Dr. Margaret Good:

Thank you, Chairman. A number of questions were raised on the bovine TB eradication programme at the meeting of this committee on 25 October looking for clarity in some areas in respect of testing. Unlike perhaps in engineering, there is no test in medicine whereby there can be 100% accuracy all of the time so it is good to try to aim for clarity.

Bovine TB is a disease caused by a living organism, Mycobacterium bovis,which affects living animals and while there are two biological entities going on, there is not 100% certainty. There are a variety of reasons for that such as the differences in the strain of the mycobacteria that are affecting the animal or even the number of bugs that got in and invaded the animal, the animal's own immune ability and the animal's genetics.

However, the Minister’s veterinary inspectors working on this disease use their veterinary training and epidemiological skills as the basis for decision making. Let us look at the tests we use for the bovine TB eradication programme. Fundamentally, we have two basic tests. We have a skin test and a blood test and I will cover both. The tuberculin skin test has been around for more than 100 years. Farmers among the committee members will know it is bothersome. It requires two visits, the cattle have to be assembled and put through the crush and handling facilities, they have to restrained twice, once for clipping, measuring and injection and then again 72 hours later for palpation of the lumps, measurement of them and assessment of them.

It is a somewhat subjective test with a degree of inter-operator variability in test performance and can be subject to human error. Nevertheless, despite repeated attempts to find a better test - a laboratory test would be much desired - we have had many false hopes but we have not yet found a better test. The skin test, in all of its forms, is still the most widely used field surveillance test for the detection of cattle with TB worldwide. The same is true in humans where they use a variation of exactly the same tests as we use.

In Ireland we have many environmental mycobacteria and other diseases which can interfere with the test. They cause cross reactions and can give rise to false positives and false negatives. To minimise the impact of the environmental mycobacteria we use the most specific of the skin tests. I will come back to what I mean by that. We use the single intradermal comparative tuberculin test, SICTT, where two injections of tuberculins, purified protein derivatives, one from M.bovis, which is what causes TB, and the other one from M.avium, which is a close relation but does not cause TB in cattle. That is to knock out the cross reaction. The injections are into the mid-neck region in cattle and the response to both injections is compared 72 hours later. Since the injections are just protein they cannot cause disease and because they are sterile they will not cause any other reactions unless the animal has met the TB organism and been sensitised, or alas, one of the cross-reacting ones also might sensitise. From here on I will call that the skin test.

We also have a blood test so when we compare tests we use terms such as test sensitivity, which is the ability of a test to correctly identify infected animals. The sensitivity of the skin test in this country has a sensitivity of approximately 85%. When we are in an infected herd we can improve the sensitivity by applying a severe interpretation to the test. Repeating the test multiple times also improves sensitivity. We also talk about test specificity, which is the ability of a test to correctly identify non-infected animals. I said that the test we use is the most specific. Our test specificity is almost 100%. It is perhaps 99.9%. We only had 0.2% of animals positive to the skin test in 2015. That tells us that the number of false positives we have has to be very low. We carry out approximately 8 million tests every year and we do get some, which is still an area of concern.

The next thing we refer to when we talk about tests is the positive predictive value, PPV. That is the odds that a positive test correctly identifies an infected animal. It is directly related to disease prevalence in that if one is in a population with a high disease prevalence, one will have a very high probability, whereas, if one is in a population with a very low disease prevalence, conversely, one will have a lower probability. As our disease prevalence falls, the risk of having false positives, the positive predictive value being low, will get higher the nearer we get to being truly TB free. We are not there yet but we can look forward to that time. In summary, on the tests, the skin test accurately performed under ideal conditions is an extremely accurate test. The number of false positives to the test use in Ireland should be low. The ability of the test to detect infected animals is good and it is improved by changing the interpretation when we are in an infected herd and repeating the tests.

Nevertheless, a proportion of infected animals can remain undetected even over multiple tests. We know that the conditions under which cattle are tested are often less than ideal, to which I am sure a number of members can attest, and that any tests performed by humans can be the subject of human error. To improve the detection of infected animals and remove them from a herd before they become reactor in a subsequent test, turn up with lesions at slaughter or spread TB to their herd mates, we also carry use a blood test known as interferon-gamma assay. We use this within eight hours of sampling. The skin and blood tests measure the same principle in the animal. The cell mediated response measures it in the live animal beside the crush by injection and the other examines the same reagents and the response in the blood in the laboratory. There should be a high degree of correlation between both tests. We need live cells in the laboratory to perform the interferon-gamma assay test and so for maximum sensitivity, we start the assay within eight hours of collection of the sample. This is challenging logistically. All the samples are tested in one laboratory to ensure consistency of assay. In other words, getting the samples taken on farm and to the laboratory so that the assay can start within eight hours can be challenging. We have two laboratories in the country that carry out the initial stage of the assay, which has helped to alleviate some of the logistical challenges.

The specificity of the skin test is almost 100% but the specificity of the interferon gamma assay is only around 97%, which in the context of 8 million tests, is low specificity. It is not suitable to replace the skin test as a screening test because that is much higher. The sensitivity when the test is carried out at eight hours is much the same as in the skin test. The benefit of doing this is that at any one time there is a small proportion of animals that will only be positive to one or other of the tests. To maximise the detection of all infected animals, we do both tests together and we get the totality or almost the totality. When we run the two tests together, the combined tests have a sensitivity of about 93%, which is a good improvement on 85%. To ensure that we maximise the positive predictive value of the test, we only use it when we have already detected TB in a herd. We use it in herds in which four or more reactors have been identified on the skin test, which animals were not bought in but acquired the disease in the herd such that there may be other animals that were in the herd that were exposed and infected. The veterinary inspector in the local office will have evaluated the epidemiological evidence of active TB and identified the cohort animals to which they have been co-exposed with the reactors already detected and so they are the animals that we reckon are at higher risk of infection.

In summary, using both tests together - the skin and eight hour gamma tests - concentrating on infected herds and on the high-risk cohorts of potentially infected animals, reduces the proportion of infected animals that might remain undetected turning up later at subsequent tests or become latent and lie dormant and either be detected later at slaughter, in which case they may restrict the herd again, which is very frustrating for a lot of farmers, or worse, cause a re-occurrence of infection and potentially spread TB again within the herd and cause another outbreak in animals or even humans. They are the two basic tests that we use. Since both tests measure the same response and should be well correlated, we have looked at using the blood tests as part of our quality assurance assay. Over the years, we have had various quality control and quality assurance measures in the programme to try to improve the quality of it. We maintain a monitor of every aspect and take steps where we can to improve the quality of the skin testing. For example, we have measures to check that the testing facilities are adequate to allow the animals to be tested accurately and safely and to ensure there is consistent injection site accuracy. We train people to do this and we also carry out quality control checks to ensure that on reading day the reactions are assessed and measured properly, with a view to more accurately identifying TB infected animals.

As a quality control, the quality assurance checks we have trialled using the blood test collected after the reactors have been disclosed. We use a 24-hour test so we can post the samples. These are animals that have been already tested and it is a quality assurance so we can afford to lose some of the sensitivity on it. The first full year of operation of the quality assurance blood test was 2015. We had 92% of the standard reactors and 72% of the standard inconclusive reactors that also failed the blood test, such that there was very good concurrence between it. When the level of agreement between the tests falls significantly below 80% to 85%, the veterinary inspector will visit the herd, if he or she has not already done so, to assess the skin reactors, examine the nature of a lump and the regressions and whether the animals have been on any medications or had any vaccinations that might have interfered with the test. The inspector will also then carry out a repeat blood test, which is usually done within eight hours, to see if there have been any mistakes. The aim is to avoid the taking of uninfected productive animals from herds unnecessarily. There is no bona fide farmer who wants to lose more animals than he or she needs to lose in order for a herd to be tested clear again.

For now, because of the greater familiarity with the skin test, the gamma assay as a quality assurance measure is still viewed with a degree of suspicion. We have the same goal ultimately in mind, which is to rid a herd of infection and to restore it to officially free status as soon as possible. To date, in exceptional cases where the quality assurance gamma has indicated that the animals are not TB infected and we have taken a decision to leave them in situand retest them after six weeks none have subsequently failed the skin test or confirmed as infected with TB. Undoubtedly, some day some will but to date none have. We are cautious about that. As I said, our goal is to get the herd TB free.