Two More Cautionary Raw Pet Food Studies To Ponder

Latest alert issued Sept 25th

#18,895

Over the summer of 2023 we saw two high-profile H5N1 outbreaks in domestic cats; one in Poland and the other in South Korea. While the exposures in Poland were inconclusive, the South Korean outbreak was quickly linked to contaminated cat food.

Over the past year we've seen numerous recalls and/or official warnings over cats sickened (or killed) by HPAI H5N1 exposure through raw cat food (see here, here, here, here, here here, here and here),

Despite long-standing warnings from veterinarians and public health officials over the dangers of feeding pets a raw food diet (see FDA Get the Facts! Raw Pet Food Diets can be Dangerous to You and Your Pet), these products remain on the shelf, and people continue to buy them.

Last January U.S. FDA announced that manufacturers of cat and dog foods need to reanalyze their food safety plans to include HPAI viruses (specifically H5N1) as a known or reasonably foreseeable hazard, but it provided no specific guidance.

Under the PCAF requirements, animal food businesses must conduct a reanalysis of their food safety plan when the FDA determines it is necessary to respond to new hazards and developments in scientific understanding.

This is not a new concern, and nor is it limited to HPAI viruses.  

In March of 2024 the Canadian Public Health Agency finalized their investigation of a multi-year outbreak of extensively drug-resistant Salmonella infections linked to raw pet food and contact with cattle:

Investigation summary

In total, 44 laboratory-confirmed cases of XDR Salmonella I 4,[5],12:i:- illness were linked to this outbreak in six provinces: Manitoba (1), Ontario (15), Quebec (24), New Brunswick (1), Nova Scotia (2), and Prince Edward Island (1).

Individuals became sick between July 2020 and January 2024. Thirteen individuals were hospitalized. No deaths were reported. Individuals who became ill were between 0 and 91 years of age. Many of the cases (43%) were in children 5 years of age or younger. About half of the cases (55%) were female. Some illnesses associated with this outbreak strain were difficult to treat with commonly recommended antibiotics.

Using a laboratory method called whole genome sequencing, it was determined that some Salmonella illnesses dating back to 2020 were caused by the same outbreak strain as the illnesses that occurred in 2023 and 2024.

The outbreak strain of Salmonella that made people sick was found in raw pet food from the home of an ill individual. The outbreak strain of Salmonella was also found in sick dogs and cattle, and some of these animals died.

The PHAC warned:

PHAC does not recommend feeding raw pet food to pets, especially in households with young children or individuals who have a compromised immune system, as they are at greater risk for more serious illness. Harmful germs such as Salmonella and E.coli have been found in commercially prepared raw pet food and treats and can be found in many raw meats and products used in homemade diets.  

Repeatedly (see 2011's Raw food diets in companion animals: A critical review) and (2019's Raw diets for dogs and cats: a review, with particular reference to microbiological hazards) vet journals have warned of the risks, yet the sale of raw pet food products continues to grow. 

Today we can add 2 more studies to the mix; one on a different avian flu risk (LPAI H9N2), and another looking at bacterial risks (Salmonella, Cronobacter and E. coli, etc.) in raw, partially cooked, or freeze-dried commercial pet food.

The first comes from South Korea, which reports finding evidence of viable LPAI H9N2 in commercial pet food. While based on a limited sampling (62 Raw meat-based diets (RMBDs)), investigators were able to isolate two live H9N2 viruses.  

Detection of H9N2 Avian Influenza Virus in Commercial Raw Meat-based Diets for Dogs and Cats in South Korea

Chang-Tae Kim a b 1, Dong-Wook Lee c 1, Da-Won Kim c, Ji-Yun Kim c, Ye-Jin Lee c, Hee-Su Kim a b, Seulgi Bae a, Jung-Hoon Kwon c

This is particularly concerning given the zoonotic potential of LPAI H9N2 (see Nature: Genetic diversity of H9N2 avian influenza viruses in poultry across China and implications for zoonotic transmission). Due to its copyright status, I've only provided the link. Click through for additional details. 

Next up, a detailed report on the risks of commercially sold raw, or freeze-dried, pet foods published last month in Communications Biology. 

Briefly, researchers from Cornell University analyzed 112 commercial cat foods—raw (frozen, freeze-dried, refrigerated) vs. conventional (canned, kibble) - using both DNA sequencing and culture-based methods.

They found a veritable smorgasbord of pathogens in the `raw, frozen, and freeze dried' products, including Salmonella, E. coli, Klebsiella, Enterobacter, Clostridium, Cronobacter, and carbapenem-resistant Pseudomonas and Stenotrophomonas.

None of the canned food, or kibble, yielded live pathogens, yet their detection was widespread across a variety of raw and freeze-dried products they tested.  First, a link and the abstract, after which I'll have a bit more.

Analysis of the microbiota of raw commercial feline diets to prioritize food safety investigations

Guillaume Reboul, Aaron C. Malkowski, Y. Tina Yu, Yunman M. Gu, Kelly L. Sams, J. Marie Umbarger, Rebecca J. Franklin-Guild, Yuhan Jin, Zhiwei Chen, Bryce J. Stanhope, Breanna R. Wendel & Laura B. Goodman

Communications Biology volume 8, Article number: 1349 (2025) 

Abstract

Microbiota sharing between people and their companion animals is a concern for development of antimicrobial resistance. To assess the risks associated with feeding raw products to cats, with an emphasis on previously understudied freeze-dried products, a collection of 112 conventional and raw products was purchased and investigated using a combination of cultivation and high-throughput sequencing techniques.

Here we show that bacterial cultures were exclusively isolated from raw foods. A total of 19 genera were cultured including Salmonella, Clostridium, Escherichia, Klebsiella, Enterobacter, and Cronobacter. Carbapenem-resistant Pseudomonas aeruginosa and Pseudomonas fulva, and Stenotrophomonas lactitubi were isolated from frozen raw products, and 6 Bacillus strains harbored carbapenemase gene bla2. Multidrug efflux pumps were highly abundant in frozen raw isolates.

Clostridium sensu stricto I genus detection predicted a raw, freeze-dried product with 95% sensitivity and 78% specificity. Genera Pseudomonas, Paraclostridium and Peptostreptococcus were associated with frozen raw food products while the Bacillus genus was associated with conventional processing. Parasite genes were exclusively detected in raw foods. 

The presence of pathogenic species and high load of resistance genes in raw commercial food products, particularly those sold on shelves at room temperature, suggests a considerable health risk to cats and the families who care for them.

        (Continue . . . )

You'll also find an accompanying press release in the Cornell Chronicle. Follow the links to read both in their entirety.

Adding insult to injury, using Mitochondrial DNA Fingerprinting, the researchers determined that the ingredients in some of the `raw or freeze-dried' products did not match their labels.  

• `Beef only' labeled products sometimes carried DNA for poultry, or vice versa. 
• Both raw and freeze-dried diets were mislabeled
• No such discrepancies were reported in canned or kibble diets

This matters because if there were a zoonotic outbreak, it could send investigators initially down the wrong rabbit hole.   

Unfortunately, we live in a world where internet influencers and slick advertising campaigns extolling the `health benefits' of a raw pet food diet carry more weight than decades of scientific research. 

Which makes me confident this won't be the last time we are forced to revisit this topic.

Preprint: The Potential of H5N1 Viruses to Adapt to Bovine Cells Varies Throughout Evolution

 
Phenotypes of the Reassortant viruses observed in this study

#18,894

Over the past 3 years we've seen a dizzying array of reports and studies showing that many clade 2.3.4.4b H5Nx viruses circulating today exhibit vastly different behaviors from the viruses that circulated prior to 2022.   

A few of many ominous changes include:

• First, we've seen numerous spillovers into mammals, often presenting with severe neurological manifestations (see Cell: The Neuropathogenesis of HPAI H5Nx Viruses in Mammalian Species Including Humans)

• Second, tens of thousands of marine mammals - in North and South America, and Europe - have been killed (see Nature Comms: Cross-species and mammal-to-mammal transmission of clade 2.3.4.4b HPAI A/H5N1 with PB2 adaptations)

• Third, nearly 1,100 dairy herds - and a smattering of other ruminants - have been infected with HPAI H5N1

• Fourth, the CDC lists 70 confirmed and 7 probable human H5N1 infections since the spring of 2024, most - but not all - with exposure to either infected cattle or poultry. 

• And hundreds of cats - both household pets and strays - have succumbed (see One Health: Outbreak of HPAI a(H5N1) Among House Cats: A Case Series Involving Oseltamivir Treatment

HPAI H5N1 is not a single viral strain, as there are literally hundreds of genotypes circulating around the world.  Some are more dangerous than others. But exactly what makes one genotype more dangerous - while others are relatively benign - remains poorly understood. 

Regular readers know we often look for a handful of mutations (e.g. PB2 mutations like E627K, D701N, Q591K, and M631L and HA mutations like Q226L and E190D) which may favor mammalian adaptation, but new ones (see Sci. Adv.: PB2 and NP of North American H5N1 Virus Drive Immune Cell Replication and Systemic Infections) continue to be discovered. 

Today we have a fascinating preprint (not yet peer reviewed) from Universities and Institutes in the UK, the Netherlands, and Italy which may end up changing the way we look for H5N1 adaptation signatures. 

Fair warning, this is a lengthy (60 page) and highly technical report, and I can only cover the high points.  First the link, and abstract, after which I'll try to break out some of the highlights. 

The potential of H5N1 viruses to adapt to bovine cells varies throughout evolution

Matthew L Turnbull, Mohammad Khalid Zakaria, Nicole S Upfold, Siddharth Bakshi, Callum Magill, Udeet Ranjan Das, Andrew T Clarke, Laura Mojsiejczuk, Vanessa Herder, Kieran Dee, Nancy Liu, Monika Folwarczna, Georgios Ilia, Wilhelm Furnon, Verena Schultz, Hanting Chen, Ryan Devlin, Jack McCowan, Alex L Young, Wai-Wai Po, Katherine Smollett, Muhammad Ahsan Yaseen, Rebecca Ros, Avanti Bhide, Bianca van Kekem, Ron Fouchier, Ana Da Silva Filipe, Munir Iqbal, Ed Roberts, Joseph Hughes, Dirk Werling, Pablo R Murcia, Massimo Palmarini
doi: https://doi.org/10.1101/2024.11.29.626120
This article is a preprint and has not been certified by peer review [what does this mean?].

Preview PDF

Abstract

Avian influenza H5N1 clade 2.3.4.4b viruses caused a global panzootic and, unexpectedly, widespread outbreaks in dairy cattle, therefore representing a pandemic threat. To inform effective control strategies, it is critical to determine whether the potential to adapt to bovine cells is a generalised feature of H5N1 viruses, or is specific to clade 2.3.4.4b, or even more restricted to specific genotypes within this clade (e.g., B3.13 and D1.1).

Using a large panel of H5N1 viruses representing >60 years of their natural history and other IAV for comparative purposes, we demonstrate that virus adaptation to bovine cells is: (i) highly variable across 2.3.4.4b genotypes, (ii) limited in viruses predating the global expansion of this clade, (iii) determined by the viral internal gene cassette, and (iv) not restricted to udder epithelial cells.

Mutations in the PB2 polymerase subunit, particularly M631L, emerge as key determinants of adaptation, although their phenotypic effects are context dependent and have limited enhanced viral polymerase activity in human cells.

Bovine B3.13 and some avian genotypes also exhibit enhanced modulation of bovine interferon-induced antiviral responses, determined by at least the viral PB2, nucleoprotein, and the non-structural protein NS1.

Our results highlight the polygenic nature of IAV host range and reveal that the potential to cross the species barrier varies during the evolutionary trajectory of H5N1, with some avian viruses more predisposed to spillover than others.

        (Continue . . . )

Essentially this study looked at whether HPAI H5N1 viruses have always had the ability to infect cattle, or if something had recently changed to enable that capability. 

They used reverse genetics to generate an array of reassortant viruses (using the HA and NA from  H1N1 Puerto Rico/8/34 (PR8)) with the internal genes from 12 different H5N1 viruses going back 60 years.  

Using bovine cell models they showed that only viruses with the internal cassette from recent H5N1 viruses (particularly genotypes B3.13 and D1.1) replicated efficiently in bovine respiratory, skin, and udder cells.

This adaptation was not driven by a single mutation, but rather by a series of changes across multiple internal gene segments (PB2, PB1, NP, and NS), which were introduced by H5's reassortment with North American LPAI viruses. 

Previously known mutations - like PB2 M631L - enhanced replication but were only fully effective when combined with these recently emerged internal gene cassettes.

These findings suggest that mammalian adaptation in H5N1 appears highly depended on permissive-compensatory changes across multiple internal gene segments - making our current surveillance based on a handful of `signature' mutations insufficient. 

It also suggest that H5's arrival in North America - and its subsequent reassortment with N.A. LPAI viruses - signaled a major turning point in its evolution. 

I've only covered the less technical highlights of this preprint, so follow the link to read it in its entirety.  

None of this means that a B3.13 or D1.1 pandemic is inevitable, only that the North American branch of the H5N1 virus has become better adapted to mammalian hosts than its Eurasian cousins. 

Where H5 goes from here is unknown.  It could fizzle - or run into an insurmountable species barrier.

But we continue to treat this virus as if it is the same pathogen that tried - and failed - to spark a pandemic 20 years ago.  It isn't, and what we may be faced with six months - or six years - from now, will almost certainly have changed even more.

A reminder that evolution never stops, and Nature always bats last.

Avian Flu's Unusually Active September in the United States

 

#18,893

Until fairly recently, HPAI H5 infections in wild birds - and spillovers into poultry - were pretty much a fall and winter phenomenon. Carried, and distributed primarily by migratory waterfowl, their regular `ride' retreats to their high latitude roosting areas by late spring.

Even after our first HPAI H5 epizootic in 2014-15, the virus disappeared abruptly in mid-June, and did not reappear in North America again until 2021 (see PNAS: The Enigma Of Disappearing HPAI H5 In North American Migratory Waterfowl).

Local wild birds, it seemed - either died when infected - or survived and gained (or had) immunity.  Either way, they were unable to provide a summer `home' for the virus.  Warmer temperatures were also thought to dampen down transmission. 

By the time the HPAI H5 virus returned to North America in late 2021, it had undergone many changes.

It was H5N1 not H5N8.  It had reassorted numerous times, and had not only expanded its avian host range in Europe, it was beginning to persist (at low levels) throughout the summer. 

During its first full year in North America, we saw continued outbreaks over the summer of 2022, with a major return of the virus in September (68 outbreaks affecting > 6 million birds). 

This persistence and early arrival were likely due to new reassortments with North American LPAI viruses (see Rapid Evolution of A(H5N1) Influenza Viruses After Intercontinental Spread to North America), and biosecurity lapses on backyard and commercial poultry operations.

In 2023 and 2024, we saw far less avian flu activity over the summer months (see chart below), with outbreaks often not starting up again until October. Notice how `dead' June-September were during 2023-2024, and the circled uptick for Sept 2025.

A little over 2 weeks ago I reported on the unusually early uptick in poultry farm infections with HPAI during the first half of September (see Canada & U.S. Report Early Fall Uptick In H5N1 Outbreaks In Poultry). 

Since then, the number of U.S. poultry involved has climbed 10-fold; from 348K to 3.83 million.  

As Hogvet51 aptly pointed out in his blog yesterday (see 20 H5 Commercial Flock Infections in September, but Nobody Can Share a Genotype...), we've no idea what genotypes have returned this fall.  Whatever the NVSL lab knows, they haven't been quick to share.

As we've seen, the fall migration can often bring abrupt changes.

• Two years ago (fall/winter 2023) we saw a new H5N1 B3.13 genotype emerge that could efficiently infect dairy cattle, while mildly infecting dozens of humans. Although not as well adapted to poultry, some spillovers were reported.

• Last fall (2024), we saw 3 new genotypes emerge (D1.1, D1.2, D1.3), with D1.1 producing more severe illness in humans, while last October Canada experienced a multi-farm outbreak  that was highly resistant to NAI antivirals.

With the U.S. government in shutdown, it remains to be seen how much data we'll be getting from the USDA in the days and weeks ahead. But even under `normal' circumstances, the release of detailed HPAI information has often come at a glacial pace.

Last March, in Nature: Lengthy Delays in H5N1 Genome Submissions to GISAID, we learned that the average delay for countries to submit non-human sequences was 7 months.

This lack of data extends to reports of mammalian infections as well, as only a handful of states (6) have found, tested, and reported incidents over the past few months (8 detections over June-Sept).

Looking back over the past 4 years, we can see a decided drop in reporting to the USDA by states since March of 2025. 

While none of this guarantees we are in for a wild HPAI ride this fall and winter - with this level of testing and reporting - a viral storm could easily be brewing, and you and I probably wouldn't see it coming.  

Brace yourself accordingly.  

Cell Reports: Dairy cow- and Avian-origin Clade 2.3.4.4b H5N1 Induce Severe Mastitis in Lactating Goats and Transmission to Suckling Goats

 

#18,892

Although quickly overshadowed by the news of the spillover of HPAI H5N1 into Texas and Kansas dairy herds (announced only 3 days later), just over 18 months ago we saw the first confirmed infection of goats with H5N1 in the United States (see Minnesota BOAH Statement On HPAI H5N1 Infected Goat Kids).

These goats were being raised on a `mixed species' farm, which was then experiencing an outbreak of H5N1 in ducks and chickens.  

These were the first detections of clade 2.3.4.4b H5N1 in ruminants in the United States, although cattle have previously been successfully infected in the laboratory with older clades (see A Brief History Of Influenza A In Cattle/Ruminants).

As the USDA explained, this spillover was from a different genotype (B3.6) than was affecting diary cattle (B3.13).

Has HPAI ever been detected in small ruminants?

In March 2024, USDA’s National Veterinary Services Laboratories (NVSL) confirmed HPAI in very young goat kids on a Minnesota farm that also had poultry infected with HPAI. The goats shared the same pasture and water source with the infected chickens and ducks before they were depopulated, while kidding. This situation led to the newborn goats being exposed to high levels of virus. Adult goats on the premises tested negative for the virus.

NVSL identified the virus infecting the goats and poultry on the farm was B3.6. This is a very common genotype that is circulating in our North American wild bird flyways and has sporadically infected poultry flocks in 2023 and 2024. The virus causing the disease in dairy cattle is B3.13. 

Two months later we learned that Alpacas - again on a `mixed-species farm'  - in Idaho, had been infected with HPAI H5N1 (see USDA: HPAI H5N1 Detected In Alpacas).  

This time, however, it was from the `bovine' B3.13 genotype, which had spilled over into the farm's poultry. 

Since then we've seen serological evidence of HPAI exposure in goats and sheep in Pakistan, last March the UK's Defra reported H5N1 Detected In Domestic Sheep with Mastitis, and in May; serological evidence of H5N1 in sheep in Norway.

Despite these findings, surveillance for HPAI H5 in cattle, sheep, goats, and other livestock is passive; meaning mild or asymptomatic carriage is unlikely to be detected.

Even when animals are symptomatic, many farmers have been reluctant to allow testing, fearing quarantines and financial losses. `Don't Test, Don't Tell' remains a popular option. 

Although nearly 1,100 herds have been reported as infected since March of 2024, the USDA's Dairy Herd Status Program website only shows 124 herds (out of an estimated 36,000) from 21 states enrolled in the voluntary herd monitoring program.  

For nearly a year it was assumed that only the `bovine' B3.13 genotype was capable of infecting dairy cows. But in February of this year two states reported outbreaks of genotype D.1 in dairy cows (note: we've also seen genotype D1.2 in pigs).  

Recent studies (see Nature: A Mathematical Model of H5N1 Influenza Transmission in US Dairy Cattle) suggest that significant under-reporting of H5N1 in dairy herds is likely.

To put it kindly, in the United States - and around the globe - surveillance and testing of livestock is far from optimal. Spillovers that were once thought unlikely, or are now considered rare, could actually be far more common than we know. 

All of which brings us to a recent study, published in Cell Reports, which finds that goats are highly susceptible to HPAI H5N1, and that they can transmit it to suckling goats. 

Due to its length, and technical nature, I've only reproduced the abstract and some excerpts. Follow the link to read it in its entirety.  I'll have a bit more after you return.

Dairy cow- and avian-origin clade 2.3.4.4b H5N1 induce severe mastitis in lactating goats and transmission to suckling goats
Tamiru N. Alkie1 ∙ Carissa Embury-Hyatt1 ∙ Anthony V. Signore1 ∙ … ∙ Samira Mubareka7,8 ∙ Richard Webby9 ∙ Yohannes Berhane yohannes.berhane@inspection.gc.ca …

Download PDF

Highlights

• Cow- and avian-origin A(H5N1) caused severe mastitis in goats 

• Infected goats shed viruses in milk, transmitting A(H5N1) to suckling kids

• Both viruses infected goat respiratory and mammary cells, which express sialic acid receptors

• Findings show underestimated A(H5N1) risk in ruminants, urging better livestock surveillance

Summary

The emergence of highly pathogenic avian influenza (HPAI) H5N1 clade 2.3.4.4b has led to unprecedented spillover and spread among US dairy cows, raising concerns about transmission to other ruminants.

We inoculated two groups of lactating goats via intramammary and respiratory routes with Cow-H5N1 (genotype B3.13) or avian-H5N1 (genotype B1.2) virus. Both groups developed severe clinical mastitis and shed viruses in milk, resulting in transmission to suckling kids. 

Viral RNA was detected in nasal and oral swabs and various tissues, and virus-neutralizing antibodies were present in serum, milk, and bronchoalveolar lavage fluid. In vitro, both viruses replicated efficiently in goat respiratory and mammary epithelial cells. Mammary tissue expresses both α2,3- and α2,6-linked sialic acid receptors.

These findings demonstrate that goats are highly susceptible to H5N1 infection, with mammary tropism facilitating transmission to offspring, and underscore the need for increased surveillance in ruminant livestock.

       (SNIP)

The detection of higher titers of infectious virus in milk from naturally infected dairy cows1 raises concerns regarding the potential zoonotic transmission through the consumption of unpasteurized milk. Standard pasteurization temperature effectively inactivate the A(H5N1) virus in milk.45,46 However, experimental studies involving raw infected cow milk in animal models have shown severe outcomes.47,48

Goat milk is an essential part of the diets in many Asian and European countries, particularly in India, Pakistan, Bangladesh, Turkey, and Italy, which are major producers and consumers, even though it is not consumed in large quantities worldwide. Acutely A(H5N1)-infected lactating goats produce milk with high virus loads that can spread to vulnerable animals or infect humans who drink unpasteurized milk.

Additionally, handling infected goats may present occupational health risks.

As the outbreaks of HPAI in dairy cows continue and new genotypes with mammalian adaptive mutations emerge, the inclusion of lactating goats or sheep in HPAI surveillance programs seems sensible. The tropism of clade 2.3.4.4b viruses for mammary glands is likely due to the expression of specific sialic acid receptors. The duplicability of clinical disease in lactating goats after A(H5N1) infection suggests their potential use as a model for studying HPAI pathogenesis and vaccine efficacy testing against these viruses. Goats offer several advantages, including feasibility and relatively shorter gestation periods.   

(Continue . . .)

Despite constant calls for more surveillance, aggressive testing, and the timely sharing of information (see here, here, here, here, here, here, and here), there are few signs that these pleas are being taken seriously by most governments, agencies, or stakeholders.

The world continues to treat the spread of HPAI as more of an economic or political concern, than a potential public health threat.

And while that optimistic assessment may hold true today, it could abruptly change with little warning. 

NPJ Vaccines: Effects of Repeated influenza Vaccination and Infection on Durable Seroprotection in Healthcare Workers

 

#18,891

I'll be rolling up my sleeve for my 20th consecutive seasonal flu shot next week, but I am keenly aware that there are some studies suggesting that with repeated flu vaccinations, comes the risk of mounting a diminished immune response.

This idea made headlines a decade ago (see Helen Branswell's report Getting a flu shot every year? More may not be better), following a preliminary report from Wisconsin’s Marshfield Clinic Research Foundation on the effectiveness of repeated flu vaccination in kids.

A similar report appeared earlier in the CMAJ (see Repeated flu shots may blunt effectiveness). However, these assumptions were based on limited data, during a time when the H3N2 subtype was undergoing rapid changes (see (CDC HAN Advisory On `Drifted’ H3N2 Seasonal Flu Virus).

Three years later (2018), the Marshfield Clinic released a follow-up report (below) which supported the receipt of the flu vaccine every year. 

Research Institute study supports annual flu vaccination for kids

Friday, October 26, 2018

A Marshfield Clinic Research Institute study recently found that getting a flu vaccination one year did not reduce vaccine effectiveness the next year in children, findings that support current recommendations for children to be vaccinated annually against influenza.

The study was conducted by Huong McLean, Ph.D., and Edward Belongia, M.D., of the Center for Clinical Epidemiology and Population Health at Marshfield Clinic Research Institute, and was published in JAMA Network Open.
(Continue . . . )

Also that year, in CMAJ Research: Repeated Flu Vaccinations Reduce Severity of Illness In Elderly, we looked at a study that found that repeated vaccinations over two or more years cut the rate of severe influenza illness in half.

In 2019, a systematic review of the literature (The impact of repeated vaccination on influenza vaccine effectiveness: a systematic review and meta-analysis) lent additional support to the idea of getting the flu vaccine every year, despite some decline in VE among people receiving back-to-back yearly flu vaccinations. 

Reassuringly, 2022's ISIRV: Immune Response From Repeated Seasonal Flu Vaccinations concluded: `Repeated vaccination provides similar or enhanced protection as compared with single vaccination in first-time vaccinees.'

Today we  have yet another study, which looked at 250 healthcare workers (median age ~38 &~78% female) from Norway over 4 consecutive (2010/11–2013/14) influenza seasons, who were divided into 4 cohorts based on vaccine receipt:

• Current & Previous Vaccinated in the current season and previous season
• Current Only Vaccinated in the current season but not the previous season
• Previous Only Vaccinated in the previous season but skipped vaccination in the current
• Unvaccinated Received no vaccine in either current or prior season

Participants were reclassified each year, depending upon vaccine receipt. All had previously received the AS03-adjuvanted pandemic A(H1N1)pdm09 vaccine in 2009. 

Key findings included:

• Unvaccinated healthcare workers showed high influenza infection rates—up to 72% infected over just four seasons.

• Vaccination significantly boosted protective antibody levels across all subtypes.

• A/H1N1 immunity grew stronger with repeated vaccination; antibody durability improves after 3+ doses.

• A/H3N2 immunity grew weaker with repeated vaccinations - with natural infection or hybrid immunity producing stronger protection.

• While repeated A/H3N2 vaccination instilled some degree of seroprotective immunity (1) Antibody levels were significantly lower compared to H1N1, and (2) Protection waned more quickly compared to H1N1

This is a lengthy, detailed, and at time technical report, so I've only reproduced the abstract and some excerpts from the conclusion below. Follow the link for a deeper dive. 

I'll have a postscript after you return.

Published: 29 September 2025
Effects of repeated influenza vaccination and infection on durable seroprotection in healthcare workers

Mai-Chi Trieu, Amit Bansal, Marianne Sævik, Sonja Ljostveit, Åsne Jul-Larsen & Rebecca Jane Cox 
npj Vaccines volume 10, Article number: 213 (2025) Cite this article

Abstract

Rapid evolution of seasonal influenza viruses necessitates annual vaccine reformulation to match circulating strains. Healthcare workers (HCWs) and high-risk groups are prioritised for annual influenza vaccination. However, repeated annual vaccination may affect immune protection. This study investigated the hemaglutination-inhibition (HI) antibody responses following influenza infection or repeated seasonal vaccination over four seasons in 250 HCWs with well-defined vaccination histories.

Unvaccinated HCWs had high infection rates, with pre-existing antibodies providing protection. Infection or hybrid immunity generated higher antibody responses to A/H3N2 viruses than vaccination alone, whereas vaccination induced more durable A/H1N1 and B virus-specific antibodies. Vaccination boosted seroprotective antibodies, irrespective of previous vaccination histories. Moreover, repeated vaccination with the same virus for more than three consecutive seasons blunted antibody responses, while updating vaccine strains improved immunity.

Annual influenza vaccination of HCWs should be strengthened to increase uptake, but next-generation influenza vaccines must improve vaccine immunogenicity, particularly against A/H3N2 viruses.

        (SNIP)

In conclusion, seasonal influenza vaccination was effective in preventing influenza infection and boosting antibodies to above seroprotective levels in HCWs.

Repeated annual vaccination increased antibody persistence, which correlated with lower antibody responses to subsequent vaccinations. Repeated vaccinations with the same viruses more than thrice blunted antibody responses to vaccines, whereas updating the vaccine viruses likely improved antibody responses to vaccines

Our findings highlight the need to improve vaccine immunogenicity, particularly against A/H3N2 viruses, and demonstrate that repeated annual vaccination may have complex effects on antibody responses. More research is required to conclusively determine whether repeated annual vaccination has any negative effect on protection and the mechanism behind. 

We support continuing the current recommendation of annual influenza vaccination in HCWs. However, more refined vaccine strategies are essential to mitigate any potential negative effects of repeated vaccination while maintaining the benefits of seasonal vaccines.

        (Continue  .  .  .)

While this is a single study, and it deals with a panoply of influenza viruses from nearly 15 years ago, it provides additional reassurance on the practice of getting an annual flu vaccine. 

It also reminds us that while `good', today's flu vaccines are not great. 

They don't protect as well, or as broadly, as we'd like. They are most useful in blunting the effects of infection, rather than preventing it (see CDC Hopes to `Reset' Flu Vaccine Expectations With New Campaign).

The human immune system is stunningly complex, and only partially understood. Individual responses to vaccination can vary depending upon age, health, and past (and first) influenza exposures (see Nature: Declan Butler On How Your First Bout Of Flu Leaves A Lasting Impression).

Add in rapidly evolving Influenza A viruses, with H3N2 being particularly labile (see The Enigmatic, Problematic H3N2 Influenza Virus), it is little wonder that decades of promises that a `universal flu vaccine' was close, have fallen short. 

But just as seatbelts don't guarantee you'll walk away unscathed from a collision, current seasonal flu vaccine can markedly reduce your chances of a severe infection. 

So, while we wait for a `better' vaccine solution - for me, at least - some protection is better than none. 

Taiwan CDC Statement On Increased Antiviral Resistance in Seasonal H1N1 Cases in 2025

Credit NIAID

 #18,890

While spontaneous `resistance' to oseltamivir (aka `Tamiflu') develops in roughly 1%-2% of treated patients, for the past 18 months we've been watching a slow increase in the number of antiviral resistant H1N1 viruses around the world.

These resistant viruses, however, are generally thought to be less biologically `fit' than their susceptible counterparts, and are only rarely transmitted to others. 

Of course, that can always change - as we saw in 2008 - when in the course of a less than a year the (pre-pandemic) H1N1 virus defied all expectations and went from being 99% susceptible to oseltamivir to 99% resistant (see CIDRAP article With H1N1 resistance, CDC changes advice on flu drugs). 

While an influenza antiviral crisis seemed unavoidable, in an unlikely Deus Ex Machina moment a new swine-origin H1N1 virus - one that happened to retain its sensitivity to Tamiflu - swooped in as a pandemic strain in the spring of 2009, supplanting the older resistant H1N1 virus.

Since 2009 we've been closely watching for signs of increased antiviral resistance in seasonal flu - and while a few isolated pockets have occurred - none have established a foothold. 

• Four years ago, in EID Journal: Cluster of Oseltamivir-Resistant & Antigenically Drifted Influenza A(H1N1)pdm09 Viruses, Texas, USA, January 2020, we looked at a cluster of resistant seasonal flu viruses which emerged just before COVID shut down influenza around the globe.

• In 2016, in Eurosurveilance: A(H1N1)pdm09 Virus With Cross-Resistance To Oseltamivir & Peramivir - Japan, March 2016 we looked at an elevated number of NAI resistant viruses with `permissive mutations' circulating in Japan.

• In 2014's Eurosurveillance: Community Cluster Of Antiviral Resistant pH1N1 in Japan, we looked at a cluster of six genetically similar resistant viruses in Sapporo, Japan  - but without epidemiological links. 

But in March of 2024 we saw a worrisome report in The Lancet - Global Emergence of Neuraminidase Inhibitor-Resistant Influenza A(H1N1)pdm09 Viruses with I223V and S247N Mutations - which reported a much higher incidence of oseltamivir resistance among samples tested in Hong Kong in 2023.

Unlike the near total collapse of oseltamivir effectivess in 2008 - which was due to a an H275Y mutation in the NA - this was due to  I223V and S247N, and was not as profound of a loss in effectiveness. 

Four months later (see Viruses: Increase of Synergistic Secondary Antiviral Mutations in the Evolution of A(H1N1)pdm09 Influenza Virus Neuraminidases) researchers at the Robert Koch-Institute that reported seeing an uptick in permissive secondary mutations (NA-V241I and NA-N369K) in the NA or circulating H1N1 viruses that they believe may enable resistant strains to transmit more efficiently.

Earlier this summer we looked at a report from Spain which reported an increase in the number of the NA-S247N mutation in seasonal flu viruses collected in Spain over the 2023-2024 flu season (see Virus Research: A 15-year Study of Neuraminidase Mutations and the Increasing of S247N Mutation in Spain).

Today, Taiwan's CDC is responding to reports of a substantial increase in H1N1 resistant viruses in 2025.  While it doesn't report causitive genetic changes (e.g. H275Y or I223V S247N) - or the degree of resistance - it does report an incidence of 6.5% in 2025. 

Whether this is a localized fluke - or an early indication of  a trend - is to early to say.  Obviously we'll be watching for any similar reports as this year's flu season progresses.

I've posted a translation below.

The CDC responded to media reports about the emergence of resistance to influenza antiviral drugs, stating that the resistance rate is still low and limited to influenza A (H1N1). The current epidemic has shifted to influenza A (H3N2), and the impact is assessed to be limited. 
 
Release Date：2025-09-29 Regarding today's (29th) media report that a frontline physician group reported that "the incidence of resistance to oral and injectable influenza antiviral drugs has increased, resulting in a decrease in efficacy", the Centers for Disease Control (CDC) explained as follows:

1. According to CDC monitoring data, approximately 6.5% of the A H1N1 virus strains tested this year (2025) have resistance-related mutations to the oral antiviral drug Tamiflu or oseltamivir and the injectable antiviral drug Peramivir, which is slightly higher than last year (the statistics at the end of last year were 3.2%). No resistance to the above-mentioned drugs was detected in A H3N2 and B influenza virus strains this year. In addition, no resistance-related mutations were detected for the oral self-paid antiviral drug Baloxavir this year, regardless of A H1N1, A H3N2 or B influenza.

Second, the CDC assesses that the aforementioned drug resistance rate remains low and is limited to influenza A (H1N1). Furthermore, since entering the influenza season in Taiwan in September, the prevalence of community-circulating influenza viruses has recently shifted to influenza A (H3N2), which has not detected drug resistance (51.3% in the past four weeks, higher than influenza A (H1N1) at 43.4%), with the prevalence of influenza A (H1N1) continuing to decline.

Third, starting October 1st of this year, publicly funded influenza antiviral medication will be expanded to cover high-risk groups, in addition to the original year-round use. Anyone experiencing flu-like symptoms, after evaluation by a physician, can receive medication without rapid testing. Publicly funded influenza vaccination will also be available on the same day. The prevalent viruses have also shifted to influenza A (H3N2). The aforementioned drug resistance in some influenza A (H1N1) strains has limited impact on clinical treatment and epidemic control. The CDC will continue to closely monitor the prevalence of influenza strains and drug resistance, and will adjust relevant prevention and control strategies on a rolling basis.

The CDC once again reminds the public not to underestimate the severity of influenza. It is essential to maintain good respiratory hygiene and cough etiquette, including frequent hand washing. If you experience flu-like symptoms such as fever and cough, wear a mask and seek medical attention promptly. If you are sick, rest at home. For information on the use of publicly funded pharmacies, a list of contracted medical institutions, and influenza prevention and control information, please visit the CDC's global information website (https://www.cdc.gov.tw) or call the toll-free epidemic prevention hotline 1922 (or 0800-001922).