Skip to content
You can now search across every topic, entity and event.What's new
Pandemics and Biosecurity
2JUN

Why H5N1 hid in cow udders

3 min read
09:17UTC

University of Pittsburgh scientists found H5N1 binds receptors abundant in udder tissue but scarce in airways, explaining why bird flu hid in US dairy herds for two years.

ScienceDeveloping
Key takeaway

H5N1 infects udders, not airways, so respiratory-keyed surveillance missed it in US dairy herds for two years.

Scientists at the University of Pittsburgh, in a team led by Suresh Kuchipudi, reported in Science Advances on 20 June that H5N1 avian influenza thrives in dairy cattle because it binds N-linked sialic acid receptors, the molecular docking points the virus needs to enter a cell, which are abundant in cow udder tissue but almost absent in the airways 1. H5N1 is a highly pathogenic bird-flu subtype that has been infecting US dairy herds since 2024; the University of Pittsburgh is a public research university in Pennsylvania.

The receptor mismatch explains a two-year blind spot. Because the virus settles in the mammary gland, it presents as mastitis, an udder inflammation, rather than as influenza. A dairy vet seeing inflamed udders reaches for antibiotics, not a flu test, so the herd reads as a routine bovine illness while H5N1 multiplies and spreads. Surveillance keyed to respiratory signals was watching the wrong tissue.

Two consequences follow. Infected cows shed high viral loads in their milk, which puts the hazard on the people doing the milking; pasteurisation destroys the virus, but raw milk does not. The receptor map also gives researchers a way to predict which tissues a flu virus will favour in which species before the next spillover, rather than reconstructing it afterwards. The finding lands while the CDC reports H5N1 in roughly 1,053 dairy herds across 17 states, and after the agency shifted its bird-flu reporting from weekly to monthly , thinning the very surveillance the paper shows was already looking in the wrong place.

Deep Analysis

In plain English

H5N1 is a strain of bird flu. It has been spreading among US dairy cattle since at least 2024, infecting more than 1,000 herds across 17 states. Until June 2026, scientists did not understand why a bird flu virus, which typically causes respiratory illness, was making cows sick through their udders rather than their lungs. A team at the University of Pittsburgh has now worked out the reason. Cow udder tissue contains a specific type of receptor (N-linked sialic acid receptors) that H5N1 can latch onto. The cow's lungs have almost none of these receptors, so the virus goes straight to the udder. This explains why dairy farmers saw signs of mastitis (a common udder infection) rather than coughing and sneezing. It also explains why early surveillance, which was looking for respiratory illness, missed the outbreak for approximately two years. The virus in infected milk is destroyed by pasteurisation; raw milk is unsafe.

Deep Analysis
Root Causes

The two-year surveillance gap reflects a structural mismatch between the pathogen's presentation and the surveillance system's design. The CDC's National Veterinary Services Laboratories and USDA APHIS were both watching for respiratory illness in dairy cattle, the expected presentation of an avian influenza virus, while H5N1 B3.13 was causing mastitis, dropping milk production, and reducing udder function.

Mastitis is a routine, economically significant dairy-cattle disease; an economic incentive exists for farms not to report it. The absence of a mandatory milk-sampling surveillance protocol (which the Kuchipudi paper implicitly recommends) meant the reservoir remained invisible until bulk milk testing began in 2024.

The USDA's decision to end mandatory interstate pre-movement testing on 1 May 2026 (context: ) removed the one surveillance instrument that had been generating real-time herd data, arriving one month after the receptor mechanism was identified but months before any replacement surveillance protocol was in place.

What could happen next?
  • Opportunity

    The receptor-mapping framework can now be applied prospectively to other influenza strains to predict tissue tropism in novel hosts before an outbreak, potentially reducing the surveillance-design lag that allowed H5N1 to circulate undetected for two years.

    Medium term · Assessed
  • Risk

    CDC's May 2026 shift to monthly H5N1 reporting (ID:3829) reduced the cadence of surveillance that the Kuchipudi paper shows was already watching the wrong tissue, compounding the data gap rather than closing it.

    Immediate · Reported
  • Consequence

    Mandatory milk-sampling surveillance for influenza A across US dairy herds becomes the logical policy response to the finding; without it, the next novel strain with udder tropism replicates the two-year blind spot.

    Short term · Assessed
First Reported In

Update #8 · Ebola reaches France through a screening blind spot

University of Pittsburgh / Science Advances· 25 Jun 2026
Read original
Causes and effects
This Event
Why H5N1 hid in cow udders
Surveillance watched the airways while the virus sat in the udder, so a respiratory-keyed system was structurally blind to a mammary infection.
Different Perspectives
Germany (evacuation recipient)
Germany (evacuation recipient)
Germany received the Bundibugyo outbreak's third international medical evacuation on 13 July, a US humanitarian worker infected in Bunia on 10 July. The evacuation, following a French doctor's 24 June departure and May's first US case, tests whether isolation and biocontainment protocols scale beyond DR Congo's own borders.
Pennsylvania Department of Public Health
Pennsylvania Department of Public Health
PDPH retested and retracted a false-positive measles wastewater signal on 6 July, then confirmed and publicised a real airport exposure from 4 July, with commissioner Palak Raval-Nelson stressing there is no broad threat to the general public. The national count, 2,231 cases across 42 states by 9 July, is on pace to beat 2025's 2,289-case record before September.
World Health Organization
World Health Organization
WHO published its first dedicated Blueprint on fungal disease and antifungal resistance on 1 July, estimating more than 300 million people suffer serious fungal disease annually. The Blueprint names the gap in WHO's own AMR strategy rather than waiting for an external audit to force the admission.
Africa CDC
Africa CDC
Africa CDC issued a formal 11 July appeal for responder protection, training and psychosocial support after health-worker infections tripled from 34 to 112 in a month. The appeal repeats June's unmet call for a rapid Bundibugyo diagnostic test, showing the ask has shifted from tools to basic safety and pay.
Front-line health workers, Ituri Province
Front-line health workers, Ituri Province
Health workers in Ituri Province walked off the job or threatened to strike over unpaid hazard pay and delayed salaries, even as responder infections tripled to 112 with 35 dead. Their absence narrows the isolation workforce the CDC's model says must reach 70% coverage to avoid a 20,000-case worst case.
ECDC
ECDC
ECDC co-published the isolation and contact-tracing figures behind WHO's DON612, tracking Ituri's isolation rate rising from 35 to 44 percent while still rating EU/EEA import risk as very low. Brussels backs the WHO line against travel restrictions, the position France's own contact-tracing response, not the US entry ban, actually validated.