PSXIII-40 Monitoring in-line milk progesterone profiles prior to first breeding to predict reproductive performance in Holstein cows.

Abstract

We investigated if milk progesterone (P4) profiles determined by an in-line milk P4 analysis system (Herd NavigatorTM, DeLaval Inc.) could be used to predict subsequent reproductive performance. Specifically, we evaluated associations of (1) commencement of luteal activity (C-LA), (2) first luteal phase length, (3) first P4 peak, and (4) number of luteal phases prior to first breeding, with pregnancy at first AI (P/1stAI) and cumulative pregnancy by 150 (P/150) d in milk (DIM). Milk P4 concentrations (ng/mL) were measured approximately every 2 d from ~21 DIM until pregnancy in 1,354 lactations (1,190 cows). Variations in P4 were used to define luteal function, such as C-LA, first luteal phase length, and pregnancy. All AI occurred within 5 d of P4 declining below 5 ng/mL. Variables were categorized into quartiles and data analyzed using GLIMMIX and PHREG procedures of SAS. First AI occurred at 70 ± 17 DIM. Overall P/1stAI and P/150 were 25.6 and 62.3%, respectively. Parameters associated with decreased probability of P/1stAI were first luteal phase length ≥ 17 vs. < 17 d (Odds Ratio [OR]: 0.13; P < 0.01), first P4 peak ≤ 17 vs. > 17 ng/mL (OR: 0.13; P < 0.01), and having one vs. two luteal phases prior to first AI (OR: 0.53; P < 0.01). The likelihood of P/150 was increased in cows that had C-LA before 50 than after 50 DIM (Hazard ratio: 0.64; P < 0.01), and in cows that had a first luteal phase ≥ 17 vs. < 17 d long (Hazard ratio: 0.58; P < 0.01). In summary, prolonged first luteal phase (≥ 17 d), low first P4 peak (≤ 17 ng/mL) and having only one luteal phase prior to first breeding reduced P/1stAI. Furthermore, a delayed C-LA (> 50 DIM) and a prolonged first luteal phase reduced likelihood of P/150.