This story originally appeared on STAT, a health and medicine website that provides ambitious coverage of the coronavirus. Go here for more stories on the virus. Try STAT Plus for exclusive analysis of biotech, pharma, and the life sciences. And check out STAT’s COVID-19 tracker.
How early did local transmission of the SARS-CoV-2 virus begin in the United States? For the second time this week, scientists have proposed a new estimate. This one, from scientists at the Centers for Disease Control and Prevention, suggests that transmission likely began in late January or early February on the West Coast and that the virus spread undetected for more than a month.
Scientists from the agency, and research groups they collaborate with, concluded community transmission in the U.S. began with a single unidentified imported case from China.
Kristian Andersen, an expert in viral genomes from Scripps Research, said the new study underscores that early chains of transmission were missed in the United States, which as a consequence lost an opportunity to prevent spread of the virus from taking root here.
The CDC team drew on multiple streams of data, including retrospective testing of nearly 11,000 samples collected through CDC-led influenza surveillance networks.
If COVID-19 cases were being mistaken for flu in the early days of the outbreak, one would have expected to have seen some positive tests among those samples. But the first that was positive was collected from a patient in Washington state on Feb. 25 — two days before authorities in California reported finding an infection in a woman who had not traveled outside the country and who was not a known contact of people who returned to the U.S. infected with the virus.
CDC Director Robert Redfield pointed to the work as evidence that the agency — which has been savaged for the early failure of a COVID-19 test that left the country with limited capacity to look for the virus in the early weeks of the outbreak — had always been alert to the virus’ movements in the United States.
“We were never blind when it came to surveillance for Coronavirus 19,” he told reporters during a briefing on the new findings, published Friday in the CDC’s online journal Morbidity and Mortality Weekly Review. “The reality is the surveillance systems that CDC had developed over the years for respiratory viral diseases, particularly influenza-like illness, really did give us eyes on this disease as it began to emerge.”
Redfield’s statement seemed to ignore the fact that the flu surveillance samples were not tested for COVID-19 in real time, when they might have helped public health departments track and halt spread of the new virus.
“That’s a preposterous statement,” Andersen told STAT. “It’s a sad fact that the United States missed the boat on getting adequate testing set up early enough to be able to stop the virus in its tracks — it’s likely one of several reasons we have by far the most cases of any country in the world.”
As of Friday, the U.S. has recorded more than102,000 deaths — 28% of the global total.
The CDC findings challenge to a degree those released earlier this week in a preprint article from Michael Worobey, a professor of evolutionary biology at the University of Arizona. Preprint articles have not yet undergone the normal peer-review process at a scientific journal.
Worobey and colleagues argued that the person who started the first chain of transmission to take root in the country — in Washington state — entered the U.S. around mid-February, either from Asia directly, or via British Columbia, Canada.
Their estimate, based on analysis of genetic sequences of viruses retrieved from infected people, in turn challenged work from Trevor Bedford, a computational biologist at Fred Hutchinson Cancer Research Institute, that suggested the Washington state outbreak began with the first COVID-19 case diagnosed in the country, a man who returned from China on Jan. 15.
But the CDC scientists came to a different conclusion after analyzing data from four sources: genetic sequences; retrospective testing of stored influenza surveillance samples; a review of emergency room records from counties that had early cases of COVID-19; and the retrospective discovery of two people in California who died from Covid-19 in early and mid-January, and an infected cruise ship crew member who departed from San Francisco on Feb. 11. Public health officials only later realized those individuals had been infected with the virus.
The combination of those data sources led the CDC team to conclude community transmission in the U.S. began in late January or early February. Worobey estimated local transmission began in the U.S. in Washington state around Feb. 13, though it could have been as early as Feb. 7 or as late as Feb. 19.
“The take-home message from the northern California thing seems to be that yes, there definitely were early cases and some early community transmission, but when you just fast forward a little bit in time, none of it seems to have established itself into a sustained transmission cluster like you see in Washington or later in New York,” Worobey told STAT.
One of the authors of the CDC study, Greg Armstrong, said the agency’s scientists disagreed with how firm the two earlier estimates were in assessing when transmission began in the United States. “We thought there was a considerable amount of uncertainty still, given how few data were available,” said Armstrong, who is CDC’s deputy incident manager for the response.
The estimates of when the virus actually started transmitting in the U.S. could change again, Armstrong said, noting pathologists continue to investigate suspicious deaths from early in the year and to send samples to the CDC for analysis. “It is quite possible we will find additional cases in the future,” he said.
Get Boston.com’s e-mail alerts:
Sign up and receive coronavirus news and breaking updates, from our newsroom to your inbox.