The coronavirus may have infected a small number of people in the United States as early as Dec. 13, more than a month earlier than researchers had thought, according to scientists who analyzed blood samples taken from American Red Cross donations.
The researchers noted that they could not say whether the apparent infections were in travelers who had caught the virus in other countries, or whether the infections led to wider community transmission.
Before this new report, the earliest documented infection in the country was on Jan. 19 in someone who had traveled to China. Although other genetic studies have suggested the possible presence of the virus earlier than that date, the new study found that blood donations from nine states sent to the Centers for Disease Control and Prevention carried coronavirus antibodies — protein markers of past exposure to the virus, or perhaps to one very similar to it.
At least one prominent virus researcher was wary of how the findings were interpreted online and in news reports. Trevor Bedford, an epidemiologist at the University of Washington who has been deeply involved in genetic studies of how, when and where the virus has spread, said in a series of tweets that he thought the study could be identifying people who had antibodies to other human coronaviruses, which cause common colds, although he did not rule out that it may have picked up some cases of travelers infected in other countries.
In the new report, which was released online Monday and has been accepted for publication in the journal Clinical Infectious Diseases, Dr. Sridhar V. Basavaraju of the C.D.C. and others reported that they had looked for antibodies that reacted to the specific virus that has caused the pandemic, SARS-CoV-2. They used samples taken from blood donations that the American Red Cross had gathered in nine states.
Susan L. Stramer, a virologist at the American Red Cross and one of the authors of the paper, said the blood samples had been gathered originally to test for exposure to mosquito-borne illnesses like West Nile virus. The C.D.C. analyzed the samples for evidence of coronavirus exposure.
Dr. Stramer noted that the antibody tests are not for the virus itself and do not offer much useful information to the person whose blood is tested. Antibodies can hang around in the blood well after the virus has left the body. But these blood markers can be useful, she said, for monitoring broad patterns of disease.
One issue with testing is that antibodies to certain coronaviruses, such as those that cause common colds, may also respond to other viruses in the same family, like SARS-CoV-2.
In the new tests on more than 7,000 samples, 106 showed coronavirus antibodies. The researchers narrowed this down to 84 that had antibodies that would attack, or “neutralize” SARS-CoV-2 to some degree. One of these samples showed very effective neutralization, Dr. Stramer said. And another sample showed a reaction to a part of the spike protein that is very specific to SARS-CoV-2. “So for two samples at least, we believe they probably represent true infections.” Those people could have been travelers who were infected outside of the U.S., however.
In future studies, Dr. Stramer said, researchers will look at earlier years to see whether blood samples, as expected, would not show antibodies to SARS-CoV-2.