Ultrasonographic estimation of fetal weight: development of new model and assessment of performance of previous models

Routine ultrasound at 32 vs 36 weeks' gestation: prediction of small-for-gestational-age neonates

OBJECTIVE

To evaluate and compare the performance of routine ultrasonographic estimated fetal weight (EFW) and fetal abdominal circumference (AC) at 31 + 0 to 33 + 6 and 35 + 0 to 36 + 6 weeks' gestation in the prediction of a small-for-gestational-age (SGA) neonate.

METHODS

This was a prospective study of 21 989 singleton pregnancies undergoing routine ultrasound examination at 31 + 0 to 33 + 6 weeks' gestation and 45 847 undergoing routine ultrasound examination at 35 + 0 to 36 + 6 weeks' gestation. In each case, the estimated fetal weight (EFW) from measurements of fetal head circumference, AC and femur length was calculated using the Hadlock formula and expressed as a percentile according to The Fetal Medicine Foundation fetal and neonatal population weight charts. The same charts were used for defining a SGA neonate with birth weight < 10^th and < 3^rd percentiles. For each gestational-age window, the screen-positive and detection rates, at different EFW percentile cut-offs between the 10^th and 50^th percentiles, were calculated for prediction of delivery of a SGA neonate with birth weight < 10^th and < 3^rd percentiles within 2 weeks and at any stage after assessment. The areas under the receiver-operating characteristics curves (AUC) in screening for a SGA neonate by EFW and AC at 31 + 0 to 33 + 6 and at 35 + 0 to 36 + 6 weeks' gestation were compared.

RESULTS

First, the AUCs in screening by EFW for a SGA neonate with birth weight < 10^th and < 3^rd percentiles delivered within 2 weeks and at any stage after screening at 35 + 0 to 36 + 6 weeks' gestation were significantly higher than those at 31 + 0 to 33 + 6 weeks (P < 0.001). Second, at both 35 + 0 to 36 + 6 and 31 + 0 to 33 + 6 weeks' gestation, the predictive performance for a SGA neonate with birth weight < 10^th and < 3^rd percentiles born at any stage after screening was significantly higher using EFW Z-score than AC Z-score. Similarly, at 35 + 0 to 36 + 6 weeks, but not at 31 + 0 to 33 + 6 weeks, the predictive performance for a SGA neonate with birth weight < 10^th and < 3^rd percentiles born within 2 weeks after screening was significantly higher using EFW Z-score than AC Z-score. Third, screening by EFW < 10^th percentile at 35 + 0 to 36 + 6 weeks' gestation predicted 70% and 84% of neonates with birth weight < 10^th and < 3^rd percentiles, respectively, born within 2 weeks after assessment, and the respective values for a neonate born at any stage after assessment were 46% and 65%. Fourth, prediction of > 85% of SGA neonates with birth weight < 10^th percentile born at any stage after screening at 35 + 0 to 36 + 6 weeks' gestation requires use of EFW < 40^th percentile. Screening at this percentile cut-off predicted 95% and 99% of neonates with birth weight < 10^th and < 3^rd percentiles, respectively, born within 2 weeks after assessment, and the respective values for a neonate born at any stage after assessment were 87% and 94%.

CONCLUSIONS

The predictive performance for a SGA neonate of routine ultrasonographic examination during the third trimester is higher if, first, the scan is carried out at 35 + 0 to 36 + 6 weeks' gestation than at 31 + 0 to 33 + 6 weeks, second, the method of screening is EFW than fetal AC, third, the outcome measure is birth weight < 3^rd than < 10^th percentile, and, fourth, if delivery occurs within 2 weeks than at any stage after assessment. Prediction of a SGA neonate by EFW < 10^th percentile is modest and prediction of > 85% of cases at 35 + 0 to 36 + 6 weeks' gestation necessitates use of EFW < 40^th percentile. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.

Magnetic resonance imaging for prenatal estimation of birthweight in pregnancy: review of available data, techniques, and future perspectives

Fetuses at the extremes of growth abnormalities carry a risk of perinatal morbidity and death. Their identification traditionally is done by 2-dimensional ultrasound imaging, the performance of which is not always optimal. Magnetic resonance imaging superbly depicts fetal anatomy and anomalies and has contributed largely to the evaluation of high-risk pregnancies. In 1994, magnetic resonance imaging was introduced for the estimation of fetal weight, which is done by measuring the fetal body volume and converting it through a formula to fetal weight. Approximately 10 studies have shown that magnetic resonance imaging is more accurate than 2-dimensional ultrasound imaging in the estimation of fetal weight. Yet, despite its promise, the magnetic resonance imaging technique currently is not implemented clinically. Over the last 5 years, this technique has evolved quite rapidly. Here, we review the literature data, provide details of the various measurement techniques and formulas, consider the application of the magnetic resonance imaging technique in specific populations such as patients with diabetes mellitus and twin pregnancies, and conclude with what we believe could be the future perspectives and clinical application of this challenging technique. The estimation of fetal weight by ultrasound imaging is based mainly on an algorithm that takes into account the measurement of biparietal diameter, head circumference, abdominal circumference, and femur length. The estimation of fetal weight by magnetic resonance imaging is based on one of the 2 formulas: (1) magnetic resonance imaging-the estimation of fetal weight (in kilograms)=1.031×fetal body volume (in liters)+0.12 or (2) magnetic resonance imaging-the estimation of fetal weight (in grams)=1.2083×fetal body volume (in milliliters)ˆ0.9815. Comparison of these 2 formulas for the detection of large-for-gestational age neonates showed similar performance for preterm (P=.479) and for term fetuses (P=1.000). Literature data show that the estimation of fetal weight with magnetic resonance imaging carries a mean or median relative error of 2.6 up to 3.7% when measurements were performed at <1 week from delivery; whereas for the same fetuses, the relative error at 2-dimensional ultrasound imaging varied between 6.3% and 11.4%. Further, in a series of 270 fetuses who were evaluated within 48 hours from birth and for a fixed false-positive rate of 10%, magnetic resonance imaging detected 98% of large-for-gestational age neonates (≥95th percentile for gestation) compared with 67% with ultrasound imaging estimates. For the same series, magnetic resonance imaging applied to the detection of small-for-gestational age neonates ≤10th percentile for gestation, for a fixed 10% false-positive rate, reached a detection rate of 100%, compared with only 78% for ultrasound imaging. Planimetric measurement has been 1 of the main limitations of magnetic resonance imaging for the estimation of fetal weight. Software programs that allow semiautomatic segmentation of the fetus are available from imaging manufacturers or are self-developed. We have shown that all of them perform equally well for the prediction of large-for-gestational age neonates, with the advantage of the semiautomatic methods being less time-consuming. Although many challenges remain for this technique to be generalized, a 2-step strategy after the selection of a group who are at high risk of the extremes of growth abnormalities is the most likely scenario. Results of ongoing studies are awaited (ClinicalTrials.gov Identifier # NCT02713568).

Prediction of small-for-gestational-age neonates at 35-37 weeks' gestation: contribution of maternal factors and growth velocity between 32 and 36 weeks

OBJECTIVE

To assess the additive value of fetal growth velocity between 32 and 36 weeks' gestation to the performance of ultrasonographic estimated fetal weight (EFW) at 35 + 0 to 36 + 6 weeks' gestation for prediction of delivery of a small-for-gestational-age (SGA) neonate and adverse perinatal outcome.

METHODS

This was a prospective study of 14 497 singleton pregnancies undergoing routine ultrasound examination at 30 + 0 to 34 + 6 and at 35 + 0 to 36 + 6 weeks' gestation. Multivariable logistic regression analysis was used to determine whether addition of growth velocity, defined as the difference in EFW Z-score or abdominal circumference (AC) Z-score between the early and late third-trimester scans divided by the time interval between the scans, improved the performance of EFW Z-score at 35 + 0 to 36 + 6 weeks in the prediction of, first, delivery of a SGA neonate with birth weight < 10^th and < 3^rd percentiles within 2 weeks and at any stage after assessment and, second, a composite of adverse perinatal outcome, defined as stillbirth, neonatal death or admission to the neonatal unit for ≥ 48 h.

RESULTS

Multivariable logistic regression analysis demonstrated that significant contributors to the prediction of a SGA neonate were EFW Z-score at 35 + 0 to 36 + 6 weeks' gestation, fetal growth velocity, by either AC Z-score or EFW Z-score, and maternal risk factors. The area under the receiver-operating characteristics curve (AUC) and detection rate (DR), at a 10% screen-positive rate, for prediction of a SGA neonate < 10^th percentile born within 2 weeks after assessment achieved by EFW Z-score at 35 + 0 to 36 + 6 weeks (AUC, 0.938 (95% CI, 0.928-0.947); DR, 80.7% (95% CI, 77.6-83.9%)) were not significantly improved by addition of EFW growth velocity and maternal risk factors (AUC, 0.941 (95% CI, 0.932-0.950); P = 0.061; DR, 82.5% (95% CI, 79.4-85.3%)). Similar results were obtained when growth velocity was defined by AC rather than EFW. Similarly, there was no significant improvement in the AUC and DR, at a 10% screen-positive rate, for prediction of a SGA neonate < 10^th percentile born at any stage after assessment or a SGA neonate < 3^rd percentile born within 2 weeks or at any stage after assessment, achieved by EFW Z-score at 35 + 0 to 36 + 6 weeks by addition of maternal factors and either EFW growth velocity or AC growth velocity. Multivariable logistic regression analysis demonstrated that the only significant contributor to adverse perinatal outcome was maternal risk factors. Multivariable logistic regression analysis in the group with EFW < 10^th percentile demonstrated that significant contribution to prediction of delivery of a neonate with birth weight < 10^th and < 3^rd percentiles and adverse perinatal outcome was provided by EFW Z-score at 35 + 0 to 36 + 6 weeks, but not by AC growth velocity < 1^st decile.

CONCLUSION

The predictive performance of EFW at 35 + 0 to 36 + 6 weeks' gestation for delivery of a SGA neonate and adverse perinatal outcome is not improved by addition of estimated growth velocity between 32 and 36 weeks' gestation. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.

Prediction of small for gestational age neonates: screening by maternal factors, fetal biometry, and biomarkers at 35-37 weeks' gestation

BACKGROUND

Small for gestational age (SGA) neonates are at increased risk for perinatal mortality and morbidity; however, the risks can be substantially reduced if the condition is identified prenatally, because in such cases close monitoring and appropriate timing of delivery and prompt neonatal care can be undertaken. The traditional approach of identifying pregnancies with SGA fetuses is maternal abdominal palpation and serial measurements of symphysial-fundal height, but the detection rate of this approach is less than 30%. A higher performance of screening for SGA is achieved by sonographic fetal biometry during the third trimester; screening at 30-34 weeks' gestation identifies about 80% of SGA neonates delivering preterm but only 50% of those delivering at term, at a screen-positive rate of 10%. There is some evidence that routine ultrasound examination at 36 weeks' gestation is more effective than that at 32 weeks in predicting birth of SGA neonates.

OBJECTIVE

To investigate the potential value of maternal characteristics and medical history, sonographically estimated fetal weight (EFW) and biomarkers of impaired placentation at 35⁺⁰- 36⁺⁶ weeks' gestation in the prediction of delivery of SGA neonates.

MATERIALS AND METHODS

A dataset of 19,209 singleton pregnancies undergoing screening at 35⁺⁰-36⁺⁶ weeks' gestation was divided into a training set and a validation set. The training dataset was used to develop models from multivariable logistic regression analysis to determine whether the addition of uterine artery pulsatility index (UtA-PI), umbilical artery PI (UA-PI), fetal middle cerebral artery PI (MCA-PI), maternal serum placental growth factor (PlGF) and soluble fms-like tyrosine kinase-1 (sFLT) would improve the performance of maternal factors and EFW in the prediction of delivery of SGA neonates. The models were then tested in the validation dataset to assess performance of screening.

RESULTS

First, in the training dataset, in the SGA group, compared to those with birthweight in ≥10th percentile, the median multiple of the median (MoM) values of PlGF and MCA-PI were reduced, whereas UtA-PI, UA-PI, and sFLT were increased. Second, multivariable regression analysis demonstrated that in the prediction of SGA in <10th percentile there were significant contributions from maternal factors, EFW Z-score, UtA-PI MoM, MCA-PI MoM, and PlGF MoM. Third, in the validation dataset, prediction of 90% of SGA neonates delivering within 2 weeks of assessment was achieved by a screen-positive rate of 67% (95% confidence interval [CI], 64-70%) in screening by maternal factors, 23% (95% CI, 20-26%) by maternal factors, and EFW and 21% (95% CI, 19-24%) by the addition of biomarkers. Fourth, prediction of 90% of SGA neonates delivering at any stage after assessment was achieved by a screen-positive rate of 66% (95% CI, 65-67%) in screening by maternal factors, 32% (95% CI, 31-33%) by maternal factors and EFW and 30% (95% CI, 29-31%) by the addition of biomarkers.

CONCLUSION

The addition of biomarkers of impaired placentation only marginally improves the predictive performance for delivery of SGA neonates achieved by maternal factors and fetal biometry at 35⁺⁰-36⁺⁶ weeks' gestation.

Fetal weight estimation at term - ultrasound versus clinical examination with Leopold's manoeuvres: a prospective blinded observational study

BACKGROUND

Fetal weight estimation is of key importance in the decision-making process for obstetric planning and management. The literature is inconsistent on the accuracy of measurements with either ultrasound or clinical examination, known as Leopold's manoeuvres, shortly before term. Maternal BMI is a confounding factor because it is associated with both the fetal weight and the accuracy of fetal weight estimation. The aim of our study was to compare the accuracy of fetal weight estimation performed with ultrasound and with clinical examination with respect to BMI.

METHODS

In this prospective blinded observational study we investigated the accuracy of clinical examination as compared to ultrasound measurement in fetal weight estimation, taking the actual birth weight as the gold standard. In a cohort of all consecutive patients who presented in our department from January 2016 to May 2017 to register for delivery at ≥37 weeks, examination was done by ultrasound and Leopold's manoeuvres to estimate fetal weight. All examiners (midwives and physicians) had about the same level of professional experience. The primary aim was to compare overall absolute error, overall absolute percent error, absolute percent error > 10% and absolute percent error > 20% for weight estimation by ultrasound and by means of Leopold's manoeuvres versus the actual birth weight as the given gold standard, namely separately for normal weight and for overweight pregnant women.

RESULTS

Five hundred forty-three patients were included in the data analysis. The accuracy of fetal weight estimation was significantly better with ultrasound than with Leopold's manoeuvres in all absolute error calculations made in overweight pregnant women. For all error calculations performed in normal weight pregnant women, no statistically significant difference was seen in the accuracy of fetal weight estimation between ultrasound and Leopold's manoeuvres.

CONCLUSIONS

Data from our prospective blinded observational study show a significantly better accuracy of ultrasound for fetal weight estimation in overweight pregnant women only as compared to Leopold's manoeuvres with a significant difference in absolute error. We did not observe significantly better accuracy of ultrasound as compared to Leopold's manoeuvres in normal weight women. Further research is needed to analyse the situation in normal weight women.

Prediction of small-for-gestational-age neonates at 35-37 weeks' gestation: contribution of maternal factors and growth velocity between 20 and 36 weeks

OBJECTIVES

To evaluate the performance of ultrasonographic estimated fetal weight (EFW) at 35 + 0 to 36 + 6 weeks' gestation in the prediction of delivery of a small-for-gestational-age (SGA) neonate and assess the additive value of, first, maternal risk factors and, second, fetal growth velocity between 20 and 36 weeks' gestation in improving such prediction.

METHODS

This was a prospective study of 44 043 singleton pregnancies undergoing routine ultrasound examination at 19 + 0 to 23 + 6 and at 35 + 0 to 36 + 6 weeks' gestation. Multivariable logistic regression analysis was used to determine whether addition of maternal risk factors and growth velocity, the latter defined as the difference in EFW Z-score or fetal abdominal circumference (AC) Z-score between the third- and second-trimester scans divided by the time interval between the scans, improved the performance of EFW Z-score at 35 + 0 to 36 + 6 weeks in the prediction of delivery of a SGA neonate with birth weight < 10^th and < 3^rd percentiles within 2 weeks and at any stage after assessment.

RESULTS

Screening by EFW Z-score at 35 + 0 to 36 + 6 weeks' gestation predicted 63.4% (95% CI, 62.0-64.7%) of neonates with birth weight < 10^th percentile and 74.2% (95% CI, 72.2-76.1%) of neonates with birth weight < 3^rd percentile born at any stage after assessment, at a screen-positive rate of 10%. The respective values for SGA neonates born within 2 weeks after assessment were 76.8% (95% CI, 74.4-79.0%) and 81.3% (95% CI, 78.2-84.0%). For a desired 90% detection rate of SGA neonate delivered at any stage after assessment, the necessary screen-positive rate would be 33.7% for SGA < 10^th percentile and 24.4% for SGA < 3^rd percentile. Multivariable logistic regression analysis demonstrated that, in the prediction of a SGA neonate with birth weight < 10^th and < 3^rd percentiles, there was a significant contribution from EFW Z-score at 35 + 0 to 36 + 6 weeks' gestation, maternal risk factors and AC growth velocity, but not EFW growth velocity. However, the area under the receiver-operating characteristics curve for prediction of delivery of a SGA neonate by screening with maternal risk factors and EFW Z-score was not improved by addition of AC growth velocity.

CONCLUSION

Screening for SGA neonates by EFW at 35 + 0 to 36 + 6 weeks' gestation and use of the 10^th percentile as the cut-off predicts 63% of affected neonates. Prediction of 90% of SGA neonates necessitates classification of about 35% of the population as being screen positive. The predictive performance of EFW is not improved by addition of estimated growth velocity between the second and third trimesters of pregnancy. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.

A Comparison of Prediction of Adverse Perinatal Outcomes between Hadlock and INTERGROWTH-21st Standards at the Third Trimester

Little is known about the clinical value of the Hadlock and INTERGROWTH-21^st EFW standards for predicting adverse perinatal outcomes (APOs) in the third trimester. The purpose of this study was to study the association between low estimated fetal weight percentile (EFWc) in the third trimester and the risk of APOs and compare predictions of APOs between Hadlock and INTERGROWTH-21^st EFW standards. A prospective cohort of 690 singleton pregnancies with ultrasonography performed in the third trimester between March 2015 and March 2016 in China was conducted. EFW and the corresponding EFWc were measured using the Hadlock and INTERGROWTH-21^st standards, respectively. Cox proportional hazard models were used to assess the relationship between low EFWc (i.e., <5 percentile, P5) and the risk of APOs. Compared with fetuses with ≥P5 of the EFWc, fetuses with <P5 of the EFWc were much more likely to have an APO, with adjusted hazard ratios of 35.0 (95% confidence interval, 13.9-88.5) and 17.5 (7.7-39.6) for the Hadlock and INTERGROWTH standards, respectively. The Hadlock-EFWc had a higher predictive accuracy for APOs than the INTERGROWTH-EFWc, with area under the receiver operating characteristic curve of 0.94 (0.92-0.95) and 0.90 (0.87-0.92), respectively (P=0.007). The cutoff value for the INTERGROWTH-EFWc was percentile 11.61 with a sensitivity and specificity of 87.9% and 80.5%, respectively. For the Hadlock-EFWc, the corresponding sensitivity and specificity were 93.9% and 81.2%, with a cutoff value of percentile 8.65. Fetuses with low EFWc (i.e., <P5) were associated with an increased risk of APOs. APOs were more accurately predicted when EFWc was measured by the Hadlock standard than by the INTERGROWTH-21^st standard.

First-trimester and combined first- and second-trimester prediction of small-for-gestational age and late fetal growth restriction

OBJECTIVE

To develop a first-trimester or combined first- and second-trimester screening algorithm for the prediction of small-for-gestational age (SGA) and late fetal growth restriction (FGR).

METHODS

This was a retrospective study of women with singleton pregnancy, who underwent routine first-, second- and third-trimester ultrasound assessment. Late FGR was defined, at ≥ 32 weeks' gestation in the absence of congenital anomalies, as either (i) estimated fetal weight (EFW) or birth weight (BW) < 3^rd centile, or (ii) EFW < 10^th centile and either uterine artery mean pulsatility index (UtA-PI) > 95^th centile or cerebroplacental ratio (CPR) < 5^th centile. Neonates with BW < 10^th centile, regardless of prenatal parameters, were defined as SGA. The predictive effectiveness of maternal and first- and second-trimester factors was tested using logistic regression and receiver-operating characteristics curve analyses.

RESULTS

A total of 3520 fetuses were included (late FGR, n = 109 (3.1%); SGA, n = 292 (8.3%)). Of the late FGR cases, 56 (1.6%) fulfilled the antenatal criteria (EFW < 3^rd centile or EFW < 10^th centile plus abnormal UtA-PI or CPR) and were defined as prenatally detected late FGR. A first-trimester screening model (comprising conception method, smoking status, maternal height, pregnancy-associated plasma protein-A (PAPP-A) and UtA-PI) could predict 50.0% of the prenatally diagnosed and 36.7% of the overall late FGR fetuses for a 10% false-positive rate (FPR). A model combining first- and second-trimester screening parameters (conception method, smoking status, PAPP-A, second- trimester EFW, head circumference/abdominal circumference ratio and UtA-PI) could predict 78.6% of the prenatally detected, and 59.6% of the overall late FGR fetuses, for a 10% FPR (area under the curve 0.901 (95% CI, 0.856-0.947) and 0.855 (95% CI, 0.818-0.891), respectively). The prediction of SGA was suboptimal for both first-trimester and combined screening.

CONCLUSIONS

A simple model combining maternal and first- and second-trimester predictors can detect 60% of fetuses that will develop late FGR, and 79% of those fetuses that will be classified prenatally as late FGR, for a 10% FPR. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.

Comparison of the Hadlock and INTERGROWTH formulas for calculating estimated fetal weight in a preterm population in France

BACKGROUND

Accurate estimation of fetal weight is needed for growth monitoring and decision-making in obstetrics; the INTERGROWTH project developed an estimated fetal weight formula to construct new intrauterine growth standards.

OBJECTIVE

We sought to compare the accuracy of the Hadlock and INTERGROWTH formulas for the estimation of fetal weight among preterm infants.

STUDY DESIGN

Using the EPIPAGE 2 population-based study of births between 22-34 weeks of gestation, we included 578 nonanomalous singleton fetuses with an ultrasound-to-delivery interval <2 days. We used abdominal circumference, head circumference, and femur length to calculate estimated fetal weight with Hadlock formula and abdominal and head circumferences to calculate estimated fetal weight according to INTERGROWTH. The mean percentage errors and the proportions of estimated fetal weight measures within ±10% of birthweight were compared between the 2 methods.

RESULTS

Mean (SD) gestational age and birthweight were 29.1 (SD 2.7) weeks and 1219 (SD 489) g. Mean (SD) percentage errors for Hadlock and INTERGROWTH were significantly different: -0.7 (SD 10.1) and -3.5 (SD 11.6), respectively (P < .001), and more infants were classified within ±10% of their birthweight with Hadlock compared to INTERGROWTH (68.7% vs 57.8%, P < .001). The INTERGROWTH formula overestimated birthweight at 22-23 weeks compared to Hadlock [mean errors of 18.8 (SD 13.6) vs 5.5 (SD 10.2)] and underestimated birthweight >28 weeks: at 29-31 weeks, mean errors were -5.8 (SD 10.9) for INTERGROWTH and -0.6 (SD 10.4) for Hadlock.

CONCLUSION

Hadlock estimated fetal weight formula was more accurate than INTERGROWTH formula for fetuses delivered between 22-34 weeks of gestation. Our results support continued use of Hadlock formula in France and raise questions about the applicability of INTERGROWTH intrauterine growth standards.

Comparison of logistic regression with machine learning methods for the prediction of fetal growth abnormalities: a retrospective cohort study

BACKGROUND

While there is increasing interest in identifying pregnancies at risk for adverse outcome, existing prediction models have not adequately assessed population-based risks, and have been based on conventional regression methods. The objective of the current study was to identify predictors of fetal growth abnormalities using logistic regression and machine learning methods, and compare diagnostic properties in a population-based sample of infants.

METHODS

Data for 30,705 singleton infants born between 2009 and 2014 to mothers resident in Nova Scotia, Canada was obtained from the Nova Scotia Atlee Perinatal Database. Primary outcomes were small (SGA) and large for gestational age (LGA). Maternal characteristics pre-pregnancy and at 26 weeks were studied as predictors. Logistic regression and select machine learning methods were used to build the models, stratified by parity. Area under the curve was used to compare the models; relative importance of predictors was compared qualitatively.

RESULTS

7.9% and 13.5% of infants were SGA and LGA, respectively; 48.6% of births were to primiparous women and 51.4% were to multiparous women. Prediction of SGA and LGA was poor to fair (area under the curve 60-75%) and improved with increasing parity and pregnancy information. Smoking, previous low birthweight infant, and gestational weight gain were important predictors for SGA; pre-pregnancy body mass index, gestational weight gain, and previous macrosomic infant were the strongest predictors for LGA.

CONCLUSIONS

The machine learning methods used in this study did not offer any advantage over logistic regression in the prediction of fetal growth abnormalities. Prediction accuracy for SGA and LGA based on maternal information is poor for primiparous women and fair for multiparous women.

Fetal Medicine Foundation fetal and neonatal population weight charts

OBJECTIVE

To develop fetal and neonatal population weight charts. The rationale was that, while reference ranges of estimated fetal weight (EFW) are representative of the whole population, the traditional approach of deriving birth-weight (BW) charts is misleading, because a large proportion of babies born preterm arise from pathological pregnancy. We propose that the reference population for BW charts, as in the case of EFW charts, should comprise all babies at a given gestational age, including those still in utero.

METHODS

Two sources of data were used for this study. For both, the inclusion criteria were singleton pregnancy, dating by fetal crown-rump length at 11 + 0 to 13 + 6 weeks' gestation, availability of ultrasonographic measurements of fetal head circumference (HC), abdominal circumference (AC) and femur length (FL) and live birth of phenotypically normal neonate. Dataset 1 comprised a sample of 5163 paired measurements of EFW and BW; ultrasound examinations were carried out at 22-43 weeks' gestation and birth occurred within 2 days of the ultrasound examination. EFW was derived from the HC, AC and FL measurements using the formula reported by Hadlock et al. in 1985. Dataset 2 comprised a sample of 95 579 pregnancies with EFW obtained by routine ultrasonographic fetal biometry at 20 + 0 to 23 + 6 weeks' gestation (n = 45 034), 31 + 0 to 33 + 6 weeks (n = 19 224) or 35 + 0 to 36 + 6 weeks (n = 31 321); for the purpose of this study we included data for only one of the three visits per pregnancy. In the development of reference ranges of EFW and BW according to gestational age, the following assumptions were made: first, that EFW and BW have a common median, dependent on gestational age; and second, that deviations from the median occur in both EFW and BW and these deviations are correlated with different levels of spread for EFW and BW, dependent on gestational age. We adopted a Bayesian approach to inference, combining information from the two datasets using Markov Chain Monte-Carlo sampling. The fitted model assumed that the mean log transformed measurements of EFW and BW are related to gestational age according to a cubic equation and that deviations about the mean follow a bivariate Gaussian distribution.

RESULTS

In the case of EFW in Dataset 2, there was a good distribution of values < 3^rd , < 5^th , < 10^th , > 90^th , > 95^th and > 97^th percentiles of the reference range of EFW according to gestational age throughout the gestational age range of 20 + 0 to 36 + 6 weeks. In the case of BW, there was a good distribution of values only for the cases delivered > 39 weeks' gestation. For preterm births, particularly at 27-36 weeks, BW was below the 3^rd , 5^th and 10^th percentiles in a very high proportion of cases, particularly in cases of iatrogenic birth. The incidence of small-for-gestational-age fetuses and neonates in the respective EFW and BW charts was higher in women of black than those of white racial origin.

CONCLUSION

We established a BW chart for all babies at a given gestational age, including those still in utero, thereby overcoming the problem of underestimation of growth restriction in preterm birth. BW and EFW charts have a common median but differ in the levels of spread from the median. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.