Fifty-nine percent of all transmission came from people without symptoms, under the model’s baseline scenario. That includes 35 percent of new cases from people who infect others before they show symptoms and 24 percent that come from people who never develop symptoms at all.

“The bottom line is controlling the covid-19 pandemic really is going to require controlling the silent pandemic of transmission from persons without symptoms,” said Jay C. Butler, the CDC deputy director for infectious diseases and a co-author of the study. “The community mitigation tools that we have need to be utilized broadly to be able to slow the spread of SARS-CoV-2 from all infected persons, at least until we have those vaccines widely available.”

The emergence of a more contagious variant, first detected in the United Kingdom and since found in several U.S. states, throws the significance of those guidelines into even starker relief. “Those findings are now in bold, italics and underlined,” Butler said. “We’ve gone from 11-point font to 16-point font.”

“It’s certainly confirmatory, but it’s nice to see confirmation,” said epidemiologist Richard Menzies, who directs the McGill International TB Centre in Canada and was not affiliated with this research. “These are pretty believable, solid results.”

Many factors influence how the coronavirus spreads. The researchers took an admittedly uncomplicated approach — Butler called it “a fairly simple mathematical model” — and used that to assess several scenarios, varying the infectious period and the proportion of transmission from people who never develop symptoms.

The model consistently predicted asymptomatic spread accounted for roughly half of viral transmission. “I was a bit surprised how well it held up under a broad range of base assumptions,” Butler said, such as shifting the timing of peak contagiousness from four days after infection to five or six.

But Muge Cevik, an infectious-disease expert at Scotland’s University of St. Andrews, argued some of the model’s assumptions are flawed. Cevik said the best estimate for the relative contagiousness of people who do not have symptoms, vs. those who do, was 35 percent, drawing from a review of the scientific literature published in September.

The study authors instead estimated, at baseline, that people without symptoms were 75 percent as contagious. That figure, Butler said, came from their own literature analysis including peer-reviewed and preprint research. “I have no doubt that we’ll still have people say, ‘Well, what if you did X, Y or Z?’ But hopefully being able to adjust the variables will help address some of those concerns, such as what we did,” Butler said.

Cevik also noted the study does not account for the environment where the spread occurs. “Maybe asymptomatic transmission is important, but it may be much more important in long-term care facilities and households,” she said. “That might mean that we need to do much more targeted testing for high-risk populations,” as opposed to mass screening.

Whether vaccines stop coronavirus transmission is not yet certain, and was not a scenario addressed in this model. “The data on the impact of the vaccines on asymptomatic infection are very limited,” Butler said, though he anticipates more information in coming months.

The clinical trials for the mRNA vaccines, authorized in December, concluded the vaccinations are highly capable of preventing symptomatic illness. But those trials did not determine whether vaccinated people are able to spread the pathogen.

“If they were asymptomatic but equally contagious, then that’s going to have quite an impact on the epidemic,” Menzies warned. That is why it is so important to keep testing people, he said, especially those who were vaccinated and then exposed to the virus.