A comparison of methods for the diagnosis of bovine subclinical mastitis within New Zealand dairy herds : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in veterinary clinical science at Massey University
During the 1986-1987 and 1987-1988 dairy seasons, milk samples were taken from cows within three New Zealand dairy herds. Individual quarter foremilk samples were taken using aseptic precautions, and whole udder, composite samples were obtained using a milk meter, at monthly intervals. The three herds differed with respect to herd size, milking shed design, use of teat spraying, and the incidence of mastitis, both past and present. The bacteriological status of each quarter was determined, and the ability of the following parameters to distinguish between infected and uninfected quarters, and between infected and uninfected cows was ascertained. (1) Somatic cell count (2) Sodium concentration (3) Potassium concentration (4) Electrical conductivity (5) pH (6) Lactose concentration (7) N-acetyl-β-d-glucosaminidase activity (8) α1-antitrypsin concentration With the exception of the antitrypsin concentration, the concentration of each parameter in the milk changed during the course of milk removal. The effect of this variation on the composition of the composite milk sample is discussed. The stage of lactation at which the sample was taken exerted a significant effect on the level of each parameter in quarter foremilk samples. Similar effects were observed with composite samples, although statistical significance was not reached in every instance. Both infected and uninfected udder quarters were affected by the stage of lactation. The age of the cow exerted a significant effect on the levels of a number of the parameters within quarter foremilk samples. The effect of age of the cow on the level of each parameter within composite milk samples was generally not significant. The bacteriological status of the udder quarter exerted a significant effect on the level of each parameter in at least one of the three herds. The effect of the bacteriological status on the pH and on the potassium concentration of the milk was smaller in degree than was the effect on the remaining parameters. The herds differed with respect to bacteriological findings. Herd A showed a lower incidence of infection than did herds B or C The incidence of infection with minor pathogens was highest within herd B, while the incidence of infection with major pathogens was highest within herd C. The incidence of infection tended to increase with the age of the cow. While each of the eight parameters showed high specificity, the sensitivity was generally lower. An exception to this finding was the somatic cell count which showed both high sensitivity and high specificity. The somatic cell count was able to correctly classify more than 75 % of quarter foremilk samples in each of the three herds, being more consistent in this respect than were any of the other parameters, the diagnostic abilities of which varied between herds. The addition of second parameter to a model containing the somatic cell count generally did not increase the diagnostic accuracy of the system. The threshold value was found to vary both between herds, and between stages of the lactation, this variation being greatest for the somatic cell count. Providing that the threshold is adjusted for the stage of lactation, the somatic cell count is able to accurately predict the infection status throughout the dairy season. The selected threshold should take into account the level of bacterial infection within the particular herd to which it will be applied.