SE0213: An epidemiological study of sheep scrapie to determine means of natural transmission and role of PrP genotype

SID5 report collated in June 2006 for Defra by Miss S.C. Tongue, Epidemiology Workgroup, Centre for Epidemiology and Risk Analysis, VLA – Weybridge.

The main aims of SE0213 at its end point in 2005/06 were:

  • To further the understanding of the epidemiology of scrapie within scrapie affected flocks
  • To quantify the effect of and investigate the relationship between possible risk factors, including PrP genotype and parental scrapie status.
  • To determine the effect of introducing resistant rams in a scrapie-affected flock.
  • To estimate the prevalence of infection in culled sheep of known genotype from naturally infected scrapie flocks

The scientific objectives set out for SE0213 (due for completion in December 2006) were as follows:

  1. Monitor incidence of clinical scrapie in well-recorded flocks to provide estimates of annual age-specific incidence.
  2. Description of clinical signs and distribution of onset for all confirmed cases
  3. Determine the effect of genotype and the incubation period in various breeds on the risk of developing clinical scrapie
  4. Determine the effect of factors on the risk of developing clinical scrapie. Risk factors to include whether born inside or outside, hygiene at lambing time, possible exposure to infected placental tissue, clinical disease in parents and whether genotype is a confounder of association with parental infection status.
  5. Determine whether genetically resistant rams can be used to decrease the incidence of disease in affected flocks.
  6. Use collected information to give feedback and advice on control policies to farmers, local Vets and Regional Laboratories and to keep the fund-holder fully informed of progress.
  7. Estimate the prevalence of infection in culled sheep of known genotype from naturally infected scrapie flocks of known clinical incidence
  8. Determine what future analyses are feasible with the data available

The extent to which the objectives have been met are as follows:

  1. Completed – see report on analysis.
  2. Data collected but analysis unable to be completed due to funding end date of March 2006 (target for objective was 31/12/2006).
  3. Completed – see report on analysis.
  4. Completed – see report on analysis
  5. Initial analysis complete – write up in progress, not yet completed (target for objective was 31/12/2006).
  6. Continuous throughout project, see Appendix ‘Outputs’.
  7. Completed – see publication: Tongue et al. Prevalence of scrapie infection in cull animals from 14 scrapie-affected flocks in Great Britain. Veterinary Record (2005) 157, 480-482.
  8. In progress – see report: (target for objective was 31/12/2006).

The details of the methods used and the results obtained are included in this attached report, as is a brief discussion of the results and their reliability. The main implications of the findings are also outlined, along with possible future work.

This project has contributed throughout its lifespan to the formulation of evidence-based policy for the control of scrapie in sheep.

Analysis of SE0213 data: Report prepared in May 2006 for Defra

By Linda Hoinville, Brian Henderson and Sue Tongue

1. EXECUTIVE SUMMARY

2. INTRODUCTION

2.1 Aims of analyses

3. Analyses

3.1 Decisions about which farms to include in the analyses

3.2 Calculation of incidence of clinical scrapie in flocks included in analysis

3.3 Analysis of incidence by calendar year, birth cohort and age in each flock

3.4 Assessing the impact of PrP genotype on the incidence of clinical scrapie

3.4.1 The impact of PrP genotype on the risk of developing clinical scrapie

3.4.2 Variation in the impact of PrP genotype between flocks

3.4.3 The impact of PrP genotype in breeds in which the risk of scrapie occurrence in ARQ homozygous animals was estimated to be less than one percent

3.4.4 The impact of PrP genotype in breeds in which the risk of scrapie occurrence in ARQ animals was estimated as more than one per cent

3.5 Assessing the impact of other individual animal level risk factors on the occurrence of clinical scrapie

4. DISCUSSION

5. IMPLICATIONS OF THESE FINDINGS

6. ACKNOWLEDGEMENTS

1. EXECUTIVE SUMMARY

This report summarizes the epidemiological analyses of data collected to investigate the occurrence of scrapie on selected flocks as part of project SE0213. The main aims of this study were to describe the occurrence of clinical scrapie on these flocks and to assess the impact of PrP genotype and other individual animal level risk factors, particularly parental scrapie status, on the risk of developing clinical disease.

The data collected over a ten year period from scrapie affected flocks has been checked for inconsistencies and analyses have been conducted to address the main objectives of the study. These analyses include calculation of the incidence rate of scrapie on each flock in order to describe the occurrence of disease and survival analysis and conditional logistic regression of sub-sets of the data to determine the impact of PrP genotype and other individual animal level risk factors on the risk of developing clinical scrapie.

The main findings are:

Confirmation that PrP genotype influences the occurrence of clinical scrapie with variation in the effect of some PrP alleles between flocks

  • The strong association between PrP genotype and the development of clinical scrapie was confirmed by our analyses with the ARR allele providing a strong protective effect in all flocks and a very low risk of clinical disease in animals carrying this allele.
  • The impact of other PrP alleles on the risk of developing clinical scrapie varies between flocks and is determined more by the risk of scrapie occurrence in homozygous ARQ animals than the prevalence of the VRQ allele in the breeds of sheep on the farm. This suggests that the variation in impact of the PrP allele may be more related to the nature of the agent infecting the flock than the breed of sheep.

Evidence for maternal transmission of scrapie but not paternal transmission

  • The offspring of ewes which later develop clinical scrapie are at significantly increased risk of developing clinical disease and this association may be stronger on those farms on which animals lamb outside and on farms that do not remove after-births from the lambing accommodation
  • The offspring of rams which later develop clinical scrapie are not at significantly increased risk of developing clinical disease

The detailed data collected in this study provide a unique opportunity to investigate the occurrence of scrapie on a number of heavily affected flocks. This will allow a detailed description of the occurrence of scrapie on these flocks and investigation of the factors influencing the occurrence of scrapie. Possible areas for future analysis include further investigation of the association between individual level risk factors and the occurrence of scrapie and the further work to determine the impact of the introduction of genetically resistant rams on the occurrence of clinical disease. A description of the clinical presentation and strains of scrapie present in these flocks could provide further evidence for strain variation in the scrapie agent. These data could also be used in mathematical studies to examine the dynamics of the scrapie agent in affected flocks. A concept note for further work using these data will be submitted when the initial analyses described in this report have been finalised and submitted for peer-reviewed publication

2. INTRODUCTION

2.1 Aims of analyses

The main aims of these analyses of the data collected from scrapie affected farms for project SE0213 were to

  • Describe the occurrence of scrapie on affected farms
  • Determine the impact of PrP genotype on the risk of developing scrapie
  • Determine the impact of individual animal level risk factors on the risk of developing scrapie particularly maternal and paternal scrapie status

3. Analyses

3.1 Decisions about which farms to include in the analyses

Having reviewed the data available from all of the 70 farms recruited to SE0213 a decision was made about whether the data from each of the farms should be included in the analysis. A summary of the number of farms recruited to the study and whether or not the data should be included in the analyses by the year in which farms were recruited is provided in Table 1. The reasons for exclusion of farms are also shown in this table.

Table 1: Farms included in analysis by year of recruitment

Year recruited / Data not provided
by farmer / Data inadequate / Flock slaughter during FMD outbreak / No confirmed cases while flock in study / Farm
included in analysis (%) / Total flocks
recruited in this year
1994 / 5 (100) / 5
1995 / 3 (100) / 3
1996 / 3 / 6 (67) / 9
1997 / 1 / 2 (67) / 3
1998 / 2 / 4 (67) / 6
1999 / 1 / 5 (83) / 6
2000 / 5 / 7 / 3 / 4 / 11 (37) / 30
2001 / 3 / 1 / 2 / 2 (25) / 8
Total / 14 / 9 / 3 / 6 / 38 (54) / 70

3.2 Calculation of incidence of clinical scrapie in flocks included in analysis

The final data set includes 32,580 animals 981 of which were pathologically confirmed as having scrapie during the study period.

Table 2: Summary of characteristics and incidence of clinical scrapie in the 38 flocks included in analyses

Flock code / Year recruit / Years in study / Flock size
(est) / Main breed of female sheep / Animals in analysis / Confirm cases in analysis / Overall incidence – cases / 100 sheep years at risk (95% ci)
1 / 1994 / 4 / 695 / Swaledale / 1494 / 46 / 1.3 (1.0 - 1.8)
2 / 1994 / 5 / 136 / Charollais / 969 / 10 / 0.5 (0.3 - 0.9)
3 / 1994 / 2 / 830 / Swaledale / 1758 / 18 / 0.6 (0.4 – 1.0)
4 / 1994 / 10 / 1700 / Finn Dorset X / 5578 / 302 / 1.4 (1.3 - 1.6)
5 / 1994 / 1 / 138 / Charollais / 150 / 4 / 2.5 (0.9 – 6.7)
6 / 1995 / 2 / (149) / Charollais / 150 / 4 / 2.3 (0.9 - 6.1)
7 / 1995 / 8 / (592) / Welsh / 2368 / 57 / 0.7 (0.5 - 0.9)
8 / 1995 / 8 / (407) / Welsh / 1638 / 29 / 0.6 (0.4 - 0.8)
9 / 1998 / 4 / 1070 / Mixed / 1963 / 59 / 1.3 (1.0 - 1.7)
10 / 1998 / 4 / 1060 / Swaledale / 1006 / 4 / 0.2 (0.06 - 0.4)
11 / 1998 / 5 / (62) / Charollais / 156 / 1 / 0.2 (0.02 - 1.4)
12 / 1999 / 5 / 358 / Mixed / 914 / 111 / 4.6 (3.8 - 5.5)
13 / 1999 / 3 / 50 / Texel / 60 / 3 / 1.6 (0.6 - 5.8)
14 / 1999 / 5 / 300 / Mixed / 463 / 18 / 1.2 (0.8 - 1.9)
15 / 1999 / 3 / 944 / Swaledale / 503 / 16 / 1.6 (1.0 - 2.5)
16 / 1999 / 4 / 424 / Swaledale / 661 / 45 / 3.2 (2.4 - 4.3)
17 / 2000 / 4 / 206 / Mixed / 302 / 10 / 1.3 (0.7 - 2.5)
18 / 2000 / 2 / 125 / Texel / 130 / 2 / 1.0 (0.2 - 3.9)
19 / 2000 / 6 / 32 / Mixed / 57 / 1 / 0.5 (0.07 - 3.6)
20 / 2000 / 5 / 24 / GFDM / 75 / 1 / 0.6 (0.09 - 4.5)
21 / 2000 / 5 / 180 / Poll Dorset / 412 / 7 / 0.6 (0.3 - 1.4)
22 / 2000 / 2 / (457) / Cross breed / 457 / 4 / 0.6 (0.2 - 1.7)
23 / 2000 / 4 / (9) / Wiltshire / 18 / 2 / 4.7 (1.2 – 19)
24 / 2000 / 3 / (87) / Bleu de Maine / 131 / 1 / 0.4 (0.06 - 3.2)
25 / 2000 / 4 / (313) / Mixed / 720 / 11 / 0.6 (0.3 – 1.0)
26 / 2000 / 5 / (403) / Mixed / 994 / 1 / 0.04 (0.01- 0.3)
27 / 2000 / 2 / (202) / Mixed / 359 / 3 / 0.5 (0.2 – 1.7)
28 / 2001 / 2 / (609) / Friesland / 609 / 1 / 0.1 (0.01 - 0.7)
29 / 2001 / 1 / (362) / Texel / 366 / 5 / 1.6 (0.7 - 3.8)
30 / 1996 / 8 / 980 / Clun cross / 3027 / 33 / 0.3 (0.2 - 0.5)
31 / 1996 / 5 / 271 / Cambridge / 637 / 10 / 0.6 (0.3 - 1.3)
32 / 1996 / 7 / 736 / Vendeen / 423 / 44 / 4.3 (3.2 - 5.8)
33 / 1996 / 7 / 230 / Easy care / 580 / 24 / 1.1 (0.8 - 1.7)
34 / 1996 / 4 / (406) / Swale/mule / 813 / 23 / 1.3 (0.9 – 2.0)
35 / 1996 / 3 / 247 / Suffolk cross / 441 / 24 / 2.9 (2.0 - 4.4)
36 / 1997 / 5 / (538) / Mixed / 1383 / 22 / 0.6 (0.4 - 0.9)
37 / 1997 / 2 / (412) / Bleu de Maine / 412 / 21 / 3.1 (2.0 - 4.7)
38 / 1998 / 7 / (114) / Mixed / 403 / 4 / 0.3 (0.1 – 0.9)
Total / 32580 / 981 / 1.1 (1.0-1.2)

* Flock sizes in brackets were not provided on farm questionnaire but were estimated from data available

The incidence rate of clinical scrapie (confirmed cases per 100 sheep years at risk) was calculated in each of the 38 flocks included in the analysis. The incidence rate and 95% confidence intervals for the overall incidence in each flock is shown in Table 2 and Figure 1.

Figure 1: Incidence rate per 100 sheep years at risk in each of flocks included

The overall incidence of confirmed clinical scrapie cases per 100 sheep years at risk was less than 2 in the majority of flocks. Eight flocks had an overall incidence of between 2 and 5 confirmed cases per 100 sheep years at risk. The very wide confidence limits for incidence in flock 23 is because this was a very small flock with a high incidence of scrapie (4.7 cases per 100 sheep years at risk).

3.3 Analysis of incidence by calendar year, birth cohort and age in each flock

A more detailed analysis of the occurrence of clinical scrapie in different calendar years, birth cohorts and age groups has been conducted for each flock. The incidence of pathologically confirmed cases of clinical scrapie on one large flock is shown here as an example.

Figure 2: Incidence of pathologically confirmed cases of clinical scrapie by calendar year, birth cohort and age on one flock

a) Incidence per 100 sheep years at risk by calendar year on flock 4

b) Incidence per 100 sheep years at risk by birth cohort on flock 4

c) Incidence per 100 sheep years at risk by age in years on flock 4

Analysis of the age specific incidence of clinical cases in animals born in each year (Figure 3) suggests that the high incidence of scrapie occurring in animals over 7 years of age is the result of exposure to scrapie at an older age. The high incidence in older animals occurs only in those animals born in late 1980’s when the scrapie epidemic had only just started in this flock, these animals were not exposed to a high level of infection until the mid 1990’s so did not develop scrapie until late in life. All animals included in this study have been monitored until they were at least nine years of age and the high incidence of scrapie in older animals has not occurred in any animals born after 1989.

Figure 3: Age specific incidence of clinical scrapie in each birth cohort in flock 4

3.4 Assessing the impact of PrP genotype on the incidence of clinical scrapie

3.4.1 The impact of PrP genotype on the risk of developing clinical scrapie

This project was set up as a cohort study to monitor the occurrence of clinical scrapie in all animals present on selected farms when the flock was recruited and in animals born in these flocks or purchased while the farm was included in the study. PrP genotype is an important determinant of scrapie occurrence but for cost reasons it was not possible to determine the PrP genotype of all animals included in the study. To determine the influence of PrP genotype on the risk of scrapie occurrence and to control for any confounding effect of PrP genotype on the impact of other risk factors the study was designed to collect PrP genotype information from

  • All scrapie suspects and two age matched unaffected animals from the same flock. This provided data for analysis as a case-control study.
  • All animals born in a single birth season at the time the flock was recruited to the study. This provided data for analysis as a cohort study.

Although the sampling of animals on farms was not always conducted according to the sampling protocol designated at the start of the study for various reasons (e.g. farmers requested sampling of other animals, no scrapie cases occurred in animals sampled as part of the cohort study, unaffected animals were not age matched to cases) it has been possible to create two data sets for analysis containing

  • All confirmed cases of scrapie and up to two age and breed matched controls from the same birth cohort in the same flock, referred to as case-control population in this report. This population included cases and controls from 32 of the 38 flocks included in the analysis. Four flocks were excluded because there were no confirmed scrapie cases of known breed and genotype and two flocks were excluded because there were no suitable control animals available.
  • Animals born in a birth season in which clinical scrapie cases occurred and at least 75% of the animals were PrP genotyped, referred to as cohort population in this report. This population included animals from 22 of the flocks included in the case control study. All flocks that were included in the case control study but excluded from the cohort study had no birth cohorts in which there was a confirmed case of scrapie and at least 75% of animals had been genotyped.

The impact of PrP genotype on the risk of developing clinical scrapie was investigated using survival analysis of the cohort population data and conditional logistic regression of the case control population data. Survival curves by PrP genotype in the cohort population are shown in Figure 4. This graph shows that the risk of developing clinical scrapie is strongly influenced by PrP genotype with genotypes falling into three groups

  • Genotypes in which no or very few scrapie cases occurred in this population– ARR/ARR, ARR/VRQ, ARR/ARH, ARR/AHQ, ARH/ARH, AHQ/VRQ, AHQ/ARH
  • Genotypes in which between 25% and 50% of animals surviving to 8 years of age developed clinical scrapie – ARQ/ARQ, AHQ/AHQ, AHQ/ARQ, ARQ/VRQ
  • Genotypes in which more than 75% of animals surviving to 5 years of age developed clinical scrapie – VRQ/VRQ, VRQ/ARH

Figure 4: Proportion of each PrP genotype surviving by age in years

The impact of PrP genotype on clinical scrapie occurrence was also assessed using the hazard ratio (HR) from a Cox proportional hazards model in the cohort population (stratified for flock and birth season) and using the odds ratio (OR) from a conditional logistic regression analysis of the case-control population (matched for age, breed, flock and birth season). The results of these analyses are shown in Table 3. In all analyses the risk of scrapie occurrence was compared to the risk of clinical scrapie occurrence in animals with the ARQ/ARQ (wild type) genotype. The HR or OR indicate how many times more likely an animal with each genotype is to develop clinical scrapie than an animal of the ARQ/ARQ genotype. The HR and OR could only be estimated when sufficient numbers of affected and unaffected animals of a genotype were present in the population. The 95 percent confidence intervals indicate the likely range of values for the HR or OR and if these limits do not contain the value 1 there is less than a 5% probability that the risk of clinical scrapie occurrence in these animals is equal to that in ARQ/ARQ animals. An HR or OR of less than one indicates that the risk of clinical scrapie in this genotype is less than that or ARQ/ARQ animals and a value greater than one indicates that the risk of clinical scrapie is increased compared to ARQ/ARQ animals.

The analyses summarised in Table 3 demonstrate that VRQ/VRQ animals are at significantly greater risk of developing scrapie than ARQ/ARQ animals and ARQ/VRQ animals are at slightly increased risk although this increase is only statistically significant in the case control population. Animals of the ARH/VRQ genotype also appear to be at increased risk but this result is not statistically significant in either population because of the low prevalence of this genotype.

Animals of the AHQ/AHQ and AHQ/ARQ genotype appear to be at less risk of developing scrapie than ARQ/ARQ animals although this reduction is only statistically significant in the case control population. It also appears that animals of the AHQ/VRQ genotype are at reduced risk of developing scrapie as there were no cases of this genotype although 4.6% of controls in the case control population were of this genotype.