RESUMO
OBJECTIVE: To investigate whether abdominal circumference growth velocity (ACGV) improves the prediction of perinatal outcome in small-for-gestational-age (SGA) fetuses beyond that afforded by estimated fetal weight (EFW) and cerebroplacental ratio (CPR). METHODS: A cohort of 235 singleton SGA fetuses at 36-38 weeks' gestation was examined. ACGV, EFW and CPR centiles were calculated. ACGV centile was determined using data from a large database of 19-21- and 36-38-week scans in an unselected population. Binary variables of ACGV < 10th , EFW < 3rd and CPR < 5th centiles were defined as abnormal. Two composite adverse outcomes (CAO) were explored: CAO-1 defined as at least one of umbilical artery pH < 7.10, 5-min Apgar score < 7 or neonatal unit admission, and CAO-2 that included in addition hypoglycemia, intrapartum fetal distress and perinatal death. Univariate and multivariate logistic regression analyses were performed to analyze the relationship between the three risk factors and their predictive value for CAO. The change in screening performance afforded by adding ACGV to EFW and CPR was assessed and receiver-operating characteristics (ROC) curves were calculated. RESULTS: ACGV < 10th centile was an independent risk factor for CAO. The sensitivity, specificity, positive and negative likelihood ratios of a predictive model based on EFW < 3rd centile and CPR < 5th centile were, respectively, 51%, 70%, 1.71 and 0.69 for CAO-1 and 41%, 70%, 1.39 and 0.83 for CAO-2. After addition of ACGV < 10th centile to the model, the respective values were 82%, 46%, 1.54 and 0.38 for CAO-1 and 71%, 47%, 1.34 and 0.62 for CAO-2. Using continuous variables, the areas under the ROC curves improved marginally from 0.669 (95% CI, 0.604-0.729) to 0.741 (95% CI, 0.677-0.798) for CAO-1 and from 0.646 (95% CI, 0.580-0.707) to 0.700 (95% CI, 0.633-0.759) for CAO-2 after addition of ACGV to the model. CONCLUSIONS: ACGV is a risk factor for adverse neonatal outcome that is independent of EFW and of CPR, although any improvement in the prediction of adverse outcome is not statistically significant. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.