In this study we compared the pregnancy outcome of 576 pregnancies after prenatal diagnosis with that of 540 pregnancies without prenatal diagnosis in our microinjection programme. Amniocentesis was suggested for singleton pregnancies (n = 465) and chorionic villus sampling (CVS) was proposed for twin pregnancies (n = 111 pregnancies, 222 fetuses). A total of 365 patients with singleton pregnancies and 175 patients with twin pregnancies who did not undergo prenatal diagnosis were selected as controls. Compared with the controls, the odds ratios in the amniocentesis group for preterm delivery, low birthweight, very low birthweight and fetal loss were 0.97 [95% confidence interval (CI): 0.60-1.57], 1.27 (95% CI: 0.78-2.06), 1.57 (95% CI: 0.53-4.66) and 0.86 (95% CI: 0.32-2.37) respectively. Compared with the controls, the odds ratios in the CVS group for preterm delivery, low birthweight, very low birthweight and fetal loss were 0.89 (95% CI: 0.61-1.30), 1.03 (95% CI: 0.74-1.45), 0.79 (95% CI: 0.41-1.53) and 0.47 (95% CI: 0.17-1.30) respectively. We concluded that, in this series of intracytoplasmic sperm injection (ICSI) pregnancies, prenatal testing did not increase the preterm-delivery, the low-birthweight, or the very low-birthweight rates as compared with those of the controls. In the prenatal diagnosis group, the fetal loss rate was comparable to that of the control group. Larger prospective controlled studies are needed in order to inform patients reliably about the risks and the advantages of prenatal testing in ICSI pregnancies.