Executive summary from DEFRA funded report on Johne's disease
Thursday 22 November 2001
The purpose of this report was to assess the surveillance and control of Johne's disease in farm animals with a view to recommending appropriate systems of surveillance and control for Great Britain.
This was to be carried out by considering not only what had been achieved in this country, but also to take into consideration the international situation. Surveillance and control was reviewed for various countries, notably Australia, the Netherlands and the USA. An assessment of the losses attributable to paratuberculosis and the costs of the proposed surveillance and control was to be made. Deficiencies in the current knowledge base would be identified and recommendations made.
The first step in this exercise was to review the literature on paratuberculosis to establish the current limitations in understanding the disease and how these impact on diagnosis, and control. This review demonstrated that the long time course of the disease from infection to shedding the infective organism in the faeces, and eventually the development of clinical disease made research into paratuberculosis extremely difficult and expensive. Large gaps in the understanding of both the host parasite interaction and the epidemiology of the disease exist. The causal organism, Mycobacterium avium subspecies paratuberculosis (Map) is able to survive for lengthy periods in the environment, but the importance this has for the epidemiology of the infection is unknown (Chapter 8, page 54). The prolonged time interval between infection and development of the disease also means that diagnostic tests are limiting in sensitivity. These deficiencies impact adversely on surveillance, but have a greater effect on control programmes, rendering eradication of the disease extremely difficult.
Fundamental to the control of any disease is the need to identify the reservoirs of infection and to determine their significance for maintaining infection on farms or spreading the infection to previously uninfected farms. At present the importance of interactions between the different species of domestic ruminants and camelids is unknown and opinion on the possible significance appears to differ between countries. In GB where sheep and cattle populations mix on many farms this question is critical to control. A range of species of wildlife has also been found to be infected and to develop lesions of paratuberculosis. Once again the significance of this for control of paratuberculosis remains to be established (Chapter 5, page 31).
Australia and the USA have carried out work to develop and validate diagnostic tests (Chapter 4, page 22). The ability of the tests to detect the presence of infection (sensitivity) in these situations has been around 50%. In contrast, the number of uninfected animals falsely identified as positive is low (specificity of close to 100%). The poor sensitivity limits the value of the diagnostic tests for a test and cull programme for the eradication of paratuberculosis, although they may be effective in demonstrating absence of infection at the herd level. The differing climatic conditions and therefore possible difference in the rate of exposure of domestic ruminants to subspecies of Mycobacterium avium other than the subspecies that causes paratuberculosis (Map) may mean that the test has inferior specificity in GB conditions. False positives may then become a problem and so adversely affect both surveillance and control programmes. Therefore it is necessary to validate the two commercially available absorbed enzyme linked immunosorbant assays (ELISA) for the detection of antibody to this disease for GB conditions. Alternatively there is the option to develop and validate a new antibody ELISA. The same drawbacks exist for bacteriological tests and therefore automated culture systems should be validated for individual samples and for pooled faeces or bulk milk. The over thirty-month slaughter scheme, part of the bovine spongiform encephalopathy control measures, currently offers a readily available population in the target age range for the validation of these tests against the gold standard of necropsy (Chapter 8, page 55).
The current passive surveillance systems are capable of demonstrating changes in the annual trend of diagnoses, but cannot provide information on which to base an estimate of national prevalence. Such an estimate is necessary in order to determine which form of control is appropriate and to calculate the resource to be allocated. We therefore recommend that once tests have been adequately validated a national survey be carried out. We envisage a national survey that is repeated at five-year intervals (Chapter 9, page 57). Our view at this time is the absorbed ELISA for the detection of serum antibody should be used for cattle and all animals over two years of age in a selection of both beef and dairy herds should be tested. A similar study and interval is required for sheep, goats and camelids using the agar gel diffusion test (AGIDT) for the detection of serum antibody. (However it is possible to modify the ELISA for use with sera from species other than cattle and as ELISA systems can be readily automated and the AGIDT cannot be, then development of the former for sheep may be indicated). In deer younger animals are affected and consequently it would be appropriate to survey the disease by carrying out abattoir surveys, using the polymerase chain reaction (PCR) test for the presence of the organism in the mesenteric lymph nodes (Chapter 9, page 59).
Vaccination has been shown to reduce the number of clinical cases in an infected dairy cow herd and to deliver a positive cost benefit. Vaccination has also been useful in the control of the disease in sheep, but no study has demonstrated that vaccination is of value in beef cow herds. However vaccination does not significantly reduce the number of cattle that are infected in all herds and vaccinated animals that are sold to other herds will remain an effective route for the spread of the disease (Chapter 6, page 39) For this reason vaccination should only be considered as a control method in herds that do not sell stock for breeding purposes. This is unless vaccinated animals can be readily identified as such and the risk they pose made clear to prospective purchasers. In addition there are issues of safety for the operator when administering live vaccines. Vaccinated animals will also react to the avian component of the comparative intra-dermal tuberculosis test and therefore have potential to cause difficulty in the interpretation of tuberculosis herd tests.
For the control of the disease we recommend using similar methodology to that used in USA and Australia, assurance programmes based on annual testing of adult stock to demonstrate absence of disease. These are supported by biosecurity systems in the form of management rules on hygiene, the prevention of feeding calves milk or colostrum from cows other than their dam and the control of added animals. (Such programmes already exist in GB.) Herds that are free of infection can then sell accredited breeding stock as free of paratuberculosis. This would have the effect of allowing the purchase of animals, with a low risk of introducing infection to the herd and so reduce the spread of the disease. In our opinion it is necessary to have wide support for this approach from within the industry, from the breed societies and the selling agents. Secondly the purchasers of the stock must demand assurances that they are buying stock that are free of infection. To achieve the latter, information on the disease must be provided in a clear way to the industry.
Such an approach will stigmatise infected herds and it is therefore necessary to offer test and cull programmes for infected herds. Currently there is no information available that demonstrates that test and cull is successful. This is largely because such schemes have not yet had sufficient time to produce results. Eradication of the disease from a herd is essentially a process that will take many years. There is also clear opinion that if these programmes are to achieve success improved hygiene and management of calves must accompany them. In dairy herds relatively straightforward new recommendations about colostrum feeding can be made, although much of the advice on colostrum contradicts that given in the past. This is not the case in the beef herd where calves remain with the cows in an infected environment during the period when they are most susceptible to infection. Relatively little improvement can be made other than to change the time of calving to summer, calve outside, fence off water courses and provide mains water for drinking. Embryo transfer and artificial insemination are techniques that may be used to salvage genetic material from healthy animals within an infected herd, however neither can be guaranteed to produce material that is free from infection with Map (Chapter 6, page 44).
In sheep, because the cost of screening is high in relation to the value of the animal, neither assurance programmes nor test and cull programmes are recommended. This position may change if it is found that either the disease is widespread in the national sheep flock or a particular problem within pedigree breeds where the value of the animals is higher. However at present paratuberculosis is thought to be less prevalent in sheep than in cattle. In infected flocks vaccination is advocated. However the absence of a commercially produced vaccine licensed for sheep in this country is an impediment to this. The vaccination currently used is a cattle vaccine that is considerably more expensive than equivalent vaccines available on mainland Europe. Vaccination is also considered to be the most suitable method for the control of the disease in goats and deer. Were these species shown to be reservoirs of infection for cattle this advice would not apply as vaccination could not be expected to remove the risk infected animals would pose to cattle.
By using modeling techniques we have calculated average current losses for an infected 100 cow dairy herd to be ¿2600 per annum. The depressive effect the disease has on milk yield is the critical factor in the dairy herd. For our standard dairy herd we estimated a milk yield depression of 10%, varying this by plus or minus 5% provided a range of ¿1700 to ¿3600 per herd. These calculations are dependent upon limited within herd prevalence data and the many assumptions discussed in Chapter 12. In beef herds the reduction in lifetime calf output is the major factor determining the losses incurred and ¿1617 is the estimated loss for the average 100 cow beef herd. This estimate is also limited by the lack of within herd prevalence data as well as the assumptions discussed in Chapter 11. These estimates are obviously affected by the current value of the output and level of subsidy support in the case of the beef herd. The cost of the disease is therefore higher in pedigree and high genetic merit animals. The national loss can only be guessed at as no national estimate of herd prevalence is available, but using prevalence figures from other countries and the limited regional data that is available within GB a figure of 17.5% is offered as a reasonable estimate. If we vary this herd prevalence estimate by plus or minus 10% we can provide a range of losses. This gives annual losses due to paratuberculosis of ¿ 9.8 million (range ¿4.2 to 15.4 million)for the dairy herd at a milk yield depression of 10% and ¿3.1 million (range ¿1.33 to 4.88 million) for the beef herd. Clearer estimates of these costs will be possible once more accurate national prevalence figures are available.
A cost benefit analysis of the control options is not possible as there is no figure for time scale of eradication through test and cull or control through vaccination. Herds that are free of infection and enrol in an assurance programme have no immediate cost benefit unless the animals that are sold gain a premium over stock sold without a declaration of status. There is also the benefit of reduced risk of introducing infection. This cannot be estimated until national prevalence figures are known. From a national perspective we cannot at this time offer estimates of how successful assurance programmes will be in reducing the spread of the disease to previously uninfected herds. However we have constructed a framework that can be used for sensitivity analyses on the cost benefit of control programmes and we have demonstrated the results that could be expected using conservative estimates of efficacy for a test and cull programme and vaccination (Chapter 13). We would like to emphasise that our limited experience to date is that a test and cull programme will perform better than the estimates used in the examples given. It would therefore be prudent to extend the limited sensitivity analyses that have been carried out.
G Caldow and GJ Gunn
SAC Veterinary Science Division