Preprint: Robustly Quantifying Uncertainty in International Avian Influenza A(H5N1) Infection Fatality Ratios

 
Based on WHO Data

#19,128

While the popular press likes to quote the scary `roughly 50% CFR (case fatality rate)' of H5N1 in humans (based on WHO reporting of 477 deaths among 993  confirmed cases (48%)), confirmed cases are usually restricted to those sick enough to be hospitalized and lucky enough to be tested for the virus. 

It is assumed that some number of mild (or asymptomatic) cases occur but are never detected.  It is also likely that some number of serious (or even fatal)  cases go unrecognized, and unreported. 

We've also seen dozens of strains of HPAI H5N1 spillover to humans over the past 30 years, with varying levels of apparent virulence in humans. As the chart above illustrates, some regions of the world have reported far lower fatality rates than others.

All of which makes it exceedingly difficult to make blanket statements about the likely impact of an H5N1 pandemic. Particularly since we can't even say, with any degree of certainty, how many people have died from COVID.   

Suffice to say, anything over a 5% IFR (Infection Fatality Rate) would be catastrophic, and would exceed the death toll of the 1918 influenza pandemic by a wide margin. 

While there are a great many unknowns that must be considered, today we've a preprint from researchers at the UK Health Security Agency and the University of Manchester, which attempts to model a more realistic IFR number from a future H5N1 pandemic. 

While they have come up with a significantly lower IFR estimate (median 15.3%), this is simply their best guess - based on their model and assumptions - and they warn the reality could be significantly higher or lower (range 0.5% to 64.2%).

Although this is quite a wide range, they state: good preparedness for a potential A(H5N1) pandemic implies adopting scenario planning under our central estimate, as well as for IFRs as high as 70%.

This is, admittedly, much higher than my personal guesstimate (5%-7%), which is based primarily on wishful thinking. 

Since I'm notoriously out of my depth when it comes to statistics and modeling, I'll simply post the abstract and urge you to read the report in its entirety.  I'll have a bit more after the break. 

Robustly Quantifying Uncertainty in International Avian Influenza A(H5N1) Infection Fatality Ratios

Leonardo Gada, Mwandida, Kamba Afuleni, Michael Noble, Thomas House, Thomas Finnie
doi: https://doi.org/10.64898/2026.04.22.26351373

Abstract

Knowing the mortality rates associated with infection by a pathogen is essential for effective preparedness and response. Here, harnessing the flexibility of a Bayesian approach, we produce an estimate of the Infection Fatality Ratio (IFR) for A(H5N1) conditional on explicit assumptions, and quantify the uncertainty thereof. We also apply the method to first-wave COVID-19 data up to March 2020, demonstrating the estimates that could be obtained were the model available then.

Our analysis uses World Development Indicators (WDI) from the World Bank, the A(H5N1) WHO confirmed cases and deaths tracker by country (2003-2024), and COVID-19 cases and deaths data from John Hopkins University (January and February 2020). Since infectious disease dynamics are typically influenced by local socio-economic factors rather than political borders, individual countries are placed within clusters of countries sharing similar WDIs relevant to respiratory viral diseases, with clusters derived by performing Hierarchical Clustering. To estimate the IFR, we fit a Negative Binomial Bayesian Hierarchical Model for A(H5N1) and COVID-19 separately. We explicitly modelled key unobserved parameters with informative priors from expert opinion and literature.

By modelling underreporting, our analysis suggests lower fatality (15.3%) compared to WHO's Case Fatality Ratio estimate (54%) on lab-confirmed cases. However, credible intervals are wide ([0.5%, 64.2%] 95% CrI). 

Therefore, good preparedness for a potential A(H5N1) pandemic implies adopting scenario planning under our central estimate, as well as for IFRs as high as 70%. Our approach also returns a COVID-19 IFR estimate of 2.8% with [2.5%, 3.1%] 95% CrI which is consistent with literature.

Key Messages 

1. We adopted a disease-agnostic and adaptable Bayesian model, embedding scientific knowledge on A(H5N1) in the priors informed by published literature, to estimate the Infection Fatality Ratio (IFR) of avian influenza A(H5N1). 

2. Accounting for underreporting of cases and deaths, we estimate the IFR of avian influenza A(H5N1) at 15.3%, albeit with wide uncertainty ([0.5%, 64.2%] 95% Credible Intervals). 

3. Due to the uncertainty in the estimate, good preparedness for a potential A(H5N1) pandemic implies adopting scenario planning under our central estimate, as well as for IFRs as high as 70%.  

Anyone looking for certainty or serenity from this preprint will likely come away disappointed, but it does remind us that we can't assume the next pandemic will be like the last one . . . or like any that we've seen in living history.

Most pandemic plans `top out' at a 1918-level event, with many anticipating a far more moderate (1957 or 1968) scenario.  Any IFR/CFR over 2% would far exceed these parameters.

In our current climate of pandemic denial, the author's advice to prepare for a 15% IFR - but have contingencies for up to 70% -  are likely to fall on deaf ears. Frankly, I'm not sure how any government would plan for a > 15% IFR, even if they could summon the political will to do so. 

Since even a 2% IFR would be devastating, and bring many critical services to the breaking point, I can only recommend that people (and businesses) take individual pandemic planning seriously.  

Which is why I've already got my supply of masks, hand sanitizer, and OTC meds in the hall closet, and have stayed current with all of my vaccines. If you aren't similarly prepared, you may want to revisit:

#Natlprep 2025: Personal Pandemic Preparedness

Med Hx: The ‘Spanish’ influenza pandemic: New evidence for influenza outbreaks in England and France prior to 1918

Credit CDC

#19,127

For the past 20 years the most widely accept theory has been that the 1918 pandemic likely began in Kansas (United States) and was carried to Europe by troops off to WWI.  But much of the data was anecdotal, and not everyone concurred.

Among the early contrarians was Professor John Oxford, who in 2001 published The so-called Great Spanish Influenza Pandemic of 1918 may have originated in France in 1916. 

In 2006, PBS published Origins of the 1918 Pandemic: The Case for France by Vikki Valentine, which would help send me down a rabbit hole in 2007 when I penned a long blog - citing U.S. Navy records, and old (1917) Lancet articles - that were highly suggestive of a highly virulent flu-like illness spreading in Europe as early as 2015. 

A brief excerpt from my blog Beginnings and Ends:

The files of The Lancet indicate that a more or less widespread epidemic occurred in England in the spring of 1915. The disease does not appear to have been as prevalent in 1916 as in 1915, but in 1917, among the military forces, cases of so-called "purulent bronchitis" occurred which were fundamentally the same as the rapidly fatal cases of influenzal pneumonia so frequently seen at the height of the pandemic.

An epidemic of purulent bronchitis was reported from a British Army base in northern France in January, 1917, whilst an epidemic of influenza was in progress. This outbreak began in December, 1916. Later, in the spring of 1917, similar cases of purulent bronchitis were treated at Aldershot, England.

These cases are noteworthy because they seem to have been similar in all respects to the fatal types of influenzal pneumonia so commonly seen in all parts of the world during the autumn of 1918. 

In 1918, influenza was still believed due to a bacterial infection (Influenza A was only isolated as a virus in 1933), and so - despite the clinical similarities - no virological confirmation that the wave of illnesses reported in 2015-2017 was due to the same virus as the 1918 pandemic exists.

 

Today we revisit this historical conundrum, with a new paper by Douglas Gill and John Oxford, that provides an updated - and far more detailed - historical hypothesis on the possible pre-1918 emergence of the `Spanish flu'. 

While we are left with something less than a `smoking gun', this 20-page narrative is a fascinating read, and very much worth your time. And with a little luck, it may serve to reopen the debate on the still murky origins of the worst influenza pandemic on record.  

 

The ‘Spanish’ influenza pandemic: new evidence for influenza outbreaks in England and France prior to 1918 

Douglas Gill1 and John Oxford
Published online by Cambridge University Press: 22 April 2026

Abstract 

The Spanish influenza pandemic of 1918 caused well over fifty million deaths. The epicentre undoubtedly was China, where gene mixing of different virus strains occurred amongst aquatic, migrant birds. But where and when did the virus first infect (or spill over to) a human being?

We take, as our starting point, a paper demonstrating that an infection causing the same symptoms as the influenza virus was widespread in New York during the winter of 1917–1918. The authors of that paper went on to suggest that the virus had probably reached North America from Europe, in the context of troop movement during World War I. Our own researches have focussed on this point.

We show that outbreaks of serious respiratory disease, local in nature but causing unusual patterns of mortality, were indeed reported by scientists and doctors in army hospitals in England and in France, well before the first wave of the pandemic had arrived. We use the records of these hospitals, now held in the National Archives, to trace the progress of this disease amongst the individuals who fell ill. We examine contemporary reactions to this minor epidemic – an epidemic, we suggest, which acted as a herald wave of the pandemic yet to come.

The latter part of our paper addresses the second question, as to how troop movement across the North Atlantic, once the United States had entered into war, may well have enabled the virus to spread from Europe to North America.

        (SNIP)

And it has enabled us to demonstrate that if indeed, as we suspect, the spillover of the H1N1 influenza virus, from a bird to a human being, took place in northern France in 1916, then a pathway certainly existed whereby that virus could cross the North Atlantic to New York. Once landed at that port, it could spread in a slow but certain fashion through the crowded cities and the burgeoning encampments of a great nation, now girding up for war. It might mutate, it might change the way it spread, but once it had adapted, it was then ready to take up the role of driving the subsequent pandemic.

        (Continue . . . ) 

 

iScience: Multiplex Serological Profiling Reveals Diverse Avian and Mammalian Influenza A Virus Exposure in Swine (in Cambodia)

 

#19,126

While the spillover of HPAI H5N1 into dairy cattle has garnered considerable attention over the past two years, historically, the susceptibility of swine to diverse influenza A viruses has generated the most pandemic concern. 

In May of 2023, in Netherlands: Zoonoses Experts Council (DB-Z) Risk Assessment & Warning of Swine As `Mixing Vessels' For Avian Flu, we looked at growing concerns in Europe that avian H5N1 could increase its pandemic threat by spreading (and evolving) in farmed swine.

Only days later, a report out of Italy confirmed an H5N1 spillover event at a `mixed species' farm (poultry & swine), and the subsequent seroconversion of the majority of the pigs tested on that farm (see Study: Seroconversion of a Swine Herd in a Free-Range Rural Multi-Species Farm against HPAI H5N1 2.3.4.4b Clade Virus).

While HPAI detections in swine have been limited, we've seen scattered evidence that H5N1 can infect pigs, albeit often asymptomatically. A few past reports include:

• WHO H5N1 detected in pigs in China (2004)
• EID Journal: Asymptomatic H5N1 In Pigs (2010)
• An Unusual Report Of H5N1 in Pigs (Indonesia 2016)
• Sci. Rpts.: Evidence Of H5N1 Exposure In Domestic Pigs - Nigeria (2018)
• A 2023 Microorganisms report on H5, H7, and H9 in Pigs in Senegal (2023)

Of course H5N1 isn't the only concern, in addition to avian H5, H7, and H9:

A novel H1N1 virus, circulating in pigs, sparked a pandemic in 2009, and a number of H1, H2, and H3 viruses with zoonotic potential have been identified in swine over the years (see J. Gen. Virology: Evaluation of Pandemic Potential of the Genotype 4 (G4) Swine Influenza Virus).

But testing is extremely limited; PCR testing can only pick up current infection (which may only last days or weeks), while antibody testing (HI/MN/NP-ELISA) can often infer past infections, but are labor intensive and can generally only screen for one subtype at a time. 

While RT-PCR can only tell you `what is', and serology only `what was', most countries aren't aggressively testing with either method.

All of which brings us to a new report, published in iScience, which provides a proof-of-concept that a multiplex immunoassay panel can detect a wide range of influenza A subtype antibodies in one pass. 

While much of this report will be of interest primarily to lab geeks, the testing took place in Cambodia between 2020 and 2022, which has been the site of a renewed outbreak of HPAI H5N1 since 2023 (see Cambodian MOH Announces 4th Human H5N1 Case of 2026).

While this testing pre-dates this recent resurgence, it revealed a complex panoply of influenza A viruses have circulated in Cambodia pigs, including a faint H5 signal (see graphical abstract below).  

The Abstract, and a brief excerpt, from the study follow.  I'll have a bit more after the break.

Multiplex Serological Profiling Reveals Diverse Avian and Mammalian Influenza A Virus Exposure in Swine

 Foong Ying Wong 1 12, Peter Cronin 1 12, Rong Zhang 1 12, Arata Hidano 2 8, Jurre Y. Siegers 3, Dolyce HW Low 1, Hannah Holt 2, William Leung 2, Dina Koeut 4, Bunnary Seng 4, Sovanncheypo Chao 4, Ty Chhay 5, Sothrya Tum 4, San Sorn 4, Monidarin Chou 6, Kimrong Bun 3, Phalla Y 3, Leangyi Heng 3, Marcus G. Mah 1, Giselle GK Ng 1…Yvonne CF Su 1 13 ∗∗Show more
 
https://doi.org/10.1016/j.isci.2026.115743 
Under a Creative Commons license 

Highlights

• Multiplex microsphere immunoassay reveals antibodies to diverse influenza lineages

• H1N1/pdm09 dominates seroprevalence among pigs in Cambodia

• European avian-like swine antibodies indicate ongoing zoonotic risk

• Limited AIV exposure; one pig serum neutralized HPAI H5N1 clade 2.3.4.4b virus

SUMMARY

Animal origin influenza viruses pose significant pandemic threats, with swine serving as key hosts. Serological surveillance in pigs remains limited in regions with intense human–animal–avian contact. Between March 2020 and July 2022, we collected 4,089 pig serum samples in Cambodia, of which 1,321 (32.5%) were influenza A virus (IAV) seropositive by ELISA. 

We developed a multiplex microsphere immunoassay comprising a broad panel of hemagglutinin (HA) and neuraminidase (NA) antigens. Seroprevalence was highest for pandemic H1N1/pdm09 (35.3–45.7%) and lower for classical swine H1 (16.4–17.7%) and European avian-like swine H1 (∼15%) lineages. 

Lineage-specific or shared exposures to multiple lineages were observed, indicating complex infection histories. H3 responses varied by clade, while antibodies to avian H5, H7, and H9 were detected in <8% of IAV-positive sera.

These findings underscore ongoing zoonotic risk posed by diverse IAVs circulating in pigs and highlights the need for integrated serological and genomic surveillance.

        (SNIP)

DISCUSSION  

Pandemic influenza A viruses have historically caused significant mortality and morbidity in humans.  These viruses typically originate from animal reservoirs and acquire the capacity for sustained human-to-human transmission through sequential adaptation events 47.

The pandemic H1N1/pdm09 influenza virus exemplifies this process, highlighting the critical role of swine as an intermediate host in the emergence of reassortant strains that had circulated undetected in pigs for nearly a decade before its detection in  humans 47,48, largely due to insufficient surveillance.

While genomic surveillance is indispensable for tracing viral origin and identifying emerging variants, multiplex serological assays provide complementary insights by revealing prior exposure histories and population-level immunity in both animal and human  hosts.

In this study, we developed a 23-plex microsphere immunoassay targeting a broad diversity of HA and NA antigens representing multiple influenza lineages. 

Analysis of NP-positive pig sera collected in Cambodia between 2020 and 2020 demonstrated seroreactivity to a diverse range of H1 lineages,  including pandemic H1N1/2009 (pdm09), classical swine (CS), European avian-like (EA) swine and pre-2009 human seasonal lineages. These findings are consistent with our previous genomic surveillance in Cambodia 16, and illustrates the complex multi-lineage influenza ecology within swine populations.     

       (Continue . . . )

Although this multiplex immunoassay isn't a substitute for PCR testing, it could provide valuable information on the types of viruses circulating in the wild, and levels of herd immunity.

While the last (2009) swine-origin pandemic virus was relatively mild, there are no guarantees that the next one will have similar virulence. Recent studies suggest that community immunity levels to many circulating swine variant viruses are quite limited.

Some recent blogs on swine-origin viruses include:

Vet. Microbiology: Continuous Evolution of Eurasian Avian-like H1N1 Swine Influenza Viruses with pdm/09-derived Internal Genes Enhances Pathogenicity in Mice

J. Gen. Virology: Evaluation of Pandemic Potential of the Genotype 4 (G4) Swine Influenza Virus using Ex Vivo and In Vitro Cultures of the Human Respiratory Tract

Emerg Microbes Infect: A fatal Case of Acute Encephalitis Associated with a Novel influenza H3N2 Recombinant Virus Possessing Human-origin H7N9 Internal Genes

ANSES Reports A `New' Swine Flu Virus Has Taken Over Other Genotypes in France

A reminder that Nature has more than one way at its disposal to spark the next pandemic. 

Cambodian MOH Announces 4th Human H5N1 Case of 2026

 

#19,125

With thanks for the head's up from @E_A_Karlsson, we have the following announcement from Cambodia's Ministry of Health on their 4th confirmed HPAI H5N1 human infection of 2026 - and 38th in just over 3 years - this time involving a 66  y.o. woman from Svay Rieng province who had recent contact with sick chickens and was confirmed positive on March April 21st.

The MOH announcement, along with the translation, follow:

      (Translation)

Kingdom of Cambodia
Nation Religion King
Ministry of Health

Press Release

On Bird flu case in 66-year-old woman

The Ministry of Health of the Kingdom of Cambodia would like to inform the public that there is 1 case of bird flu in a 66-year-old woman who was confirmed to be positive for the H5N1 avian influenza virus on April 21, 2026 by the National Institute of Public Health. The patient is a resident of Trapaing Thkov village, Pong Teuk commune, Romduol district, Svay Rieng province. The patient has been placed in isolation at the hospital and is receiving intensive medical care. Investigations revealed that from April 2 to 13, there were sick and dead chickens in the village and the patient’s house, which were used for cooking.

The emergency response teams of the national and sub-national ministries of health have been collaborating with the provincial agriculture departments and local authorities at all levels to actively investigate the outbreak of bird flu and respond according to technical methods and protocols, find sources of transmission in both animals and humans, and search for suspected cases and contacts to prevent further transmission in the community, as well as distribute Tamiflu to close contacts and conduct health education campaigns for citizens in the affected villages.

The Ministry of Health would like to remind all citizens to always pay attention to and be vigilant about bird flu because H5N1 bird flu continues to threaten the health of our citizens. We would also like to inform you that if you have a fever, cough, runny nose, or difficulty breathing and have a history of contact with sick or dead chickens or ducks within 14 days before the onset of symptoms, do not go to gatherings or crowded places and seek consultation and examination and treatment at the nearest health center or hospital immediately. Avoid delaying this, which puts you at high risk of eventual death.

How it is transmitted: H5N1 bird flu is a type of flu that is usually spread from sick birds to other birds, but it can sometimes be spread from birds to humans through close contact with sick or dead birds. Bird flu in humans is a serious illness that requires prompt hospital treatment. Although it is not easily transmitted from person to person, if it mutates, it can be contagious, just like seasonal flu.

1/2

Address: Lot No. 80, Samdech Pen Nut Street (289)
Sangkat Boeung Kak 2, Khan Toul Kork, Phnom Penh
Phone: (+855) 23 885 970
Email: info@moh.gov.kh
Website: www.moh.gov.kh 

In February of 2023 an older clade of H5N1 (2.3.2.1.x) reemerged in Cambodia's population after a 9 year absence, spilling over into 6 humans in 2023, 10 people in 2024, and 18 people in 2025.

Cambodia's recent cases are due to a new reassortment of an older clade of the H5N1 virus (recently renamed 2.3.2.1e) - which appears to be spreading rapidly through both wild birds and local poultry.

Unlike the newer clade 2.3.4.4b H5Nx viruses - which have shown  much lower mortality rates in the United States - this older clade has proved fatal in about 40% of the cases reported by Cambodia since 2023.

There is currently no evidence to suggest human-to-human transmission of this H5 virus, with most cases reporting recent contact with sick or dead poultry.

As we discussed last July in Cambodia: Food Insecurity, Food Safety & H5N1 - despite repeated warnings to the public not to prepare or cook sick/dead poultry - scarce resources and hunger can sometimes drive people to take risks.

Last October Dr. Erik Karlsson (@eakarlsson.bsky.social‬) from the Institut Pasteur du Cambodge, Phnom Penh, Cambodia, - along with a long list of colleagues -  published a brief letter in the NEJM on first 16 cases. 

Among them were 4 `clusters' (Father/Daughter, two neighbors, 2 siblings, and 2 cousins living in the same household), 6 fatalities, and 12 cases < 18 years of age.

While human-to-human transmission was considered, the available evidence suggests a `shared' exposure instead, with infected poultry the most likely source of infection.   

Due to copyright, I can only post  the title and link, but I would urge you to go to the NEJM site to read it in its entirety.

 Highly Recommended. 

Resurgence of Zoonotic Highly Pathogenic Avian Influenza A(H5N1) Virus in Cambodia
Published October 22, 2025
N Engl J Med 2025;393:1650-1652
DOI: 10.1056/NEJMc2504302
VOL. 393 NO. 16 Copyright © 2025

HK CHP Reports 5 Recent H9N2 Cases on the Mainland

 

#19,124

After going a couple of months without any reports, last week Hong Kong's CHP announced 2 new H9N2 cases on the Mainland, and this morning has announced 3 more.

First last weeks announcement:

Followed by today's:

As is typical, we only get the barebones details from the CHP reports. The WHO Western Pacific Regional Office (WPRO) has published slightly more information in their latest 2 weekly reports, but are also somewhat lacking in details. 

For the first 2 cases:

Human infection with avian influenza A(H9N2) virus

From 27 March to 9 April 2026, two new cases of human infection with avian influenza A(H9N2) virus were reported to WHO in the Western Pacific Region.

The first case was a male child under five years of age from Guangdong Province, China, who was hospitalized with bronchopneumonia and severe pneumonia and was laboratory-confirmed with influenza A(H9N2) on 4 February 2026. 

He had no direct contact with live poultry but may have had indirect exposure through a family member working at a poultry farm. He recovered and was discharged on 12 February. 

The second case was a 63-year-old male from Guangxi Province with underlying conditions. Following contact with sick poultry in late January, he developed symptoms on 5 February 2026 and was laboratory-confirmed with influenza A(H9N2) on 11 February. He recovered and was discharged on 17 February.

The three cases reported today were described as:

Human infection with avian influenza A(H9N2) virus

From 10 to 16 April 2026, three new cases of human infection with avian influenza A(H9N2) virus were reported to WHO in the Western Pacific Region. All three cases were children under 10 years of age from China.

The first case is a male with comorbidities from Guangdong Province, with symptom onset on 23 February 2026. He was hospitalised with severe pneumonia on 23 February and was discharged on 25 March. 

The second case is a female from Yunnan Province, with symptom onset on 3 March; the third case is a male from Jiangxi Province, with symptom onset on 20 March. Both cases developed mild symptoms only and did not require hospitalisations. 

The first and second cases had exposure to poultry, whereas no clear direct poultry exposure was identified for the third case. However, environmental samples from live poultry stalls in the market routinely visited by the family member of the third case tested positive for H9. 

No additional cases were reported from close contacts of the cases. Since December 2015, a total of 162 cases of human infection with avian influenza A(H9N2), including two deaths (both with underlying conditions), have been reported to WHO in the Western Pacific Region. Of these, 159 were reported from China, two were from Cambodia, and one was from Viet Nam.

 What we can piece together from the combined reports is that:

• 4 of the 5 cases were in young children, while 1 was in a 63 y.o. man.  

• 2 were characterized as having `severe' pneumonia, 1 other was hospitalized (M, 63) but no details were offered, while 2 others developed mild symptoms and were not hospitalized

• 3 of the 5 cases reportedly had direct contact with sick poultry, while 2 did not

• Both cases without poultry contact may have had indirect contact via family members who worked with poultry, or visited a LBM (Live Bird Market)

While we've seen cases with no reported direct contact to poultry before, it is a bit unusual to have two cases where `indirect exposure' via a family member is mentioned as a possibility.   

H9N2 is typically described as being a mild infection, but here 3 of the 5 were hospitalized, with two cited as having `severe pneumonia'.  While there may have been comorbidities that impacted these patients, none were mentioned.

We've seen increasing concerns from Chinese scientists that H9N2 is evolving, and is acquiring mammalian adaptations (see EM&I: Enhanced Replication of a Contemporary Avian Influenza A H9N2 Virus in Human Respiratory Organoids). The authors wrote:

In summary, we demonstrated that a recent H9N2 virus is more adapted to humans, and is able to replicate to high titres in both upper and lower human respiratory tract which may confer higher person-to-person transmissibility and virulence.

Last October, in China CDC Weekly: Epidemiological and Genetic Characterization of Three H9N2 Viruses Causing Human Infections, we looked at a local CDC investigation into 3 pediatric cases which were reported last April from Changsha City, Hunan Province, China.

Their report found a number of indicators of increased mammalian adaptation within the virus, including an enhanced ability to infect upper respiratory (α2,6-sialic acid) tract receptors, and a number of HA protein mutations, including; H191N, A198V, Q226L, and Q234L.

While trying to predict the source of the next pandemic is a mug's game, H9N2 constantly ranks in the top 10 zoonotic influenza A viruses the CDC has pegged as having some pandemic potential. 

Making reports like today's well worth our attention.

Preprint: Infection of the Bovine Mammary Gland by Avian H5N1 Subclade 2.3.4.4b Influenza viruses

#19,123

Since the first outbreak of HPAI H5N1 was reported in U.S. dairy cattle a little over 2 years ago, surveillance and testing of mammalian livestock has focused almost exclusively on lactating dairy cows.  

One notable (and rare) exception was a May 2024 press release (Updates on H5N1 Beef Safety Studies) where the USDA that `. . . viral particles were detected in tissue samples, including muscle, from one cow. To date, samples from 95 cows have tested negative for viral particles.'

The USDA's FSIS (Food Safety Inspection Service) also conducted a 1 year slaughterhouse surveillance program (ended Sept 2025) which tested 837 muscle samples for the virus.  They report detecting one. 

While reassuring, the USDA reports 33 to 34 million head of cattle are processed each year in the United States, which makes this sampling less than 1 in 40,000 beef cattle slaughtered each year.

As we discussed last week, in EID Journal: Highly Pathogenic Avian Influenza A(H5N1) Virus RNA in Bovine Semen, California, USA, 2024, the number of beef cattle far exceed dairy cattle in the United State. The logistics of testing beef cattle are formidable, and there has been little enthusiasm for pursuing any serious national surveillance program. 

But, when you combine the above report with last year's Virology: Detection of Antibodies Against Influenza A Viruses in Cattle) - which reported that bulls and steers were just as likely to carry antibodies to (non-HPAI H5) IAV as cows and heifers (and a more recent report on H1N1 in a European (male) Bison) - there are legitimate reasons to believe we aren't seeing the full H5N1 picture.

The surveillance and testing of cattle outside of the United States has been even more limited, despite scattered reports of H5N1 in  European Sheep (see here and here) and H5N1 antibodies detected in cattle in the Netherlands.

It was only last October that WOAH embraced an 11-page OFFLU technical document (see OFFLU Guidelines for High Pathogenicity Avian Influenza Virus Risk Mitigation in Cattle) which - among other things - finally made HPAI in cattle a `reportable disease'.

But surveillance remains passive, and testing largely up to the discretion of the livestock owner. Something that many farmers find unappealing, fearing the quarantining of their herds and the stigma of infection. 

All of which brings us to a new preprint, where researchers at the MRC-University of Glasgow Centre for Virus Research tested mammary explants from both beef (Aberdeen Angus & Limousin) and dairy (Holstein Friesian) cattle for susceptibility to (3 older European strains) of HPAI H5N1.

They also tested susceptibility to other influenza A viruses (see chart at top of blog).  All HPAI viruses were attenuated reassortant viruses deemed safe for use in BSL-2 labs.

Basically, they report:

• Explants from all 3 breeds showed infected cells after H5N1 exposure, mainly in teat tissue.

• Both avian (α2,3) and mammalian (α2,6) sialic acid receptors appear on cow mammary epithelial cells, potentially enabling viral mixing or adaptation.

• HPAI infection varied by individual cow and virus strain but all breeds were susceptible; seasonal human H1N1/H3N2 infected less efficiently than H5N1.

The authors concluded:

Our findings show that mammary tissue from common cattle breeds is permissive to clade 2.3.4.4b H5N1 infection. These results support the view that bovine mammary epithelium represents a susceptible tissue environment and suggest that IAV infection in cattle may be more widespread than previously assumed.

 Accordingly, cattle should be incorporated into H5N1 surveillance frameworks, particularly during poultry outbreaks or incursions of infected wild birds. Surveillance strategies that combine PCR-based detection of active infection with serology for prior exposure may provide a more complete picture of virus circulation at both individual- and herd-level. 

Overall, integrating cattle into preparedness planning is a precautionary and evidence-based response to the expanding host range of clade 2.3.4.4b377H5N1 viruses. 

I've reproduced the abstract and impact Statement below.  The full report runs 24 pages, so follow the link to read it in its entirety. 

Infection of the bovine mammary gland by avian H5N1 subclade 2.3.4.4b influenza viruses
 Rebecca A. Ross,  Sarah K. Walsh,  Hannah Montgomery,  Hanting Chen,  Edward Hutchinson,  Pablo R. Murcia
doi: https://doi.org/10.64898/2026.04.16.718897
This article is a preprint and has not been certified by peer review 

Preview PDF

Abstract

The emergence of the panzootic clade of highly pathogenic avian influenza H5N1 (2.3.4.4b) in 2020 marked a major expansion in the host range of influenza A viruses (IAVs), raising concerns about further cross-species transmission events and zoonotic spillover. Introduction of 2.3.4.4b viruses into U.S. dairy herds has resulted in widespread circulation, accompanied by reduced milk yield, mastitis, and high viral loads in milk. Notably, virus circulation in dairy cattle represents a novel route for mammalian adaptation and transmission that has already led to more than 40 human cases in the U.S. since 2024. 

Here, we investigated whether avian clade 2.3.4.4b viruses could infect mammary tissue from Aberdeen Angus, Holstein Friesian, and Limousin cattle, three breeds commonly farmed in Europe, the Americas, and Oceania. Using mammary gland explants, we inoculated tissues with attenuated reassortant viruses expressing the haemagglutinin and neuraminidase glycoproteins of three 2.3.4.4b viruses that predated the emergence of H5N1 in US cattle: A/chicken/England/053052/2021 (AIV07), A/chicken/Scotland/054477/2021 (AIV09), and A/chicken/England/085598/2022 (AIV48). 

Infected epithelial cells were identified using immunohistochemistry in explants from both the teat and gland cistern for all three breeds following infection with AIV09 and AIV48, indicating that mammary tissue from each of the three tested cattle breeds cattle is permissive to H5N1 infection. Lectin staining showed expression of both α2,3-linked and α2,6-linked sialic acids in the mammary tissue of all donors showing that all three breeds have the potential to support infection with both avian-adapted and mammalian adapted IAVs. 

Together, these findings demonstrate that mammary glands from both beef and dairy cattle breeds are permissive to infection with avian-adapted and mammalian-adapted H5N1 viruses and highlight the potential for this tissue to act as a mixing vessel for IAV reassortment, underscoring the need to include cattle in ongoing H5N1 surveillance and risk-assessment frameworks.

Impact Statement 

The emergence of highly pathogenic avian influenza H5N1 in dairy cattle has expanded the recognised host range of influenza A viruses. Further, the ability of the virus to infect the mammary gland and transmit via milk revealed a novel interface for transmission to humans and animals. Although sustained circulation in US dairy herds has been reported, the susceptibility of mammary tissue from other breeds (including beef cattle) commonly used in different countries has been largely unexplored.

Here, we show that avian-origin H5N1 viruses can infect tissues derived from the mammary gland of three common cattle breeds (Aberdeen Angus, Holstein Friesian, and Limousin). Virus was detected in epithelial cells from both dairy and beef breeds, indicating that H5N1 can infect multiple breeds. Receptor profiling showed abundant α2,3-linked and α2,6-linked sialic acids, consistent with a tissue environment that may support infection with both avian-adapted and mammalian-adapted viruses. These findings demonstrate that multiple cattle breeds are permissive to H5N1 infection and strengthens the evidence base for including cattle in H5N1 surveillance and risk-assessment frameworks.