Why This Year's Flu Shot Replaced All Three of Its Strains

Somewhere between 24 million and 47 million people in the United States caught the flu last season. Between 310,000 and 710,000 of them ended up in a hospital bed. That range comes from the Centers for Disease Control and Prevention, and it is unusually wide because the season itself was unusually messy: a mutated flu strain that the vaccine wasn't built to stop.

That's why the flu shot rolling out at pharmacies this month is not a minor update. It is a near-total rebuild. Regulators replaced all three virus strains in the 2026-2027 vaccine, an uncommon move that reflects how badly last season's shot missed its target.

The strain responsible has a name only virologists love: influenza A(H3N2) subclade K. It carries multiple mutations in the protein flu viruses use to latch onto human cells, and those mutations were different enough from the vaccine's H3N2 component that antibodies built by vaccination sometimes failed to recognize the real virus. Researchers call this antigenic drift. Flu viruses mutate constantly, and occasionally they drift far enough, fast enough, to outrun a vaccine that was locked in months earlier.

The numbers show what that drift cost. Vaccine effectiveness against the strains circulating last season came in at 14% to 48% among children, 22% to 34% among adults 18 to 64, and 30% to 41% among adults 65 and older, according to CDC data presented to federal advisers in March. Ninety children died of flu-related causes last season; of those whose vaccination status was known, 85% had not been fully vaccinated.

Not every country saw the same collapse. A study spanning nine European countries found overall vaccine effectiveness of 25% to 45%, with children faring far better: 47% to 72% protected. In England, vaccine effectiveness against flu-related hospital admissions held at 72% to 75% for children and teenagers, though it dropped to 32% to 39% for adults. Canadian researchers, meanwhile, put protection against medically-attended illness from subclade K and related H3N2 viruses at around 40%. The pattern that emerges across all of it: subclade K didn't erase the vaccine's benefit anywhere, but it dented it everywhere, and the dent was deepest in working-age and elderly adults.

So the FDA's Vaccines and Related Biological Products Advisory Committee didn't tinker at the edges this year. On March 12, the committee voted 7-0, twice, to swap out every component of the trivalent shot. Egg-based vaccines will now target an A/Missouri/11/2025 H1N1 virus, an A/Darwin/1454/2025 H3N2 virus, and a B/Tokyo/EIS13-175/2025 Victoria-lineage virus; cell-based and recombinant vaccines get a slightly different H3N2 and B pairing tuned to how those platforms grow virus. All three replace what was in last year's shot.

This is how the system is supposed to work, even when it looks like an admission of failure. Flu vaccine composition gets picked roughly six months before the vaccine reaches a pharmacy shelf, based on which viruses are circulating in the Southern Hemisphere and which mutations look poised to spread north. That lead time is unavoidable, since vaccines take months to manufacture at scale, and it is also the exact window in which a virus like subclade K can drift past the strain regulators picked.

Egg-based production adds its own wrinkle. Most seasonal flu vaccines are still grown in chicken eggs, and the H3N2 virus tends to pick up extra mutations adapting to that environment, changes that can make the egg-grown vaccine strain look slightly different to the immune system than the wild virus it's meant to fight. It's part of why the FDA is leaning harder on cell-based and recombinant vaccines, which skip eggs entirely, as manufacturing capacity allows.

None of this means last year's shot was worthless. Even in the worst-hit adult groups, vaccinated people were still meaningfully less likely to end up hospitalized than unvaccinated people with the same exposure. Protection against the flu B component, largely unaffected by subclade K, stayed intact the whole season. A mismatched vaccine is a weaker shield, not a broken one.

What's harder to fix is the second number sitting next to the case counts: distributed flu vaccine doses have dropped 23% since the 2019-2020 season, even as hospitalization rates among adults hit their third-highest mark since 2010-2011. CDC officials who presented the data to the advisory committee said they don't fully know why vaccine effectiveness ran lower than usual this past season (surveillance sites differ in the populations and strains they capture) and cautioned that the numbers were preliminary. A newly reformulated vaccine won't do much good sitting in a fridge instead of an arm, and the shots built around Missouri, Darwin, and Tokyo strains are already shipping to clinics for the season that starts in earnest next month.