Objective: To determine the risk of Down syndrome in fetuses with sonographic markers using the Bayes theorem and likelihood ratios.
Methods: We prospectively evaluated the midtrimester sonographic features of fetuses with Down syndrome and compared them with euploid fetuses. Patients were referred for an increased risk of aneuploidy and evaluated for the presence of structural defects, a nuchal fold, short long bones, pyelectasis, an echogenic intracardiac focus, and hyperechoic bowel. All fetuses underwent amniocentesis at the time of sonographic assessment. The sensitivity, specificity, and likelihood ratios for markers were calculated both as nonisolated and isolated findings.
Result: There were 164 fetuses with Down syndrome and 656 euploid fetuses. The presence of any marker resulted in sensitivity for the detection of Down syndrome of 80.5% with a false-positive rate of 12.4%. The absence of any markers conferred a likelihood ratio of 0.2, decreasing the risk of Down syndrome by 80%. As an isolated marker, the nuchal fold had an "infinite" likelihood ratio for Down syndrome; a short humerus had a likelihood ratio of 5.8, whereas structural anomalies had a likelihood ratio of 3.3. Other isolated markers had low likelihood ratios because of the higher prevalence in the unaffected population. The likelihood ratios for the presence of 1, 2, and 3 of any of the markers were 1.9, 6.2, and 80, respectively.
Conclusions: Although an isolated marker with a low likelihood ratio may not increase a patient's risk of Down syndrome, the presence of such a marker precludes reducing the risk of aneuploidy. Clusters of markers appear to confer a higher risk.