Expert advisers to the Food and Drug Administration recommended authorization Tuesday of the first coronavirus pill to prevent high-risk people from developing severe illness in a divided vote that reflects the complicated mix of benefits and risks involved with a new and easy mode of treatment.

Merck and Ridgeback Biotherapeutics developed the drug, molnupiravir, as a five-day regimen to be taken at home within five days of onset of coronavirus symptoms. The FDA is not bound by the 13-to-10 vote but typically follows its external advisers’ recommendations. The drug could have an immediate impact on the pandemic if authorized — just as the ominous new omicron variant has emerged, jolting the world with the prospect of a longer and more complicated pandemic.

Merck has pledged to manufacture treatment courses for 10 million patients by the end of the year, and the United States has preordered enough medicine to treat 3.1 million people. Monoclonal antibodies, another treatment, are harder to administer but are more effective, which led experts to debate the benefits and drawbacks of an easy-to-use medicine that would offer an additional tool to reduce the strain on hospitals as a potential winter surge looms.

“I voted yes. This was clearly a very difficult decision,” said Michael Green, a professor of pediatrics at the University of Pittsburgh School of Medicine. “Should an alternative oral agent become available that had a better safety profile and equal to or better efficacy profile, the agency might reconsider its authorization.”

When preliminary data on the drug was released in early October, molnupiravir appeared to cut risk of death and hospitalization in half, a finding so persuasive that the trial was halted at its midpoint. After continued follow-up of patients, the drug appeared less impressive, with risk of death and hospitalization reduced by 30 percent — an unexplained erosion that was another factor in some committee members’ hesitancy to endorse the treatment.

“I just want to emphasize that I think it’s a pretty minimal benefit,” said Sally Hunsberger, a statistician at the Biometrics Research Branch of the National Institute of Allergy and Infectious Diseases, who voted no. “I don’t really think we know what groups this is benefiting, if it’s benefiting.”

With the efficacy of the drug far from a slam dunk, the safety profile of the pill became a central point of debate. Molnupiravir works by scrambling the genome of the coronavirus, resulting in catastrophic errors that disable it.

Some committee members said they feared that the drug’s mechanism could result in a more fearsome version of the virus.

“I voted no. It was an easy vote for me to vote no. … There were more questions than answers. I think the potential for this drug to drive some very challenging variants into the public is a major, major concern,” said James E.K. Hildreth, president of Meharry Medical College.

Molnupiravir has not yet been tested against the omicron variant, but because it works in a different way than vaccines and monoclonal antibodies that train their firepower on the coronavirus spike protein, it is expected to hold up against a wide range of variants.

Merck scientists have been testing the drug against every new variant, and have begun working on omicron, said Nicholas Kartsonis, senior vice president of clinical research for infectious diseases and vaccines at Merck. “We expect, based on what we know about the omicron variant, that molnupiravir would be effective against this particular variant,” he said.

But one committee member cautioned against assuming that the drug would be more variant-proof than vaccines or other treatments.

“If we take a look at the omicron variant and see the number of mutations that virus has, I think in many ways we don’t really understand which direction the virus may even be going in terms of changing,” said W. David Hardy, a scientific and medical consultant at Charles R. Drew University of Medicine and Science, who voted yes. “To assume this drug … is going to work when the monoclonals don’t is a big jump. We have no assurance of that.”

Even people who voted in favor of the drug pointed out that it would not be first in line if other options existed.

“If there is an alternative therapy that’s efficacious, like monoclonal antibodies currently or a future medication, that would be the priority,” said Uma M. Reddy, an obstetrician and maternal fetal medicine specialist at Yale School of Medicine, who voted yes.

A second oral pill regimen — it was developed by Pfizer — is also being reviewed by the FDA. That drug could offer advantages that will make it the choice of physicians: It was more effective at preventing hospitalization and death, and doesn’t carry safety concerns about its potential to cause mutations. In contrast, those fears led many committee members to say they would not prescribe the Merck drug to pregnant patients.

“I don’t think I would want to take this drug, not knowing the effect it could have on my unborn child,” said Roblena E. Walker, a consumer representative on the panel who is chief executive of public-health-monitoring organization EMAGAHA, who voted yes.

Much of the expert committee’s debate Tuesday focused on the many unknowns about safety, which are intertwined with how molnupiravir works.

There are two concerns about potential unintended consequences: first, that molnupiravir could cause errors in the genome in the patient’s own cells, causing cancer, genetic diseases or birth defects. The company requested authorization for adults, but in documents filed with the FDA, said the pill should not be recommended for patients who are pregnant or breastfeeding, and said people of childbearing age should use contraception during the treatment.

While experts said they wouldn’t recommend the drug to pregnant patients, there was a debate about whether they should be ineligible for the drug altogether.

“I don’t think you can ethically say it’s okay to give this drug in pregnancy, obviously. But at the same time, I’m not sure you can ethically tell a pregnant woman who has covid-19 that she can’t have the drug if she’s decided that’s what she needs,” said Janet D. Cragan, a medical officer in the Division of Birth Defects and Infant Disorders at the Centers for Disease Control and Prevention.

The short duration of treatment and a gold standard laboratory test in rats designed to flag drugs that have the potential for mutagenesis — the process in which genetic change happens — suggested there was not a major risk to adult patients.

The concern that the drug could cause genetic mutations “in a clinical setting appears to be low,” said Robert H. Heflich, director of the Division of Genetic and Molecular Toxicology at the FDA’s National Center for Toxicological Research.

The second concern is that the drug might induce genetic changes to the virus that fuel the rise of new, threatening variants. FDA reviewers showed data demonstrating that in seven people treated with the drug, viral samples taken after treatments contained worrisome mutations, although the patients cleared the virus.

“Even if the probability is very low, 1 in 10,000 or 1 in 100,000, that this drug would induce an escape mutant for which the vaccines we have do not cover, that would be catastrophic for the whole world,” Hildreth said.

But FDA reviewers said those mutations also emerge naturally.

“It is unclear to us if molnupiravir would have a substantial impact on the evolutionary patterns that are already happening,” said Patrick R. Harrington, senior clinical virology reviewer at the FDA.

Other questions that arose, without resulting in satisfying answers, included why the drug didn’t appear to work in people with type 2 diabetes. It also was less effective against the delta variant.

There are already three treatments, called monoclonal antibodies, authorized in the United States to help keep high-risk people from ending up in the hospital. Monoclonals are cumbersome to administer and their uptake has often been uneven. Unlike pills that can be picked up at a pharmacy and taken at home, monoclonals are given through lengthy infusions or injections that are less convenient — and new variants can clearly pose a threat to their use.

Monoclonal-maker Regeneron released a statement Tuesday announcing that although direct tests have not yet been completed, modeling of the mutations in the omicron variant suggest its antibody cocktail would be less effective at blocking the variant.

Molnupiravir would be the first antiviral in pill form. An intravenous drug, remdesivir, is authorized for hospitalized patients, but its benefits are debated, with one review finding that it has “little or no effect” on deaths among hospitalized covid-19 patients.

“As we enter the winter months, another surge is imminent, potentially in the setting of emerging new variants of concern,” Kartsonis said. “We remain in dire need of novel, effective, well-tolerated and conveniently administered therapies to treat covid-19 in the outpatient.”