Use of clinical parameters for differentiation of Gram-positive and Gram-negative mastitis in dairy cows vaccinated against lipopolysaccharide core antigens

Objective-To determine whether clinical parameters could be used to differentiate clinical mastitis (CM) caused by gram-positive bacteria from CM caused by gram-negative bacteria in dairy cows vaccinated against lipopolysaccharide core antigens. Design-Case series. Animals-143 episodes of CM in 86 dairy cows in a single herd. Procedure-Cows were examined at onset of CM, and 24 clinical parameters including rectal temperature, heart rate, rumen contraction rate, degree of dehydration, various udder and milk characteristics, lactation number, stage of lactation, and season of year were recorded, Milk production and milk constituent concentrations before onset of CM were obtained from Dairy Herd Improvement Association records. Values for cows with gram-negative CM were compared with values for cows with gram-positive CM. Logistic regression was used to identify important predictors of gram-negative CM. Results-64 (45%) CM episodes were caused by gram-negative bacteria and 79 (55%) were caused by gram-positive bacteria. Rumen contraction rate was significantly lower and milk protein percentage before onset of CM was significantly higher in cows with gram-negative, rather than gram-positive, CM. Logistic regression indicated that CM was more likely to have been caused by gram-negative bacteria if it developed during the summer, milk was watery, or rumen contraction rate was low. Sensitivity and specificity of the final regression model were 0.58 and 0.80, respectively. Predictive value of a positive result was 0.74 when proportion of CM episodes caused by gram-negative bacteria was assumed to be 50%. Clinical implications-Results suggest that clinical observations do not allow accurate prediction of CM pathogens and should not be the sole criteria for deciding whether cows with CM are treated with antibiotics.