Flu Vaccines — What the Labels Say About the Shots Recommended Every Year

The Package Insert series

In March 2026, the FDA-approved prescribing information for three of the most-administered seasonal flu vaccines in the United States — Afluria, Flucelvax, and Fluzone — received a new warning. Section 5.4 was added to all three labels on the same date: Febrile Seizures. The warning is based on postmarketing observational studies from the 2023-2024 and 2024-2025 respiratory seasons, which found a 97% increase in relative risk for febrile seizures in the first day after standard-dose quadrivalent vaccination in children aged 6 months through 4 years, and a 194% increase after the trivalent. The Flucelvax label states the data “suggest a causal relationship.” The warning did not exist a year ago. It was added one month before this installment.

Across the nine flu vaccine labels covered in this installment, the pattern is not about one product. Every label admits the vaccine has not been evaluated for carcinogenic or mutagenic potential, or for impairment of fertility. Every label states the vaccine “may not protect all recipients.” Every label concedes that antibodies against one strain “confer limited or no protection” against another — which matters because the product is reformulated every year based on a strain guess made months before the season begins, and the current year’s reformulation is not tested for efficacy before being administered. One adult product, Fluad Quadrivalent, has been on the market under accelerated approval since 2020 on condition that continued approval “may be contingent upon verification and description of clinical benefit in a confirmatory trial.” The FluMist label reports that, in a daycare transmission study, at least one vaccine strain was isolated from 80% of FluMist recipients for a mean of 7.6 days, including one confirmed case of recovery in an unvaccinated daycare contact.

These are statements from the FDA-approved labels, written by the manufacturers, reviewed by the regulators, and filed with the agency.

This installment covers nine products — the full US-licensed flu vaccine market for the 2025-2026 season: Fluzone and Fluzone High-Dose (Sanofi), Fluarix and Flulaval (GSK), Flucelvax, Fluad, and Afluria (Seqirus), Flublok (Protein Sciences/Sanofi), and FluMist (MedImmune/AstraZeneca). The labels analyzed span formulas from 2019 through 2025-2026. Beginning with the 2024-2025 season, the FDA transitioned all US flu vaccines from quadrivalent to trivalent formulations, dropping the B/Yamagata lineage after it had not been detected in circulation since March 2020 — a reformulation that was itself implemented without new efficacy or safety trials. Where the labels cited here describe quadrivalent products, the findings carry over to the current trivalent reformulations, which have not themselves been trial-tested before market release. Everything that follows is drawn from the prescribing information for these products.

The Package Insert is a paid subscriber series. Each installment reads the FDA-approved labels for a major drug or vaccine class and reports what the documents actually say — including what they admit has not been studied. If you’re currently considering a flu shot, have a child under 5, are pregnant, or are over 65, this is the installment to start with.

Below the paywall:

• The warning added last month — the March 2026 febrile seizure data, including the age group, the time window, and the number the label describes as “suggestive of a causal relationship”

• What the vaccines are approved to do — the difference between preventing disease and producing antibodies, and why several products never had to prove the first

• The annual reformulation problem — why the product on the shelf this winter shares no efficacy data with itself

• Accelerated approval without confirmation — Fluad’s five-year gap, and the non-inferiority cascade that underlies High-Dose and Flublok

• Adverse reactions from the trials — what percentage of recipients reported what, across the nine labels

• Postmarketing reactions — Guillain-Barré, encephalomyelitis, transverse myelitis, optic neuritis, Bell’s palsy, Stevens-Johnson syndrome, thrombocytopenia, vasculitis, pericarditis, Leigh syndrome exacerbation: conditions that appeared only after approval

• FluMist as special case — the shed virus, the daycare transmission study, the 3.53 relative risk of asthma events in children under 5

• What’s in the vial — thimerosal and mercury content in the multi-dose formulation, plus beta-propiolactone, sodium taurodeoxycholate, neomycin, polymyxin B, and the residuals the labels disclose

• What the labels admit has not been studied — pregnancy, lactation, immunocompromised persons, long-term, and co-administration with other vaccines

• What the labels, read together, are saying — a plain-language synthesis of the nine-label picture, suitable for sharing with family members weighing the shot

• Questions to take to your doctor — a consolidated list at the end, formatted to print

A note on language

This series reads what the FDA-approved prescribing information actually says, in the terms the labels themselves use. Those terms come from conventional pharmacology and immunology — antibodies, immune response, autoimmune, immunogenicity, and similar. I do not accept that vocabulary as a description of what the body is actually doing, and readers of my other work know the reasons.

I use the terms in these installments for a specific reason: the strongest critique of a pharmaceutical label is almost always the label’s own words, measured by the label’s own methods, admitted by the label’s own manufacturer. Softening the terminology loses the force of the admission. When a label reports that 51% of patients developed antibodies and then admits it cannot characterize the effect of those antibodies, the reader should see both halves in the label’s own language. The measurement and the admission belong together.

This installment engages with that vocabulary more heavily than most, because the flu vaccine labels rest on germ-theory assumptions — circulating viruses, strains, transmission, antigenic drift. I report what the labels say in the labels’ own terms. Whether the framework producing the claims is the right framework is a separate question, discussed elsewhere.