145: Midtrimester sonographic findings for the prediction of Down syndrome in a population at risk

Abstract

SYNDROME IN A POPULATION AT RISK ELISA POZZI, ANNA LOCATELLI, STEVEN WEINER, ANNA BIFFI, ALESSANDRO GHIDINI, PATRIZIA VERGANI, University of Milano-Bicocca, Obstetrics and Gynecology, Monza, Italy, George Washington University, Rockville, Maryland OBJECTIVE: The accuracy of genetic sonogram in algorithms of risk assessment for trisomy 21 (T21) is not established in prospective series. We calculated the adjusted effect sizes of midtrimester sonographic findings of T21 in a population at risk because of age and we then applied the prediction formula to a prospective series. STUDY DESIGN: Single-center cohort study conducted between 2001 and 2006 in women at risk for T21 undergoing genetic sonogram between 15 and 22 weeks= gestation. Univariate and logistic regression analysis were used to relate sonographic findings performed during the period 2001-2005 to occurrence of T21. Using the eventual model, predicted probabilities of Down syndrome were calculated for the cohort 2006. Subjects were classified as test positive or negative using a cut-off of 1/ 270 and sensitivity and false positive rates were calculated. RESULTS: Prevalence of Down syndrome was 27/1487 (1.8%). In the most parsimonious multivariable model adjusted for gestational age and maternal age, nuchal fold thickness (NFT) *5 mm, presence of renal pelvic dilation, absent midphalanx of the 5th finger, malformations, and heart defects were independently associated with T21, as were the following interactions: gestational age * NFT *5 mm and maternal age * aneuploid-associated anomalies or heart defects. The formula inclusive of these parameter estimates was prospectively used to calculate a woman=s adjusted risk of carrying a T21 fetus . Utilizing risk cut-off point of 1/270, the sensitivity of a genetic sonogram was 100% with a false positive rate of 22%. CONCLUSION: Genetic sonogram has adequate accuracy to be incorporated into current management algorithms for risk assessment of T21 in women at increased risk because of age.