Identification of nonpregnant beef cows based on CL size vs. luteal blood perfusion at 21 days after timed artificial insemination

The aim was to evaluate the efficiency of two different ultrasonographic systems, Doppler mode vs. Two-dimensional mode (B mode), to identify the pregnancy status of beef cows and heifers on day 21 (D21) after Timed Artificial Insemination (TAI). The experiment was performed on a commercial beef farm in central-west region of Brazil using 1895 Nelore heifers and cows. All females had ovulation synchronized for a TAI that was performed on D0. Twenty-one days after the TAI, all animals had their ovaries evaluated by ultrasound for pregnancy diagnosis based on the size of the corpus luteum (CL). Using B mode ultrasonography, females without a CL or with a CL ≤ 10 mm in diameter were considered nonpregnant, whereas females with a CL > 10 mm in diameter were considered potentially pregnant. After the B mode examination, the Doppler mode was turned on, and the CL was evaluated by the subjective percentage of blood perfusion in the total area of the CL. Using Doppler mode, females were considered nonpregnant if they had no CL or the CL had ≤25 % of the total area with detectable blood perfusion, whereas animals with >25 % blood perfusion in the CL were considered potentially pregnant. The results for each method (potentially pregnant or nonpregnant) were later compared with the gold standard technique, which was a pregnancy diagnosis on D33 after TAI using ultrasound with visualization of an embryonic heartbeat. The accuracy was determined using the 2 × 2 contingency table approach. The area under the curve using the receiver operating characteristic curve for Doppler mode and B mode were 0.929 and 0.902 (P < 0.01), respectively. There were almost no false negatives (designated non-pregnant but later pregnant at D33) with either technique (0.2 % vs. 0.3 %; P = 0.65 for Doppler mode vs. B mode, respectively). False positives (designated pregnant but non-pregnant on D33) were greater for B mode compared to Doppler (19.1 % vs. 14.0 %; P < 0.01). This resulted in Doppler mode having similar high values as B mode for Negative Predictive Value (99.9 vs. 99.6 %; P = 0.85) and Sensitivity (99.8 vs. 99.7 %; P = 0.86) but there were differences in Specificity (86 vs. 80.9 %; P < 0.01), Positive Predictive Value (88 vs. 84.3 %; P < 0.01), and Accuracy (93.0 vs. 90.4 %; P < 0.01). In conclusion, evaluation of CL blood perfusion by Doppler produced greater accuracy in the early identification of nonpregnant heifers and cows on D21 after TAI than measurement of CL diameter with B mode ultrasound; although both had over 90 % accuracy in identifying pregnant and nonpregnant females.