331
Receiver Operator Curve (ROC) characteristics for a pregnancy test based on Milk Progesterone (mP4) concentration before or one week after Timed AI (TAI)

Tuesday, March 17, 2015
Grand Ballroom - Posters (Community Choice Credit Union Convention Center)
Laura J Wilsdorf , University of Missouri, Columbia, MO
Scott E Poock , University of Missouri, Columbia, MO
Matthew C Lucy , University of Missouri, Columbia, MO
Abstract Text:

Timed AI programs have increased reproductive efficiency in dairy herds. Refinements to TAI have improved effectiveness but conception rates (CR) to TAI remain below 50%. The low TAI CR is partially explained by cows that either: 1) do not have a corpus luteum when PGFis administered; 2) fail to undergo luteolysis after PGF; or 3) fail to ovulate after GnRH and TAI. Two experiments were conducted to test the hypothesis that mP4 before or 1 week after TAI could be used to identify cows that will not become pregnant because of one of the criteria stated above.  In Exp. 1, Holstein cows (n=223) that were either first AI or greater were treated with the “Ovsynch56” protocol for TAI (GnRH, 7d, PGF, 56h, GnRH, 16h TAI). mP4 concentrations were measured -3, 0, and +7 d relative to TAI by using an automated analyzer (FT Multilyser; Förster Technik; Engen, Germany). Compared with nonpregnant (NP; n=141) cows, pregnant (P; n=82) cows had greater mP4 on d -3 (15.7+0.7 vs. 12.3+0.5 ng/mL; P<0.001) and d +7 (17.2+0.8 vs. 13.4+0.6 ng/mL; P<0.001). mP4 on d0 was similar for P and NP cows (8.5+0.3 ng/mL). When mP4 concentrations were used as a test to predict pregnancy, the area under the ROC was 0.64+0.04 for d -3 and 0.67+0.04 for d +7. Both ROC were different (P<0.05) from 0.5 meaning that mP4 around TAI was predictive of pregnancy. The ROC for d0 (0.55+0.04) did not differ from 0.5 (could not be used as a pregnancy test). In Exp. 2 Holstein and Guernsey cows (n=93) were subjected to Ovsynch56 for first TAI. mP4 concentrations were measured on d -4 (evening before PGF) and d -2 relative to TAI. Compared with NP (n=53), P (n=40) cows tended to have greater mP4 on d -4 (27.3+1.4 vs. 23.8+1.2 ng/mL; P<0.06). mP4 on d -2 was similar for P and NP cows (10.9+0.6 ng/mL). When mP4 concentrations were used as a test to predict pregnancy, the area under the ROC was 0.58+0.06 for d -4 and 0.54+0.04 for d -2. Neither ROC area differed from 0.5 (P>0.05). In summary, mP4 testing around TAI was predictive of pregnancy in Exp1 but not in Exp2. The timing of the test relative to PGF and TAI may be critical toward the development of a test to predict pregnancy outcome before or shortly after TAI.

Keywords: milk progesterone, pregnancy, timed AI