ABSTRACT
The goal of this study was to evaluate the sensitivity and specificity of two cowside tests for subclinical ketosis in dairy cows. The tests utilize milk and urine samples, respectively. One hundred and eighty-five cows, one to sixty days postpartum, were sampled for milk, urine, and blood. Subclinical ketosis was defined with serum beta-hydroxybutyrate measurements. The sensitivity and the specificity of both tests at different beta-hydroxybutyrate levels were estimated. When subclinical ketosis was defined at beta-hydroxybutyrate levels of 1.4 mmol/L and higher, the milk test had sensitivity of 90% and specificity of 96%. The urine test lacked specificity (values < 67%), but sensitivity was 100% at beta-hydroxybutyrate levels of 1.4 mmol/L upward. Both the milk and urine test can be used to monitor subclinical ketosis in a herd. Milk testing is preferred, because of the easy obtainability of milk combined with the overall better test characteristics.