How prenatal head ultrasound reference ranges affect evaluation of possible fetal microcephaly

Different fetal ultrasound (US) nomograms of the head circumference (HC) have been established; however, comparisons between the detection rates of microcephaly among US nomograms are few and inconsistent. We aimed to compare the prenatal diagnostic rate of fetal microcephaly (FM) among four widely used US nomograms of the fetal HC, when applied to the same group of fetuses.

We retrospectively identified singleton pregnancies complicated by fetal HC < 5th percentile for gestational age (GA) by US, without other risk factors for FM and with normal fetal brain MRI. Raw values of HC by US were converted to z-scores using four nomograms (Chervenak = A, Hadlock = B, Gelber = C, Papageorghiou = D). Z-scores value of the HC were classified as normal, possible normal, or microcephaly if values were >-2, ≤ -2 and >-3, or ≤ -3, respectively and compared among the four nomograms.

Fifty one fetuses at a mean (±SD) GA of 28 (±4) weeks were included. The four nomograms resulted in different z-score values of the fetal HC for the same subject (p < .001) and none of them showed 100% agreement. Reference C and D showed the highest agreement in classifying subjects as normal, possible normal, or with microcephaly (simple Kappa = 0.8915, % agreement = 94.1%), while A and B had the lowest agreement (simple Kappa = 0.0977, % agreement = 51.0%).

Despite the use of similar prenatal cutoff z-score values of the fetal HC, the four nomograms led to different diagnostic rates of FM. More consistent diagnostic criteria are therefore needed to define FM, especially in the absence of other risk factors for FM and normal fetal brain MRI, since the prenatal diagnosis can affect pregnancy management.