The absence of evidence doesn’t mean that it doesn’t have benefit.”