Researchers in Denmark reported on Wednesday that surgical masks did not protect the wearers against infection with the coronavirus in a large randomized clinical trial.
The study, published in the Annals of Internal Medicine, did not contradict growing evidence that masks can prevent transmission of the virus from wearer to others. But the conclusion is at odds with the view that masks also protect the wearers — a position endorsed just last week by the Centers for Disease Control and Prevention.
Critics were quick to note the study’s limitations, among them that the findings depended heavily on participants reporting their own test results and behavior, at a time when both mask-wearing and infection were rare in Denmark.
Coronavirus infections are soaring throughout the United States, and even officials who had resisted mask mandates are reversing course. Roughly 40 states have implemented mask requirements of some sort, according to a database maintained by The New York Times.
Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, advocates a national mask mandate, as does President-elect Joseph R. Biden Jr.
“I won’t be president until January 20th, but my message today to everyone is this: wear a mask,” Mr. Biden recently wrote on Twitter.
From early April to early June, researchers at the University of Copenhagen recruited 6,024 participants who had been tested beforehand to be sure they were not infected with the coronavirus.
Half were given surgical masks and told to wear them when leaving their homes; the others were told not to wear masks in public.
At that time, 2 percent of the Danish population was infected — a rate lower than that in many places in the United States and Europe today. Social distancing and frequent hand-washing were common, but masks were not.
About 4,860 participants completed the study. The researchers had hoped that masks would cut the infection rate by half among wearers. Instead, 42 people in the mask group, or 1.8 percent, got infected, compared with 53 in the unmasked group, or 2.1 percent. The difference was not statistically significant.
“Our study gives an indication of how much you gain from wearing a mask,” said Dr. Henning Bundgaard, lead author of the study and a cardiologist at the University of Copenhagen. “Not a lot.”
Dr. Mette Kalager, a researcher at Telemark Hospital in Norway and the Harvard School of Public Health, was persuaded. The study showed that “although there might be a symbolic effect,” she wrote in an email, “the effect of wearing a mask does not substantially reduce risk” for wearers.
Other experts were unconvinced. The incidence of infections in Denmark was lower than it is today in many places, meaning the effectiveness of masks for wearers may have been harder to detect, they noted.
Participants reported their own test results; mask use was not independently verified, and users may not have worn them correctly.
“There is absolutely no doubt that masks work as source control,” preventing people from infecting others, said Dr. Thomas Frieden, chief executive of Resolve to Save Lives, an advocacy group, and former director of the C.D.C., who wrote an editorial outlining weaknesses of the research.
“The question this study was designed to answer is: Do they work as personal protection?” The answer depends on what mask is used and what sort of exposure to the virus each person has, Dr. Frieden said, and the study was not designed to tease out those details.
“An N95 mask is better than a surgical mask,” Dr. Frieden said. “A surgical mask is better than most cloth masks. A cloth mask is better than nothing.”
The study’s conclusion flies in the face of other research suggesting that masks do protect the wearer. In its recent bulletin, the C.D.C. cited a dozen studies finding that even cloth masks may help protect the wearer. Most of them were laboratory examinations of the particles blocked by materials of various types.
Susan Ellenberg, a biostatistician at the University of Pennsylvania Perelman School of Medicine, noted that protection conferred by masks on the wearer trended “in the direction of benefit” in the trial, even if the results were not statistically significant.
“Nothing in this study suggests to me that it is useless to wear a mask,” she said.
Dr. Elizabeth Halloran, a statistician at Fred Hutchinson Cancer Research Center in Seattle, said the usefulness of masks also depends on how much virus a person is exposed to.
“If you show this article to a health care provider who works in a Covid ward in a hospital, I doubt she or he would say that this article convinces them not to wear a mask,” she said.
But Dr. Christine Laine, editor in chief of the Annals of Internal Medicine, described the previous evidence that masks protect wearers as weak. “These studies cannot differentiate between source control and personal protection of the mask wearer,” she said.
Dr. Laine said the new study underscored the need for adherence to other precautions, like social distancing. Masks “are not a magic bullet,” she said. “There are people who say, ‘I’m fine, I’m wearing a mask.’ They need to realize they are not invulnerable to infection.”