Purpose: to appraise the value of ultrasonographic parameters for the diagnosis of fetal Down syndrome (T21), in order to permit its use in routine clinical practice. Methods: this is a prospective cohort study using various ultrasonographic parameters for the prediction of T21. A total of 1662 scans were evaluated in the cohort study and 289 examinations were analyzed as a differential sample to test the normality curve from October 1993 to November 2000. The statistical analysis was based on the calculation of intra- and interobserver variations, the construction of normality curves for the studied parameters, as well as their validity tests, and the calculation of sensitivity, specificity, relative risk, likelyhood ratio and posttest predictive values. Results: among 1662 cases, 22 fetuses (1.32%) with T21 were identified. The normality curves were built for nucal fold thickness, femur/foot ratio and nasal bone length. Renal pelvis had a semiquantitative distribution and the proposed cutoff level was 4.0 mm. Sensitivity, specificity, false positive rate, relative risk and likelyhood ratio for nucal fold measurements above the 95th percentile were 54.5%, 95.2%, 4.9%, 20.2 and 11, respectively. For nasal bone measurements below the 5th percentile, 59.0%, 90.1%, 9.0%, 13.4 and 6.5. For femur/foot ratio below the 5th percentile, 45.5%, 84.4%, 15.6%, 3.7 and 2,6. For renal pelvis greater than 4.0 mm, 36.4%, 89.2%, 10.9%, 4.5 and 3.4. For absent fifth finger middle phalanx, 22.7%, 98.1%, 1.9%, 13.2 and 11.9. For the presence of major malformations, 31.8%, 98.7%, 1.3%, 27.2 and 24,8. After calculating the probability rates and the incidence of T21 in different maternal ages, a table for posttest risk using ultrasonographic parameters was set up. Conclusions: normality curves and indices for the assessment of risk for fetal Down syndrome on a population basis were established by the utilization of different maternal ages and by multiplying factors proposed by the authors. It was not possible to establish a normality curve for renal pelvis measurements, because of their semiquantitative distribution.
Trisomy 21; Down syndrome; Prenatal diagnosis; Screening
