An international meeting of Biogen leaders at a Boston hotel in February led to roughly 20,000 cases of COVID-19 in four Massachusetts counties by early May, far more than the 99 previously identified, according to three scientists involved in a new study.
After examining nearly all the confirmed early cases of the illness in the area by changes in the genetic makeup of coronaviruses as they pass from one person to another, the researchers were able to assess the broader impact of the “super-spreading event” at the Marriott Long Wharf hotel.
What makes the estimate all the more startling is that this version of the virus was one of more than 80 that invaded the state between late January and early May, the researchers found. But it sparked a viral forest fire.
The 64-page study has yet to be peer reviewed for publication, and the researchers stressed that their estimate ― which doesn’t appear in the paper but they shared in interviews ― is an extrapolation based on viruses isolated from 772 local patients. But they believe their calculations are sound.
“I’m confident that the scale for measuring this event is in the tens of thousands,” said Dr. Jacob Lemieux, an infectious disease physician at Massachusetts General Hospital and one of three scientists at the Broad Institute of MIT and Harvard who spoke about the study they wrote with 50 other researchers.
Caroline Buckee, a professor of epidemiology at the Harvard T.H. Chan School of Public Health who wasn’t involved in the study, said that assuming the coronaviruses researchers analyzed were a representative sample, she had no trouble believing the estimate of 20,000 conference-related cases.
“Super-spreading is really a key component of how we ended up with an epidemic of this gravity,” said Buckee, who received a copy of the study before it appeared Tuesday on a preprint server. “If you think about it, all the cases in the whole world originated from one case. That’s the nature of the exponential growth of epidemics.”
The research team analyzed the genetic sequences of the virus that caused COVID-19 in the 772 patients, almost all from Essex, Middlesex, Norfolk, and Suffolk counties. By examining mutations in the genetic code that naturally occur as the virus makes copies of itself — subtle changes that act like a passport stamp showing where the pathogen has been — the sleuths identified more than 80 distinct SARS-CoV-2 genomes of viruses that infected the Boston area in the first five months of the year. Most of the viruses came from elsewhere in the United States and Western Europe, the scientists said.
But one virus with a unique genetic signature had an outsize impact. Some 289 of the 772 patients, or more than a third, were infected with a virus traceable to the meeting held on Feb. 26-27 by Cambridge biotech Biogen.
In a remarkable sign of how the virus can spread unpredictably and take a disproportionate toll on society’s most vulnerable members, the 289 conference-related cases included 122 people living in Boston-area homeless shelters and employees who work there, the study says. It’s unclear what path the virus took to get there.
By multiplying the proportion of conference-related viral genomes in each of the four counties by the total number of coronavirus infections in Essex, Middlesex, Norfolk, and Suffolk, the scientists estimate that 20,000 infections could be linked to the Marriott event.
The study never mentions Biogen by name. It simply refers to an “international business conference held in Boston from February 26-27.”
“We’re not trying to point fingers,” said Bronwyn MacInnis, director of pathogen surveillance at the Broad and one of the authors interviewed. “Some [viral] introductions fizzle out, others light fires. The circumstances of this event ― the fact that it happened so early in the epidemic and the timing of where we were with COVID in the public consciousness ― meant it had a disproportionate effect. It happened in a city that wasn’t ‘woke.‘ ”
In response, Biogen said the outbreak at the meeting happened almost six months ago, when the understanding of COVID-19 was limited. But the company did not dispute the study or the estimate of 20,000 conference-related cases.
“We never would have knowingly put anyone at risk,” it said in a statement. “When we learned a number of our colleagues were ill, we did not know the cause was COVID-19, but we immediately notified public health authorities and took steps to limit the spread.”
In April, the company joined a group that includes the Broad, Brigham and Women’s Hospital, and Mass General that built a collection of biological and medical data about the virus.
“The world today has a much greater understanding of how easily and quickly this virus can be transmitted, and we are proud to contribute through this collaboration to the global effort to overcome COVID-19,” Biogen said.
The study was a collaboration of researchers from several institutions, including the Broad, Mass General, and the Massachusetts Department of Public Health. In a separate effort, the health department used contact tracing to tally at least 99 coronavirus cases stemming from the conference before officials stopped counting Biogen-related infections in late March.
In addition to Lemieux and MacInnis, Stephen Schaffner, a senior computational biologist at the Broad who worked on the study, spoke with the Globe about the findings.
The research team sequenced the genome for SARS-CoV-2 viruses isolated from 772 patients in Greater Boston — as many as scientists could obtain — mostly between March 4 and May 9. The patients had sought treatment at Mass General, the Boston Health Care for the Homeless Program, and other health care facilities.
The 80 or more distinct coronavirus genomes identified by researchers included one that apparently arrived in late January when a traveler returned from Wuhan, China, the epicenter of the pandemic at the time. That patient was quarantined in Massachusetts, and the researchers found no evidence that that patient infected anyone.
But the first big cluster of cases in Massachusetts appears to have stemmed from the Biogen gathering, which attracted about 175 company leaders from the United States and overseas.
Some 37 percent of the 772 genomes that researchers sequenced had the same genetic signature. They included sequences of viruses isolated from 28 people known to have attended the conference, worked at the hotel, or had direct contact with those individuals. Those 28 provided viral samples when they sought medical treatment shortly after the conference, many of them at MGH.
The Marriott gathering was a perfect storm for super-spreading, the scientists said. The event drew attendees from far-flung places. They packed into hotel elevators and onto escalators, sat near each other at meetings, and socialized over meals and drinks. The conference happened before many Americans took precautions against COVID-19.
“Because it happened early in the epidemic, it had the chance to spread widely before extensive capacity, shut-downs, social distancing and masking were in place,” the three scientists said in an e-mail.
The patients with conference-related viral genomes included 122 who lived in homeless shelters in Greater Boston or work with those people. Many sought treatment through the health care program for the homeless. The study does not identify the shelters.
The three authors said they have no idea how the virus spread from the conference to homeless shelters. “But it illustrates that we are all connected, and that more vulnerable populations are disproportionately affected by COVID-19 ― once it enters homeless communities, the circumstances are ripe for it to spread extensively,” they said in the e-mail.
The scientists emphasized that their estimate of 20,000 conference-related cases in four counties is a “back-of-the-envelope calculation.” The total number of people infected with the virus that spread from the conference could be far higher because the researchers didn’t look at cases outside the four counties or in other parts of the country or abroad. The outbreak at the conference is believed to have led to a cascade of cases in Tennessee, North Carolina, Indiana, New Jersey, Washington, D.C., and overseas.
“A precise estimate of the total number involved [globally] is not possible with current data and may not be possible at all,” the scientists said. “If tens of thousands of individuals seems large, it is important to point out that it is in [the] context of a pandemic that has infected tens of millions of people.”
The three scientists offered some caveats. It’s possible, they said, that viral genomes from patients who sought care at Mass General and other health care providers weren’t representative of coronaviruses in the counties that the researchers used to make their extrapolation.
It’s also possible that the version of the virus they believe ricocheted at the conference had spread undetected from another local super-spreader event beforehand and then infected people at the Marriott, confusing researchers about the true origin. But, the scientists said in their e-mail, “we see no evidence for that.”
Dan Hartl, a Harvard biology professor who specializes in population genetics and wasn’t involved in the study, said he found the estimate of 20,000 cases credible, “although I grant you it’s eye-popping.” There probably wasn’t anything extraordinary about the conference-related virus that caused it to spread so widely, he added.
“You say, ‘Oh, my gosh, this one virus seems to have spread much more than the others — is there something special about it?‘ ” he said. “Not necessarily. It just got lucky and spread in one of these super-spreader events.”
Jonathan Saltzman can be reached at email@example.com