The influence of milk yield and milk composition on the diagnosis of Mycobacterium avium ssp. paratuberculosis (MAP) by milk ELISA in the context of the total IgG secretion patterns in milk throughout lactation and serum concentrations were investigated. A 2-yr trial was performed in which 1,410 dairy cows were sampled monthly and MAP milk ELISA status and milk yield and composition were determined. Data were analyzed by mixed model analysis. Milk yield was found to significantly influence ELISA results expressed as sample-to-positive (S/P) ratios. For each 5-kg increase in milk, the S/P ratio has to be multiplied by 0.89; therefore, high milk yield can change the MAP milk ELISA outcome of a cow in early infection from positive to negative. Parity influenced ELISA outcome significantly, indicating that cows with a parity >1 are more likely to be identified by milk testing. Also, herd was an important predictor, showing that herd prevalence influences the milk ELISA strongly. Other factors influencing the S/P ratios were protein concentration, somatic cell count, and days in milk. The IgG concentration and mass excreted per day were determined longitudinally in a subset of 41 cows of which samples and data of a complete lactation were available. Again, the IgG concentration in milk was mainly influenced by milk yield. The total IgG mass secreted per day in milk was found to be relatively constant, with a mean of 8.70 ± 5.38 g despite an increasing IgG concentration in serum at the same time. The variation of IgG concentration in milk can be mainly attributed to dilution through changes in milk yield. This supports the assumption that concentrations of MAP-specific antibodies are influenced by changes in milk yield similarly. In conclusion, we confirmed that antibody concentrations, and therefore MAP ELISA outcome, were influenced by milk yield, herd, and parity. To enhance performance, milk ELISA tests should either be performed in early or late lactation, when milk yield is low. From a management perspective, sampling should be done during early lactation before cows are bred again. Based on the slow progressive infection dynamics, only first-parity cows should be preferentially tested at the end of their first lactation to avoid false-negative results.