The human teratogenic potential of chlordiazepoxide is debated. To study the effects on the fetal development of very large doses of chlordiazepoxide that were used for a suicide attempt during pregnancy, self-poisoned pregnant women were identified from patients in a toxicological inpatient clinic in Budapest, Hungary. Comparisons were made between congenital abnormalities, intrauterine fetal development, and cognitive-behavioral status of the exposed children born to mothers who attempted suicide with chlordiazepoxide alone or in combination with other drugs during pregnancy and their sib controls. Of 1044 women with self-poisoning during pregnancy between 1960 and 1993, 88 (8.4%) used chlordiazepoxide with or without other drugs for suicide attempt; 35 of these women delivered live-born infants. Doses of chlordiazepoxide taken ranged between 20 and 300 mg, with a mean of 117 +/- 86 mg. Of 35 exposed children, six (17.1 %) were affected with congenital abnormalities compared with three (13.6%) of their 22 sibs (OR with 95% CI: 1.3 (0.3-4.4). Of 18 pregnant women who attempted suicide between the 4th and 12th postconceptional week, the period most sensitive to congenital malformation, four delivered live-born children affected with a congenital abnormality (atrial septal defect type II, complex defect of respiratory system, mild pyelectasis because of the stenosis of ureteropelvic junction, congenital inguinal hernia). Two other children had fetal alcohol syndrome and unrecognized multiple congenital abnormality including talipes equinovarus, deformation type, and four minor anomalies. The pregnancy age-specific mean birth weight indicated intrauterine fetal growth retardation, which was confirmed by a dose-response relationship and by the higher rate of low birth-weight newborns. Cognitive status and behavioral scale of exposed children did not indicate neurotoxic effects. Very large doses of chlordiazepoxide used for suicide attempts during pregnancy did not induce a higher rate of congenital abnormalities but were associated with dose-dependent intrauterine growth retardation.