H5N1 Has Pandemic Potential, But Critical Questions About Likelihood Remain

In 1997, a year after H5N1 first emerged, poultry farms in China and Hong Kong reported eighteen human infections. A third of those people died. That early warning sign—stark and undeniable—set the tone for three decades of scientific alarm about a virus that seemed to be waiting for its moment.

The virus has been patient. It has spread across species with unsettling ease: tens of millions of wild and farmed birds, then raccoons and foxes, then seals and sea lions, and most recently, dairy cattle in the United States. Over the past thirty years, H5N1 has infected more than eight hundred people worldwide, with a documented death rate hovering around fifty percent among confirmed cases. The Lancet Infectious Diseases warned in May that the virus's expanding host range, its capacity to jump between mammals and humans, its global reach, and the sheer scale of bird outbreaks all point toward genuine pandemic potential. To many observers, such a catastrophe feels not just possible but inevitable.

Yet the reality is more complicated. Despite three decades of circulation, H5N1 has never managed to sustain human-to-human transmission. The virus mutates readily—influenza does—but it has not yet acquired the specific genetic changes that would allow it to bind efficiently to human respiratory receptors and spread from person to person. Louise Moncla, an avian flu researcher at the University of Pennsylvania, frames the uncertainty plainly: Are we watching the opening act of a pandemic, or witnessing repeated spillovers that will never quite ignite? There is no way to know. That uncertainty is precisely why scientists are watching the virus's evolution with intense focus.

The recent spread among U.S. dairy cattle has provided researchers with a wealth of genetic material to analyze. Multiple teams, including Moncla's, concluded that H5N1 likely jumped from birds to cattle as early as late 2023 and circulated undetected for months before the first cases surfaced in March. The current lineage, known as 2.3.4.4b, carries some mutations that may enhance replication in mammals—concerning, but not yet catastrophic. The mutations scientists would truly fear, the ones that would enable easy human transmission, have not appeared. A Texas farm worker infected in March developed eye inflammation and broken blood vessels but no respiratory symptoms. His household contacts remained healthy. The virus found a dead end.

But evolutionary dead ends can become highways with a handful of genetic changes. Jennifer Guthrie, a pathogen genomics specialist at Western University, notes that influenza's capacity for mutation means the situation could shift at any moment. The longer the virus circulates, the greater the chance of that shift occurring. The problem is that nobody knows how many people are actually infected. Farm workers in dairy states are falling ill, but many are reluctant to seek testing. Some lack regular access to healthcare. Some are migrants facing language barriers or immigration concerns. State health labs have sent roughly twenty-five human samples to the CDC, with another hundred workers under monitoring, but the true number of infections remains hidden. That gap in knowledge is dangerous. Without a complete picture of how many people are infected and how severely, scientists cannot determine whether H5N1 is still blocked by evolutionary barriers or whether those barriers are beginning to crack.

The death rate question haunts epidemiologists. The World Health Organization once estimated H5N1's case fatality rate at sixty percent—a figure that would make a pandemic catastrophic beyond measure. Timothy Sly, an influenza researcher at Toronto Metropolitan University, spent years analyzing surveillance data and blood tests suggesting the true rate was closer to fourteen to thirty-three percent. Since then, more infections have been documented, many of them mild or asymptomatic, suggesting the actual death rate could be even lower. Yet even a lower rate would be devastating. The 1918 flu pandemic, with an estimated case fatality rate of two and a half percent, killed between twenty-three and fifty million people in a few years. If H5N1 achieved sustained human transmission with any significant death rate, Sly warns, the impact would be catastrophic.

What makes this moment urgent is not certainty but the absence of it. European health authorities have called for enhanced surveillance, rapid data sharing, careful oversight of poultry and fur farming, and vaccination of at-risk populations. Canada announced plans to expand avian flu monitoring, including testing of retail milk. But Sly and others argue these measures do not go far enough. Farm workers should receive regular blood tests. Countries cannot assume geographic borders will protect them. The virus does not recognize the forty-ninth parallel. The risk of moving from a handful of sporadic cases to something of international concern is, as the CDC's principal deputy director acknowledged, not insignificant. A virus can circle the globe before public health even has time to respond. That is the real danger: not what H5N1 is now, but what it could become if surveillance falters and the virus acquires just the right mutations. As Sly puts it, "It doesn't seem possible, until suddenly it is possible."