Paratuberculosis impairs productivity of infected dairy cows because of reduced milk production and fertility and enhanced risk of culling. The magnitude of the milk yield depression in individual cows is influenced by factors such as parity, the stage of the disease and the choice of test used. The objectives of this case–control study were to substantiate the influence of the different levels of the within-herd prevalence (WHP) on individual milk yield of fecal culture (FC)-positive cows (FC+) compared with FC-negative herd-mates (FC−), and to estimate the magnitude of the deviation of the milk yield, milk components and somatic cell count (SCC) in an FC-based study. Of a total of 31 420 cows from 26 Thuringian dairy herds tested for paratuberculosis by FC, a subset of 1382 FC+ and 3245 FC− with milk recording data were selected as cases and controls, respectively. The FC− cows were matched for the same number and stage of lactation (±10 days in milk) as one FC+ from the same herd. Within a mixed model analysis using the fixed effects of Mycobacterium avium ssp. paratuberculosis (MAP) status, lactation number, days in milk, prevalence class of farm and the random effect of farm on milk yield per day (kg), the amount of fat and protein (mg/dl) and lactose (mg/dl) as well as the SCC (1000/ml) were measured. On the basis of least square means, FC+ cows had a lower test-day milk yield (27.7±0.6 kg) compared with FC− (29.0±0.6 kg), as well as a lower milk protein content and a slightly diminished lactose concentration. FC status was not associated with milk fat percentage or milk SCC. In FC+ cows, reduction in milk yield increased with increasing WHP. An interaction of FC status and farm was found for the test-day milk yield, and milk protein percentage, respectively. We conclude that the reduction in milk yield of FC+ cows compared with FC− herd-mates is significantly influenced by farm effects and depends on WHP class. Owners of MAP-positive dairy herds may benefit from the reduction in WHP not only by reducing number of infected individuals but also by diminishing the individual losses in milk production per infected cow, and therefore should establish control measures. [ABSTRACT FROM PUBLISHER]