Prediction of late-onset fetal growth restriction using a combined first- and second-trimester screening model in South Chinese infants: a retrospective study
Abstract Background: Prediction models for early and late fetal growth restriction (FGR) have been established in many high-income countries. However, prediction models for late FGR in China are limited. This study aimed to develop a simple combined first- and second-trimester prediction model for screening late-onset FGR in South Chinese infants. Methods: This retrospective study included 2258 women who had singleton pregnancies and received routine ultrasound scans as training dataset. A validation dataset including 565 pregnant women was used to evaluate the model in order to enable an unbiased estimation. Late-onset FGR was defined as a birth weight < the 10th percentile plus abnormal Doppler indices and/or a birth weight below the 3rd percentile after 32 weeks, regardless of the Doppler status. Multivariate logistic regression was used to develop a prediction model. The model included the a priori risk (maternal characteristics), the second-trimester head circumference (HC/AC) / abdomen circumference (HC) ratio and estimated fetal weight (EFW). Results: Ninety-three fetuses were identified as late-onset FGR. The significant predictors for late-onset FGR were maternal age, height, weight, and medical history; the second-trimester HC/ AC ratio; and the EFW. This model achieved a detection rate (DR) of 52.6% for late-onset FGR at a 10% false positive rate (FPR) (area under the curve (AUC): 0.80, 95%CI 0.76-0.85). The AUC of the validation dataset was 0.65 (95%CI 0.54-0.78). Conclusions: A multivariate model combining first- and second-trimester default tests can detect 52.6% of cases of late-onset FGR at a 10% FPR. Further studies with more screening markers are needed to improve the detection rate.