To evaluate the accuracy and usefulness of predicting birth weight by measuring fetal thigh circumference by ultrasound and to compare with other conventional methods like Johnson's and Hadlock's.

In 100 pregnant women, ultrasonic measurements of mid-thigh circumference, along with BPD, FL and AC were done within 48 hours before delivery. Birth weights were estimated by Johnson's clinical method, Hadlock's and Vintzileos’ method. Statistical analysis of various ultrasound birth weight formulae in different weight categories was done and compared with each other, and also with clinical method.

In the present study, Vintzileos’ method was found to be better than Johnson's and Hadlock's in predicting birth weight in the categories < 2500 gm, and 2500 to 3000 gm. Between 3000 and 3500 gm, it was better than Johnson's method, while the results were comparable to Hadlock's method. In the category > 3500 gm, all three methods were comparable to each other, however it could be because of the small sample size involved (n = 6).

Incorporating fetal thigh circumference measurements along with biparietal diameter, femur length and abdominal circumference, significantly improved the accuracy of birth weight estimations by ultrasound. There was a good correlation between ultrasound measurements and actual postnatal measurements of thigh circumference (r² = 0.71).