An evaluation of a progesterone-based diagnostic as an aid to re-insemination decisions in a seasonal, pasture-grazed dairy cow herd

摘要

Progesterone-based diagnostic tests are commercially available to predict pregnancy status 19 to 23 days following an insemination, thus providing decision-making support for artificial breeding. The study objective was to evaluate the likely performance of such a test in a pasture-grazed, seasonal dairy farm. Progesterone was determined in milk samples collected twice weekly from 553 cows at the Lincoln University Dairy Farm. Receiver operating characteristic analysis was applied to these data to determine sensitivity, false positive rate and the positive predictive value using a range of progesterone threshold values for detecting non-pregnant cows. The gold standard was non-pregnancy as measured by ultrasonography. At an optimised threshold of 4.1 ng/mL, the sensitivity, false positive rate and positive prediction rate were 78%, 2.6% and 96% for detecting non-pregnant cows on Days 19 to 23. A major source of inaccuracy was the 28 cows with elevated levels of progesterone in milk on Days 19 to 23after first insemination that were subsequently diagnosed non-pregnant. Another source of error was the inherent variance in progesterone concentration profiles among individuals. At a current cost of about $5 per test, this technology would have limited appeal for detecting non-pregnant cows in dairy systems where a satisfactory level of oestrus detection performance can be achieved.