Milk samples are becoming more used as a diagnostic specimen for assessment of occurrence of antibodies to Mycobacterium avium subsp. paratuberculosis (MAP). This study assessed the effect of days in milk (DIM) and milk yield on testing positive in a commercial MAP specific milk antibody ELISA among 222,774 Danish Holstein cows. Results showed that odds of testing positive on 1-2 DIM were 9-27 times higher than the rest of lactation, where the chance of testing positive varied less. The reason is most likely a high concentration of non-specific antibodies in colostrum. Consequently, samples from the first couple of DIM should be excluded from MAP testing until further information on their significance is established. Milk yield also had a significant effect on odds of testing positive due to its diluting effect. Inclusion of milk yield in the interpretation of test results could improve the diagnostic value, resulting in more predictable patterns corresponding to progression of infection.