OBJECTIVE: To derive a formula for calculating fetal weight in small for gestational age (SGA) fetuses and to determine prospectively whether the use of such a targeted formula reduces birth weight prediction errors. METHODS: Standard ultrasonic measurements were made in 159 SGA fetuses within 7 days of delivery. Three classes of fetal weight formulas (linear, quadratic, and cubic) were fitted to the data using stepwise regression analysis. Birth weight predictions using these three formulas were then compared prospectively with five previously reported formulas in 187 SGA fetuses. RESULTS: R2 was 0.97 for each of the three derived formulas. The 95% prediction intervals were comparable for the three formulas (eg, cubic model -11.6, 17.8%), and none were statistically superior to previous formulas. Each of the formulas evaluated prospectively had a systematic error and, with the exception of the present study's linear formula, all had percentage errors that varied systematically over the range of actual birth weights. CONCLUSION: Clinically useful birth weight predictions can be made in SGA fetuses, although no particular formula estimates birth weight significantly more accurately than any other.
OBJECTIVE: To derive a formula for calculating fetal weight in small for gestational age (SGA) fetuses and to determine prospectively whether the use of such a targeted formula reduces birth weight prediction errors. METHODS: Standard ultrasonic measurements were made in 159 SGA fetuses within 7 days of delivery. Three classes of fetal weight formulas (linear, quadratic, and cubic) were fitted to the data using stepwise regression analysis. Birth weight predictions using these three formulas were then compared prospectively with five previously reported formulas in 187 SGA fetuses. RESULTS: R2 was 0.97 for each of the three derived formulas. The 95% prediction intervals were comparable for the three formulas (eg, cubic model -11.6, 17.8%), and none were statistically superior to previous formulas. Each of the formulas evaluated prospectively had a systematic error and, with the exception of the present study's linear formula, all had percentage errors that varied systematically over the range of actual birth weights. CONCLUSION: Clinically useful birth weight predictions can be made in SGA fetuses, although no particular formula estimates birth weight significantly more accurately than any other.