Paxlovid is not Pez

and you shouldn't be able to just walk up to the pharmacy and get it

The recent move by the FDA to approve dispensation of the antiviral medication Paxlovid by state-licensed pharmacists is a mistake on a number of levels. It will increase the risk of complications and antiviral resistance and it is, to put it charitably, based much more in politics than in evidence.

First a brief review: Paxlovid is a combination medication, including the antiviral drug nirmatrelvir and the drug, ritonavir, which is included to reduce the speed at which enzymes in the liver remove nirmatrelvir from the body. This creates a higher effective drug level of nirmatrelvir, but it has the side effect of increasing the drug levels of a number of other drugs. More about that later.

Paxlovid was approved by the FDA under an Emergency Use Authorization based on impressive results in terms of reduced risk of hospitalization or death in an unvaccinated population who had risk factors for progression to severe COVID. The list of risk factors is pretty exhaustive, including obesity and hypertension, and well over half of US adults have one or more. 

The population that showed that large benefit: at risk, unvaccinated patients, is dramatically different from the population that the FDA has authorized for treatment, with the main difference being that the requirement that patients be unvaccinated was lifted without evidence showing that the benefit extended to vaccinated or prior infected patients.

Now, under political pressure that kicked off with President Biden’s January 2022 State of the Union address, the FDA has authorized pharmacists to prescribe Paxlovid without a physician's involvement. This approval comes with no new data, but six weeks after the Biden Administration pushed for expedited access. Pfizer, the juggernaut pharma that has already received over $100 billion in payments from the government, has recently applied for a broad new drug approval (NDA) for use in all patients over the age of 12 with COVID risk factors, regardless of vaccination status. 

There are a number of issues that make this problematic:

Unproven benefit

The most recent NDA is an attempt to conflate a successful study with a failed one. That’s right. Failed. Pfizer has tried to spin this, but the EPIC-SR (Evaluation of Protease Inhibition for COVID-19 in Standard-Risk Patients) study that looked at lower-risk patients, including those who had been vaccinated, did not show a benefit in its primary outcome measure. For reasons that are unclear, the FDA generously allowed Pfizer to conduct research looking at the “novel endpoint” (Note: this means it is a made-up endpoint created by Pfizer because they thought it would be easy to meet) of “Time to Sustained Alleviation of All Targeted COVID-19 Signs/Symptoms.” This is a much lower goal than preventing severe illness, and it was chosen because, in a standard-risk population, severe illness from COVID is rare. In short, the FDA allowed Pfizer to pick its own goal line, and the trial still fell short. 

Pfizer tries to salvage this with a secondary endpoint analysis showing “a non-significant 57% relative risk reduction in hospitalization or death.” Do you know what the FDA used to call a “non-significant 57% relative risk reduction?” No benefit. The whole point of having the significance assessment, especially when you are allowing lots of secondary endpoints, is to reduce the risk you are fooled by chance. A non-significant benefit has a higher risk of being a statistical fluke, so you don’t get to claim it as a “real” benefit. 

(Pfizer's response, by the way, was telling. After the interim assessment showed no significant benefit, they closed enrollment. That’s right, apparently getting more accurate info was more likely to harm their chances. They quit looking.)

The final indignity Pfizer is asking the FDA to accept is an “integrated analysis” of their high-risk and low-risk studies. This is, I am sorry, utter bullshit. You don’t get to run a 2236-person study of high-risk patients and an early-stopped study of 1153 low-risk patients, combine them, and declare that on average everyone benefits. A study of 2000 lung cancer patients combined with 1000 healthy Olympic athletes might show on average there was an overall benefit to surgically removing a lung from everyone. That would only be a little bit less legitimate than what Pfizer is suggesting.

Increased risk

I noted that Paxlovid is actually a two-drug combo. The first, nirmatrelvir, is a “protease inhibitor” that helps stop the virus from growing, and the second, ritonavir was added to cause side effects with other medication.

OK, not really, ritonavir is also a protease inhibitor, but it was added because it blocks CYP3A, an enzyme produced by the liver. Reducing the activity of CYP3A increases the time that nirmatrelvir stays in the system, increasing its effect. The problem is that CYP3A4 is also crucial in the metabolism of a laundry list of other drugs. Many of these drugs are taken to treat exactly the sorts of illnesses that lands a person in the high-risk category in the first place. 

Saying Paxlovid is not appropriate for on-demand access at the pharmacy is not a shot at pharmacists. As an oncologist, I’d hate to have to admit the number of times a pharmacist pointed out a potential medication interaction to me. They are critical members of a team for medication dispensation, and they are experts in their domain. 

But they are also human, and just one, frequently overburdened, link. When you look at the language the FDA used to ensure drug interactions with Paxlovid do not occur, it does not fill me with confidence in its checks and balances:

“Sufficient information is available, such as through access to health records, patient reporting of medical history, or consultation with a health care provider in an established provider-patient relationship with the individual patient, to obtain a comprehensive list of medications (prescribed and non-prescribed) that the patient is taking to assess for potential drug interaction.” (Emphasis added)

If I was making a list of drugs for which I would not want the last line of defense be whether a patient is certain about their medications, it would start with meds like Paxlovid. I mentioned that the pharmacist was part of a team. Well, the prescriber (doctor, PA, APRN) writing the order is also on that team. They can provide perspective and information about the patient’s case that the patient themselves might not disclose. The FDA is putting the pharmacist in a bind by increasing expectations while removing one of the important safeguards on a potentially dangerous drug.  The American Medical Association has objected based on these safety issues. 

Overuse

I recall a day, as recent as 2019, when we were collectively concerned about increasing resistance to antibiotics. It never got as much attention, but the CDC was also concerned with preserving the effectiveness of its antiviral medications against influenza. Resistance is an issue because the more an antiviral is used, the greater the chance that any surviving virus has some higher-than-average ability to tolerate the medication. As that virus reproduces, it passes the resistance on to its progeny. 

It is particularly concerning that we are already seeing reports of COVID infections that “rebound” soon after stopping Paxlovid. Currently, there is no recommendation to take a second course of Paxlovid in the event of rebound, the actions of director of the National Institute of Allergy and Infectious Diseases Anthony Fauci notwithstanding. It is too early to say if this will be the start of widespread antiviral resistance, but it should be considered a huge warning sign. 

By increasing the impression that Paxlovid treatment should be the norm, rather than the exception, we increase usage in situations where the risk of a bad outcome is already very low. Higher prescription rates add pressure on the virus to evolve in ways that evade the antiviral effect. One way to look at it is that for any antibacterial or antiviral agent, you only get so many uses as a society before the treatment starts to become ineffective. Advising when and on whom to use those doses so they are most likely to be effective is supposed to be the role of the FDA. Encouraging its use indiscriminately is not good stewardship of this medication.

In summary, we have seen a political push for greater use of a medication with intrinsic risk but little proven benefit as it is currently used, in a situation when speed of access risks compromised safety. My recommendation is that we reexamine this approach, reserve Paxlovid for those who are truly at high risk, and demand that Pfizer conduct clinical trials demonstrating safety and efficacy in the populations it is petitioning the FDA for approval to treat.