Use of on-farm milk progesterone information to predict fertility outcomes in dairy cows subjected to timed artificial insemination.

The objectives of this study were to evaluate the use of a qualitative on-farm milk progesterone test to predict non-pregnancy in dairy cows. Lactating Jersey cows (n = 752) were subjected to the 5-d Cosynch-72 protocol for timed artificial insemination (AI; d -8 GnRH, d -3 and -2 PGF2α, d 0 GnRH and timed AI). Milk was sampled on d -3, 0, 7, and 28 relative to timed AI, and progesterone concentrations were assessed using a lateral flow immunochromatographic test. Samples were classified into 3 groups indicative of high (hP4; test line not visible or lighter than reference), intermediate (iP4; test line similar to reference), and low (lP4; test line darker than reference) progesterone concentrations. Blood was sampled from a subset of cows (n = 50) on d -3, 0, 7, and 28 relative to timed AI, and plasma progesterone concentrations were determined by RIA. Cows were observed daily for signs of estrus based on removal of tail paint. Pregnancy was diagnosed by ultrasonography on d 34 and 62 after AI. Plasma progesterone concentrations across all time points were greater for hP4 (3.13 ± 0.20 ng/mL) followed by iP4 (1.12 ± 0.27 ng/mL) and lP4 (0.38 ± 0.23 ng/mL). Cows in lP4 on d -3 had lesser pregnancy per AI (P/AI) compared with iP4 and hP4 (17.4, 38.3, and 37.2%, respectively). For measurements performed on the day of AI (d 0), lP4 cows had greater P/AI compared with hP4 and iP4 (34.8, 0.0, and 15.6%, respectively), and the risk of pregnancy loss tended to be greater for iP4 compared with lP4. Cows in lP4 on d 7 after AI had lesser P/AI than those in iP4 and hP4 (12.0, 34.0, and 37.7%, respectively). Cows classified as lP4 on d 28 had the least P/AI on d 62 followed by iP4 and then hP4 (0.8, 9.2, and 59.4%, respectively) and were at the greatest risk for pregnancy loss (lP4 = 74.6%, iP4 = 8.4%, hP4 = 7.1%). Sensitivity and specificity to predict non-pregnancy on d 62 were 0.86 and 0.32 (d -3), 0.95 and 0.15 (d 0), 0.93 and 0.23 (d 7), and 0.99 and 0.53 (d 28), respectively. On-farm milk progesterone profiling using a lateral flow immunochromatographic test was able to identify cows without functional corpus luteum and to predict fertility outcomes following timed AI.