We hear it often, "I tested my horse for EPM". Or, "I want to run the EPM test". We don't mean to nit pick, but there aren't any antibody tests for EPM and it is unlikely there will ever be one.
There are antibody tests to detect sarcocystosis but, alas, no valid tests for EPM. That is because EPM is a disease that is difficult to diagnose and an EPM diagnosis is often exclusionary. The signs that are usually attributed to disease are likely a result of the inflammation caused by the parasite in the intestine, the initial site of infection. Parasites in the central nervous system (CNS) can cause lesions, but for the majority of horses with a putative diagnosis of EPM, the signs are due to inflammation from a gut infection. A diagnostic test for EPM must identify the horses with parasites in the CNS. An exclusionary diagnosis is made when other diseases are ruled out and what you are left with is EPM as a most likely possibility.
A definitive diagnosis of EPM is made by post mortem exam. A diagnostic test for EPM would have to relate post-mortem exam with test positive results. Diagnostic tests are licensed by USDA. The first step to licensing a test is to communicate with the Center for Veterinary Biologics. There is a specific process for communicating with CVB, you don't just pick up the phone! The requirements, core guidance, for a diagnostic kit is found in VSM 800.73.
Here lies the heart of the problem of making a diagnostic test for EPM, the test must be validated to demonstrate that the test is scientifically sound, reliable, reproducible, and fit for the intended use. An EPM test is intended to identify a horse with Sarcocystis neurona in the central nervous system. That requires a validated assay.
Assay validation
Assay validation studies are required to demonstrate the diagnostic accuracy of the kit, analytical sensitivity, analytical specificity, and ruggedness. There is a requirement of interlaboratory comparison studies.
Diagnostic accuracy (sensitivity and specificity) is estimated by comparing with a reference test, a known positive, in a sample set. This is the first problem with EPM. The "gold standard", reference test, is a post mortem sample from a horse that has parasites in the CNS by natural infection. It is difficult to isolate or demonstrate S. neurona in the central nervous system of a horse. Because organisms were difficult to find researchers began using histopathology lesions showing "inflammation consistent with cases of EPM". And then when organisms were not recovered or identified the "standard" language became that the lesions look like the inflammatory lesions that are associated with S. neurona. This isn't good enough for CVB, you have to isolate the organism. That is the reference standard they require. Standards can come from natural or experimental infections. The next step is a sample set.
The impossibility of a sample set from natural EPM cases
The sample set is a set of at least 20 positive specimens from animals, selected randomly, from an intended target population. In practice that isn't feasible for EPM, the sample set from a group of available specimens is acceptable. The sample set has to have positive and negative samples. Organisms are hard to find in brain and spinal tissues, if organisms aren't found, is it good enough to say the sample is truly negative? Maybe not if you attend some EPM Society meetings. Reports of false negative cases made some researchers gun shy when they did find organisms on some horses. So hurdle #1 is finding enough cases of naturally diseased horses, all of them had to be identified by post-mortem identification of parasites in the tissues.
One has to estimate the prevalence of disease-positive samples. And if a sample was test negative, but was in realty positive, the test method is considered undeniably fallible. Most post-mortem samples are negative for parasites and that makes an EPM test is undeniably fallible. Finding a sample set is impossible using natural infections. Not that someone involved in the industry doesn't have 20 samples, but you need enough serum/CSF from those cases for the life of the test to include as the reference standard positive.
There are sera collected from horses with natural infections out there. Often the serum samples described in multiple papers in literature came from the same set of horses. These samples are precious. Another game changer is that the only histologically confirmed natural cases of S. neurona from horses are from SAG 1 and SAG 5 strains. There are no reported cases of a SAG 6 infection in a horse that has been verified by histopathology or isolation! Don't forget, we need 20 of each serotype. Possibly the largest hurdle for a diagnostic test for EPM is the need for positive samples for each serotype of S. neurona.
The impossibility of a sample set from experimental EPM infections
Another way to get a sample set is to experimentally infect the horses. We infected clinically normal horses with a SAG 1 serotype. Overall, in our studies we induced 75 infections. To make a valid assay sample set we would have to isolate the organism from the nervous system from 20 of the SAG 1 challenged horses.
It might be possible to create enough infections with the SAG 5 serotype of S. neurona, but no one has experimentally infected a horse with the SAG 5 serotype yet. In the other model of EPM the researchers never isolated organisms from any of the horses they challenged. They didn't create any samples to validate an assay. They "verified" the infection by histopathology consistent with EPM, but no organisms were found. The assay requires that organisms are recovered, 20 times.
Disease would need to be created for the SAG 6 serotype. No one has isolated SAG 6 S. neurona ever from a horse, but we know that SAG 6 S. neurona does infect horses and causes natural disease. The SAG 6 organism needs to be isolated from a horse, grown in the lab and then used in the experimental model to create disease. Twenty times.
The requirement for validating a diagnostic test is 20 samples, that is 60 infections with parasites demonstrated by post mortem exam including two strains that no experimental infections have been accomplished.
The stage of disease and cross-reactivity
CVB also considers that the stage of disease that a test detects may differ from an intended target population. When the horse produces antibody after infection, about day 17, in most cases the horse eliminates the organism. An experimentally challenged horse that is diseased shows signs at 10 days after infection. That indicates there is a week when the test may read as a false negative. The validated test must select the animal in which the organism made it to the central nervous system. We don't know how long that takes. Or when the organism is eliminated. Antibody levels, serum or CSF antibodies, don't relate to the CNS-disease status. We published evidence that it is the duration of infection that relates to antibody levels, in serum and CSF, not the severity of clinical signs. Our data was obtained using three, ACVIM, equine neurologists and a very experienced statistician.
There is published evidence that Sarcocystis neurona evades immune detection by changing the expression of its surface antigen. The diagnostic work-around for this issue is using antigens that don't change over time. Yet, if the organism stops displaying the serotype specific antigens, then the test will suffer from cross-reactivity issues.
It is what it is
The current commercially available tests measure antibodies against S. neurona. The presence of antibodies indicate infection, not disease. These tests are valid, it is easy to make a sample set by vaccinating horses with recombinant antigens. All serotypes can be used to make the required standard positive and negatives. But detecting antibody is not detecting disease. We showed that antibody titer is a measure of the duration of infection. Serum antibodies can be quantitated.
The art of veterinary practice is using knowledge of the pathophysiology of disease in the horse, the biology of Sarcocystis in horses, and how to use the available tests.
Comments