Between September 1st 1990 and Juli 31st 1993, 5071 pregnant women were screened prospectively by the "triple-test", including maternal serum alpha-fetoprotein, human chorionic gonadotropin and unconjugated oestriol in order to detect chromosomal anomalies and open neural tube defects. The serum samples were collected in collaboration with the obstetricians of the region of West-Mecklenburg and North-West-Brandenburg. Laboratory testing using radioimmunoassays was performed between weeks 15 and 20 of gestation, all serum specimens being investigated in only one institution. The original alpha-software from Wald et al. was the basis for calculating the statistical risk for Down's syndrome. Pregnant women with a high risk for Down's syndrome (cutoff > or = 1:250) were taken care of in a special outpatient clinic including procedures like amniocentesis and fetal blood sampling. Amongst 5071 pregnant women, 21 fetal anomalies were seen. Five cases of Down's syndrome, three of trisomy 18, one trisomy 13, two cases of triploidy and four cases of open neural tube defects, one 46 xy/45 x mosaic karyotype and one case of gastroschisis could be diagnosed correctly. One case of trisomy 21, one case of trisomy 18 and two open neural tube defects showed false negative results. Using the cutoff of 1:250 for prenatal detection of Down's syndrome and performing ultrasound routinely to determine gestational age, the sensitivity of the "triple-test" was 83.33% having a specificity of 92.68%. The predictive value of a positive test for prenatal diagnosis of Down's syndrome was 1.33%.(ABSTRACT TRUNCATED AT 250 WORDS)