Quantifying the occurrence of early embryonic mortality on three equine breeding farms

This prospective field study was designed to describe the incidence of early embryonic mortality (EEM) and factors associated with the cause of EEM on three equine breeding farms in Ontario during the 1989 breeding season. Early embryonic mortality was defined as the loss of a single embryo during the first 40 days of pregnancy (day 0 = day of ovulation or last breeding). Pregnancy diagnoses and subsequent embryonic losses were observed by serial trans-rectal ultrasonography between days 12-20 (PD1) and 21-30 (PD2), and by trans-rectal ultrasonography or palpation per rectum between days 31-40 (PD3). Information on pregnancy status of a mare (or cycle) at 40 days after the last breeding was recorded when available. Nonpregnancy rates were calculated on a per cycle basis, to account for mares with no ultrasonic evidence of an embryo at the initial pregnancy examination. Embryonic mortality rates per cycle were calculated cumulatively (EMR((40))) for the entire 40 day embryonic period and during the specific time periods when a pregnancy diagnosis took place (EMR((PD1)), EMR((PD2)), EMR((PD3))). Embryonic mortality rates were also calculated on a per mare basis for mares experiencing EEM on either their first (EMR((f))) or any (EMR((a))) breeding cycle. Per cycle mare withdrawal rates were calculated cumulatively for the entire 40 day embryonic period (MWR((40))), and at each specific pregnancy diagnosis time period (MWR((PD1)), MWR((PD2)), MWR((PD3))) to account for those breeding cycles in which mares were not able to be observed for the entire forty days of the embryonic period. Records from a total of 699 mares involving 1014 breeding cycles were examined and analyzed. Per cycle risk rates for nonpregnancy (NP) were 36.4%, 45.0%, and 22.1%, for farms 1,2 and 3, respectively. Per cycle EMR((40)) ranged from 8-17%. Per cycle MWR((40)) ranged from 56.5-98.9%, indicative of a high rate of mare withdrawal from the study for the duration of the "embryonic" period. Significant differences (p < 0.05) in EMR((f)) and EMR((a)) per mare existed between two farms, indicative of a farm effect on EMR. Multivariable forward stepwise logistic regression analyses revealed that mares bred on foal heat were 1.9 times more likely than mares not bred on foal-heat to experience EEM (p = 0.008).