Bird flu, specifically the H5N1 strain, has killed roughly half of all confirmed human cases since 2003. Out of 954 people confirmed infected worldwide through December 2024, 464 died, a case fatality rate of 49%. That makes it far deadlier than seasonal flu, which kills less than 0.1% of those infected. The reasons come down to how the virus attacks the body, where it lands in the lungs, and how violently the immune system reacts to it.
The Virus Hits Deep in the Lungs
Seasonal flu viruses are well adapted to human airways. They latch onto receptors concentrated in the upper respiratory tract, your nose and throat, which is why a normal flu often starts with a sore throat and runny nose. Bird flu works differently. H5N1 preferentially binds to a different type of receptor found deeper in the respiratory system, in the cells lining the small air sacs of the lungs called pneumocytes. That’s where oxygen exchange happens.
This deep-lung tropism is a big part of why bird flu is so dangerous. Instead of causing an upper respiratory infection that your body can manage relatively easily, the virus sets up shop in the most critical part of your breathing apparatus. The primary site of H5N1 replication in humans is the pneumocyte, and this leads directly to severe pneumonia, fluid buildup in the lungs, and acute respiratory distress syndrome.
The Immune System Overreacts
The virus itself causes serious damage, but the immune response it triggers may be even more destructive. H5N1 infections are characterized by what researchers call a “cytokine storm,” an exaggerated flood of inflammatory signaling molecules. Normally, these molecules coordinate the immune response in a controlled way. In H5N1 infections, the body produces excessive levels of them, particularly three key inflammatory proteins driven mainly by immune cells called macrophages and certain white blood cells.
This runaway inflammation is responsible for many of the lethal complications: massive fluid accumulation in the lungs (pulmonary edema), bleeding in the air sacs, widespread tissue destruction, and a condition where immune cells start attacking the body’s own blood cells. The severity of infection is closely tied to this inflammatory dysregulation. In other words, it’s not just the virus killing tissue directly. It’s the body’s own defense system, pushed into overdrive, that causes catastrophic collateral damage.
The Virus Struggles in Human Cells, but That’s Not Reassuring
One reason H5N1 hasn’t caused a pandemic is that it replicates inefficiently in human cells. Influenza viruses need a host protein called ANP32A to copy their genetic material. The avian version of this protein is 33 amino acids longer than the mammalian version, and avian-adapted flu viruses depend on that longer protein. When H5N1 enters human cells, its replication machinery can’t use the shorter human version of ANP32A nearly as well. Other aspects of the virus’s gene processing are also disrupted in mammalian cells, further reducing how effectively it can multiply.
This is a significant barrier to efficient human infection, but it’s not a permanent one. Specific genetic mutations can help the virus overcome it. The most studied is a change in one of the virus’s replication proteins (known as PB2 E627K), which enhances the virus’s ability to replicate in mammalian cells and increases infectivity in animal models. This mutation has already been detected in H5N1 viruses found in cows, minks, and foxes. Other mutations that improve mammalian adaptation have also been identified in circulating strains.
Why the Death Rate Might Be Misleading
The 49% fatality rate comes with an important caveat: it’s based only on confirmed cases, which tend to be the sickest people, those who showed up at hospitals with severe pneumonia. Milder infections likely go undetected, meaning the true fatality rate could be lower. Recent evidence supports this. Since March 2024, the United States has reported 70 human infections linked to a bird flu outbreak in dairy cattle. Most of those 70 cases involved mild illness, primarily eye infections and mild respiratory symptoms, with only one death (a fatality rate of about 1.4%).
The difference matters. The older cases from Southeast Asia and Egypt, which drove the 49% figure, involved a different pattern of exposure and possibly different viral genetics. The current clade circulating in U.S. cattle (called B3.13 genotype) has generally caused milder disease in the dairy workers exposed to it. However, a different version of the same virus lineage (the D1.1 genotype) caused severe illness in two people in late 2024, including one death in Louisiana. So the severity depends partly on which specific genetic variant someone encounters.
Treatment Works Best When It’s Early
Antiviral treatment significantly improves survival, but timing is critical. In an observational study comparing treated and untreated H5N1 patients across multiple countries, antiviral treatment was significantly associated with survival. The benefit was greatest when treatment started within two days of symptom onset. After that window, the virus has already caused enough lung damage and triggered enough inflammation that antivirals alone may not be enough to reverse the cascade.
Where the Risk Stands Now
The CDC’s risk assessment tool currently places H5N1 in the “moderate risk” category for both its potential to emerge as a pandemic virus and its potential public health impact. The immediate risk to the general public remains low. Most human cases have involved direct contact with infected birds or cattle, and sustained person-to-person transmission has not been documented.
The concern is evolution. H5N1 has been circulating in wild birds and poultry for over two decades, and since 2021 the 2.3.4.4b clade has spread to an unprecedented range of mammals, including dairy cows, sea lions, minks, and foxes. Every new mammalian host gives the virus more opportunities to acquire the handful of mutations it would need to spread efficiently between people. Researchers are tracking specific genetic changes that would enhance human receptor binding, improve replication in mammalian cells, and enable airborne transmission. Some of those changes have already appeared individually in animal isolates, though the full combination has not yet been found in a single virus.