Abstract Title
Modelling the risk of reassortment between H5N1 avian influenza and seasonal influenza in US farmworkers
Abstract
In March 2024, avian influenza (AI) was reported for the first time in US cattle, with increasing spread among farms to date. Arrival of the 2024/25 seasonal influenza (SI) season rears pandemic preparedness challenges over possible reassortment of AI. This study aims to quantify the risk of reassortment, a known driver of historical influenza A pandemics, in a US farm worker co-infected with SI and AI this coming season.
AI and SI data were collected from USDA, WAHIS and WHO FluNet. A Bayesian stochastic probabilistic model with 10,000 simulations was applied. Human SI and livestock AI forecasts were used to estimate likelihood of an infected farmer encountering an infected animal. Livestock outbreaks were forecast with Holt-Winters exponential smoothing, while SI cases were forecast using a custom-built ensemble model comprising five ARIMA forecasts. Modelled probabilities of spillover likelihood (resulting in coinfection in an already-infected farmer) and recombination were taken from literature estimates.
As of 1 November 2024, 412 outbreaks have been reported in US cattle. We forecast 9.8% [8.8%, 11.0%] of farmworkers in the US will get infected with SI and a further 223 [84, 362] outbreaks of AI in livestock will occur over the winter period. The reassortment risk increases from 0.207% [0.049%, 0.358%] in spring/summer to 2.515% [1.074%, 3.959%] in winter, representing over a tenfold (1118%) increase.
Although AI reassortment is a low-probability event, its potential to generate highly pathogenic, transmissible strains is of high impact and warrants vigilance and continued risk assessment through monitoring, biosecurity, and pandemic preparedness.
AI and SI data were collected from USDA, WAHIS and WHO FluNet. A Bayesian stochastic probabilistic model with 10,000 simulations was applied. Human SI and livestock AI forecasts were used to estimate likelihood of an infected farmer encountering an infected animal. Livestock outbreaks were forecast with Holt-Winters exponential smoothing, while SI cases were forecast using a custom-built ensemble model comprising five ARIMA forecasts. Modelled probabilities of spillover likelihood (resulting in coinfection in an already-infected farmer) and recombination were taken from literature estimates.
As of 1 November 2024, 412 outbreaks have been reported in US cattle. We forecast 9.8% [8.8%, 11.0%] of farmworkers in the US will get infected with SI and a further 223 [84, 362] outbreaks of AI in livestock will occur over the winter period. The reassortment risk increases from 0.207% [0.049%, 0.358%] in spring/summer to 2.515% [1.074%, 3.959%] in winter, representing over a tenfold (1118%) increase.
Although AI reassortment is a low-probability event, its potential to generate highly pathogenic, transmissible strains is of high impact and warrants vigilance and continued risk assessment through monitoring, biosecurity, and pandemic preparedness.
Co-Author(s)
Samuel Cutler, MSc1
Selina Kim, MSc 1
Will Gilks, PhD1
Matt Linley, PhD1
Jacqueline Buchanan, MSc1
Affiliation:
1 Airfinity, LLC
Abstract Category
Avian influenza in mammals, pandemic preparedness, and one health