Search for a command to run...
Gram-positive (GP) and gram-negative (GN) bacteria are the most common pathogens responsible for mastitis infections. Gram-negative infections typically elicit a more intense inflammatory response than GP infections, potentially influencing both cow behavior and milking performance before the onset of clinical mastitis (CM). Automatic milking systems (AMS) can detect early signs of CM through indicators like deviations in milk yield, electrical conductivity, milk color, and milking behavior. However, the relationship between specific pathogen types and changes in milking behavior has not been widely explored. The objective of this study was to investigate whether milking behavior and performance during the week preceding CM diagnosis are associated with the type of pathogen involved (i.e., GN, GP, mixed infection [MX], and no-growth [NG]), and how this association varies across different lactation stages. The study considered data from 49,899 milking events in 3,467 lactations from September 2023 to June 2024 in a commercial dairy in Colorado, United States. Cows were grouped in a freestall barn system with 62 automatic milking units and were allowed to visit the milking station at intervals based on their lactation stage. To facilitate comparisons among pathogen categories, the lactations were divided into 3 stages based on the heathy control group's average milk peak DIM: prepeak (1-72 DIM), postpeak (73-144 DIM), and late lactation (145 DIM to dry-off). Depending on the pathogen isolated from milk, CM cases were classified into 4 categories: GN, GP, MX infections, and NG. The analyses were conducted separately for each stage of lactation, and a healthy control group (HLT) was identified as a reference for the comparisons using the average DIM of each period (prepeak: 36 DIM, postpeak: 103 DIM, and late lactation: 243 DIM). Multivariate mixed linear and logistic regression models for repeated measures were fitted to calculate estimated marginal means and probabilities, respectively, including pathogen category as the main predictor and using parity and milk yield as covariates. To assess the effect of time before CM diagnosis on the dynamics of the variables in the study, data were analyzed at 3 time points: -7, -4, and 0 d relative to the CM diagnosis. A total of 431 quarter milk samples were submitted for bacteriological analysis from cows affected by CM, resulting in 50.0%, 23.4%, 16.2%, and 10.4% GN, GP, NG, and MX infections, respectively. Overall, cows with CM caused by GN and GP pathogens showed significant alterations in milking behavior and performance, which in some cases were evident 7 d before diagnosis. This study provides novel insights into how specific mastitis-causing pathogen groups affect cow milking behavior and performance in AMS.