Influenza A is generally considered the worst type of flu because it’s the only type capable of causing pandemics and is responsible for the large majority of flu hospitalizations and deaths each season. But the answer gets more nuanced when you look at specific subtypes, and one subtype in particular, H3N2, has a reputation for producing the most severe seasonal flu years.
Influenza A vs. Influenza B
Flu viruses come in two main types that infect people each winter: influenza A and influenza B. Influenza A dominates most seasons and causes far more total illness. During the 2024-25 season, influenza A viruses peaked at 30.4% of respiratory specimens tested, while influenza B peaked at just 4.3%. That pattern holds year after year: A circulates more widely and infects more people.
Here’s where it gets interesting, though. A CDC study spanning eight flu seasons found that among adults who were actually hospitalized, influenza B caused equally severe outcomes as influenza A. Length of hospital stay, ICU admission rates, and the proportion of deaths were all comparable between the two types. So influenza B isn’t a “mild” flu. It simply infects fewer people overall, which means its total burden is smaller.
In children, influenza A is associated with the majority of pediatric flu deaths in nearly every season. Of the 280 pediatric flu deaths reported during the 2024-25 season, 86% were linked to influenza A and 14% to influenza B. Since pediatric flu deaths became nationally reportable in 2004, influenza A has been responsible for more child deaths than influenza B in all but two seasons.
Why H3N2 Seasons Tend to Be Worse
Not all influenza A is equal. The two main subtypes circulating in humans are H1N1 and H3N2, and H3N2 seasons consistently correlate with more severe outcomes at the population level. H3N2 is also harder to vaccinate against. In the 2024-25 season, the flu vaccine reduced outpatient visits from H1N1 by 42-72% depending on age group, but effectiveness against H3N2 in outpatient settings was often not statistically significant, hovering around 16-25% for some populations. Against H3N2 hospitalizations, the vaccine performed better (42-55%), but still generally trailed its performance against H1N1.
The reason H3N2 is so difficult to pin down is that it mutates rapidly. All flu viruses undergo “antigenic drift,” meaning small genetic changes accumulate as the virus copies itself, gradually altering the proteins on its surface that your immune system recognizes. H3N2 drifts faster than H1N1 or influenza B, which means the vaccine strain chosen months before flu season may no longer match well by the time you’re exposed. That mismatch between vaccine and circulating virus is a major reason some flu seasons hit harder than others.
What Makes Influenza A Uniquely Dangerous
Beyond its seasonal toll, influenza A is the only type that causes pandemics. This is because of a process called “antigenic shift,” where the virus undergoes a sudden, major change rather than the gradual drift that happens every year. Shift occurs when a flu virus from an animal population (typically birds or pigs) gains the ability to infect humans. Because the resulting virus carries surface proteins people have never encountered, most of the population has no pre-existing immunity.
Every flu pandemic in the last century was caused by influenza A. The 1918 pandemic, caused by an H1N1 virus with genes of avian origin, infected roughly one-third of the world’s population and killed at least 50 million people. The 1957 pandemic was caused by an H2N2 virus, and the 1968 pandemic by H3N2. Influenza B lacks this pandemic potential because it doesn’t circulate in animal reservoirs the same way, so it can’t undergo antigenic shift.
How the Flu Actually Kills
The flu itself can be deadly, but in many severe cases, the real danger comes from what follows. When the virus infects and damages the lining of your airways, it creates an opening for bacteria that normally wouldn’t cause harm. Up to 75% of flu patients who develop pneumonia have a confirmed bacterial co-infection. During the 1918 pandemic, an estimated 95% of deaths were due to secondary bacterial pneumonia.
The virus also triggers an intense inflammatory response. Infected cells release a flood of immune signaling molecules that, in severe cases, can spiral out of control and damage the body’s own tissues. This inflammatory cascade, combined with the bacterial infections that follow, is what turns a respiratory virus into a life-threatening illness. Post-mortem studies across multiple pandemics have found evidence of bacterial infection in up to 80% of fatal flu cases, with the most common culprits being common respiratory bacteria that exploit the damage influenza leaves behind.
During the 2009 H1N1 pandemic, secondary bacterial pneumonia was identified in 29-55% of deaths. In the 1968 pandemic, England and Wales saw a 55% increase in respiratory deaths, with bacterial co-infection as a major contributor. This pattern repeats in every major flu event: the virus weakens the body’s defenses, and bacteria finish the job.
The 2024-25 Season in Context
The most recent flu season was classified as high severity, the worst since 2017-18 by multiple measures. The cumulative hospitalization rate reached 128.3 per 100,000 people, the highest recorded since the 2010-11 season. Pediatric flu deaths hit 279, the highest number ever reported during a single seasonal epidemic. Both H1N1 and H3N2 circulated at roughly equal levels (53% and 47% of subtyped influenza A, respectively), which is unusual since one subtype typically dominates.
The peak percentage of doctor visits for flu-like illness reached 7.9%, surpassing every season since the 2009 pandemic. Peak weekly flu-related deaths hit 2.8% of all deaths, the highest since surveillance data were incorporated in 2015-16. This season demonstrated that even without a novel pandemic virus, seasonal influenza A can produce devastating results when conditions align.
Avian Flu: A Different Category of Risk
Highly pathogenic avian influenza, particularly H5N1, exists in a category apart from seasonal flu. Globally, the mortality rate from H5N1 infections in humans has historically been around 50%. That number reflects a small sample of cases, mostly involving direct contact with infected birds, and the virus does not yet spread efficiently between people. In the United States, only one of 70 documented human H5N1 infections has resulted in death, suggesting the global fatality rate may overstate the risk in settings with modern healthcare.
H5N1 is not currently a seasonal threat. But it represents the type of antigenic shift event that public health officials monitor closely: an animal-origin influenza A virus that could, with the right mutations, gain the ability to spread between humans and trigger a pandemic in a population with no immunity.