Risk assessment of the introduction of highly pathogenic avian influenza into Spain by the arrival of migratory wintering birds
PhD Thesis defense by Marta Martínez Avilés at the Faculty of Veterinary Medicine of the Complutense University of Madrid
June 24th, 2010
The intercontinental spread of H5N1 highly pathogenic avian influenza (H5N1 HPAI) at the end of 2005 from Southeast Asia to Europe and Africa, reached an unprecedented spread causing for massive mortalities in birds, many of them wild species. Aquatic wild birds, natural reservoirs of influenza virus type A, were victims in Europe of an epidemic the first wave of which coincided with the autumn migration from Northeastern to Southwestern Europe. The spread of H5N1 HPAI among wild aquatic birds in Europe and the progressive and continuous detection of outbreaks increased the alarms leading to huge efforts in time, sampling and costs in avian influenza surveillance.
Spain’s strategic situation in relation to the migratory flyways of these birds translates in an important population of wintering waterbirds. This fact, added to the representativeness of the avian production industry in Spain, increased the concern about the continuous spread of the disease and its possibility to reach our country.
Previously, EFSA had published a qualitative assessment of the risk of introduction of AI by migratory birds in Europe, which we took as a starting point in our investigations. Within this framework, a semiquantitative model was developed for Spain based on data from our own country, on our birds and farms, selecting the information relative to migration season, behaviour and bird movement of aquatic bird species. The result, expressed by provinces, allows obtaining a first level identification of the risk of introduction and its consequences by type of avian production.
Once the information was processed and analyzed, a more complex but also more precise methodology was raised to assess the risk of introduction by migratory water birds through a spatial quantitative analysis. The detailed analysis of the Migratory Species Office’s (MARM) database of wild birds’ movements and the use of spatial epidemiology techniques allowed estimating a relative risk of introduction for each of the Spanish veterinary units. This allowed evaluating and improving the design of the surveillance scheme.
Among the most noteworthy results, both methods identified the Southwest corner of Spain as the area with the highest relative risk of introduction of wild wintering waterbirds. Together with the Northeastern region, they represent the areas in which the local impact of the introduction of H5N1 HPAI, estimated with the semiquantitative method, would be higher, that could be because of their importance as wintering sites but also because of the spatial distribution of certain outdoor avian production systems. With the quantitative spatial model, there is also a northern higher risk area corresponding to the provinces of Alava, La Rioja, Soria and to some of the veterinary units of Zaragoza and Huesca mainly.
Another interesting result is the high risk index identified for aquatic wild birds such as Larus ridibundus, Anser anser or Anas crecca which can help to target surveillance sampling towards those areas in which these species are abundant during the wintering period. However, the scientific knowledge about the disease in wild birds is still limited, and of the 3 species mentioned, for the moment, only Anas crecca has been pointed as a possible spreader of AI virus, increasing the importance of this species for surveillance.
Results were also obtained for the assessment of the wild bird surveillance efforts in the European Union (EU) during 2005-2007. In both Spain and The Netherlands, the number of wild birds sampled was apparently much higher than its probability of H5N1 HPAI occurrence compared with other EU countries. However, results at the veterinary unit level revealed that this apparent oversampling was concentrated in specific areas, while others were completely neglected.
The two AI outbreaks occurred in Spain, one in a H5N1 HPAI positive crested grebe (Podiceps cristatus) in 2006 in Salburua (Vitoria), and another more recent one in a layer hen farm positive to H7N7 HPAI in Almoguera (Guadalajara) in October 2009, have allowed the validation of our results. The former was located in one of the areas at higher relative risk of introduction, while the latter was in a high relative risk area.
The results obtained in this Thesis demonstrate that a sampling scheme based only on the abundance of aquatic wild birds is not enough to early detect the circulation of a specific subtype. On the contrary, an efficient early detection system should be based on epidemiological tools such as risk assessment or spatial analysis techniques that consider both the disease incidence as well as the biological information of vectors, wild birds in this case, which is in summary what is being demonstrated in this Thesis.