Evaluation of the Azotest® strip as recommended for the estimation of milk urea nitrogen concentrations in individual cow, milk line and bulk tank samples

The objectives of this study were to determine inter-reader agreement and to describe the accuracy of the Azotest(R) strip for estimating urea nitrogen concentrations in milk samples from individual cows (n = 214), milk lines (n = 41) and bulk tanks (n = 41). Samples were split, with one portion used for milk urea nitrogen (MUN) analysis using the Azotest(R) strip, and the second portion submitted to an accredited diagnostic laboratory for MUN analysis using wet chemistry analysis as the gold standard test. There was excellent inter-reader agreement (Kappa = 0.87) in this study for distinguishing between individual cow samples with either normal or high MUN values. However, the overall accuracy of the Azotest(R) strip was poor, with MUN results from the Azotest(R) strip being different than those from wet chemistry analysis (P < 0.01). There was poor agreement between the two test methods (Kappa = 0.12 - 0.14); average Azotest(R) results were approximately 5 mg/dl units higher than the wet chemistry analysis method for either individual cow, milk line or bulk tank samples. Forty-five percent of individual cow samples, and 45% of milk line and bulk tank samples were Correctly categorized as low, normal or high. The sensitivity and specificity of the Azotest(R) strip for detecting high MUN concentrations (vs normal or low) in individual cow samples was 98.7 and 33.1%, respectively, and in pooled milk line and bulk tank samples was 1001 and 21.1%, respectively. For this data set, the predictive value of a positive test result when looking for high MUN measures (vs normal or low) was only 33.1 to 35.7%. Thus, approximately two-thirds of samples identified as high by the test strip, actually had a normal MUN concentration. The sensitivity and specificity of the Azotest(R) strip for detecting low MUN concentrations, (vs normal or high) in individual cow samples was 0 and 100%, respectively. The test failed to correctly identify any truly low MUN samples as being low, and overestimated the MUN concentration in a high proportion of truly low and normal samples. Given this high degree of test inaccuracy in these data, we conclude that the Azotest(R) strip will not be beneficial to commercial dairy producers or dairy consultants. The authors recommend that producers wishing to monitor MUN submit milk samples to an accredited diagnostic laboratory for MUN analysis.