Prenatal diagnosis of cystic fibrosis was performed in 200 pregnancies with a 1-in-4 risk, and was based on significant modifications in amniotic fluid taken at 17, 18, 19 weeks of pregnancy, of six enzymatic assays: gamma-glutamyl-transpeptidase, aminopeptidase M, and alkaline phosphatase (total and isoenzymes). On the basis of normal values, normal outcome was predicted in 135 pregnancies reaching term, all the babies were normal. On the basis of significantly abnormal enzymatic values, an affected fetus was predicted in 56 pregnancies, 53 were terminated, and 3 went to term; the infants were affected. There were discrepancies in enzymatic values in nine cases, in eight cases normal outcome was predicted, six babies were normal and two were affected; in one case an affected baby was predicted, the pregnancy went to term and the baby is normal. Criteria giving evidence for cystic fibrosis in fetuses have been described: macroscopic observation of a typical meconium ileus, significant increase of albumin content in the meconium, and PAS-positive mucus-like material in some pancreatic acini. Using these criteria, diagnosis of cystic fibrosis has been confirmed in all the examined fetuses. The recurrence rate of cystic fibrosis was 22.5% in 147 diagnoses in which the index case had cystic fibrosis without a history of meconium ileus at birth, but was 47.5% when the index case had meconium ileus. The results of the study suggest that prenatal diagnosis of cystic fibrosis can be performed with an accuracy of 98%.