In-line milk progesterone monitoring as a tool for precision reproductive management

Abstract

In-line milk progesterone (IMP4) monitoring (Herd Navigator, DeLaval) is a technology that automatically detects onset of cyclicity, estrus, and pregnancy. Sampling starts at 20 DIM and occurs on average every 2 d until pregnancy. Estrus is detected based on a decline in progesterone (P4) concentrations below a threshold, and pregnancy is assessed from 30 to 55 d after AI in cows without return to estrus. Here, we review the potential of IMP4 as a tool for reproductive management. In a series of observational studies with up to 158,961 IMP4 records from 4,353 AI events, we characterized predictors of pregnancy per AI (P/AI) and investigated IMP4 profiles in cows returning to estrus. Some of the predictors included prolonged luteal phase before AI and suboptimal P4 levels at different time points before and after AI. Over one-third of cows had at least one characteristic of P4 profile unfavorable to P/AI, but with low predictive abilities. Among nonpregnant cows, 5% returned to estrus by 17 d after AI, 64% between 18 and 24 d, 16% between 25 and 30 d, and 15% between 31 and 55 d. This represents 85% of cows that are not pregnant 55 d after AI returning to estrus before 30 d, when first pregnancy diagnosis occurs in many dairies. Monitoring IMP4 might be used to identify subgroups of cows with different predicted P/AI to develop selective breeding strategies or targeted interventions. It can also aid in identifying nonpregnant cows early for timely reinsemination.

Summary: In-line milk progesterone (IMP4) monitoring automatically samples and measures milk progesterone (P4) and detects the onset of cyclicity, estrus, and pregnancy status. Recent studies characterized markers of IMP4 before and after artificial insemination (AI) associated with pregnancy failure, such as prolonged luteal phase before AI and suboptimal P4 levels at different time points before and after AI. Among cows returning to estrus by 55 days after AI, IMP4 detected 69% of cows returning to estrus before 24 days, 16% between 25 and 30 days, and 15% between 31 and 55 days. This means that 85% of cows that were not pregnant at 55 days could be detected returning to estrus before 30 days, which is the typical time of first pregnancy diagnosis at many dairies. Monitoring IMP4 can be used to identify cows with different predicted probabilities of pregnancy per AI and to detect early return to estrus for timely reinsemination.

Figure 1

Frequency of samples automatically taken and analyzed by the in-line milk progesterone (IMP4) monitoring (Herd Navigator, DeLaval) biomodel. The blue symbols represent raw (unadjusted) milk progesterone concentrations, and the black/gray symbols represent adjusted concentrations. Sampling frequency varies according to the estimated stage of the estrous cycle (postpartum anestrous, estrous cycling, or potentially pregnant) and increases when an estrus event (decline in IMP4 below 5 ng/mL) is expected. Briefly, sampling occurs every 7 d (q7d) until onset of cyclicity, then every 2 to 3 d (q2–3d) during luteal phases and then once or twice daily (q<1d) when estrus is expected. If the cow receives AI at estrus and does no return to estrus until 24 d, sampling occurs every 1 to 2 d (q1–2d) until 32 d and every 5 d (q5d) until 57 d (Bruinjé et al., 2019).

Figure 2

Associations of IMP4 variables obtained before AI (panels A–C) and after AI (panels D–F) with the predicted pregnancy per AI (P/AI) in 4,353 AI events in 4 herds, as described elsewhere (Bruinjé et al., 2019). The variation in the associations was attributed to different herds (each represented in a different color), and other covariates in multivariable models. Tables show the comparisons between categories defined based on receiver operating characteristic cutpoints (panels A–C, E, F) or based on sampling frequency (panel D). P4 = progesterone; LP = luteal phase; peak P4 = maximum P4 in the LP preceding AI; nadir P4 = P4 at onset of estrus; time to onset of LP = interval to increase in P4 to ≥5 ng/mL after AI. Ref. = referent.

Figure 3

Distribution of time to return to estrus following 3,218 AI events in cows that did not result in a pregnancy by 55 d, from a total of 4,353 AI events, from 4 dairy herds in a dataset described in Bruinjé et al. (2019). Estrus events immediately preceding AI and at return to estrus were detected by IMP4 monitoring (Herd Navigator, DeLaval). The proportions above each category are of primiparous and multiparous cows, respectively.