Risk of H5N1 combining with seasonal flu to create a more dangerous pathogen increases five-fold this winter

New modelling by Airfinity shows the incoming winter flu season brings a nearly five-fold increase in the risk of an influenza reassortment event in the United States. 

Reassortment can occur when two strains of a virus infect the same cell simultaneously, creating the opportunity for the virus to exchange gene segments to become a new hybrid virus, that can potentially become more pathogenic. Reassortment has been the mechanism behind most, if not all, historical Influenza A pandemics, including most recent with the 2009 H1N1 pandemic strain.

Today’s modelling looks at the risk of a reassortment event in US farmer workers this winter. It shows the risk increases nearly five-fold, by 479% in winter as seasonal flu cases rise. 

The H5N1 avian influenza outbreak in US cattle has continued for approximately 10 months and shows no signs of ending before the colder weather begins. The USDA reported tally of infected herds has reached 298, with the current hotspot for outbreaks in California, and the previous hotspot in Colorado. 

Globally, 30 human infections of H5N1 have been reported this year, 18 of these have been in the United States. The cases outside the US were in Cambodia, Vietnam and Australia (with the latter case likely infected in India before travelling to Australia). These all involved people with direct exposure to sick birds. 

The H5N1 outbreak in cattle remains isolated within the US and has not been reported in any other country. Of the human cases reported in the US, nine had exposure to infected birds, eight had exposure to infected dairy cows, while one infected individual in Missouri had no known animal exposure, highlighting the risk of undetected transmission. H5N1 does not currently spread effectively between humans, however, a reassortment event has the potential to change how the virus spreads, posing a greater risk to humans. 

Part of the US response to the outbreak has been to increase vaccination of farm workers against seasonal flu. However early data suggests the effectiveness of the 2024 flu vaccine to be as low as 35% which may limit the impact of this strategy.

Unlike COVID-19, avian influenza is not a novel pathogen and has long been considered a global health threat with pandemic potential. For H5N1, there are approved diagnostics, vaccines and treatments, how effective they are will depend on how the virus changes over time. 

There are currently 18 pandemic vaccines approved around the world. In the US, three vaccine candidates have been approved from CSL Seqirus, Sanofi and ID Biomedical Corporation of Quebec. Our procurement data shows that 700 million pandemic vaccines doses have been secured by governments since 2019. Seven influenza antivirals have been approved globally, including Tamiflu which has been widely stockpiled by governments. 

Speaking to Airfinity during a webinar on the issue, Florian Krammer, Professor at Icahn School of Medicine at Mount Sinai, said: “We need to stockpile vaccines, we need to seriously think about vaccinating cows and figure out if that works. We need to lower the barrier to testing. Right now everything goes through the CDC, I think every clinical laboratory in the US should have the ability to test for an H5N1 infection. And the last point is really about information. It’s not just about cows, we’ve found H5N1 in New York City birds, people get exposed to this virus and there are no information campaigns. People should know what H5N1 is and that they shouldn’t touch sick or dead birds.”

Angie Rasmussen, a virologist at the Vaccine and Infectious Disease Organization at the University of Saskatchewan in Canada said: “We need to expand testing, without sufficient testing it is impossible to understand the scale and scope of it. We need to first prioritise getting vaccines to people who are at high risk of occupational exposure as well as strongly considering vaccinating cows. We should also expand our surveillance efforts and not limit it to species we already know are susceptible, we need to try and find places where essentially the virus could hide, develop and spread, which is how the cattle outbreak got so out of control in the first place.”

Hanna Nohynek, Chief Physician at the Finnish Institute for Health and Welfare, told the audience about Finland’s program to deploy H5N1 vaccines. She said: “We had prepared to vaccinate 10,000 people. The vaccinations started at the very best time ahead of summer vacations. There was intensive campaigning to reach these people, so they knew the vaccine was available and free of charge. But I have to say that until last week, out of the 10,000 people we were hoping to reach, we have only vaccinated about 500. So there still remains a lot to be done.”

Bob Kadlec, former Assistant Secretary of Preparedness and Response at the U.S. Department of Health & Human Services (HHS) told the webinar audience: “Let’s do what Finland has done (and vaccinate farm workers against H5N1), expand testing in animals and humans extensively, immunisation in cows, initiate dose sparing vaccine trials in humans and offer doses to high-risk workers, at least in the agricultural community to limit exposure.”

ENDS

Methodology

A probabilistic model was developed to estimate the risk of farm workers in the US being co-infected, and subsequently having a reassortment event occur, with both seasonal flu and H5N1 (bird flu) during the winter season. The model first predicts the number of farm workers likely to be infected with seasonal flu using machine learning. This calculates the probability of flu infection on any given day, taking into account how long the infection lasts. At the same time, the model predicts outbreaks of bird flu in livestock and calculates the chance that animals are infected with bird flu during the same period. Combining these forecasts with data on how frequently farm workers interact with animals, the model determines the probability of a worker coming into contact with an infected animal while also having the flu. The model further uses scientific literature to estimate the likelihood of an infected worker catching bird flu via a spillover event when coming into contact with an infected animal and then experiencing co-infection with both viruses. It then assesses the chance that these two flu strains mix, leading to a reassortment event that could create a new flu strain. The overall risk for the winter is compared with the risk for the summer months, using historical data on flu cases in both humans and animals. This comparison was used to calculate how much greater the risk of reassortment is in winter than in summer.

Limitations

Forecasted seasonal influenza cases are based on reported data, but are subject to testing rates and may not account for all infections. Forecasted avian influenza outbreaks in livestock are based on reported outbreaks, however, it may be that not all outbreaks have been detected and reported. The model takes averages of livestock numbers in farms, however, the size of different farms can vary significantly, though the model assumes a good distribution of outbreaks across both large and smaller farms.