By browsing our site you accept the installation and use cookies on your computer. Know more

Menu Bienvenue sur le site mihmes

Home page

Task 2a: Understanding the role of the different types of transmission between herds in the regional spread of pathogens: application to between-herd diffusion of Cb

Domestic animal populations are structured into herds which are not homogenously in contact. Their direct contacts are rather heterogeneous (determined by geographical proximity, type of herds, etc) and can be summarized through networks. Host localisation and between-herd distance influence the regional pathogen spread, especially for indirect transmission which mostly depends on meteorological conditions. How space is represented in models of pathogen spread at large scales depends on knowledge of the epidemiological system, especially host localisations, between-host contacts, and host movements. The underlying spatial structure of the host population shapes the simulated patterns of disease spread produced by metapopulation, continuous-space, and network-based models (see Keeling and Rohani, 2008 for a review).

Numerous spatio-temporal models exist in population dynamics and epidemiology (Tilman et Kareiva, 1997). However, the relative contribution of modelling assumptions concerning between-herd spatial transmission (e.g. related to herd density, host population structure and animal movements, wind or vector borne) to the predicted infection dynamics remains unclear. This issue will be addressed here within a theoretical modelling framework.

As an application, we will determine the relative roles of animal movements and airborne dispersion in between-herd Cb spread. A longitudinal field study will be performed in a limited area of Western France to describe the spatio-temporal distribution of infected herds and of isolated strain variants. Concomitantly, animal movements occurring between surveyed herds will be described. The infection status of each herd will be determined based on the detection of shedder animals (using PCR). In each herd detected as infected, the bacterial load in environmental samples will be quantified and considered as a proxy of the airborne material that could be released from an infected herd considered as a source. Furthermore, Cb isolates will be typed using multiple loci variable number of tandem repeats analysis (MLVA) (Arricau-Bouvery et al., 2006) to identify plausible genetic clusters of closely related strains infecting herds, and to follow in space and time the transmission of these strains between herds. A similar study design will be implemented on the Isle of Gotland (Sweden) by partner 6. Two complementary approaches will be used. First, spatial statistics will highlight putative clusters in the surveyed areas. Our hypothesis is that transmission by contiguity (e.g. wind) does not lead to distributions resembling those correlated with the animal movement network. Distributions correlated to the animal movement network therefore will indicate that animal movements most likely play a major role. In contrast, when some (putative) clusters are observed, the role of wind will be indirectly confirmed (or infirmed) by assessing whether these clusters continue to be observed after adjustment for animal movements. Second, a modelling approach coupling the within-herd model (developed in Courcoul et al., 2010 and adapted in WP1) and a between-herd model based on real cattle movements between explicitly geolocalized herds (initially developed by partner 6 for verotoxin Escherichia coli transmission) will be implemented. Simulated infection trajectories in space and time will be compared to those observed in the longitudinal study, allowing the contribution of cattle movements to the between-herd infection dynamics to be weighed depending on the degree of matching between simulated and observed data. Concerning this last point, appropriate criteria for the comparison should be proposed, since the sparsity of data does not permit an inference-based analysis. Finally, the direct contribution of the wind will be explored by using existing air dispersion models (Wallensten et al., 2010).