OBJECTIVE: To evaluate the degree to which prenatal knowledge of fetal anomalies and sociodemographic characteristics determined outcome of 53,000 pregnancies. METHODS: Pregnancies were consecutively evaluated at a university hospital between 1984 and 1997. The severity of anomalies was graded by using an ordinal scale, in which 0 was no anomalies, 1 was no impact on quality of life, 2 was little impact but possibly requiring medical therapy, 3 was serious impact on quality of life even with optimal medical therapy, and 4 was incompatible with life. RESULTS: The abortion rates for grades 1 and 3 anomalies increased from 0.9% to 72.5%, and 0.9% to 37.1% for central nervous system and non-central nervous system anomalies, respectively (P < .001). Multiple logistic regression showed that mothers without a high school education were more likely than those who completed high school to abort a normal pregnancy (odds ratio [OR] 1.62, 95% confidence interval [CI] 1.07, 2.45). In the 452 pregnancies in which there was one grade 3 anomaly, logistic regression also showed that the abortion rate decreased by 6% per year as maternal age decreased (OR 0.94, 95% CI 0.91, 0.97). CONCLUSIONS: The severity of anomalies directly correlates with abortion rates, but at similar degrees of severity, central nervous system anomalies are more likely to lead to abortion. Maternal level of education inversely correlates with likelihood of termination of a normal pregnancy, whereas maternal age directly correlates with pregnancy termination when serious anomalies are present. Serious congenital anomalies may disproportionately affect children from families with the youngest mothers because these mothers are likely to continue these pregnancies.