The effectiveness of unconjugated estriol (uE3) as a serum marker for the detection of Down syndrome (DS) during the 2nd trimester of pregnancy was evaluated. A population of 18,764 normal singleton pregnancies was screened for alpha-feto-protein and human chorionic gonadotropin. In 9,311 women, uE3 was added. Using a risk of 1:250 at term as a cutoff value, the false-positive rates were 4.1 and 4.3% without and with uE3, respectively. The detection rates in 47 DS serum samples, some of which were studied retrospectively, were 66% without uE3 and 57% with uE3. In 12 of 25 younger women and in 19 of 22 older women, DS was detected without uE3. The uE3 contributed to the detection of 4 additional DS pregnancies (1 in the young and 3 in older women). On the other hand, 8 DS pregnancies (3 in younger women and 5 in older women) escaped detection. In our sample the addition of uE3 lowered the detection rate of DS pregnancies with only a small and insignificant effect on the false-positive rate. Our results call for special caution in the addition of markers for risk calculations. We suggest that pregnancies with a calculated risk of > 1:250 following maternal serum alpha-fetoprotein and human chorionic gonadotropin markers tests should be regarded as high-risk pregnancies, even in cases in whom the addition of uE3 lowers the risk beneath the cutoff value.