CDC Finds Flu Vaccines Had Below-Average Effectiveness This Season

The 2025-2026 flu season is winding down, but the damage assessment is sobering. According to the CDC's interim vaccine effectiveness estimates published in the Morbidity and Mortality Weekly Report on March 7, 2026, this year's flu shots reduced the risk of adults needing to visit a doctor for influenza by just 22% to 34% — one of the worst performances in over a decade [1][2]. The culprit was a familiar adversary in a new disguise: an antigenically drifted variant of H3N2 known as subclade K that emerged months after vaccine formulations had been locked in.

But the subpar vaccine isn't just a story about viral evolution outrunning human prediction. It is also an early warning about the consequences of dismantling the global surveillance infrastructure that makes flu vaccines possible in the first place.

The Numbers: A Brutal Season by Any Measure

The CDC estimates that between October 1, 2025, and February 28, 2026, influenza caused at least 27 million illnesses, 350,000 hospitalizations, and 22,000 deaths in the United States [2][3]. The cumulative hospitalization rate of 78.2 per 100,000 population is the third highest recorded since the 2010-2011 season, with a peak weekly rate second only to the worst seasons on record [4].

Among the most devastating statistics: at least 101 children have died from influenza this season. Approximately 85% of those pediatric deaths with known vaccination status occurred in children who had not been fully vaccinated [4].

Source: CDC / Flu Vaccines Work

Data as of Mar 15, 2026CSV

Vaccine effectiveness varied significantly by age group and virus type. Among children and adolescents aged 6 months to 17 years, the vaccine reduced outpatient visits by 38-41% and hospitalizations by 41% [1]. For adults, the numbers were grimmer: effectiveness against outpatient visits ranged from just 22% to 34%, and against hospitalization was approximately 30% [1].

The disparity between age groups is not unusual — children's immune systems tend to respond more robustly to vaccination — but the adult numbers are particularly low even by historical standards.

The Subclade K Problem

The story of this season's vaccine shortfall begins in a WHO conference room in February 2025, when an expert panel selected the virus strains that would go into the Northern Hemisphere's 2025-2026 flu shots. The H3N2 component was based on a clade J.2 reference virus. At the time, it was the best match available [5].

Then, in June 2025, CDC scientists identified a new H3N2 variant — clade 2a.3a.1, subclade J.2.4.1, quickly dubbed "subclade K" — that carried seven new mutations, including three at key hemagglutinin sites involved in immune evasion [6][7]. By the time flu season arrived, subclade K had become the dominant circulating strain. According to the CDC's genetic characterization, 83% of H3N2 samples tested during the 2025-2026 season were subclade K — a virus antigenically distinct from the one targeted by the vaccine [1].

Laboratory studies showed the mismatch was substantial: subclade K viruses demonstrated a greater than 32-fold reduction in reactivity with antibodies raised against the egg-propagated vaccine strain, and at least an eightfold reduction against cell-propagated vaccine viruses [6].

"Although it's not a complete match with what's circulating, this vaccine appears to have the capacity to produce antibodies that will likely provide protection," said Scott Hensley, an immunologist at the University of Pennsylvania, whose team studied antibody responses in 76 vaccinated adults. Their research found that antibodies effective against subclade K specifically rose from 11% to 39% after vaccination — significant, but far below the 71% cross-reactive response against H3N2 generally [8].

Source: CDC MMWR / Interim VE Estimates 2025-26

Data as of Mar 7, 2026CSV

An Ancient Problem with No Easy Fix

The flu vaccine has always been a gamble. Unlike vaccines for measles or polio, which target stable viruses and deliver near-perfect protection, flu shots must be reformulated annually to chase a rapidly mutating target. The CDC has tracked vaccine effectiveness since the 2004-2005 season, and the overall numbers tell a humbling story: effectiveness has ranged from as low as 10% (2004-2005) to as high as 60% (2010-2011), with most seasons falling in the 30-50% range [9].

The H3N2 subtype is the perennial troublemaker. It mutates faster than H1N1 or influenza B, and the egg-based manufacturing process that produces most flu vaccines can introduce additional changes that further reduce the match between the vaccine strain and circulating viruses [10]. This season's vaccine performed notably better against influenza B, with effectiveness of 45-71% in children and 63% in adults — a stark contrast to the H3N2 numbers [1].

"The winter respiratory virus season is slowly coming to a close, and we're all very grateful for that," said Dr. William Schaffner, a Vanderbilt University vaccine expert. He emphasized that even a 30% effective vaccine prevents tens of thousands of hospitalizations and thousands of deaths — but acknowledged that the numbers are "not what we would wish for" [2].

The Surveillance Breakdown

If subclade K's emergence was bad luck, the erosion of the system designed to detect such threats was a policy choice.

An NPR investigation in November 2025 revealed that the CDC received 60% fewer flu virus samples from international sources between February and July 2025 compared to the same period the previous year. Only 12 countries submitted samples by July, down 65% from 2024 levels [11]. The decline was directly linked to President Trump's January 2025 withdrawal from the WHO, which removed approximately $1 billion from the organization's budget and disrupted the global network of collaborating centers that normally process samples from over 150 countries [11].

"When those viruses aren't coming in, we don't know what to put in the vaccine, and you're going to have less effective vaccines," warned a former CDC official quoted in the NPR report [11].

The problems extended domestically as well. CIDRAP reported that the CDC stopped posting standard weekly respiratory illness surveillance data in September 2025, with no clear timeline for resumption — a blackout that left public health officials in the dark as flu season began [12].

"This is not the time to be flying blind into the respiratory virus season," said Dr. Danuta Skowronski of the BC Centre for Disease Control [12]. Dr. Joshua Petrie of the Marshfield Clinic echoed the concern: "The influenza season can take off quickly, and it might sneak up on us if we aren't watching nationally" [12].

Policy Shifts and Vaccination Rates

The surveillance crisis unfolded alongside a notable shift in federal vaccination policy. In January 2026, the Trump administration stopped broadly recommending flu shots for all children, saying instead that vaccination decisions should be left to parents and family physicians [2].

The policy change appears to have had a measurable effect. Childhood vaccination coverage dropped to 48% this season, down from 52% in 2024 [2]. Adult vaccination rates held roughly steady at 46.5% [2]. The lower childhood vaccination rate is particularly concerning given that 85% of pediatric flu deaths this season occurred among children who were not fully vaccinated.

What Comes Next: The 2026-2027 Vaccine

On February 27, 2026, the WHO issued its strain recommendations for the 2026-2027 Northern Hemisphere flu season, directly addressing the subclade K problem. The new H3N2 component will be based on A/Darwin/1454/2025, a subclade K reference virus — a move that should significantly improve the vaccine's match to circulating strains if subclade K continues to dominate [13][14].

The FDA's Vaccines and Related Biological Products Advisory Committee has already voted to recommend this update for U.S. vaccines [15].

But some experts caution that catching up to last year's virus provides no guarantee against next year's surprises. H3N2's rapid mutation rate means that a new drift variant could emerge at any time, restarting the cycle.

The Promise — and Uncertainty — of Next-Generation Vaccines

The recurring mismatch problem has intensified interest in technologies that could break the annual prediction cycle. In February 2026, the FDA agreed to review Moderna's mRNA-based flu vaccine after initially raising concerns about its efficacy in older adults [16]. The mRNA platform's key advantage is speed: vaccines can be manufactured far more quickly than traditional egg-based shots, potentially allowing mid-season updates when mismatches are detected.

Meanwhile, researchers at the NIH's Vaccine Research Center are conducting Phase 1 trials of a universal flu vaccine candidate designed to provide broad protection against multiple influenza strains [17]. Animal studies have shown promise, with mRNA vaccines encoding 20 different influenza types providing protection even against strains not included in the formulation.

A European regulatory body has also recommended Moderna's combination COVID-influenza mRNA vaccine for adults over 50 — a potential milestone in simplifying respiratory virus protection [16].

These technologies remain years from widespread deployment. For now, the world continues to rely on a prediction-based system that, as this season starkly demonstrates, is only as good as the data feeding it.

The Bigger Picture

The 2025-2026 flu season illustrates a troubling convergence: a virus that evolved faster than expected, a vaccine system structurally constrained by decades-old manufacturing timelines, and a surveillance apparatus weakened by political decisions made thousands of miles from any laboratory.

The season's 22,000 estimated deaths are not an abstraction. Each represents a failure of a system that, even in its best years, offers only partial protection. The question facing public health officials is whether the infrastructure necessary to make even that partial protection possible will survive intact into the next flu season — and the one after that.

For the 2026-2027 season, the WHO's updated strain recommendations offer reason for cautious optimism. But as Arnold Monto, an epidemiologist at the University of Michigan, observed about the reduced sample submissions: the fewer data points informing these decisions, "the less confident we can be that we're making the right call" [11].