Aims (1) To identify which ultrasound parameters can be most accurately measured and which best reflect ovine fetal weight during late pregnancy. (2) To formulate regression equations for fetal weight estimation and to evaluate their accuracy.

Methods Singleton pregnancies of known gestational age were established using embryo transfer in 32 adolescent ewes who subsequently underwent nutritional manipulation to produce fetuses of variable size. Ultrasound measurements of biparietal diameter, abdominal circumference (AC), trunk diameter, renal volume (RV), femur length, tibia length and placental size were taken at 126–133 days gestation (term=145d) and compared with fetal/placental weight and physical measurements at late-gestation necropsy (n=19) or following spontaneous delivery (n=13).

Results Nutritional manipulation resulted in a wide range of necropsy/birth weight (1.72–6.26 kg). Of all parameters measured the AC and RV correlated most strongly with physical measurements (r=0.78–0.83), relevant fetal organ weights (r=0.75–0.80) and necropsy/birth weight (r=0.79–0.84). Regression equations derived from AC and RV performed best in the prediction of fetal weight with equivalent absolute and percentage mean prediction errors (396–656 g and 11.1–15.4%). Error using AC alone was significantly higher for the largest fetuses (p=0.036). In contrast error using RV (individually or in combination) was consistent across the entire range of fetal size.

Conclusions The ultrasound parameters of choice for the evaluation of fetal growth in late-gestation sheep pregnancy are the AC and RV. These may be used for fetal weight estimation with a similar degree of accuracy as the human.