Within-flock transmission of H7N1 highly pathogenic avian influenza virus in turkeys during the Italian epidemic in 1999-2000

Are all avian influenza outbreaks in poultry the same? The predicted impact of poultry species and virus subtype

AIMS

Outbreaks of avian influenza in poultry farms are currently increasing in frequency, with devastating consequences for animal welfare, farmers and supply chains. Some studies have documented the direct spread of the avian influenza virus between farms. Prevention of spread between farms relies on biosecurity surveillance and control measures. However, the evolution of an outbreak on a farm might vary depending on the virus strain and poultry species involved; this would have important implications for surveillance systems, epidemiological investigations and control measures.

METHODS AND RESULTS

In this study, we utilized existing parameter estimates from the literature to evaluate the predicted course of an epidemic in a standard poultry flock with 10,000 birds. We used a stochastic SEIR simulation model to simulate outbreaks in different species and with different virus subtypes. The simulations predicted large differences in the duration and severity of outbreaks, depending on the virus subtypes. For both turkeys and chickens, outbreaks with HPAI were of shorter duration than outbreaks with LPAI. In outbreaks involving the infection of chickens with different virus subtypes, the shortest epidemic involved H7N7 and HPAIV H5N1 (median duration of 9 and 17 days, respectively) and the longest involved H5N2 (median duration of 68 days). The most severe outbreaks (number of chickens infected) were predicted for H5N1, H7N1 and H7N3 virus subtypes, and the least severe for H5N2 and H7N7, in which outbreaks for the latter subtype were predicted to develop most slowly.

CONCLUSIONS

These simulation results suggest that surveillance of certain subtypes of avian influenza virus, in chicken flocks in particular, needs to be sensitive and timely if infection is to be detected with sufficient time to implement control measures. The variability in the predictions highlights that avian influenza outbreaks are different in severity, speed and duration, so surveillance and disease response need to be nuanced and fit the specific context of poultry species and virus subtypes.

Estimation of introduction time window of highly pathogenic avian influenza virus into broiler chicken farms during the 2020 - 2021 winter season outbreak in Japan

When an infectious disease occurs in an area, early detection of infected farms is important to respond quickly and contain the outbreak on a small scale. Estimating the time window for the introduction of the infection is important for its prevention and control. The aim of this study was to estimate the farm-specific time window from the introduction of the highly pathogenic avian influenza (HPAI) virus into poultry farms using field data from the HPAI H5N8 outbreak in the 2020-2021 winter season in Japan. Daily mortality data from 12 broiler chicken farms during the outbreak were used for the analysis. A mathematical model (Susceptible-Exposed-Infectious-Removed, SEIR model) was applied to generate the within-flock transmission of HPAI. The model-predicted mortality was fitted to the observed excess mortality data induced by HPAI to estimate the farm-specific transmission rate and the time of virus introduction. The estimated value of the transmission rate in each farm was 1.449 day^-1 in median (min: 0.661 day^-1, max: 3.387 day^-1). The time window from the introduction of the virus to notification in each farm was estimated at 14.0 days in median (min: 8.6 days, max: 24.1 days) in the deterministic model. In addition, in the stochastic model considering the randomness of transmission in the early phase of the outbreak, the upper value of 95 % credible interval of the time window ranged from 12 to 34 days, with a median of 21 days. The results suggest that although one to three weeks had elapsed on most farms until notification after the virus introduction, the time window could exceed three weeks considering the stochasticity of disease transmission. As for the potential farm characteristics affecting within-flock transmission, the transmission rate was smaller (p-value=0.02) and the estimated time window from introduction to notification was longer (p-value=0.02) when birds were older. This study provides reliable information for setting up a tracing period for a potential source farm and enhancing the efforts for early detection.

Estimating epidemiological parameters using diagnostic testing data from low pathogenicity avian influenza infected turkey houses

Limiting spread of low pathogenicity avian influenza (LPAI) during an outbreak is critical to reduce the negative impact on poultry producers and local economies. Mathematical models of disease transmission can support outbreak control efforts by estimating relevant epidemiological parameters. In this article, diagnostic testing data from each house on a premises infected during a LPAI H5N2 outbreak in the state of Minnesota in the United States in 2018 was used to estimate the time of virus introduction and adequate contact rate, which determines the rate of disease spread. A well-defined most likely time of virus introduction, and upper and lower 95% credibility intervals were estimated for each house. The length of the 95% credibility intervals ranged from 11 to 22 with a mean of 17 days. In some houses the contact rate estimates were also well-defined; however, the estimated upper 95% credibility interval bound for the contact rate was occasionally dependent on the upper bound of the prior distribution. The estimated modes ranged from 0.5 to 6.0 with a mean of 2.8 contacts per day. These estimates can be improved with early detection, increased testing of monitored premises, and combining the results of multiple barns that possess similar production systems.

Estimating the introduction time of highly pathogenic avian influenza into poultry flocks

The estimation of farm-specific time windows for the introduction of highly-pathogenic avian influenza (HPAI) virus can be used to increase the efficiency of disease control measures such as contact tracing and may help to identify risk factors for virus introduction. The aims of this research are to (1) develop and test an accurate approach for estimating farm-specific virus introduction windows and (2) evaluate this approach by applying it to 11 outbreaks of HPAI (H5N8) on Dutch commercial poultry farms during the years 2014 and 2016. We used a stochastic simulation model with susceptible, infectious and recovered/removed disease stages to generate distributions for the period from virus introduction to detection. The model was parameterized using data from the literature, except for the within-flock transmission rate, which was estimated from disease-induced mortality data using two newly developed methods that describe HPAI outbreaks using either a deterministic model (A) or a stochastic approach (B). Model testing using simulated outbreaks showed that both method A and B performed well. Application to field data showed that method A could be successfully applied to 8 out of 11 HPAI H5N8 outbreaks and is the most generally applicable one, when data on disease-induced mortality is scarce.

A review of knowledge discovery process in control and mitigation of avian influenza

In the last several decades, avian influenza virus has caused numerous outbreaks around the world. These outbreaks pose a significant threat to the poultry industry and also to public health. When an avian influenza (AI) outbreak occurs, it is critical to make informed decisions about the potential risks, impact, and control measures. To this end, many modeling approaches have been proposed to acquire knowledge from different sources of data and perspectives to enhance decision making. Although some of these approaches have shown to be effective, they do not follow the process of knowledge discovery in databases (KDD). KDD is an iterative process, consisting of five steps, that aims at extracting unknown and useful information from the data. The present review attempts to survey AI modeling methods in the context of KDD process. We first divide the modeling techniques used in AI into two main categories: data-intensive modeling and small-data modeling. We then investigate the existing gaps in the literature and suggest several potential directions and techniques for future studies. Overall, this review provides insights into the control of AI in terms of the risk of introduction and spread of the virus.

Estimating within-flock transmission rate parameter for H5N2 highly pathogenic avian influenza virus in Minnesota turkey flocks during the 2015 epizootic

Better control of highly pathogenic avian influenza (HPAI) outbreaks requires deeper understanding of within-flock virus transmission dynamics. For such fatal diseases, daily mortality provides a proxy for disease incidence. We used the daily mortality data collected during the 2015 H5N2 HPAI outbreak in Minnesota turkey flocks to estimate the within-flock transmission rate parameter (β). The number of birds in Susceptible, Exposed, Infectious and Recovered compartments was inferred from the data and used in a generalised linear mixed model (GLMM) to estimate the parameters. Novel here was the correction of these data for normal mortality before use in the fitting process. We also used mortality threshold to determine HPAI-like mortality to improve the accuracy of estimates from the back-calculation approach. The estimated β was 3.2 (95% confidence interval (CI) 2.3-4.3) per day with a basic reproduction number of 12.8 (95% CI 9.2-17.2). Although flock-level estimates varied, the overall estimate was comparable to those from other studies. Sensitivity analyses demonstrated that the estimated β was highly sensitive to the bird-level latent period, emphasizing the need for its precise estimation. In all, for fatal poultry diseases, the back-calculation approach provides a computationally efficient means to obtain reasonable transmission parameter estimates from mortality data.

Controlling highly pathogenic avian influenza outbreaks: An epidemiological and economic model analysis

Outbreaks of highly pathogenic avian influenza (HPAI) can cause large losses for the poultry sector and for animal disease controlling authorities, as well as risks for animal and human welfare. In the current simulation approach epidemiological and economic models are combined to compare different strategies to control highly pathogenic avian influenza in Dutch poultry flocks. Evaluated control strategies are the minimum EU strategy (i.e., culling of infected flocks, transport regulations, tracing and screening of contact flocks, establishment of protection and surveillance zones), and additional control strategies comprising pre-emptive culling of all susceptible poultry flocks in an area around infected flocks (1 km, 3 km and 10 km) and emergency vaccination of all flocks except broilers around infected flocks (3 km). Simulation results indicate that the EU strategy is not sufficient to eradicate an epidemic in high density poultry areas. From an epidemiological point of view, this strategy is the least effective, while pre-emptive culling in 10 km radius is the most effective of the studied strategies. But these two strategies incur the highest costs due to long duration (EU strategy) and large-scale culling (pre-emptive culling in 10 km radius). Other analysed pre-emptive culling strategies (i.e., in 1 km and 3 km radius) are more effective than the analysed emergency vaccination strategy (in 3 km radius) in terms of duration and size of the epidemics, despite the assumed optimistic vaccination capacity of 20 farms per day. However, the total costs of these strategies differ only marginally. Extending the capacity for culling substantially reduces the duration, size and costs of the epidemic. This study demonstrates the strength of combining epidemiological and economic model analysis to gain insight in a range of consequences and thus to serve as a decision support tool in the control of HPAI epidemics.

Emergence of a highly pathogenic avian influenza virus from a low-pathogenic progenitor

Avian influenza (AI) viruses of the H7 subtype have the potential to evolve into highly pathogenic (HP) viruses that represent a major economic problem for the poultry industry and a threat to global health. However, the emergence of HPAI viruses from low-pathogenic (LPAI) progenitor viruses currently is poorly understood. To investigate the origin and evolution of one of the most important avian influenza epidemics described in Europe, we investigated the evolutionary and spatial dynamics of the entire genome of 109 H7N1 (46 LPAI and 63 HPAI) viruses collected during Italian H7N1 outbreaks between March 1999 and February 2001. Phylogenetic analysis revealed that the LPAI and HPAI epidemics shared a single ancestor, that the HPAI strains evolved from the LPAI viruses in the absence of reassortment, and that there was a parallel emergence of mutations among HPAI and later LPAI lineages. Notably, an ultradeep-sequencing analysis demonstrated that some of the amino acid changes characterizing the HPAI virus cluster were already present with low frequency within several individual viral populations from the beginning of the LPAI H7N1 epidemic. A Bayesian phylogeographic analysis revealed stronger spatial structure during the LPAI outbreak, reflecting the more rapid spread of the virus following the emergence of HPAI. The data generated in this study provide the most complete evolutionary and phylogeographic analysis of epidemiologically intertwined high- and low-pathogenicity viruses undertaken to date and highlight the importance of implementing prompt eradication measures against LPAI to prevent the appearance of viruses with fitness advantages and unpredictable pathogenic properties.

IMPORTANCE

The Italian H7 AI epidemic of 1999 to 2001 was one of the most important AI outbreaks described in Europe. H7 viruses have the ability to evolve into HP forms from LP precursors, although the mechanisms underlying this evolutionary transition are only poorly understood. We combined epidemiological information, whole-genome sequence data, and ultradeep sequencing approaches to provide the most complete characterization of the evolution of HPAI from LPAI viruses undertaken to date. Our analysis revealed that the LPAI viruses were the direct ancestors of the HPAI strains and identified low-frequency minority variants with HPAI mutations that were present in the LPAI samples. Spatial analysis provided key information for the design of effective control strategies for AI at both local and global scales. Overall, this work highlights the importance of implementing rapid eradication measures to prevent the emergence of novel influenza viruses with severe pathogenic properties.

Quantitative transmission characteristics of different H5 low pathogenic avian influenza viruses in Muscovy ducks

EU annual serosurveillance programs show that domestic duck flocks have the highest seroprevalence of H5 antibodies, demonstrating the circulation of notifiable avian influenza virus (AIV) according to OIE, likely low pathogenic (LP). Therefore, transmission characteristics of LPAIV within these flocks can help to understand virus circulation and possible risk of propagation. This study aimed at estimating transmission parameters of four H5 LPAIV (three field strains from French poultry and decoy ducks, and one clonal reverse-genetics strain derived from one of the former), using a SIR model to analyze data from experimental infections in SPF Muscovy ducks. The design was set up to accommodate rearing on wood shavings with a low density of 1.6 ducks/m(2): 10 inoculated ducks were housed together with 15 contact-exposed ducks. Infection was monitored by RNA detection on oropharyngeal and cloacal swabs using real-time RT-PCR with a cutoff corresponding to 2-7 EID50. Depending on the strain, the basic reproduction number (R0) varied from 5.5 to 42.7, confirming LPAIV could easily be transmitted to susceptible Muscovy ducks. The lowest R0 estimate was obtained for a H5N3 field strain, due to lower values of transmission rate and duration of infectious period, whereas reverse-genetics derived H5N1 strain had the highest R0. Frequency and intensity of clinical signs were also variable between strains, but apparently not associated with longer infectious periods. Further comparisons of quantitative transmission parameters may help to identify relevant viral genetic markers for early detection of potentially more virulent strains during surveillance of LPAIV.

Implications of within-farm transmission for network dynamics: consequences for the spread of avian influenza

•

Cross-scale dynamics were investigated for avian influenza in British poultry.

•

Transmission risk is dependent on the assumed within-flock transmission mode.

•

Transmission risk may not scale with transmissibility or flock size.

•

Transmission risk corresponds with between-farm impact for 28% of farms.

•

These results have implications for targeted disease control at the farm-level.

The importance of considering coupled interactions across multiple population scales has not previously been studied for highly pathogenic avian influenza (HPAI) in the British commercial poultry industry. By simulating the within-flock transmission of HPAI using a deterministic S-E-I-R model, and by incorporating an additional environmental class representing infectious faeces, we tracked the build-up of infectious faeces within a poultry house over time. A measure of the transmission risk (TR) was computed for each farm by linking the amount of infectious faeces present each day of an outbreak with data describing the daily on-farm visit schedules for a major British catching company. Larger flocks tended to have greater levels of these catching-team visits. However, where density-dependent contact was assumed, faster outbreak detection (according to an assumed mortality threshold) led to a decreased opportunity for catching-team visits to coincide with an outbreak. For this reason, maximum TR-levels were found for mid-range flock sizes (~25,000–35,000 birds). When assessing all factors simultaneously using multivariable linear regression on the simulated outputs, those related to the pattern of catching-team visits had the largest effect on TR, with the most important movement-related factor depending on the mode of transmission. Using social network analysis on a further database to inform a measure of between-farm connectivity, we identified a large fraction of farms (28%) that had both a high TR and a high potential impact at the between farm level. Our results have counter-intuitive implications for between-farm spread that could not be predicted based on flock size alone, and together with further knowledge of the relative importance of transmission risk and impact, could have implications for improved targeting of control measures.

Enhanced replication of swine influenza viruses in dexamethasone-treated juvenile and layer turkeys

Frequent transmission of swine influenza viruses (SIVs) to turkeys has been reported since 1980s. Experimental studies also showed that SIVs can infect turkeys with varying replication and transmission efficiency depending on the strain. However, host factors involved in infection/replication efficiency remain unclear. To investigate whether the immune status of turkeys might play a role in the susceptibility of turkeys to SIVs, we studied the replication efficiency of two recent SIVs (human-like H1N2 and triple reassortant (TR) H3N2) in dexamethasone-treated turkeys. The viruses were inoculated intranasally in both dexamethasone-treated and untreated control juvenile and layer turkeys. Amount of virus shedding was monitored at 2, 4, and 7 days post inoculation (DPI). Additionally, passage of both viruses was attempted in dexamethasone-treated 4-week-old turkeys. In both juvenile and layer turkeys, we were able to detect human-like H1N2 SIV only from dexamethasone-treated turkeys and no virus was detected in untreated birds. The virus shedding of the TR H3N2 SIV was also consistently higher (≈ 1 Log(10)EID(50)/ml) in dexamethasone-treated birds in both tracheal and cloacal swabs compared to untreated birds. Virus passage in dexamethasone-treated turkeys was successful up to the second passage and no virus was recovered from the third passage. These results show that potential immunosuppression due to dexamethasone treatment may enhance the transmission and adaptation of SIVs in turkeys through enhancement of virus replication, prolonged virus shedding, and possible decrease of infectious dose required to initiate infection.

Quantifying transmission of highly pathogenic and low pathogenicity H7N1 avian influenza in turkeys

Outbreaks of avian influenza in poultry can be devastating, yet many of the basic epidemiological parameters have not been accurately characterised. In 1999-2000 in Northern Italy, outbreaks of H7N1 low pathogenicity avian influenza virus (LPAI) were followed by the emergence of H7N1 highly pathogenic avian influenza virus (HPAI). This study investigates the transmission dynamics in turkeys of representative HPAI and LPAI H7N1 virus strains from this outbreak in an experimental setting, allowing direct comparison of the two strains. The fitted transmission rates for the two strains are similar: 2.04 (1.5-2.7) per day for HPAI, 2.01 (1.6-2.5) per day for LPAI. However, the mean infectious period is far shorter for HPAI (1.47 (1.3-1.7) days) than for LPAI (7.65 (7.0-8.3) days), due to the rapid death of infected turkeys. Hence the basic reproductive ratio, [Formula: see text] is significantly lower for HPAI (3.01 (2.2-4.0)) than for LPAI (15.3 (11.8-19.7)). The comparison of transmission rates and [Formula: see text] are critically important in relation to understanding how HPAI might emerge from LPAI. Two competing hypotheses for how transmission rates vary with population size are tested by fitting competing models to experiments with differing numbers of turkeys. A model with frequency-dependent transmission gives a significantly better fit to experimental data than density-dependent transmission. This has important implications for extrapolating experimental results from relatively small numbers of birds to the commercial poultry flock size, and for how control, including vaccination, might scale with flock size.

Transmission dynamics of low pathogenicity avian influenza infections in Turkey flocks

Low pathogenicity avian influenza (LPAI) viruses of H5 and H7 subtypes have the potential to mutate into highly pathogenic strains (HPAI), which can threaten human health and cause huge economic losses. The current knowledge on the mechanisms of mutation from LPAI to HPAI is insufficient for predicting which H5 or H7 strains will mutate into an HPAI strain, and since the molecular changes necessary for the change in virulence seemingly occur at random, the probability of mutation depends on the number of virus replicates, which is associated with the number of birds that acquire infection. We estimated the transmission dynamics of LPAI viruses in turkeys using serosurveillance data from past epidemics in Italy. We fitted the proportions of birds infected in 36 flocks into a hierarchical model to estimate the basic reproduction number (R₀) and possible variations in R₀ among flocks caused by differences among farms. We also estimated the distributions of the latent and infectious periods, using experimental infection data with outbreak strains. These were then combined with the R₀ to simulate LPAI outbreaks and characterise the resulting dynamics. The estimated mean within-flock R₀ in the population of infected flocks was 5.5, indicating that an infectious bird would infect an average of more than five susceptible birds. The results also indicate that the presence of seropositive birds does not necessarily mean that the virus has already been cleared and the flock is no longer infective, so that seropositive flocks may still constitute a risk of infection for other flocks. In light of these results, the enforcement of appropriate restrictions, the culling of seropositive flocks, or pre-emptive slaughtering may be useful. The model and parameter estimates presented in this paper provide the first complete picture of LPAI dynamics in turkey flocks and could be used for designing a suitable surveillance program.