Diagnostic performance of third-trimester ultrasound for the prediction of late-onset fetal growth restriction: a systematic review and meta-analysis

OBJECTIVE: The objective of the study was to establish the diagnostic performance of ultrasound screening for predicting late smallness for gestational age and/or fetal growth restriction. DATA SOURCES: A systematic search was performed to identify relevant studies published since 2007 in English, Spanish, French, Italian, or German, using the databases PubMed, ISI Web of Science, and SCOPUS. STUDY ELIGIBILITY CRITERIA: We used rrospective and retrospective cohort studies in low-risk or nonselected singleton pregnancies with screening ultrasound performed at >= 32 weeks of gestation. STUDY APPRAISAL AND SYNTHESIS METHODS: The estimated fetal weight and fetal abdominal circumference were assessed as index tests for the prediction of birthweight <10th (i.e. smallness for gestational age), less than the fifth, and less than the third centile and fetal growth restriction (estimated fetal weight less than the third or estimated fetal weight <10th plus Doppler signs). Quality of the included studies was independently assessed by 2 reviewers, using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. For the meta-analysis, hierarchical summary receiver-operating characteristic curves were constructed, and quantitative data synthesis was performed using random-effects models. The sensitivity of the abdominal circumference <10th centile and estimated fetal weight <10th centile for a fixed 10% false-positive rate was derived from the corresponding hierarchical summary receiver-operating characteristic curves. Heterogeneity between studies was visually assessed using Galbraith plots, and publication bias was assessed by funnel plots and quantified by Deeks' method. RESULTS: A total of 21 studies were included. Observed pooled sensitivities of abdominal circumference and estimated fetal weight <10th centile for birthweight <10th centile were 35% (95% confidence interval, 20-52%) and 38% (95% confidence interval, 31-46%), respectively. Observed pooled specificities were 97% (95% confidence interval, 95-98%) and 95% (95% confidence interval, 93-97%), respectively. Modeled sensitivities of abdominal circumference and estimated fetal weight <10th centile for 10% false-positive rate were 78% (95% confidence interval, 61-95%) and 54% (95% confidence interval, 46-52%), respectively. The sensitivity of estimated fetal weight <10th centile was better when aimed to fetal growth restriction than to smallness for gestational age. Meta-regression analysis showed a significant increase in sensitivity when ultrasound evaluation was performed later in pregnancy (P = .001). CONCLUSION: Third-trimester abdominal circumference and estimated fetal weight perform similar in predicting smallness for gestational age. However, for a fixed 10% false-positive rate extrapolated sensitivity is higher for abdominal circumference. There is evidence of better performance when the scan is performed near term and when fetal growth restriction is the targeted condition.