Objectives: To determine optimal volume and type of milk samples for the Paracheck ™ ELISA, a test that was developed to assay serum or plasma for antibodies against Mycobacterium paratuberculosis.

Experimental design: Samples of milk and blood were collected from 12 cows that previously tested positive by the ELISA.Of the 12 cows, 11 tested positive by faecal culture and/or histopathology and one had no bacteriological or histopathological evidence of infection. In the first experiment, duplicate samples of serum, whole milk and skim milk the 12 cows were assayed for antibodies using a commercial bovine ELISA. Each replicate was diluted 1:20 in absorbing buffer and tested as recommended by the manufac­turer. In the second experiment, 3 sets of duplicate samples of whole milk from 5 cows with paratuberculosis and one cow with no evidence of infection were assayed for antibodies. One set was diluted 1:20 in absorbing buffer (25 μL sample/475 μL absorbing buffer), the second set was diluted 1:10 (50 μL sample/450 μL absorbing buffer) and the third set was diluted 1:5 (100 μL sample/400 μL absorbing buffer).

Results: In the first experiment, there was no significant difference (P<0.05) between the mean OD values in whole and skim milk. The average coefficient of variation (CV) between OD values in whole and skim milk was 4.8%. Although there was strong positive correlation between OD values in samples of serum and that measured in corresponding samples of whole milk and skim milk, only 7 of the 11 cows with paratuberculosis gave positive reactions in both the whole and skim milk. The OD values in samples of serum were significantly higher than that measured in corresponding samples of whole milk and skim milk.

In the second experiment, the mean OD values in whole milk samples diluted 1:20, 1:10 and 1:5 were 0.217, 0.340 and 0.549, respectively. The latter was similar to the mean OD in corresponding samples of serum. Of the 5 cows with paratuberculosis, all tested positive when the test was applied to serum and samples of milk diluted 1:5. In comparison, only one of the 5 cows gave a positive reaction when samples of milk diluted 1:20 were tested. The negative control showed a slight, negligible increase in OD values when larger volumes were tested.

The one cow with no evidence of infection gave consistent negative results throughout the testing.

Conclusions: The likelihood of detecting infected animals increases when the ELISA is applied to large volume samples of milk. The whole milk is a suitable sample as the differences between the OD values in samples of whole and skim milk are negligible and similar to normally expected well-to-well variation.