Peace on Earth in Solstice Darkness as Wisconsin Dairy Case is Sequenced as a Separate H5N1 2.3.4.4b D1.1 Spillover
Just released NIH preprint documents rapid predominance of D1.1 genotype in 2025 North American H5N1 isolates, incorporating a LPAI N gene for first time
I’ve personally taken some family and personal time at Christmas-New Years, and I trust that most of you have also. It’s taken me a lifetime to more fully appreciate the traditions in my Christian faith, including inclusion of pagan celebrations of light in the winter solstice into which we have imprinted our Holiest traditions of the birth of Jesus. Signifying His arrival as bringing light to our world was certainly a Divine inspiration for Christians. Like many, I tend to suffer from seasonal affective disorder brought on by decreasing daylight of fall-early winter. Whatever our faith or philosophy, we all can take heart in the assurance that every day gets slightly longer from here until late June!
Applying that to influenza and viruses, we know that despite several more months of cold and darkness which are favorable to disease and pestilence, spring and renewal (and Easter for us Christians!) will come as surely as the earth circles the sun at a more acute angle each day! Battling our common viral foes together, we know that the ultimately effective powers of heat and sunshine that overcome microbial enemies are waiting in the wings to eventually come to our aid!
I don’t normally spend too much time on esoteric ruminations like this; however, it’s important that all of us keep our philosophical and psychological bearings moored in our daily professional battles. We must be at peace with the cyclical nature of our environment and our adaptation to it as a crucial part of how we engage our challenges. Just remember, every day is a bit brighter from here on out for the next 6 months!
Wisconsin Dairy Herd Infection
As widely suspected, the H5N1 dairy herd infection in Dodge County Wisconsin was confirmed as a newly introduced D1.1 genotype through confirmatory sequencing at NVSL: Update: Genetic Sequencing Results for Wisconsin Dairy Herd Detection of Highly Pathogenic Avian Influenza
WASHINGTON, D.C., December 19, 2025—On December 14, 2025, USDA’s Animal and Plant Health Inspection Service (APHIS) announced the first detection of highly pathogenic avian influenza (HPAI) in a dairy herd in Wisconsin. On December 17, the National Veterinary Services Laboratories (NVSL) completed whole genome sequencing and confirmed that the virus is H5N1 clade 2.3.4.4b genotype D1.1. Analysis indicates that this detection is a new spillover event from wildlife into dairy cattle, separate from previous events.
Key Points
Most detections in U.S. dairy herds have resulted from movements linked to the original spillover event that occurred in the Texas Panhandle in late 2023, involving the B3.13 strain.
In early 2025, through the National Milk Testing Strategy, USDA detected two spillover events in Nevada and Arizona dairy herds. Both were identified early, and no further herd infections occurred through animal movements. These events involved the D1.1 strain.
The Wisconsin herd, also detected through the National Milk Testing Strategy, represents a new, separate spillover event and involves the D1.1 strain. At this time, no additional dairy herds have been identified as infected in association with this event.
Taking this language at face value, I assume that this sequence is not closely related to any other D1.1 dairy isolates (from AZ or NV) and also not closely related to other D1.1 avian poultry isolates, since it was classified as a (direct?) spillover event from wildlife. My conjecture in my last column about nearby earlier poultry H5N1 D1.1 outbreaks possibly being related to this dairy outbreak was shown to NOT be the case.
Subsequently, Wisconsin DATCP has announced a resumption of enhanced testing of all Wisconsin herds under the NMTS: National Milk Testing Strategy Numbers Update:
With help from industry partners, Wisconsin is sampling milk at the farm level for the mandatory USDA Animal and Plant Health Inspection Service (APHIS) NMTS. When Wisconsin began NMTS testing in May 2025, the goal was to test each farm approximately every month. Wisconsin achieved unaffected Gold Status under the NMTS in September 2025. In October, DATCP reduced its testing frequency to 50%.
Wisconsin will resume NMTS testing at a 100% frequency statewide. DATCP will continue to utilize milk samples that are already collected through existing, routine sampling to streamline the NMTS testing process. The DATCP HPAI in Dairy Cattle webpage will continue to be updated weekly as the testing continues. To date, more than 27,000 milk samples have been tested as part of the NMTS. These tests represent nearly all licensed dairies in Wisconsin, with continuing verifications underway to ensure complete statewide reach.
First, I want to congratulate Wisconsin for exemplary transparency in its milk sampling program! I’m not aware of any other major dairy states that post weekly milk sampling numbers and results. Here is the chart from their web site through December 30th:
Wisconsin Dairy Facts and Stats | Wisconsin Cheese (an industry web site) lists 5222 dairy herds in Wisconsin as of August 1, 2025. At that inventory, Wisconsin met one herd test per month goal from July-October and were at the 50% level through November on a rolling multi-week average.
However, given all the unknowns regarding length of infected herd viral shedding, especially with D1.1, it may have been more luck than rigorous statistical bulk tank sampling that allowed WI DATCP to discover this herd. They had only tested 609 milk samples from 5222 herds in the 3 weeks prior to the positive finding. Testing a herd via PCR on an average of once every 2 months allows a lot of time for a new infection to subside below detectable levels prior to a retest, especially if ELISA serology is not utilized on milk samples as a backstop for evidence of historical herd infection.
Remember that this case was detected in a state that is showing us their books! What about the many other states testing 50% of their herds monthly with no public transparency? What are the odds that we are finding all the D1.1 infected dairy herds in the U.S., given all the infected wildlife and poultry interspersed among dairy herds that seem to remain largely clinically asymptomatic with this strain of H5N1?
Back to Questions from my Last Post (USDA Confirms Highly Pathogenic Avian Influenza in a Dairy Herd in Wisconsin):
Was the WI herd initially flagged with positive serological findings (ELISA) or positive PCR? Or both? (The news release referred to positive PCR and ELISA results - confirmed via PCR (polymerase chain reaction) and ELISA (Enzyme-Linked Immunosorbent Assay) assay a detection of highly pathogenic avian influenza (HPAI) H5 clade 2.3.4.4b in a dairy cattle herd in Wisconsin.) Is NMTS monitoring both assays across all states, given that positive PCR results may be fleeting with milder D1.1 infections? What is the “silo test sensitivity” for both ELISA and PCR at varying intervals for testing and varying herd concentrations within silo samples? These questions really get back to the issues raised by the California paper reviewed last week, striking at the heart of the assumptions in the National Milk Testing Strategy, based on B3.13 epidemiology. Will D1.1 behave similarly if it becomes the more likely dairy herd H5N1 genotype? Do we leave herds undiagnosed with exclusive periodic H5 PCR screening versus adding serological ELISA assays to screen for earlier H5 exposure?
It seems to me that any effort to seriously understand H5N1 infection levels in milk and dairy cattle will require more, not less frequent targeted bulk milk sampling as the sampling program moves forward. Intensified wild bird and infected poultry flock exposures are additional biosecurity risks for dairy herds that will only add to the stakes for intense NMTS sampling.
Ideally our national non-poultry H5N1 risk surveillance needs to move well beyond dairy herds, ad hoc wild bird and mammal sampling, and occasional cat mortality autopsies. We seem to be paralyzed against use of serological tools and exploratory diagnostic sampling to assess livestock (and human!) exposure to H5N1 in hotbed areas of infection! In a way we’ve turned our dairy herds into (ineffective) D1.1 sentinels through a skeleton NMTS because of dairy’s B3.13 history and steady milk supply, while we ignore other domestic mammalian species for targeted surveillance! If we’re truly concerned about interspecies spillover from wildlife and poultry, we need to be serologically and symptomatically targeting all susceptible groups of livestock (beef cattle, sheep, goats, swine, horses, pets, etc. in H5N1 viral hot-bed outbreak localities like northern Indiana and the turkey areas of SD-ND-MN. We are perversely using lack of validated case definitions and tests, especially serology, to discourage any sampling of previously non-diagnosed domestic livestock species for a defined foreign animal disease.
We have not depopulated H5N1-infected dairy herds, and we will not kill other livestock species, except in dire circumstances. We can generate reasonable movement restriction and release plans to match our scientific risk assessments for a virus that will clear with proper management in most mammals. We must get past the old mantra that “don’t test if you don’t know what you’ll do with a positive”.
Finally, the biggest uninvestigated fomites we may have are human mouths and noses! Do we know that we are not carrying H5N1 at low but transmissible levels in human respiratory tracts from farm to farm? Dairy to poultry or vice versa? We can’t prove a negative, but we can surely generate more data that we can generate now…ignorance of transmission risks may really be expensive here!
I was about to expand on some thoughts questioning whether we might be settling into an “endemic” situation with H5N1 in dairy where we’d all just slowly tolerate an increasingly lax testing protocol. B3.13 seems to be slowly dying out within the dairy population as within herd immunity to a well-adapted cattle virus may be smothering it while D1.1 seems to be a “non-problem” clinically for the dairy herds and non-reported as a worker health issue…
What - me worry??
Then I came across this excellent preprint released on December 23rd by lead authors Alvin Crespo-Bellido and Martha I. Nelson of the Division of Intramural Research, National Library of Medicine, Bethesda, MD USA. I consider Martha to be one of the world’s preeminent international influenza molecular biologists. This paper is a poster case for why integrated federal research support paired with open sharing of the world’s viral sequences and associated metadata is so critical for outbreak preparedness. USDA must intensify its sequence data curation pipeline to get the data into the hands of people like Martha and Alvin in a much timelier manner. I have taken the liberty of pasting portions of sections of the preprint fairly liberally; however, please read the paper in its entirety in PDF format for the best viewing experience and full context. I’ve italicized direct segments with my comments left in regular font:
Emergence of D1.1 reassortant H5N1 avian influenza viruses in North America
Summary
Since 2021, highly pathogenic avian influenza viruses (HPAIV) belonging to H5N1 clade 2.3.4.4b have caused high mortality in North American wild birds and poultry. In 2025, a new D1.1 genotype caused two human deaths and host-switched to dairy cattle. However, the evolutionary origins and dynamics of D1.1 have not been fully characterized. Here, our phylogenetic analysis of 17,516 H5N1 genome sequences uncovers how D1.1 introduced a major shift in the antigenic diversity and ecology of the H5N1 epizootic in North America. D1.1 is the first major H5N1 genotype to (a) emerge in the Pacific flyway and spread west-to-east faster than any prior genotype; (b) antigenically shift via reassortment with the North American N1 segment, displacing the previously fixed Eurasian N1; and (c) transmit to a broader range of host species than any H5N1 genotype to date, introducing mammalian adaptations.
Introduction
…A new genotype termed “D1.1” was first detected in a wild duck in western Alaska on October 6, 2024 (A/northern pintail/USA/IZ24-0555/2024(H5N1), accession PV602213; Figure 1A). In November 2024, a D1.1 virus hospitalized a teenager with no known animal exposure in British Columbia, Canada (Figure 1B). In December 2024, the D1.1 genotype caused the first fatality from avian influenza in the Americas, which occurred in an adult with underlying medical conditions who was exposed to backyard poultry in Louisiana, USA. In early 2025, D1.1 viruses were identified in dairy cattle in Nevada and Arizona, leading to sustained outbreaks involving multiple herds in both states. In Nevada, a dairy worker infected with D1.1 following exposure to raw milk experienced mild conjunctivitis symptoms similar to the dairy workers infected with B3.13 in California, Colorado, and Michigan in 2024. In March 2025, a D1.1 infection in Durango, Mexico – and the first reported H5N1 human infection in Mexico overall – led to the death of a young child with no underlying medical conditions, no travel history, nor exposure to infected animals.
There is an urgent need to track the changing ecology and genetics of H5N1 in North America. To compare D1.1 evolution to prior H5N1 genotypes, we performed a phylogenetic analysis using 17,516 H5N1 2.3.4.4b genome sequences that are available from all host species in the Americas, as of June 13, 2025. Here, we use Bayesian phylodynamic approaches to provide a comprehensive analysis of the central evolutionary processes that drive H5N1 diversity in the Americas, including (a) intercontinental migration from Asia and Europe, (b) spatial dissemination across the Americas, (c) genomic reassortment with LPAI, and (d) interspecies transmission followed by host adaptation. In particular, we examine how the evolutionary dynamics of D1.1 differ from prior H5N1 genotypes in all of these areas.
Here is an illustration taken directly from the preprint that illustrates well the relationships between the 2.3.4.4b H5N1 genotypes described to date with their approximate times of emergence in North America:
Note that D1.1 and B3.13 have diverse flyway origins and gene segment constellations, i.e. they are quite different viruses in origin and makeup. (See the pre-print in .pdf format for a much clearer view of the 8 segment origin keys) Importantly, the D1.1 viruses also appear to be more susceptible to acquiring mammalian adaptation mutations in spillover events to mammals as illustrated here:
Because the B3.13 genotype already has already acquired a fixed PB:2M631L mammalian adaptation in spilling over to cattle, there may be less pressure for further mammalian adaptations in B3.13 genotype spillover infection events from either infected dairy cattle herds or spill-back poultry flocks to mammalian contacts.
Human Cases
The preprint also contains a table of known D1.1 human cases reported through March 15, 2025, which is up to date, given the dearth of any recent reports of human infections with H5N1. The illness in Mexico City was reportedly due to an H5N2 infection with a predecessor 2022-2023 H5 2.3.4.4b gene, and the Washington fatality was diagnosed as a newly identified H5N5, briefly discussed in the preprint:
Four of the nine cases resulted in hospitalization or death, which is an overestimate, given a lack of rigorously applied screening for mild cases in human contacts. Speculation continues that D1.1 viruses are riskier for human contacts than B3.13, perhaps due to this tendency to acquire mammalian PB2 adaptions upon spillover.
Discussion
…The D1.1 genotype’s acquisition of a new N1 neuraminidase from the North American LPAI lineage by reassortment in 2024 represents a turning point in the evolution of H5N1 in the Americas. Prior to the emergence of D1.1 in mid-2024, mostly LPAI internal gene segments had been acquired by H5N1 in the Americas. The rapid expansion of D1.1 viruses fundamentally altered the genetic composition of the H5N1 epizootic in North America, and 92.8% of clade 2.3.4.4b viruses sampled January 1 – June 13, 2025, contained the LPAI N1, mainly due to the emergence of D1.1 (Figure S3C-S3D). The North American LPAI N1 segment acquired by D1.1 differs from the Eurasian N1 at numerous amino acid sites, and the effect of these changes on virus phenotype and pandemic risk requires further study. Human exposure to H1N1 seasonal influenza viruses containing an N1 segment from the 2009 H1N1 pandemic, originally sourced from Eurasian swine, is thought to provide a degree of cross-protection against most H5N1 clade 2.3.4.4b viruses circulating in birds globally. Whether the North American LPAI N1 elicits similar cross-protection needs to be examined…
The future trajectory of D1.1 in the Americas is uncertain, including whether D1.1 spreads to South America this winter with migrating wild birds…A recent human fatality involving an H5N5 virus in Washington State (reported on November 15, 2025) was surprising, given that H5N5 was introduced into North America in 2023 and was excluded from further examination in this study because it had been sparsely sampled up to this point. The human case indicates that H5N5 has reached the Pacific flyway in wild birds, even though the virus had never been detected outside the Atlantic flyway. If H5N5 is an emerging genotype that presents a new risk to humans, our surveillance is inadequate. Relatively small investments in data pipelines for avian influenza could provide a large return in tracking evolution and zoonotic risk. We are currently lagging even in D1.1 surveillance.
As the epicenter for H5N1 evolution shifts to the Americas, there is an urgent need to expand the region’s infrastructure for real-time surveillance and monitoring. The Americas present a different ecological landscape compared to Eurasia, including South American marine mammals that are highly susceptible to H5N1 disease and farming systems with long-range animal movements that sustain virus persistence in dairy cattle…Thus far, Eurasia has been a major source of H5N1 viruses for North America, and there is little evidence of D1.1, B3.2, or other dominant North American genotypes migrating back to Europe or Asia. However, H5N1 is routinely changing patterns. Understanding the risk that major evolutionary changes in H5N1 in the Americas presents to the rest of the world is a central question going forward.
Metadata and Timely Useable Information
Following is an illustration from the preprint presenting some information I had not been aware of previously. As of the time of writing of this publication, USDA has still not provided metadata for 10 positive H5N1 D1.1 cattle samples entered into the NCBI dataset. It’s inexplicable why they were reported as NCBI sequences but yet not reported to WOAH as officially diagnosed H5N1 cattle infection(s) with associated metadata. Note that 31 feline and 48 terrestrial mammal sequences also lack metadata at this point.
Back in the supplemental data, the preprint illustrates the lack of currency in assigning metadata to D1.1 wild bird samples, similar to the issues with the cattle, cats, and terrestrial mammals shown above:
The report states:
Of the 1,722 D1.1 sequences available from North American wild birds that were deposited on NCBI’s Sequence Read Archive (SRA) as of Sep 10, 2025 (Figure S8), only 8.9% (n = 154) have a corresponding GenBank entry that includes a collection date (month, day, and year) and more precise sampling location (US state). If complete metadata were available for even half of those samples, scientists would be able to track D1.1 with higher precision.
It is important that this paper states: we performed a phylogenetic analysis using 17,516 H5N1 2.3.4.4b genome sequences that are available from all host species in the Americas, as of June 13, 2025. Yet the authors are flying blind on 1000’s of those sequences’ locations and dates of sequencing!
I’m sure the budget and management constraints imposed by the current administration have hobbled staff support to augment the current systems, but has anyone stopped to really analyze the processes in information flow and recordation from the field to GISAID? 92% non-curated samples 10 months after raw sequence submission is abysmal! This is a process begging for high-level expert review or for contracting to qualified vendors, e.g. qualified NAHLN labs or high-volume university sequencing labs. NVSL’s core mission is (or should be) to serve as a national/international diagnostic and reference lab, not an ongoing high-volume H5 sequencing facility.
Conclusion
I’ve overstayed my welcome for your attention in this first delayed column of 2026. I already see signs that we’d better buckle up, with broiler farm outbreaks increasing, ducks in Lancaster County PA, and a new egg layer outbreak in Nebraska just reported. Plus Indiana BOAH just reported LaGrange County’s 66th outbreak since mid-October (2 1/2 months!) in a commercial table egg pullet farm…I have to wonder if we are beginning round 2 in some farms in the county, given that these are pullets?
So, what do we know about H5N1 D1.1, just 1 year + after it was first isolated?
It has gone from a single case in western Alaska to dominance of all H5N1 isolates in North America by mid-2025. It’s obviously very fit, well-adapted, and burrowed into wildlife reservoirs at this point.
The LPAI N1 gene now predominates in North American H5N1 isolates, with the practical effects of that change unknown for overall pandemic susceptibility from this virus. Worse, the virus seems to be adapted for initial mammalian spillover, but still quite susceptible to further known mammalian adaptations upon crossover, adding to concerns regarding increased risks for onward transmission and zoonotic potential.
D1.1 has hit early, hard and often in the fall-early winter 2025 HPAI poultry outbreak “season”, remaining active in the Dakotas/Minnesota and northern Indiana after initial establishment. The degree of flock to flock versus wild bird to flock spread has not been disclosed by officials, although the velocity of spread in the Dakotas and Indiana would strongly suggest area spread.
The Wisconsin dairy herd outbreak was disclosed as a wild bird D1.1 spillover event, with reportedly few or no clinical effects in the herd. Diagnosis was described as accomplished by PCR and ELISA testing of bulk tank samples collected via the National Milk Testing Strategy of 50% herd testing per month.
Additionally, backyard flock outbreaks have occurred nationally on a more widespread basis than usual across a wide geographical area. I will have a more complete accounting of flock counts per month next week after NVSL “catches up” on submission reporting through the 2-week disruption caused by the end of year holiday shipping and testing disruptions.
The lack of serious human cases is momentarily reassuring, but hardly definitive. Human exposures from dairy herds may be less risky and numerous, given possibly less total raw milk and cattle aerosol viral production. On the poultry side outbreaks have not yet involved many widespread mega-site egg layer outbreaks, where millions of birds are shedding virus and requiring labor-intensive depopulation efforts. Every effort must be made to minimize human exposure to this virus in all situations. The danger remains that discounting mild early undiagnosed cases could lead to an occasional severe case. Read the highlighted points in the latest King County Washington Department of Public Health guidelines for suspected hospitalized human H5N1 cases to get a sense for how missed early diagnostics could have contributed to the fatal outcome there: Updated guidance for clinical consideration of Avian Influenza - King County Washington
All-in-all, D1.1 may be establishing a new “normal” for seasonal H5N1 infections in western hemisphere poultry and mammalian wildlife. We are writing the book without a precedent to follow, and any references to how this virus “should” behave are dangerous assumptions.
Finally, the U.S. has intentionally rejected all plans for enhanced readiness response activities in both animals and humans. We have rejected emergency vaccine planning and stockpiling, as well as further research into mRNA technology for human influenza vaccines, the single most promising route for a pandemic volume influenza vaccine for both humans. We are frozen from establishing scalable early detection capabilities by policy and budget roadblocks (no COVID lessons here).
Important Research Update:
Again, kudos to Michael Coston for flagging this timely preprint in today’s column:
Here is the direct link to the preprint:
I haven’t had time to read this closely - we can digest it together, and I’ll comment on it further in my next column. However, the early bottom line is depressingly familiar for flu - very narrow antigenic cross protection in the host means lots of reinfection risk and pressure for antigenic drift in the influenza virus - where have we seen THAT before?
John