Feasibility of sonography in estimating fetal weight of low birth weight babies

Sanghamithra Reddy, Shripad Hebbar, Lavanya Rai

Research output: Contribution to journalArticlepeer-review


Background: Currently available ultrasound-based fetal birth weight estimation methods have been designed for a group of neonates with wide birth weight range and hence are faced with increased error of margin. Whenever there is a need for delivering pregnant woman with small fetus, prior knowledge of approximate fetal weight is of utmost importance for neonatal survival, and an error in this process can result in significant morbidity/mortality to the newborn baby. This necessitates need for the establishment of new birth weight formula exclusively for this subset of fetuses. Objectives: To test the accuracy of established formulae in fetuses ≤ 2000 gm at birth in singleton pregnancies. To develop new formula for this group of small fetuses delivering in our institution with maximal accuracy and reliability and to prospectively validate this formula in subsequent set of pregnant cohort. Materials and methods: The current study was done in two phases: The first phase was a formula derivation phase wherein the four major parameters [biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL)] were evaluated from a set of 128 postpartum women who delivered a neonate weighing ≤ 2 kg within 1 week of ultrasound examination. Stepwise regression analysis using birth weight as dependent parameter and fetal biometric parameters as independent parameters was used to develop the best formula for estimating fetal weight at birth. In the second phase (formula validation phase), the newly derived formula was tested for its accuracy in 31 pregnant women who gave birth to neonates weighing ≤2 kg. Results: The new formula (log10 [BW] = 1.0131 + 0.0216 × HC + 0.0448 × AC + 0.2183 × FL + 0.0001 × BPD × AC – 0.0059 × AC × FL) was superior to established birth weight formulae. In the formula derivation group, the lowest mean ± standard deviation (SD) absolute error was 130 ± 91 gm and the lowest mean absolute percentage error was 9.8 ± 7% SD for the new formula and 61.7% of weight estimates fell within ± 10% of the actual weight at birth and this percentage further increased to 83.6 and 91.4% for error of margin of ±15 and ±20% respectively. When this formula was applied in the validation group, the absolute error in grams was 102 ± 115 and absolute percentage error was 7.4 ± 7; hence 77.4% fell within 10%, 80.6% fell within 15%, 90.3% fell within 20%. Further, in the validation group, mean ± SD of estimated birth weight was 1337 ± 406 gm, which was closest to actual birth weight (1328 ± 433 gm). Conclusion: Our new formula is likely to estimate birth weight in small fetuses (≤2 kg) with reasonable accuracy and reliability. When compared to available methods of ultrasound birth weight estimation, absolute error and absolute percentage error is least with our formula indicating a good fit.

Original languageEnglish
Pages (from-to)42-48
Number of pages7
JournalInternational Journal of Infertility and Fetal Medicine
Issue number2
Publication statusPublished - 01-01-2016

All Science Journal Classification (ASJC) codes

  • Reproductive Medicine


Dive into the research topics of 'Feasibility of sonography in estimating fetal weight of low birth weight babies'. Together they form a unique fingerprint.

Cite this