Use of a preclinical test in the control of classical scrapie

Scrapie control in Great Britain (GB) was originally based on the National Scrapie Plan's Ram Genotyping scheme aimed at reducing the susceptibility of the national flock. The current official strategy to control scrapie in the national flock involves culling susceptible genotypes in individual, known affected flocks (compulsory scrapie flock scheme or CSFS). However, the recent development of preclinical test candidates means that a strategy based on disease detection may now be feasible. Here, a deterministic within-flock model was used to demonstrate that only large flocks with many home-bred ewes are likely to be a significant risk for flock-to-flock transmission of scrapie. For most other flocks, it was found that the CSFS could be replaced by a strategy using a currently available live test without excessive risk to other farmers, even if the proportion of susceptible genotypes in the flock is unusually large. Even for flocks that represent a high risk of harbouring a high prevalence of infection, there would be limited probability of onward transmission if scrapie is detected soon after disease introduction (typically less than 5 years). However, if detection of disease is delayed, the existing CSFS strategy may be the most appropriate control measure in these cases.

. Estimated survival distribution of breeding sheep. Data from McLean et al. (1999).

Modelling genetics in the flock
There are three TSE-related alleles that have been identified -ARR, AXX and VRQ. The AXX genotype refers to ARQ, AHQ and ARH genotypes. Each lamb will have a pair of alleles; one coming from each parent. In this model the proportion of offspring of each genotype will be determined by the proportion of each of the alleles in the male population and each of the alleles in the female population, as described by Mendelian genetics.

Flock types
In the UK, flock classifications are based on breed and the role they play in the industry (Truscott & Ferguson, 2009). There is also a recognized flow of animals, genetic material and hence infection based on grazing altitude (hill to lowland) as well as from pure-bred to cross-breeding flocks (Pollott & Stone, 2003;Truscott & Ferguson, 2009).
Flock parameters selected for the model are based on flock type (pure-bred, commercial or mixed), flock size (25th and 75th percentiles), the proportion of home-bred sheep (25th and 75th percentiles), the likelihood of selling sheep on to other flocks and whether scrapie has ever been identified in the flock. Due to the difference in the distribution of the number of sheep sold, purebred and commercial flocks, as defined by the correlation between flock size and flock type, were the major flock categories retained in the model. As the numbers of sheep in mixed commercial and pure-bred scrapie-positive flocks are similar to both commercial and pure-bred scrapie-positive flock types, this mixed flock type was not investigated further.
Further stratification of flock types into small and large commercial, pure-bred and mixed hill, upland and lowland flocks was also investigated. According to the postal survey data, farm type (hill, upland and lowland flocks) appears to be correlated with flock type. Independently, farm type does not appear to play a significant role in the risk of being scrapie positive and does not have an impact on the number of sheep sold on to other flocks (Hoinville et al., 2000). Therefore these subclassifications were not used. In the postal survey data, of flocks that have ever had scrapie, commercial flocks represent the majority (68%). Pure-bred flocks (14.1%) and mixed flocks (17.5%) formed smaller proportions of the GB flock demographics (Sivam et al., 2006).

Within-flock spread of classical scrapie Susceptibility to classical scrapie
Under natural conditions, a sheep's genotype is a major influence on its susceptibility to a scrapie infection (Table 1).

Incubation periods for scrapie
Mean incubation periods for scrapie are presented in Table 2.

Horizontal transmission
It is clear that scrapie can be transmitted horizontally (from sheep to sheep through direct or indirect contact) under natural conditions.
Here, the likelihood that a sheep will become infected is determined by two things: firstly how susceptible that sheep is, and secondly the total burden of infectiousness of the flock at that time.
Infectiousness of the flock is dependent upon the number of infected sheep in the flock and upon the level of infectivity that has accumulated within each infected sheep.
Total infectivity in a sheep over the course of an infection is estimated by summing over the infectivity present in each tissue in the sheep (Table 3). Total infectivity per sheep, scaled to 1 at the clinical stage, is used to define how infectious sheep of different genotypes are over the course of an infection (Fig. 2). This is calculated by assuming that each sheep reaches the limit of infectiousness by the end of its incubation period.  Fig. 2. Normalized infectivity of a sheep as a function of time since infection, for each genotype.
The total infectiousness of the flock is the estimated at any time point as the sum of each sheep's infectiousness. This value is then scaled to give within-flock epidemics of comparable sizes to natural scrapie epidemics (Fig. 3). We assume that a flock with around 1070 breeding ewes that breeds all its own replacements has around 10 clinical cases 10 years into an epidemic.

Vertical transmission
The model also accounts for the possibility of vertical transmission from ewe to lamb. It is assumed that the probability of transmission is dependent upon the genotype of the lamb and the infection stage of the ewe. The probability of infection increases as the infection stage of the ewe increases and is bounded by the value of relative susceptibility corresponding to the lambs' genotype, as described in section 2.1.

Fig. 4. A plot showing the assumed probability of vertical transmission for lambs of different
genotypes as a function of how long the ewe has been infected.

The effectiveness of disease-specific PrP tests
To estimate the overall effectiveness of a control strategy that involves disease-specific PrP testing requires an estimation of how reliable these tests are at detecting positive scrapie samples.

Live test parameters
The live preclinical test is based on the experimental support for rectal biopsies of lymphoid tissue and is described in detail by González et al. (2008). They found that the risk of a false-negative result in a preclinical rectal biopsy sample was 9.3% if the sample contained 10 follicles and that the probability of obtaining a sample containing at least 10 follicles was 87%. The risk of a false positive is believed to be negligible (L. González, personal communication) Frequency of detection of the abnormal prion protein (PrP d ) was only slightly higher in samples of palatine tonsil and retropharyngeal lymph node of infected sheep, but these tissues are much less accessible than the rectal mucosa and therefore considered unsuitable for large scale application of a live animal test (González et al., 2006). For this model, it was assumed that at least 10 follicles are retrieved in a single biopsy sample and accordingly, initial test sensitivity of 90% and specificity of 100% was chosen for the model. Test sensitivities of 70% and 35% are also investigated.
In the model, the live preclinical testing strategies are applied to sheep of different ages (greater than 20, 12 and 6 months). These ages were based on data from González et al.