Bird flu has long been a top contender for starting the next pandemic. Between 2003 and 2023, there were 882 cases reported in humans worldwide, resulting in 461 deaths. This suggested a staggering mortality rate, around 50 percent — a warning of H5N1’s deadly potential.
The vast majority of infections and deaths were before 2020. But a few years of quiet ended in 2024, when America became the epicenter for the world’s human infections. And there was something else about the U.S. cases: They were nowhere near as severe as before.
The virus had been mostly tied to poultry, but in America last year it began circulating in dairy cows as well. Dozens of farmworkers have become infected. And yet, of the 70 human H5N1 cases reported over the last year in the United States, there was only one death. All but a few of the other U.S. cases have been mild, with most infected people experiencing little more than pink eye or a cough. It’s a pattern that is puzzling epidemiologists.
The lower-than-expected death rate has not escaped the notice of the health secretary Robert F. Kennedy Jr., who has cited it as a reason to allow the virus to “run through the flocks” rather than cull and control outbreaks in poultry. That proposal invites disaster. The United States has gotten lucky so far, but that luck might not last.
The one fatality over the last year was in January, when an older adult in Louisiana died after contracting the virus from a backyard flock. Last month, two people with exposure to poultry were hospitalized. And in Canada, which has also experienced outbreaks in birds, a teenager spent weeks in intensive care. These cases are a sharp reminder that the virus hasn’t lost its lethal potential.
Epidemiologists like me are wary of drawing premature conclusions about severity. Small samples can be misleading, and most of the recent cases occurred in dairy farmers. If the virus reaches infants, older adults or people with serious health conditions, it may be more deadly.
But let’s assume that today’s H5N1 is in fact milder compared with past outbreaks in the Middle East and Asia. Why might that be?
One theory centers on how people are infected. While previous cases of severe H5N1 infections typically began with respiratory symptoms, patients with recent cases have frequently reported eye redness. This pattern suggests that many infections may have occurred through eye exposure — perhaps through splashes of contaminated milk during routine farm operations — rather than through inhalation.
Another theory has to do with the variants of H5N1 circulating in the United States. By 2021, a new version of the bird flu had emerged and begun spreading rapidly around the world. Since then, the epidemic has devastated wild bird populations and shown a troubling ability to spill over into many species, including mammals like mink, foxes and cows. The variants currently circulating in the United States are descendants of this version.
The most widespread variant in American cows, B3.13, is responsible for most cases of pink eye. Two other variants, D1.3 and D1.1, which are found mostly in wild birds and farmed poultry, are behind the three severe cases seen in the United States. This has led some to speculate that these variants might be more likely to cause hospitalizations or deaths.
Mr. Kennedy has been particularly complacent about the B3.13 variant spreading on farms. In a recent interview, he suggested that the variant is “not very dangerous to humans” and that his counterpart in the Agriculture Department should consider allowing it to circulate in hopes of identifying birds that are resistant to infection. This idea is not only dangerous, but also impractical. When an outbreak is detected, farmers can’t wait days for test results. They must act quickly to prevent the virus from spreading further.
What keeps epidemiologists up at night is a scenario where bird flu gains the ability to spread efficiently among humans. Letting the virus spread widely among animals increases its chances of evolving to do so. And if we do end up in this nightmare scenario, the virus will already have mutated — so there’s no way to know if such a variant would be milder, or much more deadly. It’s also worth remembering that a virus doesn’t need a mortality rate of 50 percent to devastate society. Some scholars estimate that the mortality rate of the 1918 flu pandemic was most likely only around 2 or 3 percent.
I fear that the apparent mildness of infections in the United States has confounded calls to act more decisively. Agricultural and health authorities were slow to look for the virus in dairy herds, reimburse affected farmers and answer important research questions. A national strategy for testing milk began in December, months after the virus was found in dairy cows.
During the early stages of Covid-19, the first Trump administration underestimated the threat, repeatedly assuring the public that the risk to the population was low. While the future course of H5N1 is more difficult to predict, the administration should adopt a “no regrets” policy, doing everything it can to prevent the worst-case possibilities and resisting the temptation to downplay potential dangers.
It’s possible that H5N1 will never make the jump to spread human to human. After all, it hasn’t done so yet. But we shouldn’t give it the opportunity.
The post America Has Been Lucky on Bird Flu — So Far appeared first on New York Times.
