Comparison of pregnancy diagnosis by milk progesterone on day 21 and day 24 postbreeding: field study in dairy cattle.

Milk samples collected at 21 and 24 d after breeding were used to diagnose cows as pregnant, questionable, or not pregnant with high, intermediate, or low progesterone in the milk. Total agreement of pregnancy diagnosis in cows by milk progesterone at 21 d postbreeding and diagnosis by return to estrus or palpation of reproductive organs was 85.8% (77.4% on pregnant progesterone diagnosis and 97.2% on not pregnant progesterone diagnosis). Total agreement of diagnosis by progesterone at 24 d postbreeding and diagnosis by estrus or palpation was 88.4% (83.5% on pregnant progesterone diagnosis and 95.3% on not pregnant progesterone diagnosis). Although total accuracy of cows diagnosed both pregnant and not pregnant by milk progesterone on d 24 postbreeding was greater than diagnosis on d 21 postbreeding when cows in estrus by d 24 were deleted from the data, other comparisons for effect of day on accuracy of diagnosis were not different. Total accuracy of diagnosis on d 24 was not increased by using milk progesterone on both d 21 and 24. Samples collected on the day of breeding indicated that 18 of 960 (1.9%) had intermediate or high progesterone. Fourteen (1.5%) additional cows had progesterone above levels expected for a cow in estrus.