The body of evidence for Paxlovid therapy


Dear Colleagues,

We have a mismatch. The evidence supporting treatment for Paxlovid is compelling for people aged 60 or over, but the older patients in the United States are much less likely to be treated. Not only was there a randomized, placebo-controlled trial of high-risk patients which showed 89% reduction of hospitalizations and deaths (median age, 45), but there have been multiple real-world effectiveness studies subsequently published that have partitioned the benefit for age 65 or older, such as the ones from Israel and Hong Kong (age 60+). Overall, the real-world effectiveness in the first month after treatment is at least as good, if not better, than in the high-risk randomized trial.

But it’s more likely in the United States for a person age 45-50 to get Paxlovid over people age 80 or older. Why? We’re doing the current survey to find out, but the most likely reasons include (1) lack of confidence of benefit; (2) medication interactions; and (3) concerns over rebound.

Let me address each of these briefly. The lack of confidence in benefit stems from the fact that the initial high-risk trial was in unvaccinated individuals. That concern can now be put aside because all of the several real-world studies confirming the protective benefit against hospitalizations and deaths are in people who have been vaccinated, and a significant proportion received booster shots.

The potential medication interactions due to the ritonavir component of the Paxlovid drug combination, attributable to its cytochrome P450 3A4 inhibition, have been unduly emphasized. There are many drug-interaction checkers for Paxlovid, but this one from the University of Liverpool is user friendly, color- and icon-coded, and shows that the vast majority of interactions can be sidestepped by discontinuing the medication of concern for the length of the Paxlovid treatment, 5 days. The simple chart is provided in my recent substack newsletter.

As far as rebound, this problem has unfortunately been exaggerated because of lack of prospective systematic studies and appreciation that a positive test of clinical symptom rebound can occur without Paxlovid. There are soon to be multiple reports that the incidence of Paxlovid rebound is fairly low, in the range of 10%. That concern should not be a reason to withhold treatment.

Now the plot thickens. A new preprint report from the Veterans Health Administration, the largest health care system in the United States, looks at 90-day outcomes of about 9,000 Paxlovid-treated patients and approximately 47,000 controls. Not only was there a 26% reduction in long COVID, but of the breakdown of 12 organs/systems and symptoms, 10 of 12 were significantly reduced with Paxlovid, including pulmonary embolism, deep vein thrombosis, and neurocognitive impairment. There was also a 48% reduction in death and a 30% reduction in hospitalizations after the first 30 days. I have reviewed all of these data and put them in context in a recent newsletter. A key point is that the magnitude of benefit was unaffected by vaccination or booster status, or prior COVID infections, or unvaccinated status. Also, it was the same for men and women, as well as for age > 70 and age < 60. These findings all emphasize a new reason to be using Paxlovid therapy, and if replicated, Paxlovid may even be indicated for younger patients (who are at low risk for hospitalizations and deaths but at increased risk for long COVID).

In summary, for older patients, we should be thinking of why we should be using Paxlovid rather than the reason not to treat. We’ll be interested in the survey results to understand the mismatch better, and we look forward to your ideas and feedback to make better use of this treatment for the people who need it the most.

Sincerely yours, Eric J. Topol, MD

Dr. Topol reports no conflicts of interest with Pfizer; he receives no honoraria or speaker fees, does not serve in an advisory role, and has no financial association with the company.

A version of this article first appeared on

Next Article: