A spatiotemporal model to assess the introduction risk of African horse sickness by import of animals and vectors in France.

BACKGROUND: African horse sickness (AHS) is a major, Culicoides-borne viral disease in equines whose introduction into Europe could have dramatic consequences. The disease is considered to be endemic in sub-Saharan Africa. Recent introductions of other Culicoides-borne viruses (bluetongue and Schmallenberg) into northern Europe have highlighted the risk that AHS may arrive in Europe as well. The aim of our study was to provide a spatiotemporal quantitative risk model of AHS introduction into France. The study focused on two pathways of introduction: the arrival of an infectious host (PW-host) and the arrival of an infectious Culicoides midge via the livestock trade (PW-vector). The risk of introduction was calculated by determining the probability of an infectious animal or vector entering the country and the probability of the virus then becoming established: i.e., the virus's arrival in France resulting in at least one local equine host being infected by one local vector. This risk was assessed using data from three consecutive years (2010 to 2012) for 22 regions in France. RESULTS: The results of the model indicate that the annual risk of AHS being introduced to France is very low but that major spatiotemporal differences exist. For both introduction pathways, risk is higher from July to October and peaks in July. In general, regions with warmer climates are more at risk, as are colder regions with larger equine populations; however, regional variation in animal importation patterns (number and species) also play a major role in determining risk. Despite the low probability that AHSV is present in the EU, intra-EU trade of equines contributes most to the risk of AHSV introduction to France because it involves a large number of horse movements. CONCLUSION: It is important to address spatiotemporal differences when assessing the risk of ASH introduction and thus also when implementing efficient surveillance efforts. The methods and results of this study may help develop surveillance techniques and other risk reduction measures that will prevent the introduction of AHS or minimize AHS' potential impact once introduced, both in France and the rest of Europe.

A quantitative risk assessment for bovine spongiform encephalopathy in Japan.

A predictive case-cohort model was applied to Japanese data to analyze the interaction between challenge and stability factors for bovine spongiform encephalopathy (BSE) for the period 1985-2020. BSE risk in cattle was estimated as the expected number of detectable cases per year. The model was comprised of a stochastic spreadsheet calculation model with the following inputs: (1) the origin and quantity of live cattle and meat and bone meal imported into Japan, (2) the age distribution of native cattle, and (3) the estimated annual basic reproduction ratio (R(0) ) for BSE. The estimated probability of having zero detectable cases in Japan in 2015 was 0.90 (95% CI 0.83-0.95). The corresponding value for 2020 was 0.99 (95% CI 0.98-0.99). The model predicted that detectable cases may occur in Japan beyond 2015 because of the assumption that continued transmission was permitted to occur (albeit at a very low level) after the 2001 ban on the importation and domestic use of all processed animal proteins for the production of animal feed and for fertilizer. These results reinforce the need for animal health authorities to monitor the efficacy of control measures so that the future course of the BSE epidemic in Japan can be predicted with greater certainty.

Derivation of a Floquet formalism within a natural framework.

Many biological systems experience a periodic environment. Floquet theory is a mathematical tool to deal with such time periodic systems. It is not often applied in biology, because linkage between the mathematics and the biology is not available. To create this linkage, we derive the Floquet theory for natural systems. We construct a framework, where the rotation of the Earth is causing the periodicity. Within this framework the angular momentum operator is introduced to describe the Earth's rotation. The Fourier operators and the Fourier states are defined to link the rotation to the biological system. Using these operators, the biological system can be transformed into a rotating frame in which the environment becomes static. In this rotating frame the Floquet solution can be derived. Two examples demonstrate how to apply this natural framework.

Calculating the time to extinction of a reactivating virus, in particular bovine herpes virus.

The expected time to extinction of a herpes virus is calculated from a rather simple population-dynamical model that incorporates transmission, reactivation and fade-out of the infectious agent. We also derive the second and higher moments of the distribution of the time to extinction. These quantities help to assess the possibilities to eradicate a reactivating infection. The key assumption underlying our calculations is that epidemic outbreaks are fast relative to the time scale of demographic turnover. Four parameters influence the expected time to extinction: the reproduction ratio, the reactivation rate, the population size, and the demographic turn-over in the host population. We find that the expected time till extinction is very long when the reactivation rate is high (reactivation is expected more than once in a life time). Furthermore, the infectious agent will go extinct much more quickly in small populations. This method is applied to bovine herpes virus (BHV) in a cattle herd. The results indicate that without vaccination, BHV will persist in large herds. The use of a good vaccine can induce eradication of the infection from a herd within a few decades. Additional measures are needed to eradicate the virus from a whole region within a similar time-span.

A meta-analysis quantifying transmission parameters of FMDV strain O Taiwan among non-vaccinated and vaccinated pigs.

Our aim was to provide additional estimates of main parameters for the transmission of foot-and-mouth disease virus (FMDV) strain O Taiwan (3/97). We used the data of previous experiments in non-vaccinated and vaccinated pigs and combined the data of experiments with the same treatment(s). First, we quantified the reproduction ratio R for the various groups using a final-size method. Our final-size results predicted that vaccination with a four-fold vaccine dose (but not with a single dose) at 1 week before inoculation (-7 dpi) would reduce R compared to the non-vaccinated group. Secondly, we used the daily results of virus excretion to quantify the transmission rate beta (by using generalized linear modelling), and the infectious period T (by using survival analysis). We used the estimates of beta and T to estimate R more precisely as compared to the final-size method and also for the groups for which a finite estimate could not be obtained using a final-size method. Our modelling results predicted that beta for non-vaccinated, for single-dose and four-fold-dose groups would be 6.1 (3.7, 10)day(-1), 2.0 (1.0, 4.0)day(-1) and 0.4 (0.1, 1.4)day(-1), T at 6.5 (5.7, 7.3), 5.3 (4.7, 6.0) and 2.3 (0.9, 5.7) days and R at 40 (21, 74), 11 (4.9, 24) and 1.0 (0.1, 7.8), respectively. These results predicted that both vaccination with a four-fold vaccine dose and with a single dose at -7 dpi would reduce beta, T and R significantly as compared to the non-vaccinated pigs, thereby showing that vaccination will reduce transmission of FMDV significantly already 1 week post vaccination.

Assessing the variability in transmission of bovine tuberculosis within Spanish cattle herds.

In Spain, despite years of efforts to eradicate bovine tuberculosis (bTB), the disease is still endemic, with some areas of high prevalence. In this context, the surveillance and control plans may need to be re-evaluated, and understanding the dynamics of bTB spread within Spanish herds may help to develop new strategies for reducing the time for detection of infected herds and for the elimination of bTB from the herds already infected. Here, we developed a compartmental stochastic model to simulate bTB within-herd transmission, fed it with epidemiological data from 22 herds (obtained from a previous work) and carried out parameter inference using Approximate Bayesian Computing methods We also estimated the "Within-herd transmission potential Number" (Rh), i.e. the average number of secondary cases generated by a single animal infected introduced into a totally susceptible herd, considering different scenarios depending on the frequency of controls. The median global values obtained for the transmission parameters were: for the transmission coefficient (ß), 0.014 newly infected animals per infectious individual per day (i.e. 5.2 per year), for the rate at which infected individuals become infectious (α), 0.01 per day (equivalent to a latent period of 97 days), and for the rate at which infected individuals become reactive to the skin test (α1), 0.08 per day (equivalent to a period of 12 days for an infected animal to become reactive). However, the results also evidenced a great variability in the estimates of those parameters (in particular ß and α) among the 22 herds. Considering a 6-month interval between tests, the mean Rh was 0.23, increasing to 0.82 with an interval of 1 year, and to 2.01 and 3.47 with testing intervals of 2 and 4 years, respectively.

Comparative Risk Analysis of Two Culicoides-Borne Diseases in Horses: Equine Encephalosis More Likely to Enter France than African Horse Sickness.

African horse sickness (AHS) and equine encephalosis (EE) are Culicoides-borne viral diseases that could have the potential to spread across Europe if introduced, thus being potential threats for the European equine industry. Both share similar epidemiology, transmission patterns and geographical distribution. Using stochastic spatiotemporal models of virus entry, we assessed and compared the probabilities of both viruses entering France via two pathways: importation of live-infected animals or importation of infected vectors. Analyses were performed for three consecutive years (2010-2012). Seasonal and regional differences in virus entry probabilities were the same for both diseases. However, the probability of EE entry was much higher than the probability of AHS entry. Interestingly, the most likely entry route differed between AHS and EE: AHS has a higher probability to enter through an infected vector and EE has a higher probability to enter through an infectious host. Consequently, different effective protective measures were identified by 'what-if' scenarios for the two diseases. The implementation of vector protection on all animals (equine and bovine) coming from low-risk regions before their importation was the most effective in reducing the probability of AHS entry. On the other hand, the most significant reduction in the probability of EE entry was obtained by the implementation of quarantine before import for horses coming from both EU and non-EU countries. The developed models can be useful to implement risk-based surveillance.

How to Assess Data Availability, Accessibility and Format for Risk Analysis?

Risk assessments are mostly carried out based on available data, which do not reflect all data theoretically required by experts to answer them. This study aimed at developing a methodology to assess data availability, accessibility and format, based on a scoring system and focusing on two diseases: Venezuelan equine encephalomyelitis (VEE), still exotic to Europe, and alveolar echinococcosis, caused by Echinococcus multilocularis (EM), endemic in several Member States (MSs). After reviewing 36 opinions of the EFSA-AHAW Panel on risk assessment of animal health questions, a generic list of needed data was elaborated. The methodology consisted, first, in implementing a direct and an indirect survey to collect the data needed for both case studies: the direct survey consisted in a questionnaire sent to contact points of three European MSs (Belgium, France and the Netherlands), and the organization of a workshop gathering experts on both diseases. The indirect survey, focusing on the three MSs involved in the direct survey plus Spain, relied on web searches. Secondly, a scoring system with reference to data availability, accessibility and format was elaborated, to, finally, compare both diseases and data between MSs. The accessibility of data was generally related to their availability. Web searches resulted in more data available for VEE compared to EM, despite its current exotic status in the European Union. Hypertext markup language and portable document files were the main formats of available data. Data availability, accessibility and format should be improved for research scientists/assessors. The format of data plays a key role in the feasibility and rapidness of data management and analysis, through a prompt compilation, combination and aggregation in working databases. Harmonization of data collection process is encouraged, according to standardized procedures, to provide useful and reliable data, both at the national and the international levels for both animal and human health; it would allow assessing data gaps through comparative studies. The present methodology is a good way of assessing the relevance of data for risk assessment, as it allows integrating the uncertainty linked to the quality of data used. Such an approach could be described as transparent and traceable and should be performed systematically.

A decision-tree to optimise control measures during the early stage of a foot-and-mouth disease epidemic.

A decision-tree was developed to support decision making on control measures during the first days after the declaration of an outbreak of foot-and-mouth disease (FMD). The objective of the tree was to minimise direct costs and export losses of FMD epidemics under several scenarios based on livestock and herd density in the outbreak region, the possibility of airborne spread, and the time between first infection and first detection. The starting point of the tree was an epidemiological model based on a deterministic susceptible-infectious-recovered approach. The effect of four control strategies on FMD dynamics was modelled. In addition to the standard control strategy of stamping out and culling of high-risk contact herds, strategies involving ring culling within 1 km of an infected herd, ring-vaccination within 1 km of an infected herd, and ring-vaccination within 3 km of an infected herd were assessed. An economic model converted outbreak and control effects of farming and processing operations into estimates of direct costs and export losses. Ring-vaccination is the economically optimal control strategy for densely populated livestock areas whereas ring culling is the economically optimal control strategy for sparsely populated livestock areas.

The foot-and-mouth disease epidemic in The Netherlands in 2001.

An outbreak of foot-and-mouth disease (FMD) in Great Britain was reported on 21 February 2001, followed by an outbreak of FMD in The Netherlands a month later. This Dutch index outbreak occurred on a mixed, veal-calf/dairy-goat farm in Oene, in the central part of The Netherlands. The most-likely route of infection was the import of Irish veal-calves to this Dutch herd via an FMD-contaminated staging point in France. With hindsight, more herds seemed to be infected by the time the index outbreak was confirmed. The regular EU control measures were implemented, in combination with pre-emptive culling of herds within 1km of each outbreak. Nevertheless, more outbreaks of FMD occurred. Most of the virus infections on those farms were "neighborhood infections". Because the situation seemed out of control locally and the destruction capacity became insufficient, it was decided to implement an emergency vaccination strategy for all biungulates in a large area around Oene to stop further spread of the virus. All susceptible animals on approximately 1800 farms in this area were vaccinated. All farms subsequently were depopulated, starting from 2 weeks after vaccination. In total, 26 outbreaks were detected (the last outbreak on 22 April 2001). In total, approximately 260,000 animals were killed.

Studies on the efficacy of hyperbaric rendering procedures in inactivating bovine spongiform encephalopathy (BSE) and scrapie agents.

The efficacy of the procedures in use at the two rendering plants in the Netherlands was assessed on a laboratory-scale using procedures that simulated the pressure cooking part of the rendering process. A pool of bovine spongiform encephalopathy (BSE)-infected brainstem from the United Kingdom and a pool of scrapie-infected brainstem from Dutch sheep were used to spike the rendering materials. The mixtures were subjected to various time-temperature combinations of hyperbaric heat treatment related to the conditions used in Dutch rendering plants in the early 1990s, and to the combination of 20 minutes at 133 degrees C required by the EU Directive on rendering of 1996. The efficacy of the procedures in inactivating BSE or scrapie infectivity was measured by titrating the materials before and after heat treatment in inbred mice, by combined intracerebral and intraperitoneal inoculations at limiting dilutions. Two independent series of experiments were carried out. The design of the study allowed for minimum inactivations of up to 2.2 log (2.0 in the second series) to be measured in the diluted infective material and 3.1 log in the undiluted material. After 20 minutes at 133 degrees C there was a reduction of BSE infectivity of about 2.2 log in the first series (with some residual infectivity detected), and in the second series more than 2.0 log (with no residual infectivity detected). With undiluted brain material there was an inactivation of about 3.0 log (with some residual infectivity detected). With the same procedure, scrapie infectivity was reduced by more than 1.7 log in the first series and by more than 2.2 log in the second series. With undiluted brain material there was an inactivation of more than 3.1 log. In each case no residual scrapie infectivity was detected. The BSE agent consistently appeared to be more resistant to heat inactivation procedures than the scrapie agent, particularly at lower temperatures and shorter times.

Risk based surveillance for early detection of low pathogenic avian influenza outbreaks in layer chickens.

Current knowledge does not allow the prediction of when low pathogenic avian influenza virus (LPAIV) of the H5 and H7 subtypes infecting poultry will mutate to their highly pathogenic phenotype (HPAIV). This mutation may already take place in the first infected flock; hence early detection of LPAIV outbreaks will reduce the likelihood of pathogenicity mutations and large epidemics. The objective of this study was the development of a model for the design and evaluation of serological-surveillance programmes, with a particular focus on early detection of LPAIV infections in layer chicken flocks. Early detection is defined as the detection of an infected flock before it infects on average more than one other flock (between-flock reproduction ratio Rf<1), hence a LPAI introduction will be detected when only one or a few other flocks are infected. We used a mathematical model that investigates the required sample size and sampling frequency for early detection by taking into account the LPAIV within- and between-flock infection dynamics as well as the diagnostic performance of the serological test used. Since layer flocks are the target of the surveillance, we also explored whether the use of eggs, is a good alternative to sera, as sample commodity. The model was used to refine the current Dutch serological-surveillance programme. LPAIV transmission-risk maps were constructed and used to target a risk-based surveillance strategy. In conclusion, we present a model that can be used to explore different sampling strategies, which combined with a cost-benefit analysis would enhance surveillance programmes for low pathogenic avian influenza.

Financial consequences of the Dutch bluetongue serotype 8 epidemics of 2006 and 2007.

This study calculates the financial consequences of the bluetongue serotype 8 (BTV8) epidemics of 2006 and 2007 in the Netherlands. We constructed a deterministic economic model that is compatible with the Dutch livestock production systems for cattle, sheep and goats. Two hundred cattle farms and 270 sheep farms were infected with BTV8 in the epidemic of 2006, whereas 30,417 cattle farms, 45,022 sheep farms and 35,278 goat farms were estimated to be infected in the epidemic of 2007. The net costs (costs minus benefits) of the BTV8 epidemic of 2006 (BT2006) was estimated at 32.4 million Euros. The net costs of the BTV8 epidemic of 2007 (BT2007) was valued at 164-175 million Euros, depending on the mortality and morbidity rates for cattle used. The losses account for 2%, 10% and 11% of the gross value of the primary production within Dutch pasture-based livestock farming that equals 1.6 billion Euros. Control measures accounted for 91% of the net costs of the BT2006, while diagnostic costs represented 7%. By contrast, for the BT2007 92% of the net costs were in the form of production losses and veterinary treatment fees, while only 6% were related to control measures. Furthermore, the control costs dropped from 29,630 in BT2006 to 10,990 in BT2007 mainly due to the costly indoor housing that was not obligatory during the BT2007 epidemic. The cattle sector suffered 88% and 85% of the net costs for the BT2006 and BT2007, respectively; the highest of all sectors.

Questionnaire survey about the motives of commercial livestock farmers and hobby holders to vaccinate their animals against Bluetongue virus serotype 8 in 2008-2009 in the Netherlands.

After a massive epidemic of Bluetongue virus serotype 8 (BTV-8) among ruminants in 2006-2007 in the European Union (EU), the Netherlands started a voluntary emergency vaccination campaign in May 2008, subsidized by the EU. At the start of a new campaign in 2009, without subsidized vaccination, we investigated by mail survey the motives of farmers and hobby holders to vaccinate against BTV-8 in 2008 and 2009. Mean vaccine uptake in 2008 was: 73% in sheep, 71% in cattle, 43% in goat farms and 67% in hobby holdings. Top-5 motives pro-vaccination were: prevention of production loss; subsidized vaccination; recommendation by practitioner; welfare reasons; contribution to the eradication campaign. Top-5 motives against vaccination were: vaccination costs; absence of clinical BT-problems; presumed low infection risk; balance between vaccination costs and loss without vaccination; bad experience with earlier vaccination campaigns. Willingness to vaccinate was significantly lower in 2009: 42% in sheep, 58% in cattle, 19% in goat farms and 49% in hobby holdings. Measures to stimulate vaccination among those that did not want to vaccinate in 2009 were: subsidized vaccination; possibility to vaccinate their own animals; more information on efficacy/safety of vaccine and why animals had to be vaccinated again; availability of a BT vaccine combined with vaccine(s) against other diseases.

Mapping the basic reproduction number (R0) for vector-borne diseases: a case study on bluetongue virus.

Geographical maps indicating the value of the basic reproduction number, R0, can be used to identify areas of higher risk for an outbreak after an introduction. We develop a methodology to create R0 maps for vector-borne diseases, using bluetongue virus as a case study. This method provides a tool for gauging the extent of environmental effects on disease emergence. The method involves integrating vector-abundance data with statistical approaches to predict abundance from satellite imagery and with the biologically mechanistic modelling that underlies R0. We illustrate the method with three applications for bluetongue virus in the Netherlands: 1) a simple R0 map for the situation in September 2006, 2) species-specific R0 maps based on satellite-data derived predictions, and 3) monthly R0 maps throughout the year. These applications ought to be considered as a proof-of-principle and illustrations of the methods described, rather than as ready-to-use risk maps. Altogether, this is a first step towards an integrative method to predict risk of establishment of diseases based on mathematical modelling combined with a geographic information system that may comprise climatic variables, landscape features, land use, and other relevant factors determining the risk of establishment for bluetongue as well as of other emerging vector-borne diseases.

Modelling the spread of phocine distemper virus among harbour seals.

Data presented in earlier publications on the 1988 epizootic among seals in North West Europe show a pattern that is somewhat inconsistent with the predictions of the standard mathematical model of epidemics. We argue that for animals living in herds or colonies, such as seals, the mutual contact behaviour is such that models for the transmission of infectious diseases should be applied with special care for the distinction between numbers and densities. This is demonstrated by using a mechanistic description of the contacts among seals, which leads to a slightly different formulation of the model. Results of the analysis of this formulation are more in line with the data. The model introduced here can be applied to epidemics among all kinds of animals living in herds and in fact to any species with constant local density, independent of the total population size (i.e., occupying a variable area). Application of the traditional formulation, using different parameters for herds of different sizes, will give equally good results for non-lethal diseases. However, especially for diseases with a low R0 and high death rates, such as the phocine distemper virus (PDV) disease, the two model formulations give quite different results. Further analysis of the model is performed to determine the most important factors influencing such an epidemic. The survival of infected animals turns out to have a disproportionately great influence on the intensity of the epidemic. Therefore in the case of the PDV epizootic we conclude that marine pollution may not only have contributed to the high death rates, but, if so, it has intensified the epizootic as well.