Evaluation of the California mastitis test as a discriminant method to detect subclinical mastitis in ewes

The reliability of the California mastitis test (CMT) has been compared in a field study with another indirect method, somatic cell count (SCC), and with direct bacteriological cultures. A total of 1366 half-udder samples from ewes in mid-lactation were tested, using animals of three different breeds. Admitting a threshold value of 300 000 cells ml(-1) to detect subclinical mastitis, prevalence of subclinical mastitis was 43%, and 39.2% had bacterial infection. Accuracy of CMT was higher when compared with SCC, with a correlation of 0.82, 87% of correctly classified samples, 4% false-positives and a predictive value of positive results of 89.6%. According to bacteriological analysis, correctly classified samples were 78.4%, 10% false-positives, and 75.5% the predictive value of positive results. Periodical sampling was carried out in an experimental flock throughout the milking period; the number of false-positives classified with CMT increased towards the end of lactation.