The predictivity of different ultrasonographic estimated weight formulas and their correlations with actual birth weight

Objective

To compare the accuracy of the eight different formulas for 2-dimensional ultrasonographic fetal weight estimations in our Labor Ward.

Methods 

This prospective cohort study was carried out at a single branch Zekai Tahir Burak Women's Health Education and Research Hospital, between October 2013 and March 2014. 1000 low pregnant women were enrolled into the stııdy. Pregnant women with maternal obesity, oligohydramnios, premature rupture of membranes and a cervical dilatation of more than 6 cm in active labor were considered as the factors that might dinterfere with measurements were excluded from the study. Fetal biometric measurements were only used if the birth occurred within 24 hours of the measurement. Only a single operator(İÖ) performed all measurements. 8 different models "Hadlock 1", "Hadlock 2", "Hadlock 3", "Hadlock 4", "Campbell", "Combs", "Ott", "Warsoff" formulas were used. In the evaluation of the accuracy of these formulas:
• Average Error: Estimated birth weight (EBW) - Actual birth weight (G) of the algebraic sum of average,
• Average Absolute Error: EBW-G average of the sum of the values
• Average Percentage(% ) Error: (EBW-G) / G x 100 of the algebraic sum of the average
• Average Absolute Error: The percentage of the actual birth weights corresponding to the absolute margin of error in IEFW-G) I / S 100 of the sum of the average of the formulas,
Statistical evaluation was performed using SPSS version 17

Results 

Of 1000 low risk pregnant women, biometric measurements of 914 were evaluated as birth occurred within 24 hours of measurements. In an analysis of eight different EBW formulas, a significant difference was detected among formulas (p <0.001), and each formula of EBW showed a significant correlation with the actual birth weight (p <0.001). Of the sonographic fetal weight estimation formulas, the lowest mean absolute percentage error was found to be present in the Hadlock formula 4 (0.97 ± 1.37). The narrowest confidence interval were also found to be in Hadlock 4.

Conclusion 

Correct estimation of fetal weight may improve maternal and fetal outcomes. Among the eight different formulas, the best results in terms of lowest mean absolute percentage error and the narrowest CI was detected in Hadlock 4 formula. Therefore, in order to obtain the best clinical results in the estimation of fetal weight, the formula determining the lowest absolute error and the lowest percentage absolute error should be determined for each population.

Keywords

Estimated fetal weight, ultrasound

	File/Dsecription
	Table 1 Distribution of the estimated mean birthweight +actual birthweight and p values according to formula
	Table 2 Correlation of the formulas according to percentage of absolute error