During a 2-year study period, 252 animals from dairy herds infected with Mycobacterium avium subsp. paratuberculosis, and 119 animals from non-infected herds were subjected to repeated blood and faecal sampling. Animals were retrospectively grouped by infection status as infected, exposed (culture negative animals from infected herds), or non-infected animals, and by age, 12-23 months (1+ year), 24-35 months (2+ years), or 36 months and older (3+ years). Samples were collected for culture of faeces, assessment of interferon-gamma (IFN-gamma) secreted by M. paratuberculosis antigen stimulated whole-blood lymphocytes (IFN-gamma test), and measurement of antibody responses against M. paratuberculosis in serum and milk by an in-house absorbed ELISA. The IFN-gamma test diagnosed higher proportions of infected and exposed animals than the antibody ELISAs. The highest sensitivity of IFN-gamma test was in infected cattle of 2+ years of age. Receiver-operating characteristic (ROC) analyses supported the assumption that the IFN-gamma test had a better performance than antibody tests of animals of 1+ and 2+ years of age. However, for animals of 3+ years all tests performed equally well. Application of single sampling compared with repeated samplings showed better performance of the IFN-gamma test by repeated samplings, and the milk antibody ELISA in animals of 3+ years of age performed significantly better with repeated sampling compared with single sampling. In conclusion, the IFN-gamma test may be applied for screening of cattle of I and 2 years of age for exposure to M. paratuberculosis and the antibody ELISAs from 3 years of age.