The diagnosis of abnormal fetal abdominal circumference is based on values >90th or <10th percentile. There are dozens of established norms that can be used to determine the percentile of a given abdominal circumference measurement, but there is no established method to determine which norms should be used. This study aimed to evaluate the applicability of 5 established abdominal circumference norms to our measurements and to determine which, if any, should be used for the diagnosis of abnormal fetal abdominal circumference. Data were pooled from 6 maternal-fetal medicine practices to conduct a cross-sectional study. The inclusion criteria were a singleton fetus at 22.0 to 39.9 weeks of gestation with cardiac activity present, complete fetal biometry measured, and examination from 2019 or 2020. For patients with >1 eligible examination during the study period, a single examination was chosen at random for inclusion. Five norms of abdominal circumference were studied: the Hadlock formula, the World Health Organization Fetal Growth Curves, the International Fetal and Newborn Growth Consortium for the 21st-Century Project; and the National Institutes of Child Health and Human Development Fetal Growth Studies (fetuses of White patients and unified standard). Using formulas relating abdominal circumference to gestational age, we calculated the z scores of abdominal circumference (standard deviations from the mean), standard deviation of the z score, Kolmogorov-Smirnov D statistic, and relative mean squared error. The 5 norms were assessed for fit to our data based on 6 criteria: mean z score close to 0, standard deviation of the z score close to 1, low D statistic, low mean squared error, fraction of values >90th percentile close to 10%, and fraction of values <10th percentile close to 10%. The inclusion criteria were met in 40,684 ultrasound examinations in 15,042 patients. Considering the 6 evaluation criteria, observed abdominal circumferences had the best fit to the World Health Organization standard (mean z score of 0.11±1.05, D statistic of 0.041, mean squared error of 0.84±1.46, 13% of examinations >90th percentile, and 7% of examinations <10th percentile). The Hadlock reference had an anomaly in its assumption of a constant standard deviation, resulting in the underdiagnosis of abnormal values at early gestational ages and overdiagnosis at late gestational ages. The International Fetal and Newborn Growth Consortium for the 21st-Century Project standard had a mean circumference smaller than all the other norms, resulting in the underdiagnosis of small circumferences and the overdiagnosis of large circumferences. Similar results were observed when restricting the analyses to a low-risk subgroup of 5487 examinations without identified risk factors for large for gestational age or small for gestational age. The diagnosis of abnormal abdominal circumference depends on the norms used to define abdominal circumference percentiles. The World Health Organization standard had the best fit for our data.
Read full abstract