Early Biometric Lag in the Prediction of Small for Gestational Age Neonates and Preeclampsia

Abstract

Objectives An early fetal growth lag may be a marker of future complications. We sought to determine the utility of early biometric variables in predicting adverse pregnancy outcomes. Methods In this retrospective cohort study, the crown‐rump length at 11 to 14 weeks and the head circumference, biparietal diameter, abdominal circumference, femur length, humerus length, transverse cerebellar diameter, and estimated fetal weight at 18 to 24 weeks were converted to an estimated gestational age using published regression formulas. Sonographic fetal growth (difference between each biometric gestational age and the crown‐rump length gestational age) minus expected fetal growth (number of days elapsed between the two scans) yielded the biometric growth lag. These lags were tested as predictors of small for gestational age (SGA) neonates (≤10th percentile) and preeclampsia. Results A total of 245 patients were included. Thirty‐two (13.1%) delivered an SGA neonate, and 43 (17.6%) had the composite outcome. The head circumference, biparietal diameter, abdominal circumference, and estimated fetal weight lags were identified as significant predictors of SGA neonates after adjusted analyses (P < .05). The addition of either the estimated fetal weight or abdominal circumference lag to maternal characteristics alone significantly improved the performance of the predictive model, achieving areas under the curve of 0.72 and 0.74, respectively. No significant association was found between the biometric lag variables and the development of preeclampsia. Conclusions Routinely available biometric data can be used to improve the prediction of adverse outcomes such as SGA. These biometric lags should be considered in efforts to develop screening algorithms for adverse outcomes.

Two‐dimensional sonographic placental measurements in the prediction of small‐for‐gestational‐age infants

“…Therefore, MPD was considered the parameter of choice for inclusion in subsequent multivariable models. While both the MPT and MaxPT variables 18 Gestational age (weeks) performed similarly in the prediction of both SGA < 10 and SGA < 5, we chose the mean thickness as the variable of choice for multivariable modeling, since a falsely increased MaxPT value is more likely to be due to tangential imaging, while the mean value may be a more consistent parameter. Inclusion of both MPD and MPT into a combined model to predict SGA yielded a slightly higher AUC (SGA < 10, 0.6302; SGA < 5, 0.6224) than MPD alone (P = 0.06 and 0.08, respectively).…”

“…As our data were limited to those included in the birth log or the ultrasound database, we were unable to collect important demographic variables such as parity and ethnicity. Our prior work showed that some of these demographic variables significantly improve the accuracy of our biometric lag model. Furthermore, other clinical variables, such as medical co‐morbidities, maternal serum analytes and diagnosis of pre‐eclampsia, were not available, probably limiting the performance of our prediction models.…”

Slow fetal growth between first and early second trimester ultrasound scans and risk of small for gestational age (SGA) birth

“…Our findings support the hypothesis that IUGR and preeclampsia, that are presumed to share the same etiology, might be two different entities and IUGR can be present without clinical signs of gestational hypertension or preeclampsia. [ 20 , 21 , 31 – 33 ]…”

“…[ 16 ] In previous research on discrepancy in biparietal diameter (BPD) growth between the first and second trimester ultrasound scans, where an association with small-for-gestational-age (SGA) birth was observed, the focus was on overall risk of SGA, and risks of preterm or term SGA were not studied. [ 17 – 21 ]…”

External validation and clinical usefulness of first‐trimester prediction models for small‐ and large‐for‐gestational‐age infants: a prospective cohort study

“…The cross-checking of citation lists yielded three additional articles. We excluded ten articles for the following reasons: algorithm not available (n = 8), 35,[38][39][40]43,45,46,55 predictors not applicable in a high-income country (n = 1), 56 and model already published in another included article (n = 1). 34 The eight eligible articles described nine prediction models aimed at predicting any SGA (n = 6), 29,33,44,[57][58][59] preterm SGA (n = 2), 60,61 and late SGA (n = 1).…”