BACKGROUND Cancer is the second leading cause of death of women during the reproductive years, and its occurrence in pregnancy is between 0.07% and 0.1%. METHODS To analyze the effect of cancer on pregnancy, we compared 21 pregnancies occurring during 30 years in women who received chemotherapy for their cancer with a control group matched for maternal age and composed of women not exposed to known teratogens or reproductive risks during pregnancy. RESULTS Of 13 women exposed to chemotherapy during the first trimester, two of five whose pregnancies continued to term had major malformations in their infants, four had spontaneous abortions, and four had therapeutic abortions. Of four women with second-trimester exposure to chemotherapy, two had normal live births, one had a stillbirth, and one had a therapeutic abortion. All four pregnancies exposed to chemotherapy during the third trimester resulted in healthy live births. Infants exposed to chemotherapy had statistically significantly lower birth weights than their matched controls (2227 +/- 558 g vs 3519 +/- 272 g, P less than .001), due to both significantly lower gestational age and substantial intrauterine growth retardation (P less than .01). The trend for higher rate of stillbirth (1/11) agrees well with 10 stillbirths among all women with cancer in pregnancy without and with chemotherapy who gave birth (n = 223), when compared with the population of Ontario (P less than .0005). CONCLUSIONS This study confirms the increased likelihood of spontaneous abortions and major birth defects when chemotherapy is used during embryogenesis, whereas such a risk is not apparent beyond the first trimester. Because of the higher risk of stillbirth and intrauterine growth retardation, women with cancer should be monitored closely by a high-risk obstetric unit to define the optimal time of delivery.