Pugh SJ, et al. - Researchers evaluated the accuracy of prediction of gestational age (GA) at birth and preterm birth classification using routinely collected anthropometry measures. In the unavailability of ultrasound measures, longitudinal FH measures alone could provide a reasonably accurate assessment of GA in low-risk, non-obese women.

Methods

• Experts conducted a prospective cohort study in United States on a total of 2,334 non-obese and 468 obese pregnant women.
• They determined the enrolment GA based on last menstrual period, confirmed by first-trimester ultrasound.
• Researchers measured the maternal anthropometry and fundal height (FH) by a standardised protocol at study visits.
• They additionally abstracted FH alone from medical charts.
• Neonatal anthropometry measurements were obtained at birth.
• Three predictor models were developed using longitudinal FH alone and with maternal and neonatal anthropometry to estimate GA at delivery,.
• In order to construct training (60%) and test (40%) sets for all predictors, repeatedly sampled observations were made.
• Longitudinal maternal anthropometry was incorporated in linear mixed models and a shared parameter model incorporated neonatal anthropometry.
• The accuracy of the model was assessed under varied scenarios.
• The main outcome measures were estimated GA at delivery.

Results

• As per data, the prediction error for various combinations of anthropometric measures ranged between 13.9 and 14.9 days.
• Findings suggested that GA was predicted by longitudinal FH alone within 14.9 days with relatively stable prediction errors across individual race/ethnicities [whites (13.9 days), blacks (15.1 days), Hispanics (15.5 days) and Asians (13.1 days)], and correctly identified 75% of preterm births.
• Results demonstrated this model to be robust to additional scenarios.