AIM: The early detection of small-for-gestational age (SGA) fetuses and newborns is pivotal in the prevention of perinatal mortality. OBJECTIVES: To compare the predictive capability of performing ultrasound-based estimated fetal weight (EFW) at 32 versus 36 weeks' gestation on the detection rate of SGA fetuses and SGA newborns at delivery, and to find a better cutoff level to consider a fetus at risk of being born small. MATERIAL AND METHODS: Nine hundred fifteen low-risk pregnant women were assessed at both 32 and 36 weeks' gestation. EFW centile was calculated in both occasions. The rate of SGA fetuses was compared. SGA fetuses were considered when both abdominal circumference (AC) and EFW were below the 10th centile from a total of 488 delivered at our Hospital. Paired comparisons between ultrasound at 32 and 36 weeks' gestation were done to predict SGA at delivery. Percentages of SGA fetuses were compared by chi-squared test. ANOVA test was used for comparing centiles among groups. Receiver operating characteristic (ROC) curve was used to find the best cutoff ultrasound centile to predict SGA at delivery. Statistical significance was previously set at 95% level (p < .05). RESULTS: Ultrasound-based EFW at 32 weeks showed 23 cases of SGA (2.5%) while at 36 weeks this rate increased up to 4% (37/915) (p < .000001). When comparing both outcomes, 2.8% of those catalogued as adequate-for-gestational age (AGA) at 32 weeks were cases of SGA at 36 weeks. In addition, 47.8% of those diagnosed as SGA were not confirmed at 36 weeks. Only 12.3% of SGA neonates were identified at 32 weeks' gestation ultrasound, while using the 36 weeks' gestation approach this rate increased up to 30.9%. So, only a low proportion of SGA neonates were SGA fetuses at any of these two gestational ages. However, the area under the curve (AUC) at 36 weeks was as high as 0.86. Being a matter of cutoff rather than a matter of choosing the correct variable, ROC analysis showed that the best cutoff for prediction having the best sensitivity (0.80) with the best specificity (0.77) was 28th centile of EFW. This represents 24.9% of the studied women (228/915). CONCLUSIONS: In general, ultrasound at 36 weeks has better performance detecting SGA fetuses than ultrasound at 32 weeks and we suggest to definitively change from 32 to 36 weeks in order to increase the detection rate of SGA fetuses. Moreover, in order to detect those fetuses who will grow below the lower level of the normal range in the last month of pregnancy, we suggest that those with EFW below the 28th centile at 36 weeks should be rescanned later in pregnancy to identify prenatally as many cases as we can of SGA newborns.
AIM: The early detection of small-for-gestational age (SGA) fetuses and newborns is pivotal in the prevention of perinatal mortality. OBJECTIVES: To compare the predictive capability of performing ultrasound-based estimated fetal weight (EFW) at 32 versus 36 weeks' gestation on the detection rate of SGA fetuses and SGA newborns at delivery, and to find a better cutoff level to consider a fetus at risk of being born small. MATERIAL AND METHODS: Nine hundred fifteen low-risk pregnant women were assessed at both 32 and 36 weeks' gestation. EFW centile was calculated in both occasions. The rate of SGA fetuses was compared. SGA fetuses were considered when both abdominal circumference (AC) and EFW were below the 10th centile from a total of 488 delivered at our Hospital. Paired comparisons between ultrasound at 32 and 36 weeks' gestation were done to predict SGA at delivery. Percentages of SGA fetuses were compared by chi-squared test. ANOVA test was used for comparing centiles among groups. Receiver operating characteristic (ROC) curve was used to find the best cutoff ultrasound centile to predict SGA at delivery. Statistical significance was previously set at 95% level (p < .05). RESULTS: Ultrasound-based EFW at 32 weeks showed 23 cases of SGA (2.5%) while at 36 weeks this rate increased up to 4% (37/915) (p < .000001). When comparing both outcomes, 2.8% of those catalogued as adequate-for-gestational age (AGA) at 32 weeks were cases of SGA at 36 weeks. In addition, 47.8% of those diagnosed as SGA were not confirmed at 36 weeks. Only 12.3% of SGA neonates were identified at 32 weeks' gestation ultrasound, while using the 36 weeks' gestation approach this rate increased up to 30.9%. So, only a low proportion of SGA neonates were SGA fetuses at any of these two gestational ages. However, the area under the curve (AUC) at 36 weeks was as high as 0.86. Being a matter of cutoff rather than a matter of choosing the correct variable, ROC analysis showed that the best cutoff for prediction having the best sensitivity (0.80) with the best specificity (0.77) was 28th centile of EFW. This represents 24.9% of the studied women (228/915). CONCLUSIONS: In general, ultrasound at 36 weeks has better performance detecting SGA fetuses than ultrasound at 32 weeks and we suggest to definitively change from 32 to 36 weeks in order to increase the detection rate of SGA fetuses. Moreover, in order to detect those fetuses who will grow below the lower level of the normal range in the last month of pregnancy, we suggest that those with EFW below the 28th centile at 36 weeks should be rescanned later in pregnancy to identify prenatally as many cases as we can of SGA newborns.