Here is what Shionogi’s SCORPIO-PEP trial actually demonstrated, in the population of 2,041 household contacts of a sick patient that the FDA used to approve ensitrelvir on Monday: 2.9% of the drug arm and 9.0% of the placebo arm developed mild symptomatic COVID by day 10. Zero people in either arm were hospitalized. Zero people in either arm died. The drug works in the sense that fewer people felt crummy. It does not, on this trial, work in the sense most people still mean when they hear “COVID prevention.”

That is the gap the headlines collapse into the 67% relative risk reduction Shionogi’s press release leads with. Both numbers come from the same column of the same table. They are not the same claim. A six-point absolute reduction in mild illness compresses to a sharper-looking 0.33 relative risk with a tight confidence interval of 0.22 to 0.49. The trial hit its primary endpoint cleanly. It also enrolled the country we now live in: more than 98% of participants had antibodies against SARS-CoV-2 at baseline, from prior infection, vaccination, or both.

So what did the FDA license a market for?

Ensitrelvir, sold as Xocova, is the second oral SARS-CoV-2 3CL protease inhibitor to reach Americans. Pfizer’s Paxlovid was the first. The mechanism is the same: block the enzyme the virus uses to chop its own proteins into working parts, and the assembly line jams. Shionogi’s chemistry is sound, and the Japanese approval timeline (emergency in November 2022, full treatment approval in March 2024, post-exposure indication this past March) gave the molecule a five-year head start before the FDA’s June 16 deadline. The agency cleared it early. The advisory committee was told what every reader of the NEJM paper can see: this is a five-day course (three tablets on day one, one tablet on each of days two through five) tested in a 2,041-person primary analysis where the severe-outcome event count was zero on both arms.

Which makes the next question the one Shionogi did not put on the slide. What does this drug buy?

In symptomatic illness it buys a 6.1-point absolute reduction, from nine cases per hundred exposed contacts to three. That is real and it is small. In hospitalizations it buys nothing the trial could measure, because there was nothing to measure. SCORPIO-PEP did not demonstrate a severe-outcome benefit. It could not. The arms were too healthy and the population too immune for the design to produce one. That is a fact about the trial, not a guarantee about every future patient. It is, however, the trial the FDA used.

The subgroup result Shionogi did not lead with is worth naming carefully. Among participants already SARS-CoV-2 positive at baseline (the group you might expect to benefit most from an early antiviral) the relative risk was 0.75, with a confidence interval of 0.46 to 1.23. That interval crosses 1.0. In plain English: the trial did not establish a symptomatic benefit in baseline-positive contacts. It does not prove the drug fails in them. It proves SCORPIO-PEP was not powered to answer the question, and the agency approved on the answer it did get.

Credit where it is owed on safety. Reported adverse-event rates were essentially identical, 15.1% on drug versus 15.5% on placebo. There were no reports of the metallic-mouth dysgeusia that became Paxlovid’s signature complaint. Without ritonavir boosting, Xocova carries fewer of the CYP3A4 drug-drug landmines that have made Paxlovid a quiet nightmare for older patients on statins and blood-pressure regimens, though the label still requires interaction screening, and contraindicates simvastatin and colchicine among others. NEJM also reported a reversible HDL reduction with ensitrelvir. The molecule is well-behaved by the standards of the class. Then the question reasserts itself: well-behaved in service of what clinical outcome?

The harder question is who pays. A course of Paxlovid currently runs around $1,470 at U.S. list. Shionogi has not announced U.S. pricing for Xocova. A molecule in the same class, launched into a less acute clinical need, has no obvious reason to come in much cheaper. Insurance will absorb most of the cost, which means premiums will, which means everyone will. That is the actual mechanism by which a 6.1-point absolute reduction in mild illness, in a 98%-immune population, becomes a national expenditure.

None of this is the agency’s failure. The FDA evaluated a clean Phase 3 with a clear primary endpoint and a tight effect estimate, and licensed it for the indication the data supported. The harder-to-fault judgment is the agency’s choice of priorities. An oral pill that drops a household contact’s odds of feeling sick from nine in a hundred to three in a hundred is a category the agency decided to populate. A pill that prevents hospitalization in the patients still at meaningful risk of one, mostly the severely immunocompromised SCORPIO-PEP did not enroll in numbers that would have answered the question, is a category nobody yet has a Phase 3 for.

The trial Shionogi should publish next is the one in patients at high baseline hospitalization risk, powered for severe outcomes. Until then, the next data point worth watching is uptake. If primary-care physicians write Xocova at scale for household contacts who would have stayed home with chicken soup and resolved fine, the agency’s choice will have nudged American medicine one more step toward treating ordinary viral exposure as a billable event. If the prescriptions cluster in the immunocompromised, where this molecule might still earn its keep, the calculus changes. Monday’s approval did not answer that question. It built the market in which the answer will be revealed.

Sources

  1. NEJM – SCORPIO-PEP: Ensitrelvir for Prevention of COVID-19 (2026)
  2. BusinessWire – Shionogi Announces FDA Approval of XOCOVA (ensitrelvir) (2026)
  3. Shionogi – FDA Approval Announcement for XOCOVA (corrections) (2026)
  4. Healio – FDA approves Xocova for COVID-19 prevention after exposure (2026)
  5. Shionogi – FDA accepts NDA for ensitrelvir as post-exposure prophylaxis (2025)