Quantification of the risk for introduction of virulent Newcastle disease virus into Spain through legal trade of live poultry from European Union countries

Abstract

Newcastle disease (ND) causes large economic losses in poultry production worldwide. Spain has reported two ND epidemics in poultry farms since 1993, the most recent in 2009. The recent increase in the number of ND epidemics reported in Spain and in other European Union (EU) member countries along with the failure to identify the source of the Spanish epidemics caused concern over the vulnerability that Spain has to the disease. Some of the epidemics recently reported in EU member states were associated with legal introduction of live poultry; the large number of susceptible species annually imported by Spain from the EU suggests that legal imports of poultry may impose a risk for the introduction of virulent ND virus (v-NDV) into the country. This article presents the results of the first quantitative assessment of the risk for v-NDV introduction into an ND-free country via legal trade of live poultry. The geographical variation of the risk and the relative contribution of exporting countries and susceptible poultry species to the risk were also estimated. The model here estimated that if prevailing conditions persist, then it would be expected that ND epidemics caused by legal trade of live poultry will occur, on average, once every 196 years in Spain. These results suggest that the risk for ND epidemics in Spain, and probably the sources of recent epidemics reported in the country, were associated with routes of entry other than legal trade of poultry.

Introduction

Newcastle disease (ND) is regarded throughout the world as one of the most important diseases of poultry due to the economic impact and social disruption that it produces on affected regions (Wakamatsu et al., 2006). ND is caused by the infection with virulent (v) strains of the serotype 1 avian paramyxovirus, which has been classified as one of the members of the Avulavirus genus and is most commonly referred to as ND virus (NDV) (Mayo, 2002a, b; OIE, 2009a).

ND is a highly infectious disease, characterized by mortality rates that can affect as much as 100% of infected flocks (Aldous et al., 2003; Wakamatsu et al., 2006; Bogoyavlenskiy et al., 2009). The disease is endemic in many countries of Central and South America, the Middle East, Africa, and Asia (USDA, 1992; Alexander, 2001; Wakamatsu et al., 2006). The broad distribution of ND imposes a high risk for epidemics in ND-free countries and regions (Herczeg et al., 1999). The disease is considered a serious threat for the poultry industry of the European Union (EU) and 230 outbreaks have been reported in 13 of the 27 EU member states over the past 5 years (2005 to 2009) (OIE, 2010a). Some of the numerous epidemics that affected the EU over the past two decades have been particularly devastating. For example, the ND epidemics that affected the UK in 1997 and Italy in 2000 resulted in >1.8 million and >2.5 million poultry destroyed, respectively (Alexander et al., 1998; Capua et al., 2002; OIE, 2010b). Other regions have also been recently devastated by ND epidemics. For example, the epidemic that affected the US in 2002 and 2003 resulted in estimated direct and indirect losses of more than US$280 million, including the depopulation of 3.21 million birds (Hietala et al., 2004).

The three ND epidemics that have been reported in the modern history of Spain took place over the past 25 years. ND was first reported in Spain in 1986, when the variant NDV termed pigeon paramyxovirus type 1 (PPMV-1 ) that affected racing pigeons on a panzootic scale was identified in the country (Alexander, 1995a). In 1993, an ND outbreak of unknown origin that affected a poultry flock in the province of Zaragoza resulted in the destruction of 125,000 broilers (OIE, 2010c). The most recent ND epidemic in Spain took place in November 2009, when a game farm located in the province of Guipuzcoa was reported to have been NDV-infected (AECA-WPSA, 2009; OIE, 2010a). By the time of depopulation, 17% of the 11,000 pheasants on the farm infected in 2009 were reported to have died as a consequence of infection with the NDV (AECA-WPSA, 2009; OIE, 2010a). Although the epidemiological investigation of the outbreak was still ongoing, the source of the 2009 outbreak remained unknown at the time of writing.

There is no active sampling for detection of ND incursions into Spain; instead, the ability of national and regional veterinary services to detect ND cases depends on the investigation of suspects and on the routine inspection of flocks and shipments. The limited extent of the three ND epidemics reported in the country suggests that, at least in the past, official intervention was effective in detecting and controlling virus spread beyond the index farm. For that reason, the consequences of ND epidemics have been relatively mild for Spain, compared with the experiences of other countries. However, an eventual ND epidemic of larger size than those previously reported in the country would have had devastating consequences for Spain. The country is considered the third and fourth largest producer of poultry meat and eggs, respectively, of the EU, which results in an estimated income of more than €2.73 billion to the country's economy (MARM, 2008a, b). Despite the apparent development of the Spanish poultry industry, the production system is still heavily dependent on the importation of poultry, which resulted, for example, in the introduction of >32 million live poultry, including chickens, turkeys, ducks, geese, and guinea fowl, from other EU member states in 2007 (Spanish Foreign Trade Statistics Database, 2008).

The apparent increase in the number of ND epidemics reported over the past 25 years in Spain and the failure to identify the source of the introductions raised concerns regarding the vulnerability that Spain has to the disease. Several routes of transmission, including migration or trade of wild birds, legal or illegal trade of live poultry and their products, such as meat and eggs, and mechanical transmission associated with movement of people and contaminated objects, may be associated with NDV introduction into free regions (Wobeser et al., 1993; Alexander, 1995b; Alexander et al., 1998; Seal et al., 1998; Capua et al., 2002; Aldous et al., 2007; Bogoyavlenskiy et al., 2009). However, it is commonly believed that the highest risk for NDV spread is associated with migration of wild birds and trade of live poultry (Wobeser et al., 1993; Capua et al., 1994; Alexander et al., 1998; Jørgensen et al., 1999; Aldous et al., 2007; OIE, 2007; Aldous & Alexander, 2008; Bogoyavlenskiy et al., 2009). For those two routes of introduction, nothing can be done to modify the patterns of bird migration and the only effective strategy to decrease the risk associated with migrations of wild birds is to increase biosecurity the farms. On the other hand, knowledge on the risk associated with legal trade of live poultry, and on the factors associated with that risk, may lead to the implementation of alternative strategies to mitigate the risk of introduction of NDV into a country. Furthermore, the combination of a free animal trade policy among EU member states, relative short distances, long latency periods for ND in certain poultry species, and the large number of susceptible species imported annually into Spain, increased the concerns regarding the risk for v-NDV introduction into Spain via the legal trade of infected poultry during the silent phase of an epidemic.

Quantitative risk assessment (QRA) was used here to quantify and to identify factors associated with the risk of v-NDV introduction into Spain through the legal importation of live poultry. Quantitative knowledge of the risk and of the factors associated with the risk for v-NDV introduction is prerequisite for a country to allocate financial and human resources to the development of policy for prevention and eventual control of ND epidemics. If the risk estimated here was relatively high, then Spain may consider the promotion of preventive measures in order to minimize the risk for the disease. In turn, if the risk estimated here was relatively low, then other measures, such as active surveillance of wild birds and increase of bio-security measures in poultry farms to decrease the risk for effective contact with wild birds, should be promoted.

The objective of the present study was to estimate whether the legal trade of live poultry imposes a high risk for the introduction of v-NDV into Spain. The geographical variation of the risk and the relative contribution of exporting countries and susceptible poultry species to the risk were also estimated. These results will help to estimate whether changes in the structure of legal trade of Spain, including the number of poultry annually imported, may have any impact on the risk for v-NDV introduction into the country. Moreover, to the best of our knowledge, no study aimed at quantifying the risk for ND introduction into EU member states such as Spain via legal trade of live poultry has been published in the peer-reviewed literature.

Materials and Methods

A QRA framework developed to assess the risk for highly pathogenic avian influenza entering Spain was modified to quantify the risk for ND introduction into the country. The specifics for the original model formulation have been described elsewhere (Sánchez-Vizcaíno et al., 2010). Briefly, the model here quantifies the annual probability of ND introduction into Spain via the legal trade of live poultry (P _ND). The assessment was developed assuming that risk is only associated with the silent phase of an ND epidemic in the country of origin of the shipment, which was defined as the period between first infection and first detection of the disease, because after official declaration of an ND epidemic all exports from the infected country will be banned. For each species s, the duration of the silent phase of an ND epidemic (b _s ) was assumed to follow a Pert distribution with minimum, most likely, and maximum values obtained from ND epidemics in the EU reported to the OIE from 1992 through 2008 (Table 1). Risk was assessed separately for each ND-susceptible species imported by Spain, such as chickens (Gallus gallus), turkeys (Meleagris gallopavo), ducks (Anas platyrhynchos), geese (Anser anser), and guinea fowl (Numida meleagris), for each EU member state that exports poultry to Spain, and for each Spanish province. The model was developed considering data collected from 1992 to 2008, and only premises with >100 poultry, which are the only ones from which Spain imports animals (Spanish Foreign Trade Statistics Database, 2008; TRACES, 2008), were considered at risk for exporting infected animals into Spain. The assessment was conducted up to the entry of at least one infected bird into an agricultural operation; spread from this initial point of entry was not considered.

Table 1.  Description and source of information of variables used to parameterize a model of the risk for introduction of Newcastle disease (ND) into Spain via legal trade of live poultry.

Notation	Variable description	Parameterization and value	Source of information
f cds	Number of species s flocks in country c from which poultry were exported into a Spanish province d during the silent phase of an ND epidemic	(j cds xe cds /365)×b s	Sánchez-Vizcaíno et al. (2010)
j cds	Number of shipments of poultry of specie s sent from a country c into Spanish province d per year	Normal (µj cds σj cds )	Spanish Foreign Trade Statistics Database (2008)
e cds	Number of flocks per shipment	NA	TRACES (2008)
b s	Duration of the silent phase of an ND epidemic in the species s	Pert (3, 14, 30) for chickens and guinea fowl, Pert (6, 17, 34) for turkeys, Pert (6, 25, 40) for ducks and geese	OIE (2010a,b,c)
P Fcds	Probability of exporting fowls from an ND-infected flock of species s from country c into a Spanish province d during the silent phase of an ND epidemic	P(A 1 )xP(A 2 )	Sánchez-Vizcaíno et al. (2010)
P(A 1 )	Probability that species s from country c become ND-infected in the lapse of one year	Gamma (α cs , β)	Martínez-López et al. (2008)
α cs	Probability of having one ND epidemic affecting species s in flocks of country c with more than 100 poultry through period of time β		Sánchez-Vizcaíno et al. (2010)
E csi	Number of ND epidemics i that affected species s in flocks of country c with more than 100 poultry and that were reported to the OIE from 1992 to 2008	NA	OIE (2010a,b,c)
S csi	Probability that species s was ND-infected during the silent phase of the epidemic i in that affected country c	NA	OIE (2010a,b,c)
t	Period of time for which information regarding the occurrence of ND epidemics was available	NA	OIE (2010a,b,c)
β	Period of time considered in the assessment	NA	NA
P(A 2 )	Probability of selecting an infected species s flock for exporting poultry into a Spanish province d during the silent phase of the ND epidemic	Beta (α 1f , α 2f ) α1f =IF s +1 α2f =TF cs −IF s +1	Vose (1997); OIE (2004b); Martínez-López et al. (2008); Sánchez-Vizcaíno et al. (2010)
IF s	Number of species s flocks with more than 100 poultry in the country c infected before the detection of the epidemic	Pert (1, 1, 28) for chickens and guinea fowl, Pert (1, 1, 18) for turkeys, Pert (1, 2, 19) for ducks and geese	OIE (2010a,b,c)
TF cs	Total number of species s flocks with more than 100 poultry in country c	Normal (µ TFcs , σ TFcs )	BMELV (2008); Agreste (2009); DEFRA (2009); FAO (2009a); KSH (2009); INS (2009); CBS (2009)
n cds	Number of species s poultry exported per flock	Normal (µn cds , σn cds )	Spanish Foreign Trade Statistics Database (2008)
	Species-specific and country-specific probability that an infected bird is introduced into an agricultural operation of Spanish province d	P(B 1 )xP(B 2 )xP(B 3 )	Martínez-López et al. (2008)
P(B 1 )	Probability that an ND infected bird of species s was exported from an infected flock	Beta (α 1w , α 2w ) α1w=IA cf's +1 α2w=Z cs −IA cf's +1	Vose (1997); OIE (2004b); Martínez-López et al. (2008);
IA cf's	Number of infected poultry of species s in a infected flock of country c	Z cs xPF	Sánchez-Vizcaíno et al. (2010)
Z cs	Average size of a species s flocks with more than 100 poultry in country c	NA cs /TF cs	Sánchez-Vizcaíno et al. (2010)
NA cs	Number of species s poultry in the flocks with more than 100 poultry of country c	Normal (µ NAcs , σ NAcs )	BMELV (2008); Agreste (2009); DEFRA (2009); FAO (2009b); KSH (2009); INS (2009); CBS (2009)
PF	Expected intra-flock ND prevalence	Pert (0.1, 0.7, 1)	Gohm et al. (1999)
P(B 2 )	Probability that an ND-infected fowl of the species s survive to the infection	Pert (0, 0.33, 0.7) for chickens and guinea fowls, Pert (0, 0.65, 0.88) for turkeys, Pert (0, 0.65, 1) for ducks and geese	Onapa et al. (2006); Lee et al. (2009); OIE (2010a,b,c)
P(B 3 )	Probability that an imported bird reaches an agricultural operation in the Spanish province d	[1−P(D s )]x[1−P(S)]
P(D s )	Probability that a bird of species s dies during the shipment	Pert (minimum, most likely, maximum)	Warriss et al. (1992); Vecerek et al. (2006); Voslárová et al. (2006)
P(S)	Probability that an imported bird is shipped into a slaughterhouse	Pert (minimum, most likely, maximum)	TRACES (2008)
NA, not applicable.

The value of P _ND was estimated as the sum of the probabilities of introduction [P(I _cds )] of at least one ND-infected bird into each Spanish province d, from each exporting country c, and via the legal importation of each susceptible species s. The value of P(I _cds ) was modelled as a multi-level binomial process of the form (OIE, 2004a; Sánchez-Vizcaíno et al., 2010):

where f _cds is the number of species s flocks in country c from which animals were exported into a Spanish province d during the silent phase of an ND epidemic, is the probability of exporting fowls from an ND-infected flock of species s from country c into a Spanish province d, n _cds is the number of species s poultry exported per flock, and is the species-specific and country-specific probability that an infected fowl is introduced into an agricultural operation of province d.

The value of was estimated as the conditional probability of two independent events [P(A _i ), i=1, 2]:

where P(A ₁ ) is the probability that species s from country c become ND infected in the lapse of 1 year, and P(A ₂ ) is the probability of selecting an infected species s flock for exporting fowls into a Spanish province d during the silent phase of the epidemic (Sánchez-Vizcaíno et al., 2010).

The value of was estimated as the product of three independent events [P(B _i ), i=1 – 3]:

where P(B ₁ ) is the probability that an ND-infected bird was exported from an infected flock, P(B ₂ ) is the probability that the infected-and-exported birds survive the infection and, P(B ₃ ) is the probability that the bird reaches an agricultural operation in Spanish province d. A sensitivity analysis was conducted to quantify whether the model was affected by the uncertainty and variation associated with the true value of the parameters. A full description of the procedures used to compute P(A ₁ ), P(A ₂ ), P(B ₁ ), P(B ₂ ), and P(B ₃ ) and to conduct the sensitivity analysis is available elsewhere (Sánchez-Vizcaíno et al., 2010). Briefly, the sensitivity analysis was conducted following a two-stage process. A multivariate regression model was used to identify the parameters of the risk assessment most influential on the value of P _ND, as suggested by values of the regression coefficient (β)≥0.1. The risk assessment model was subsequently run using alternative parameterizations (±30% of the base value) for those variables with β≥0.1, and absence of substantial changes in the estimated value of P _ND was interpreted as an indication of robustness of the results to the model parameterization. A list of distributions and their sources used to parameterize the model here is provided in Table 1.

Furthermore, an ND epidemic was simulated in each EU member state that exports poultry into Spain to quantify the value of P _ND given that an ND epidemic was declared in a commercial poultry farm of the exporting country.

The model was formulated and run in an Excel spreadsheet (Microsoft Office Professional Edition, 2007) using @ Risk version 4.5.5 (Palisade Corporation, 1996 to 2007). Results were mapped using ArcMap 9.1 (ESRI^©, 2005).

Results

The mean estimated value of P _ND was 5.11x10⁻³ (P _ND 95% probability interval (PI) =1.35x10⁻⁴ to 2.99x10⁻²), which suggests that if the conditions and parameters assumed here persist, then it would be expected that ND epidemics caused by the legal introduction of live poultry will occur, on average, once every 196 years in Spain. Most of the P _ND was concentrated for imports from France and the Netherlands (64.8% and 23.7%, respectively), whereas all other countries accounted for <5% of the P _ND. Much of the risk imposed by France and the Netherlands was associated with imports of ducks (49.8% of the French-specific P _ND) and chickens (79.3% of the Dutch-specific P _ND), respectively.

The probability of v-NDV introduction through legal import of ducks (mean=1.94x10⁻³) was, respectively, 1.16-fold, 1.41-fold, 17.8-fold, and 244.3-fold higher than the probability of introduction through chickens (1.67x10⁻³), turkeys (1.38x10⁻³), guinea fowl (1.09x10⁻⁴), and geese (7.94x10⁻⁶). France was the most probable origin of epidemics caused by the introduction of ducks (85% of the species-specific P _ND), turkeys (92.7% of the species-specific P _ND), and geese (100% of the species-specific P _ND), whereas The Netherlands and Italy were the most probable origin of epidemics associated with the introduction of, respectively, chickens (57% of the species-specific P _ND) and guinea fowl (55% of the species-specific P _ND).

The highest probabilities of v-NDV introduction into Spain were estimated for the provinces of Tarragona (mean=7.3x10⁻⁴), Gerona (6.5x10⁻⁴), Barcelona (4.4x10⁻⁴), and Lérida (2.9x10⁻⁴). These four provinces constitute the region of Catalonia, in north-eastern Spain, which concentrates 41.2% of the P _ND and that, in general, was the region at highest risk for introduction through every single species analysed here (Figure 1).

Figure 1.  Geographical variation of the total risk for introduction of Newcastle disease (ND) into Spain via legal trade of live poultry (1a) estimated using a quantitative risk assessment and discriminated by species of introduction: (1b) chickens, (1c) ducks, (1d) turkeys, (1e) guinea fowl, and (1f) geese. Risk has been categorized using quartiles as nil (~0), low (5x10⁻⁷ to 2.04x10⁻⁵), medium (2.05x10⁻⁵ to 4.46x10⁻⁵), high (4.47x10⁻⁵ to 2.13x10⁻⁴), and very high (2.14x10⁻⁴ to 7.3x10⁻⁴). Provinces with nil risk and those included within the first, second, third, and fourth quartiles are shown in white, yellow, orange, red, and brown, respectively. Provinces of Catalonia, which is the region that concentrated most of the risk, are indicated with a white outline (1a).

Results of the sensitivity analysis suggest that the P _ND was affected (β≥0.1) only by four parameters; namely, the probability of an ND epidemic in ducks, turkeys, or chickens in France, or in chickens in the Netherlands. However, 30% increments in the value of those four parameters did not result in a substantial variation of the value of P _ND, with ND epidemics being expected once every 178 to 192 years. Therefore, the model was considered robust to changes in its parameterization.

Simulation of an ND epidemic in France and the Netherlands resulted in an increase in the value of P _ND of 8.2-fold and 1.85-fold, respectively. Thus, if an ND epidemic was reported in the poultry industry of France every year, it would be expected that, on average, such epidemics would affect Spain once every 24 years (or four times every century).

Discussion

To the best of the authors’ knowledge, this is the first quantitative assessment of the risk for v-NDV introduction into an ND-free country via the legal trade of live poultry, discriminated by susceptible species imported, country of origin of the export, and administrative unit of destination. Results suggest that legal trade of poultry does not impose a substantial risk for ND introduction into Spain (P _ND=5.11x10⁻³). The low number of flocks expected to be infected before the detection of an ND epidemic in EU countries (Table 1) and the high mortality rate expected for the disease are probably the most influential factors on these results. Based on previous experiences with ND (OIE, 2010a, OIE, 2010b, OIE, 2010c), one would expect that only a few flocks would be infected during the silent phase of an epidemic in the EU, as a consequence of adequate biosecurity conditions, high sensitivity of the surveillance systems, and good diagnostic capacity. Therefore, the combination of few flocks that could potentially export ND-infected poultry into Spain before the diagnosis of the epidemic and the high mortality levels expected in susceptible domestic species resulted in the very low probability of introducing v-ND virus into Spain through legal trade of poultry estimated here.

The sources of the last two ND epidemics reported in poultry flocks of Spain remain unknown. Although trade of live poultry has been proposed as a likely source for some of the ND incidents reported in the EU (Capua et al., 1994; Aldous et al., 2007; OIE, 2007; Aldous & Alexander, 2008), the results presented here suggest that the source of the recent ND epidemics reported in Spain were probably unrelated to legal trade of live poultry. Moreover, simulation of an ND epidemic in countries that export poultry into Spain did not substantially increase the risk for ND epidemics. The low risk associated with legal imports of poultry suggests that wild bird migration, which is considered another important route for ND spread into naïve regions (Wobeser et al., 1993; Alexander et al., 1998; Jørgensen et al., 1999; Bogoyavlenskiy et al., 2009), or illegal trade of animals and products, may be the routes that impose the highest risk for introduction of v-ND into Spain. Consequently, one may suggest from the decision-making standpoint that reinforcement of biosecurity measures at the flock level would have a higher impact on the risk for ND epidemics in Spain than changes in the number of poultry imported.

Results presented here are certainly influenced by the parameterization of the model, and one may argue that the true value of at least some of the variables is highly variable or uncertain. However, it is of note that changes in the values of the parameters have not influenced the model outcomes, as indicated by the results of the sensitivity analysis, which suggests that the results presented here were not sensitive to the model parameterization. In any case, results presented here are valid only for the conditions that have been modelled, including the prevailing epidemiological conditions in Spain and the assumptions and values used to formulate and parameterize the model. It is of note that the model was parameterized using values of v-ND virulence typically observed in the EU (Table 1). However, unusually mild strains of the virus may result in a higher risk for ND introduction into Spain through legal trade of live poultry compared with the value estimated here. Since importation of ND-vaccinated animals was prohibited in the EU in 1990 (The Council of the European Communities, 1990), it has been suggested that at least two epidemics were associated with legal trade of live poultry. An ND epidemic occurred in the UK in July 2005 following the importation of ND-infected pheasants from France (Aldous et al., 2007; Aldous & Alexander, 2008). Another epidemic took place in Slovakia in January 2007, when ND-infected pigeons were imported from the Czech Republic (OIE, 2007). In both cases, the mortality and morbidity rates associated with the disease were unusually low. This atypical epidemiological condition contributed to the absence of obvious clinical signs of ND in the farms of origin in France and the Czech Republic, respectively. Subsequently, the time-to-detection of the index case during the silent phase of the epidemics in the countries of origin was probably higher than those typically expected for ND. These observations suggest that, under uncommon epidemiological conditions that lead to an increase in the duration of the silent phase of an ND epidemic, the risk of v-NDV introduction into ND-free EU member states such as Spain may be higher than the value estimated here. For those reasons, and even if the risk for v-NDV introduction into Spain estimated here was nil, the spatial quantification of the risk in terms of most likely country of origin, species of introduction, and region of Spain in which the epidemic is most likely to occur provides valuable information in terms of disease prevention and preparedness.

Interestingly, risk was substantially concentrated in terms of most likely species of introduction, of spatial distribution, and of country of origin. The highest risk was imposed by importation of ducks. In addition to the large number of ducks imported annually by Spain, and the observation that most of these shipments (99.4%) originate from France, where ND epidemics have been reported recently (Spanish Foreign Trade Statistics Database, 2008; OIE, 2010a), ducks offer conditions that may favour the spread of the disease. Specifically, ducks have higher probability of survival of NDV infections compared with other species; and, furthermore, the duration of the silent phase of an ND epidemic is likely to be longer in ducks than in any other species (Table 1) (Onapa et al., 2006; Lee et al., 2009; OIE, 2010a, OIE, 2010b, OIE, 2010c). Much (41.2%) of the risk was estimated for the four provinces of the Spanish region referred to as Catalonia, which is the region that receives the largest number of poultry shipments, including 99.1% of these shipments from countries with a recent history of ND epidemics (Spanish Foreign Trade Statistics Database, 2008; OIE, 2010a, OIE, 2010b). Moreover, Catalonia is the Spanish region with the largest density of commercial poultry farms (0.107 commercial poultry farms/km²) (MARM, 2008a, b; INE, 2010), which would probably favour ND spread if the index case was located in this region of the country. The combination of a large number of duck shipments and a recent history of ND epidemics resulted in France being the country estimated to give the highest risk (64.8% of P _ND) for introduction of v-NDV into Spain via the import of live poultry. Consequently, and although occurrence of an ND epidemic in Spain through import of live poultry was estimated to be an extremely rare event, one would expect that, if an epidemic associated with such route of introduction takes place in Spain, then it would most probably be caused by the introduction of ducks from France into Catalonia.

In the context of international trade, risk analysis, of which risk assessment is an important component, is the method of approach recommended by the OIE to provide importing countries with an objective and transparent tool for assessing disease risks associated with the importation of animals, animal products, or biological products (OIE, 2009b). Noteworthy and similarly to any risk assessment, these results will remain accurate and valid as long as the conditions, parameters, and values modelled here persist. Changes in some of those conditions, parameters, or values may result in different outcomes, and for that reason periodical updates of the model are required in order to early identify potential changes in the risk.

In conclusion, the present study, which was the first quantitative assessment of the risk of ND introduction through the legal importation of poultry published in the peer-reviewed literature, suggests that this route imposes almost no risk for v-ND introduction into Spain. If prevailing conditions persist, increase in biosecurity measures, which would decrease the risk associated with entry routes other than legal trade, would have a higher impact on decreasing the risk for v-ND epidemics in Spain than decreasing the number of poultry imported by the country.

Acknowledgements

F. Sánchez-Vizcaíno was supported by a research grant from Conselleria de Agricultura, Pesca y Alimentación (Generalitat Valenciana). The study was also supported in part by the Ministerio de Medio Ambiente y Medio Rural y Marino (MARM), the project FAU 2008-00001-C02-01and by the US National Center for Medical Intelligence (NCMI).