Prediction of birth weight by ultrasound in the third trimester

To compare the accuracy of predicted birth weight by the gestation-adjusted projection method using ultrasonographic measurements obtained just before and at term.

The study group comprised patients with singleton pregnancies who underwent sonograms between 34.0 and 36.9 weeks' gestation (period 1) and at 37 weeks and beyond (period 2). The mean error in birth weight prediction, absolute birth weight error, and signed and absolute percent errors were compared with paired t tests. Thus, each patient served as her own control.

The study included 138 patients undergoing 276 sonograms. The mean absolute error of the predicted birth weight was smaller for period 1 than for period 2 (197 +/- 167 g compared with 235 +/- 209 g, P =.019). The mean absolute percent error was 6.2 +/- 5.2% for period 1 compared with 7.4 +/- 6.3% for period 2 (P =.019). These same trends were observed when fetuses with suspected growth abnormalities were examined separately. Averaging data from both gestational periods did not improve the prediction of birth weight.

Sonograms between 34.0 and 36. 9 weeks' gestation allow for more accurate prediction of birth weight than sonograms later in gestation. Though these differences are small and not clinically significant, this study indicates that serial sonograms in the late third trimester do not improve the ability to predict birth weight, even in abnormally grown fetuses. A single sonogram between 34 and 37 weeks' gestation is recommended for prediction of birth weight.