Milk samples from dairy cows provide a ready source of material for measuring antibody responses to Mycobacterium bovis antigens. In this study, we evaluated the IDEXX enzyme-linked immunosorbent assay (ELISA) for the measurement of antibody responses to M. bovis antigens MPB70 and MPB83 in milk samples from New Zealand cattle. Test sensitivities for individual milk and serum samples were assessed in samples collected from 44 M. bovis-infected cows, and test specificities were assessed in milk samples collected from 356 cows from tuberculosis (TB)-free herds. Milk vat samples were collected from 505 herds from regions with relatively high or low prevalences of infection. The ELISA had a sensitivity of 50% and a specificity of 97.5% for milk samples, and the test sensitivities for milk and serum samples were the same. Dilution of the positive test milk samples in milk from noninfected cows at 1/10, 1/20, and 1/50 dilutions reduced the proportions of positive responses to 13/21, 9/21, and 4/21, respectively. Small differences were observed in the ELISA responses of milk samples from individual TB-free cows collected at different times during lactation. No significant differences were detected in the ELISA responses of milk vat samples collected from infected and noninfected herds. This study shows that milk samples can be substituted for serum samples for screening individual cows for M. bovis infection, and pooling of milk samples from 10 to 20 animals can result in a reduction in the sensitivity by approximately 50%. However, screening of milk vat samples is unlikely to be useful in countries with low prevalences of M. bovis in cattle and large herd sizes.
Bovine tuberculosis (TB) caused by Mycobacterium bovis is a major economic problem in many countries, and control of this chronic disease in domestic livestock has often proved difficult. Although the disease has been eradicated from cattle in a number of countries through the implementation of a "test-andslaughter" control program, this strategy has been less effective in countries which have a wildlife reservoir of M. bovis infection and has been impractical in countries where the strategy is unaffordable or socially unacceptable (1). The test-and-slaughter programs have been based on screening of animals using the tuberculin skin test, which measures delayed hypersensitivity responses to purified protein derivative (PPD) prepared from M. bovis or which, in some countries, compares PPDs prepared from M. bovis and Mycobacterium avium. More recently, the whole-blood gamma interferon (IFN-␥) test has been used in conjunction with the skin test to improve specificities by retesting skin test reactor cattle or to improve sensitivities by retesting skin test-negative animals from chronically infected herds (2, 3). Although the IFN-␥ test has made a valuable contribution in improving the accuracy of diagnosing TB in cattle, additional types of assays are needed, particularly assays which are inexpensive and which can potentially be used on pooled samples from group...