What we’re starting to learn about H5N1 in cows, and the risk to people

The H5N1 bird flu virus has been around for decades, and the damage it wreaks on chickens and other poultry is well documented. But the recent discovery that the virus has jumped into dairy cattle — whose udders seem to be where the virus either infects or migrates to — has dumbfounded scientists and agricultural authorities.

Questions for which there are pretty clear answers when it comes to birds are suddenly unsettled science in cows. How are they getting infected? Are they transmitting the virus cow to cow, or are human actions — activities that are part of the day-to-day of farming — serving as an unrecognized amplifier of viral transmission? In the interface between infected cows and humans, how might people be at risk? Does consuming milk laced with live H5 virus pose a hazard?

At this point in the H5N1-in-cows story, these are questions that don’t have solid answers, though some evidence is coming into focus. Getting answers is critical, for controlling spread among cows, if that’s still possible; stopping transmission from cows to key species like poultry or pigs; and keeping the virus from getting a foothold in the human population.

Let’s explore the little we know, and the much more we don’t know.

Are cows spreading the virus the way people transmit flu?

Wash your hands; cover your cough; wear a mask. These are things public health officials advise at the start of each flu season. That’s because the influenza viruses that typically infect humans spread primarily through the air — either through small, virus-laced particles that can be inhaled, or through larger droplets that spray out of a sick person’s mouth when they cough or sneeze.

Scientists are now trying to figure out if cows are passing around the H5N1 bird flu in the same way. So far, it seems possible, but unlikely to be the main driver of disease transmission.

The U.S. Department of Agriculture has said few tests from the respiratory tracts of infected cows have come back positive — and those that did showed there wasn’t a lot of virus present. But there is at least some evidence that H5N1 is on occasion getting deep into the respiratory tracts of cows. The USDA said in a frequently asked questions document it released last week that its National Veterinary Services Laboratories had found H5N1 in a lung tissue sample from an asymptomatic dairy cow that originated from an affected herd.

So what is the main driver of infection among cows?

While the contribution of respiratory transmission is still in question, there appears to be little doubt that a lot of spread is happening in milking parlors, where cows are strapped into the milking machines, and that in dairy cows, H5N1 seems to be primarily infecting mammary glands. The amount of virus in the udders of infected cows is off-the-charts high, making it easy to see how one cow’s infection soon becomes a herd’s problem.

“You can imagine that if such a cow is milked in a milking stall, even a few drops of milk remaining on the teat cup which is used to milk will then subsequently contaminate the teat of the next cow,” said Thijs Kuiken, a pathologist in the department of viroscience at Erasmus Medical Center in Rotterdam, the Netherlands. “So you can imagine very easily how, in the course of a few days, the virus could spread very fast among the cows being milked on that farm.”

That fits with other evidence: the fact that infections so far appear to be limited to lactating dairy cows, and that the members of a herd that are the biggest milk producers — which are also the animals that stay on milking machines the longest — are the ones most heavily impacted by the disease. “There’s still a lot that’s unknown, but what we’re seeing right now is not a respiratory disease by and large; it’s a mastitic disease,” said Jared Taylor, a professor of veterinary pathobiology at Oklahoma State University. For now, that’s somewhat reassuring, because it means that increased attention to sterilization and disinfection during milking should make a dent in the virus’s further spread. But he stressed that it’s still early and researchers are just starting to study the extent to which there could be respiratory involvement.

Taylor noted another worry: H5N1, which is notorious for its ability to evolve, is being given a huge opportunity to adapt to bovine hosts. “The concern is if it becomes effective as a respiratory pathogen in cattle, it’s more likely to become effective as a respiratory pathogen in humans,” he said.

When are infected cows most contagious?

Little is known about how a cow’s contagiousness changes over the course of an infection with H5N1. Researchers examining outbreaks among dairy cattle in Kansas and Texas found that on affected farms, incidence of the disease peaked four to six days after the first animals began showing symptoms such as lethargy, fever, reduced feed intake, and an abrupt drop in milk production. The spread tapered off within 10 to 14 days, and most animals were slowly returned to regular milking after they recovered.

But researchers have not yet systematically and repeatedly collected samples from affected animals to understand how much virus they are harboring inside their noses or shedding into milk over the course of their infection.

Those are among the types of studies that are urgently needed, Donald Prater, acting director of the Center for Food Safety and Nutrition at the Food and Drug Administration, said during an online symposium last week organized by the Association of State and Territorial Health Officials (ASTHO). “As these animals become infected, what does that profile look like for viral shedding?”

Such research would help answer the question of how much H5N1 virus a cow needs to come in contact with to get infected — the so-called “infectious dose.” Right now, that remains unknown, Prater said.

Can cows be infected but not show symptoms? Can those cows shed virus in milk or spread virus to other cows?

There’s been so little testing going on — and so little disclosed about the numerators and denominators of that testing — that it’s hard to gauge how big an issue asymptomatic infection is and what role, if any, it is playing in transmission within and among herds.

But the information USDA released about the virus having been found in the lung of an asymptomatic cow is clear proof that not all infected cows have outward signs of illness.

Rosemary Sifford, the agency’s chief veterinary officer, said during the ASTHO symposium that little is known about cow-to-cow spread via the respiratory route, but she gave an example of apparent transmission from presymptomatic animals. (A person with flu can be contagious a day before developing symptoms.)

Sifford noted that cows in a herd without symptoms tested positive after cattle were moved into it from another herd whose remaining animals then went on to develop symptoms. The positive cows in the second herd haven’t developed symptoms, she said. “We are just getting underway with those studies to give us an idea of the opportunity for viremia either ahead of or after clinical signs,” she said. “So we should have more information about that in the coming weeks.”

One such study will be taking place in the high-containment laboratories at Kansas State University’s Center of Excellence for Emerging and Zoonotic Animal Diseases. Director Jürgen Richt said his group already has the necessary approvals and hopes to begin the research in mid-May.

They will be experimentally infecting lactating and non-lactating cows with the version of H5N1 that has been found in cattle herds, both by the oral-nasal route and by directly injecting the virus into udders. The goal: To see where the virus goes in the cows and how they shed it. Is it just found in milk? Or in feces and urine too?

Some uninfected cows will be housed with the infected animals to see whether they contract the virus, Richt said.

As to whether infected cows that have no symptoms are shedding virus in their milk, the evidence of viral traces in commercially purchased milk brings that question to the forefront. Farmers are supposed to discard milk from infected cows, which reportedly looks odd — yellowish and unusually thick. But PCR testing of commercially sold pasteurized milk has shown a substantial portion of samples were positive for RNA from H5N1, indicating the presence of either viral fragments or dead viruses. (The FDA said last week about one in five samples purchased in a cross-country survey tested positive.) So either some farmers aren’t following the recommendation, or some milk isn’t noticeably altered, or some cows that aren’t known to be sick are shedding the virus in their milk.

What is it about milking parlors that allows transmission to occur?

Milking parlors are typically enclosed buildings outfitted with individual stalls arranged in a ring or in rows where dairy cows are led two to three times each day to drain their udders of milk. In a milking operation, the stimulus to secrete milk comes not from the sight or touch of a calf but is usually provided by a farm worker. That person also cleans the animal’s teats with a damp cloth and then dips them into a disinfectant solution to protect them from infectious bacteria present on a farm. The teats are then attached to the milking unit, also called a claw, which consists of a cluster of four rubber or silicon-based liners that fit snugly around each teat. The milking lasts about six to nine minutes per animal, and then each cow receives another disinfectant treatment before it’s ushered out and another animal is brought in.

The problem is that the milking equipment that comes into contact with the cow’s udders is typically not sanitized between individual animals, said Nigel Cook, a professor in Food Animal Production Medicine at the University of Wisconsin-Madison and long-time dairy cow health researcher. Rather, sanitization steps happen only two or three times a day.

“Contamination of the milking unit with milk residue cross-contaminating to the next cow would be a risk — as it would with any mastitis pathogen,” Cook told STAT via email.

Liners also have to be regularly replaced, as wear and tear and chemical exposure make them lose their elasticity, becoming rough and split. When that happens, the liners are more difficult to disinfect and can act as a reservoir for infection.

Liners, dip cups, washrags, and milkers’ gloved hands are all possible means of spreading the virus from one animal to the next. Washrags used on different animals are often laundered together before repeat use, but some dairies don’t use hot water, and researchers have found genetic traces of H5N1 on both used and clean rags using PCR testing. More work is needed to figure out which vectors are playing the biggest role, scientists told STAT.

Kuiken is a bit pessimistic about whether, once the virus has found its way into a herd, transmission can be stopped: “You can’t not milk. And you probably can’t milk so well as to prevent cow-to-cow spread. I don’t think you can do it.”

What puts people working on dairy farms at highest risk?

Only one human infection — a Texas dairy farm worker who developed conjunctivitis — has been reported, but anecdotal reports abound of other farm workers with conjunctivitis and mild respiratory symptoms. Scientists at the Centers for Disease Control and Prevention are working to figure out where the biggest risks to these workers lie, Sonja Olsen, associate director for preparedness and response in CDC’s influenza division, said during the ASTHO symposium. Is it exposure to cows? Contact with milk? Are there specific activities on farms or in slaughterhouses that put people at elevated risk of contracting H5N1?

The CDC recommends that people working with or around cattle suspected or confirmed to be infected with H5N1 wear gloves, disposable fluid-resistant coveralls, vented safety goggles or a face shield, and an N95 respirator.

Kuiken said the setup of many milking parlors is almost tailor-made to put workers in contact with viruses being shed in milk. That’s because there is typically a well — think of the area under the hoist in the garage where your car gets repaired — where workers are located while cows are being milked.

“So the milk worker is standing in a depressed area, and therefore his eyes are about at knee level — a little bit higher, maybe — with the cow. So very good for being inoculated, for eye infection,” he said.

“They can milk the cow beforehand just a little bit, and the spray of milk will come out of the teat. That’s one way of getting infected. Taking off or putting on the teat cups is one way. The cows are defecating and urinating during the milking process, or afterwards. They clean the [milking parlor] using a high-pressure hose. So there are all different ways in which these milk workers can easily become infected.”

Cook is also concerned about the possibility that the high pressure hoses that are used to spray down the parlors after milking may be aerosolizing virus that has fallen to the floor, making it easier for cows — and humans —to breathe in. He and other colleagues at the University of Wisconsin have begun deploying air monitoring devices into the milking parlors of affected farms to investigate the extent to which they can find genetic evidence of the virus in the air.

If so many dairy farm workers have been and are being exposed to H5N1, why has only one human infection been confirmed? Is that plausible?

Not likely.

Farmers, who mostly haven’t been willing to have their cows tested, haven’t been keen to have their workers tested either. The CDC’s Nirav Shah, principal deputy director, admitted as much last week. “These situations are challenging. There may be owners that are reluctant to work with public health to say nothing of individual workers who may be reluctant to sit down with somebody who identifies themself as being from the government in some way.” There have been reports that some farms may employ undocumented workers.

Dairy farm workers presenting with milder signs of illness, like conjunctivitis, also would make cases harder to spot. If only conjunctivitis is occurring, though, that presents another potential clue that people aren’t currently getting sick by breathing in H5N1, said Angela Rasmussen, a virologist who studies emerging zoonotic pathogens at the Vaccine and Infectious Disease Organization at the University of Saskatchewan in Canada. In the rare cases where people have contracted H5N1 from birds — mostly poultry farm workers and people with small backyard flocks — infections have resulted from close contact or exposure to aerosolized poultry waste. Since 2003, there have been 889 confirmed cases of human infection with H5N1; of those 52% were fatal.

“It’s entirely possible that there have been a lot of human cases from cows and people are just getting it in a different way,” Rasmussen said. “What that says about the risks to farm workers and dairy processing staff on a given farm or milking parlor is really unclear to me and it’s also something we need to urgently catch up on.”

Should more be done to find infected people?

The World Health Organization appears to be concerned about the possibility of undetected human cases. Maria Van Kerkhove, acting head of the department of epidemic and pandemic preparedness, told STAT she’d like to see, among other things, studies looking for antibodies to H5N1 in the blood of farm workers and people who’ve been in contact with farm workers, to determine if there have been unreported cases and possibly even spread from those individuals to others.

“For every human case of avian influenza since its emergence in 1996, 1997, there has always been … active case-finding, extensive testing, serologic surveys around humans. And we would certainly want to see that in any country, including in the U.S.,” said Van Kerkhove. “I’ve heard and read quite a bit about some potential additional cases. We’ve heard maybe conjunctivitis or whatnot. But I think what’s really critical now is to understand the extent of infection in humans.”

“The more we look, potentially the more we can find. And if we’re not looking, then we’re perhaps missing something that could be quite significant.”

Could people become infected by drinking milk containing H5N1 viruses?

Research is still being done to determine if all pasteurization techniques — there are multiple approaches — inactivate the virus. But the findings so far are reassuring. Scientists at St. Jude Children’s Hospital in Memphis, Tenn., have reported that while they could find genetic evidence of the virus in milk bought in stores, they could not grow live viruses from that milk, suggesting pasteurization kills H5N1.

But given the concentration of virus that researchers are seeing in milk from infected cows, they believe that raw milk — milk that has not been pasteurized — is an entirely different story.

Because so little testing has been reported, it’s really not known how many herds in how many parts of the country have been or are currently infected with H5N1. The USDA has reported 34 outbreaks in nine states since late March. It has not disclosed whether any of those dairy farms sold raw milk. Nor is it known if any raw milk producers have had bird flu infections in their cows.

If cows that produce milk destined for the raw milk market got infected with H5N1, people who consume that milk could drink a large dose of the virus, scientists say. Kuiken said he’s heard of concentrations that would be the equivalent of a billion virus particles per milliliter of milk. He told STAT he thought authorities should ban raw milk sales while the outbreak is underway.