Blinded ultrasound fetal biometry at 36 weeks and risk of emergency Cesarean delivery in a prospective cohort study of low‐risk nulliparous women

Abstract

To compare the association between risk of emergency Cesarean delivery (CD) and non-customized vs customized ultrasound estimated fetal weight (EFW) at 36 weeks' gestation, determine whether addition of ultrasound EFW to a model based on maternal characteristics alone improved prediction of emergency CD, assess the screening performance of a multivariable model using both EFW and maternal characteristics to predict emergency CD, and determine whether women at high predicted risk of emergency CD based on this model had higher risk of maternal and perinatal morbidity compared with screen-negative women. We studied 3047 low-risk (no pre-existing medical conditions or acquired complications of pregnancy) nulliparous women from the prospective Pregnancy Outcome Prediction study (Cambridge, UK) cohort, who underwent ultrasound EFW at ∼36 weeks' gestation. Both the women and their clinicians were blinded to fetal biometry results. Emergency CD was defined as delivery by Cesarean section in pregnancies in which the date of delivery had not been prearranged. Additional candidate predictors of emergency CD evaluated were maternal age, height, body mass index (BMI), weight gain, fetal abdominal circumference growth velocity and fetal sex. External validation of the predictive model was performed using routinely collected data from 55 337 births in Scotland between 2003 and 2008. Women with an estimated risk of emergency CD ≥ 40% were defined as screen positive. Blinded EFW was associated strongly with the risk of emergency CD (coefficient for increase of 1 SD in EFW, 0.39 (95% CI, 0.30-0.48); odds ratio (OR), 1.48 (95% CI, 1.35-1.62)). The coefficient for customized EFW was similar (0.42 (95% CI, 0.33-0.51); OR, 1.53 (95% CI, 1.39-1.67)); hence, for simplicity, non-customized EFW was employed subsequently. A multivariable logistic regression model combining maternal characteristics (age, height, BMI and weight gain between 12 and 36 weeks) was moderately predictive of emergency CD (area under the receiver-operating characteristics curve (AUC) = 0.68). Addition of blinded EFW to the model increased the AUC to 0.71 and improved prediction (likelihood-ratio test P< 0.0001). Based on this model, 189 (6.2%) women were screen positive and 48% of these delivered by CD. Screen-positive women had elevated risks of severe postpartum hemorrhage (relative risk (RR), 2.49; 95% CI, 1.83-3.38), any adverse neonatal outcome (RR, 1.86; 95% CI, 1.22-2.82) and severe adverse neonatal outcome (RR, 4.03; 95% CI, 1.35-12.03) compared with screen-negative women. The risks of these events were also higher compared with women who had a term CD for breech presentation. The model was similarly predictive of the risk of emergency CD and perinatal morbidity when evaluated using the dataset from Scotland. Ultrasound EFW at 36 weeks, combined with maternal characteristics, can identify women who are at increased risk of subsequent emergency CD. These women are at increased risk of maternal and perinatal morbidity compared with women at low risk of emergency CD and those having CD for breech presentation at term. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.