Paratuberculosis (Johne's disease), which is widely distributed throughout the world, is caused by Mycobacterium avium subspecies paratuberculosis (MAP). Diagnosis of subclinically infected cattle is challenging and is especially problematic in herds with low prevalence of MAP. The aim of this long-term study was the comparison of different diagnostic tests for MAP and specific antibodies in a herd with low prevalence of MAP. Three different commercially available serum-ELISA (Svanovir-ELISA, Svanova, Uppsala, Sweden; IDEXX-ELISA, IDEXX Laboratories, Maine, USA; Pourquier-ELISA, Institut Pourquier, Montpellier, France) and two milk ELISA (Svanovirm-ELISA Svanova, Uppsala, Sweden; Pourquier-ELISA, Institut Pourquier, Montpellier, France) were compared. Apart from these indirect diagnostic tests, two methods for the detection of the etiologic agent (bacteriologic culture and real-time PCR of faecal samples) were performed. In January 2005 the first and in April 2005 the second herd investigation of all animals older than 2 years (n=335) were carried out. Blood, milk and faecal samples were taken. From November 2005 until April 2006 follow up investigations were performed. For this purpose, blood-, milk- and faecal samples were monthly taken from 63 selected animals. The highest number of blood- and milk samples with a detectable antibody-level was found by the Svanovir-ELISA. There was a significant correlation between serum- and milk- Svanovir-ELISA results, whereas the agreement between ELISA and faecal culture/PCR was low. Significant correlations between Svanovir-serum-ELISA results and milk somatic cell counts could be registered. Moreover, there was significant agreement between IDEXX-serum-ELISA results with the age and number of lactations of the cows, as well as the mother's MAP-status.

Paratuberculosis (Johne's disease) in ruminants is caused by Mycobacterium avium subsp. paratuberculosis (MAP). Owing to the lack of accurate laboratory tests, diagnosis is challenging in subclinically infected cattle. To evaluate the long-term performance of serum ELISAs for the detection of paratuberculosis in a dairy herd with low MAP-prevalence, three investigations of all the cows and the consecutive testing of 33 cows suspected to be infected with MAP and 30 cows classified as MAP free were performed over a period of 22 months. Blood samples were tested by three commercial serum ELISAs, MAP shedding was detected by bacteriological culture and polymerase chain reaction (PCR). The ELISA results varied in a wide range in the herd investigations with 1.2% to 18.8% positive samples, the faecal samples were positive for MAP between 1.8% and 4.9% in the three herd investigations. Over the study period, ELISA-positive serum samples varied between 0.0% and 69.7% in MAP-suspicious and 0.0% and 17.6% in MAP-unsuspicious cows with a poor correlation between ELISAs and faecal shedding. The correlation coefficient of the optical density values of the three ELISAs varied between 0.348 and 0.61. Evidence of cow specific variations of residuals was found in all linear models. The linear mixed models showed relevant contribution of cow specific variation in explanation of the residual variances. They also showed significant effects of the explanatory ELISA, the group (MAP-suspicious or MAP-unsuspicious) and the time of sampling. It can be concluded that the choice of the laboratory test significantly influences the outcome of the testing for MAP and that none of the three ELISAs can be thoroughly recommended as single test for the early diagnosis of paratuberculosis in cattle. Test results should always be interpreted with caution to avoid erroneous decisions and the disappointment of those engaged in the abatement of paratuberculosis.