Abstract During 2025, the number of EU Member States affected by African swine fever (ASF) increased from 13 to 14, following the detection of African swine fever virus (ASFV) in wild boar in Spain. The number of ASF outbreaks in domestic pigs in the EU increased from 333 in 2024 to 585 in 2025, largely driven by Romania, accounting for 81% of outbreaks in 2025. Most outbreaks (91%) occurred in establishments with more than 100 pigs, while 26 outbreaks were reported in establishments with 1000–10,000 pigs, and 11 in establishments with more than 10,000 pigs. As in previous years, ASF outbreaks in domestic pigs showed a summer seasonality. Most outbreaks were detected through passive surveillance based on clinical suspicions (84%). Among outbreaks in farms with more than 1000 pigs (n = 37), 65% (n = 24) were identified via systematic testing of two dead pigs per week. Overall, 518,088 samples from domestic pigs were analysed in the EU in 2025. In wild boar, the number of outbreaks notified in the EU increased from 7677 in 2024 to 11,036 in 2025, marking a change from the relatively stable situation of 2022–2024. Poland and Germany accounted for 31% and 18% of the total, respectively. Similarly to previous years, a winter seasonality in wild boar was observed in several Member States. Overall, 28% of the 44,578 wild boar carcasses tested positive for ASFV by polymerase chain reaction (PCR), representing 71% of wild boar outbreaks notified in the EU. In contrast, 1% of the 531,832 hunted wild boar tested positive by PCR, representing 27% of wild boar outbreaks. Despite the higher number of ASF outbreaks, the average size of the area under restriction in the EU due to outbreaks in domestic pigs increased only slightly in 2025 (+2%), while the average size of the area under restriction due to outbreaks in domestic pigs and in wild boar remained at a similar level as in 2024.

The first epidemic of lumpy skin disease (LSD) in France was detected in June 2025, with a total of 117 outbreaks recorded by the end of the year. This fast-spreading vector-borne disease of cattle prompted the implementation of strict control measures, including total depopulation of affected herds, resulting in the culling of more than 3500 cattle. Over the course of the epidemic, this measure became increasingly unacceptable, leading to major protests. To contribute to this veterinary public health debate, we present a mathematical model, accounting for both cattle and vector populations, that compares within-herd transmission dynamics under selective or total depopulation strategies, and different vector control scenarios. The selective depopulation strategies are modelled based on a bi-daily test-and-cull approach implemented upon detection of the first clinical case in an unvaccinated herd, with imperfect diagnostic tests capable of detecting infection in asymptomatic cattle under different assumptions of test sensitivity and of time from infection to detectability. Our model shows that the selective culling strategy is insufficient to control the spread of the disease in the absence of vector population control, with the entire herd eventually becoming infected. In the best-case scenario with highly sensitive and timely diagnostic tests and an 80

Abstract Infections at the animal-human or wildlife-livestock interfaces have severe health and socio-economic consequences. Combined with empirical data, mathematical models can contribute to a better understanding of the reservoirs of these infections, which is a priority for mitigating their impact by using appropriate management interventions. Taking brucellosis in the Bargy massif (French Alps) as an example of a zoonosis at the wildlife-livestock interface, we developed and calibrated a multi-host model integrating data on direct and environment-mediated cross-species contacts from field observations. Estimates of the basic reproduction number ( R 0 ) allowed to identify the population of Alpine ibex ( Capra ibex ) and its environment as an essential host in the reservoir, driving both pathogen maintenance (within-species R 0 ≥1: 1.66, 95% credible interval: 1.42-2.03) and its transmission to livestock (between-species R 0 >0: 0.035, 0.01-0.05). Our approach can be adapted to other multi-host pathogens, which will contribute to improve the understanding and management of these complex systems.

The on-going African swine fever pandemic has been devastating to affected nations, with continued spread observed despite aggressive interventions. Suspected interspecific transmission among wild and domestic hosts further complicates control efforts, yet its role in epidemic propagation remains poorly understood. Here, we develop and calibrate a multi-host mechanistic transmission model to the first wave of the epidemic in Romania (June-December 2018), quantifying these dynamics and evaluating counterfactual management scenarios. We estimated that 60% (95% credible interval: 27-83%) of outbreak farms were linked to other outbreak farms, 27% (5.3-67%) to infected wild boar populations, and 13% (1.9-27%) to external sources. For wild boar, 39% (3.8-93%) of infected populations were estimated to have originated from outbreak farms, and 61% (7.3-96%) from other infected wild boar populations, with favorable habitat exhibiting higher susceptibility and infectivity than unfavorable habitat. Among alternative control strategies, reactive and preventive culling of domestic pig herds yielded the greatest decrease in median final epidemic size among domestic pigs. These findings provide quantitative evidence that interspecific transmission was a critical epidemic driver, and a necessary target for achieving holistic control. Our model offers a flexible, rapidly-deployable framework for informing surveillance and response policy in at-risk regions.

Avian influenza (AI) is a highly contagious viral disease affecting poultry and wild water birds, posing significant global challenges due to its high mortality rates and economic impacts. Highly pathogenic avian influenza (HPAI) outbreaks, particularly those caused by H5N1 and its variants, have surged since 1959. The HPAI H5N1 clade 2.3.4.4b viruses have notably expanded their geographical reach, affecting numerous countries, diverse avian species, and now mammals. Using an ecological niche modelling approach, this study aims to elucidate the environmental factors associated with increased HPAI H5 cases since 2020, investigate potential shifts in ecological niches, and predict new areas suitable for viral circulation. We developed ecological niche models for HPAI cases in wild and domestic birds across two distinct periods: 2015–2020 and 2020–2022. Key environmental predictors include chicken and duck population density, human density, distance to water bodies, and land cover variables. Post-2020, we observe increased relative influence of predictors such as intensive chicken population density and cultivated vegetation. Risk maps reveal notable ecological suitability for HPAI H5 circulation in Europe, Asia, and the Americas, with significant expansions of at-risk areas post-2020. Wild bird H5 occurrences appear primarily correlated with urban areas and open water regions. Our analyses also highlight a potential shift in affected wild bird species diversity, with more avian species, particularly sea birds, impacted post-2020. Overall, these results further contribute to the understanding of HPAI epidemiology and identify regions where surveillance and control measures should be prioritised.

Highly pathogenic H5Nx viruses of clade 2.3.4.4b have spread worldwide, causing major economic losses and increased human exposure. Since 2020, multiple mammalian infections have been reported, raising concerns about further adaptation to mammalian hosts. We analyzed influenza A virus sequences from the Influenza Virus Database at the National Center for Biotechnology Information to identify new mammalian adaptation markers in the polymerase complex and nucleoprotein, using recursive partitioning. These markers were grouped into “proteotypes” to assess their co-occurrence and association with host origin. This analysis revealed distinct groups of proteotypes linked to mammalian adaptation, including those seen in historical and pandemic human strains. Identified mutations were introduced alone or in combination into a 2.3.4.4b H5N8 virus to evaluate their impact on polymerase activity in mammalian cells using a minigenome assay. PB1 V336I and PB2 K702R increased polymerase activity in human cells, particularly with PB2 E627K, supporting enhanced surveillance of 2.3.4.4b H5Nx viruses. These findings highlight mutation combinations relevant for enhanced surveillance of 2.3.4.4b H5Nx viruses.

Carp edema virus (CEV), a member of the Poxviridae family, has been a significant pathogen in koi and common carp since its initial identification in Japan during the 1970s. CEV, the causative agent of Koi Sleepy Disease (KSD), can cause high mortality rates and has been reported in many countries and is often linked to the fish trade. The virus is typically detected through DNA analysis of gill tissues, where the highest viral loads are found. However, traditional sampling methods, such as gill sampling, are lethal, complicating routine surveillance, particularly in asymptomatic or high-value koi. This study aimed to evaluate nonlethal sampling methods for CEV surveillance in the koi trade. We analysed various shipping environment samples, such as shipping water and fish bag swabs, alongside gill swabs from anaesthetised fish and gills from naturally deceased fish. Using qPCR, we found that the sensitivity of environmental samples, particularly shipping water, was greater than that of direct fish samples. Latent class modelling estimated that the sensitivity associated with 1.5 mL shipping water samples was greater than 89%, making them a reliable alternative for early detection. All detected variants belonged to genogroup II. Some post-import outbreaks shared variants with earlier outbreaks or shipping environment samples, suggesting that the detected DNA generally reflected infectious particles rather than just free environmental DNA and indicating that CEV can go unnoticed for several months after importation. These findings highlight the utility of environmental samples for effective, non-invasive surveillance and improved biosecurity management in the koi trade.

Dynamic modelling of infectious diseases of importance to livestock production is a valuable tool for policy and decision makers. Mathematical and simulation models play an essential role in understanding complex systems, but parameterising these models can be challenging, especially in data-sparse environments. When parameters are unable to be estimated from epidemiological or experimental data, a time-consuming and labour-intensive literature review-to identify suitable literature-informed values-is often necessary. In service of this, here we present PARAMETRA, a parameter database for 20 pathogens of livestock, envisaged as an open-source collaborative tool for the research community to aid in the development of future transmission models of livestock pathogens. Pathogens included in the database so far were selected using a disease prioritisation exercise. Parameters of interest were selected by experts with a strong background in epidemiology and mathematical modelling. We populated the database with over 2000 individual values, covering a wide range of different parameters including transmission rates, diagnostic test efficacies, pathogen survival on surfaces, and the farm and regional level prevalences of selected diseases. Finally, we present an initial illustrative analysis of the database contents and the associated metadata of studies included. One of the principal conclusions we can draw from the data available is that in many cases research is reactive, rather than proactive, with research only tending to focus on specific diseases after outbreaks have already occurred, as is the case for African swine fever for example. This has important implications for future research moving to a more proactive approach for experimental and epidemiological studies based on observations of gaps in the data, and high-risk diseases. This publication represents the first step in development for the PARAMETRA database, which will be updated and expanded in the coming years.

The continuous spread of highly pathogenic avian influenza H5 viruses poses significant challenges, particularly in regions with high poultry farm densities where conventional control measures are less effective. Using phylogeographic and phylodynamic tools, we analysed virus spread in southwestern France in 2020 and 2021, a region with recurrent outbreaks. Following a single introduction, the virus spread regionally, mostly affecting duck farms, with an average velocity of 10.5 km/week and a peak at 27.8 km/week in mid-December. The effective reproduction number between farms (Re) exceeded 1 in December, peaking at 3.8 in mid-December. Transmission declined after late December, coinciding with extensive preventive culling, suggesting that standard 3- and 10-km zones were more effective when combined with timely, large-scale interventions. Farm infectiousness was estimated around 9 days, highlighting a critical window of undetected viral shedding and transmission risk. Duck farm density and poultry farm proximity were key drivers of virus spread, likely due to the higher virus susceptibility and transmission efficiency of ducks compared to chickens. We identified density and proximity thresholds required to maintain effective control (Re < 1). These findings offer actionable guidance to support regional biosecurity and to improve the robustness of the poultry sector to mitigate future outbreaks.

Contact tracing is commonly used to manage infectious diseases of both humans and animals. It aims to detect early and control potentially infected individuals or farms that had contact with infectious cases. Because it is very resource-intensive, contact tracing is usually performed on a pre-defined time window, based on previous knowledge of the duration of the incubation period. However, pre-defined time windows may not be always relevant, reducing the efficiency of contact tracing. In this study, we estimated the day when farms were first infected with highly pathogenic avian influenza viruses, a devastating pathogen causing severe socio-economic damage in domestic poultry. The estimation was performed by fitting a stochastic mechanistic model to observed daily mortality data from 63 infected poultry farms in France and The Netherlands, using approximate Bayesian computation. Independent of the poultry species or country, the estimates of the time of first infection ranged between 3.4 (95% credible interval—CrI: 2.6, 4.6) and 19.9 (95% CrI: 11.9, 31.3) days prior to the last observation. We developed an online application to provide real-time support to policymakers by estimating realistic ranges of dates of first infection to inform contact tracing and improve its efficiency.

The continuous spread of highly pathogenic avian influenza H5 viruses poses significant challenges, particularly in regions with high poultry farm densities where conventional control measures are less effective. Using phylogeographic and phylodynamic tools, we analysed virus spread in Southwestern France in 2020-21, a region with recurrent outbreaks. Following a single introduction, the virus spread regionally, mostly affecting duck farms, peaking in mid-December with a velocity of 27.8 km/week and an effective reproduction number between farms ( R e) of 3.8, suggesting the virus can spread beyond current control radii. Transmission declined after late December following preventive culling. Farm infectiousness was estimated around 9 days. Duck farm density was the main driver of virus spread and we identified farm density and proximity thresholds required to maintain effective control ( R e < 1). These findings offer actionable guidance to support regional biosecurity and to improve the robustness of the poultry sector to mitigate future outbreaks. ### Competing Interest Statement The authors have declared no competing interest.

Effective mortality thresholds are critical for timely reporting and management of highly pathogenic avian influenza. Using standard modelling techniques, we evaluated the performance of different mortality thresholds in mule duck flocks. Using an eightfold increase of the mortality for two consecutive days compared to the average mortality the previous week led to a perfect classification of all flocks used for validation (12 affected and 18 non-affected flocks). A fixed daily threshold of 0.25% showed a perfect sensitivity and a good specificity (3 false positives/18). Our results fill a knowledge gap and can inform HPAI surveillance policy in non-vaccinated mule ducks.

Highly pathogenic avian influenza causes substantial poultry losses and zoonotic concerns globally. Duck vaccination against highly pathogenic avian influenza began in France in October 2023. Our assessment predicted that 314-756 outbreaks were averted in 2023-2024, representing a 96%-99% reduction in epizootic size, likely attributable to vaccination.

During 2024, the number of EU Member States affected by African swine fever (ASF) decreased from 14 to 13, with Sweden regaining freedom and no new Member State becoming infected. ASF outbreaks in domestic pigs in the EU declined by 83% compared to 2023, primarily due to fewer outbreaks in Croatia and Romania, although Romania notified 66% of the 333 outbreaks in the EU. Most outbreaks (78%) occurred in establishments with fewer than 100 pigs. However, an increase in outbreaks in establishments with more than 100 pigs was observed in Italy and Poland. Like previous years, there was a clear seasonality for domestic pig outbreaks, with 51% of them notified between July and September. Most of the outbreaks in domestic pigs were detected through passive surveillance based on clinical suspicion (79.4%), while fewer outbreaks were detected through enhanced passive surveillance involving systematic testing of dead pigs (14.2%) and 6.4% through tracing contacts after outbreak detection. In wild boar, the number of outbreaks notified has remained stable since 2022 (between 7000 and 8000) with a less clear seasonality than for domestic pigs, and a winter peak observed only in Hungary, Italy, Poland and Slovakia. Overall, 29% of the 23,919 wild boar carcasses found during passive surveillance activities tested positive for ASFv by PCR, representing 70.4% of the wild boar outbreaks in the EU. In contrast, around 0.4% of the 412,753 hunted wild boar tested positive by PCR, representing 28.4% of the wild boar outbreaks. While the use of serological tests performed in wild boar decreased, the number of PCR tests remained stable. Despite the reduction in the number of outbreaks in domestic pigs, the total size of the restricted zones III in the EU remained stable, with a slight increase in restricted zones II + III in 2024.

In this study, we present a comprehensive analysis of the key spatial risk factors and predictive risk maps for HPAI infection in France, with a focus on the 2016-17 and 2020-21 epidemic waves. Our findings indicate that the most explanatory spatial predictor variables were related to fattening duck movements prior to the epidemic, which should be considered as indicators of farm operational status, e.g., whether they are active or not. Moreover, we found that considering the operational status of duck houses in nearby municipalities is essential for accurately predicting the risk of future HPAI infection. Our results also show that the density of fattening duck houses could be used as a valuable alternative predictor of the spatial distribution of outbreaks per municipality, as this data is generally more readily available than data on movements between houses. Accurate data regarding poultry farm densities and movements is critical for developing accurate mathematical models of HPAI virus spread and for designing effective prevention and control strategies for HPAI. Finally, our study identifies the highest risk areas for HPAI infection in southwest and northwest France, which is valuable for informing national risk-based strategies and guiding increased surveillance efforts in these regions.

The ongoing panzootic of highly pathogenic avian influenza (HPAI) H5 clade 2.3.4.4b has caused widespread poultry mortality and raised concerns about zoonotic pandemics and wildlife conservation. France recently adopted a preventive vaccination strategy, vaccinating domestic ducks with inactivated and mRNA vaccines. This study evaluates the impact of this campaign on reducing HPAI H5 outbreaks. Using predictive modeling based on previous outbreak data, the expected number of outbreaks in 2023-24 without vaccination was significantly higher than the observed cases, indicating a 95.9% reduction attributable to vaccination. These findings suggest that vaccination effectively mitigated the HPAI H5 outbreak in France. ### Competing Interest Statement The authors have declared no competing interest.