Am. Vintzileos et al., Indication-specific accuracy of second-trimester genetic ultrasonography for the detection of trisomy 21, AM J OBST G, 181(5), 1999, pp. 1045-1048
OBJECTIVE: The object of this study was to determine whether there are any
clinically significant indication-specific variations in the accuracy of se
cond-trimester genetic ultrasonography and to provide a risk adjustment for
fetal trisomy 21 according to the results of genetic ultrasonography.
STUDY DESIGN: From November 1, 1992, to September 30, 1998, a second-trimes
ter genetic sonogram was offered to all pregnant women who were at an incre
ased risk for fetal trisomy 21 (greater than or equal to 1:274) because of
either advanced maternal age (greater than or equal to 35 years) or abnorma
l serum biochemical profile or both of these. Outcome information included
the results of genetic amniocentesis if performed and the results of pediat
ric assessment and follow-up after birth. In determining diagnostic accurac
y of the genetic sonogram the presence of greater than or equal to 1 abnorm
al ultrasonographic marker was considered an abnormal test result.
RESULTS: A total of 1835 fetuses with known outcomes underwent genetic ultr
asonography between 15 and 24 weeks' gestation; of these 1792 had normal re
sults, 34 had trisomy 21, and 9 had other chromosomal abnormalities. The li
kelihood of fetal trisomy 21 was reduced by 80% after a normal result of ge
netic ultrasonography. The overall sensitivity, specificity, and positive a
nd negative predictive values of genetic ultrasonography for the detection
of trisomy 21 were 82%, 91%, 15%, and 99.6%, respectively. There were no si
gnificant indication-specific variations in the accuracy of second-trimeste
r ultrasonography. The sensitivity for the detection of fetal trisomy 21 ra
nged from 80% among women with advanced maternal age to 100% among women wi
th both an abnormal biochemical profile and advanced maternal age.
CONCLUSIONS: The likelihood of fetal trisomy 21 risk was reduced 80% after
a normal result of genetic ultrasonography. In addition there were no signi
ficant indication-specific variations in the detection rate of genetic ultr
asonography.