H1N1 makes up 96% of subtyped US flu cases as infections climb early in the West

The 96% figure needs its denominator. It is H1N1's share of the US positive tests that were subtyped in the week ending 26 September ^[16]. Nature reports that the early findings come from a relatively small number of positive tests, too few to predict the outbreak's trajectory ^[8]. Caitlin Rivers told NPR that the activity in California, Washington, Hawaii and Alaska looks more like November or December ^[2][3].

Sequencing asks whether the mutation is common in the people getting sick. Alex Greninger's lab at the University of Washington found it in 13 of 17 H1N1 samples from Seattle hospitals and clinics that were complete enough to check ^[17], about 76% ^[23]. That is one city and 17 genomes. The serum test asks whether the mutation could plausibly help the virus spread. Jesse Bloom's group at the Fred Hutchinson Cancer Center tested blood from 260 people in Australia and the US against 148 influenza A viruses circulating by mid-2026, in a preprint not yet peer reviewed ^[18].

"We saw that a number of strains which have the G155E mutation are, in some cases, less well recognized by pre-existing human antibodies," Bloom said ^[19]. His sentence has two qualifiers: a number of strains, and in some cases. Sera from teenagers and young adults did worst, probably because many in those groups have not met similar H1N1 viruses before ^[20]. Together the two parts show a mutation that is common in early Seattle cases and partly evades existing antibodies. They do not show that it started the season early. Discover's account, citing NPR, calls the mutated-strain idea one theory and says it is too early for conclusions ^[4].

The two reports also point at different viruses. Discover's section on the mutation is about subclade K, an H3N2 subclade the CDC first noticed in August 2025 data ^[5], and it says this season's vaccines were updated to target it ^[6]. In Nature's account, subclade K was last season's dominant variant ^[7]. On the latest subtyping, H3N2 is almost 4% of typed US cases ^[16]. Neither report says whether this season's H1N1 vaccine component was chosen with G155E in mind, and neither addresses hospital staffing.

The advice on timing stands on firmer ground. The CDC says most people need one flu dose a season, and it can be given as early as September ^[12]. Rivers, of Johns Hopkins, usually waits until mid-to-late October because the US peak tends to come in December and January. She is now thinking about a shot as soon as possible ^[22]. She wrote on 27 September that "there is something afoot" ^[21]. Demetre Daskalakis, a former director of the CDC's National Center for Immunization and Respiratory Diseases, told NPR that flu causes tens of thousands of hospitalizations and deaths every year ^[13].

How severe the season will be is a separate question. "An early start doesn't necessarily mean a more severe season," said James Hay, an epidemiologist at the University of Oxford ^[9]. Scott Hensley, a virologist at the University of Pennsylvania, said "It would be an incredibly early season" for the US if West Coast cases keep rising at their current rate ^[15]. In my view, the CDC's one-dose guidance alone justifies an earlier shot. Whether G155E explains the early start will depend on sequences from beyond Seattle and on Bloom's study holding up in review.