The slaughterhouse as data source for monitoring programmes in cattle

The encountered animal and public health related challenges over the last decades, with a globalized market and current changes in climate, there is general consensus that the (re-) emergence of infectious disease is going to be a major preoccupation of veterinary public health also in the future. To protect domestic livestock from the introduction of diseases and to facilitate trade, countries are bound by international agreements to conduct nation-wide surveillance programmes with a sound scientific basis in order provide evidence on the sanitary status of the production animals.
The contrasting scarce resources at disposition give flexible and cost effective tools and methods for animal health surveillance great relevance.
In Switzerland, the implementation of bulk tank milk testing yields a significant potential for reducing costs and effort of surveillance programmes, as on-farm blood sampling is laborious and costly. On the downside, cattle reared for meat production, i.e. roughly ⅓ of the Swiss cattle population, are not covered. An alternative for this population segment would be the sampling in slaughterhouses, provided the abattoir cattle population reflects the general cattle population in the surveyed area.
Based on the data from the Swiss animal movement database (AMD) and a feasibility study in the six biggest slaughterhouses, the aim of the present PhD project was to assess the practicability, cost-effectiveness and representativeness of sampling the non-dairy population at the slaughterhouse instead of on-farm.
A part of the study was to understand the demography and movement patterns in the Swiss cattle industry. A stratified demographic population model served to define parameters that can describe the population dynamics in dairy and non-dairy cattle. By introducing trigonometric functions to simulate the seasonal oscillations of calve birth and mortality rates, the model fitted very well to the data from the AMD. The obtained population parameters can be used to describe the population dynamic in cattle as building block for future model applications. The network analysis of the cattle movements depictured a highly interlinked industry with the properties of a scale free network and very large contact chains (i.e. chronological movement sequences). The seasonal changes in the network are linked to the traditional alpine pasturing during the summer months.
Possible risk factors for surveillance would be a high level of ingoing contacts (direct or over several farms) but as the cattle trade network is so connected and many traders operate on national level the value of the risk factors compared to random sampling should be validated in further studies. To assess also the connectivity of the network, the movements were not only investigated using network analysis techniques but also by mapping them on the road system in Switzerland. The intensity of road utilization creates a distinct pattern and allows identifying high risk areas for disease spread.
The sampling at the slaughterhouse was studies from different points of view; a transdiciplinary approach was realized to assess the practicality and implications for the veterinary service if sampling is shifted from on-farm to the slaughterhouse. By including the chief meat inspectors of the six biggest cattle slaughterhouses in Switzerland, the experience and hands-on knowledge of the people who are most concerned by the implementation of such a shift in sampling policy could be integrated from the beginning of the project.
Risk factors derived from the animal movement database for the occurrence of extended-spectrum β-lactamase (ESBL) producing Enterobacteriaceae in young cattle in Switzerland were analysed using a slaughterhouse sample. The faeces samples could be gained at the slaughterhouse without problems and represented the Swiss cattle population well. Risk factors for shedding ESBL producing Enterobacteriaceae were age, primary production type, meaning dairy compared to beef on farm of origin the number of animal movements to the farm of origin. The gained intelligence could improve management strategies in animal holdings towards a reduction of ESBL prevalence in cattle.
With a more theoretical approach, the sampling was modelled using the Swiss cattle movement data from 2012 for different scenarios. With this individual-based model, the costs and sensitivity of the sampling programme was compared for on-farm versus slaughterhouse sampling. The model showed that on animal level the slaughterhouse was cheaper than on-farm sampling with the same outcome. For conclusions on herd-level, the low herd-level sensitivity at the slaughterhouse is limiting possible sampling scenarios. For instance, to proof absence from disease with the internationally given 99% sensitivity, that the herd-level prevalence is below 0.2%, the samples need to be collected over the entire year with a limited number of samples per day to cost less than the on farm sample with the same outcome.
In conclusion, sampling at slaughterhouse level is a valid alternative to on-farm blood sampling and the slaughterhouses provide a flexible sampling location where hundreds of specimens are accessible daily. For farm-level outcomes of routine surveillance, the programme must be planed carefully and a longer sampling period is necessary to reach reasonable system sensitivities. The practical feasibility is given, as long as all stakeholders are included in the planning and implementation of surveillance programmes.
The animal movement database is a valid resource for a wide range of information related to cattle production and trade for the public health sector. The conducted network analysis and the dynamic population model gave insight to some characteristics of the Swiss cattle industry and create a basis for further exploration of disease transmission and control strategies in the cattle population.