Universal third-trimester ultrasonic screening using fetal macrosomia in the prediction of adverse perinatal outcome: a systematic review and meta-analysis of diagnostic test accuracy

Background The effectiveness of screening for macrosomia is not well established. One of the critical elements of an effective screening program is the diagnostic accuracy of a test at predicting the condition. The objective of this study is to investigate the diagnostic effectiveness of universal ultrasonic fetal biometry in predicting the delivery of a macrosomic infant, shoulder dystocia, and associated neonatal morbidity in low- and mixed-risk populations. Methods and findings We conducted a predefined literature search in Medline, Excerpta Medica database (EMBASE), the Cochrane library and ClinicalTrials.gov from inception to May 2020. No language restrictions were applied. We included studies where the ultrasound was performed as part of universal screening and those that included low- and mixed-risk pregnancies and excluded studies confined to high risk pregnancies. We used the estimated fetal weight (EFW) (multiple formulas and thresholds) and the abdominal circumference (AC) to define suspected large for gestational age (LGA). Adverse perinatal outcomes included macrosomia (multiple thresholds), shoulder dystocia, and other markers of neonatal morbidity. The risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. Meta-analysis was carried out using the hierarchical summary receiver operating characteristic (ROC) and the bivariate logit-normal (Reitsma) models. We identified 41 studies that met our inclusion criteria involving 112,034 patients in total. These included 11 prospective cohort studies (N = 9986), one randomized controlled trial (RCT) (N = 367), and 29 retrospective cohort studies (N = 101,681). The quality of the studies was variable, and only three studies blinded the ultrasound findings to the clinicians. Both EFW >4,000 g (or 90th centile for the gestational age) and AC >36 cm (or 90th centile) had >50% sensitivity for predicting macrosomia (birthweight above 4,000 g or 90th centile) at birth with positive likelihood ratios (LRs) of 8.74 (95% confidence interval [CI] 6.84–11.17) and 7.56 (95% CI 5.85–9.77), respectively. There was significant heterogeneity at predicting macrosomia, which could reflect the different study designs, the characteristics of the included populations, and differences in the formulas used. An EFW >4,000 g (or 90th centile) had 22% sensitivity at predicting shoulder dystocia with a positive likelihood ratio of 2.12 (95% CI 1.34–3.35). There was insufficient data to analyze other markers of neonatal morbidity. Conclusions In this study, we found that suspected LGA is strongly predictive of the risk of delivering a large infant in low- and mixed-risk populations. However, it is only weakly (albeit statistically significantly) predictive of the risk of shoulder dystocia. There was insufficient data to analyze other markers of neonatal morbidity.
Gordon Smith and colleagues investigate the diagnostic effectiveness of universal ultrasonic fetal biometry in predicting infant macrosomia
Author summary Why was this study done? There is a debate regarding introducing universal third-trimester screening for macrosomia. An effective screening program requires two elements: an effective test at predicting a condition and an effective intervention. There is evidence that early-term induction of labor (IOL) could reduce the rates of shoulder dystocia. However, there is no high-quality evidence regarding the diagnostic effectiveness of fetal biometry at predicting macrosomia and associated morbidity. What did the researchers do and find? We searched more than 10,000 titles and identified 41 studies including 112,034 patients that offered third-trimester ultrasounds for the prediction of macrosomia as part of universal ultrasound screening or were done in low- and mixed-risk populations. The quality of the studies was variable, and only three studies blinded the ultrasound findings to the clinicians. We found that the two most common ultrasound markers, the estimated fetal weight (EFW) and the abdominal circumference (AC), could predict the majority of macrosomic infants at birth (sensitivity >50%) with high diagnostic performance (positive LRs between 7 and 10). However, the EFW could only predict about 1 in 5 cases of shoulder dystocia (22% sensitivity) with low diagnostic performance (positive likelihood ratio of about 2). There was insufficient data to analyze other markers of neonatal morbidity. What do these findings mean? Universal third-trimester ultrasound screening will identify more pregnancies with macrosomia. However, it will not have a clinically significant effect at predicting shoulder dystocia. There is not enough evidence on the effect of ultrasound screening on neonatal morbidity. We recommend caution prior to introducing universal third-trimester screening for macrosomia, as it would increase the rates of intervention, with potential iatrogenic harm, without clear evidence that it would reduce neonatal morbidity.