BACKGROUND: Prenatal testing was offered in all pregnancies obtained after ICSI with ejaculated or non-ejaculated sperm as part of the evaluation of the safety of ICSI. METHODS: Between 1990 and 2001, a chorionic villus sampling (CVS) or amniocentesis was offered for multiple or singleton pregnancies respectively during a genetic counselling session for all couples applying for ICSI. ICSI was carried out using ejaculated, epididymal or testicular sperm. RESULTS: In total, 1586 ICSI fetuses obtained after fresh embryo transfer were tested by CVS (n = 698) or by amniocentesis (n = 888). Abnormal fetal karyotypes were found in 47 samples [3.0%; 95% confidence interval (CI) 2.2–3.9%]; 25 anomalies (1.6%; 95% CI 1.0–2.3%) were de novo. These were 10 sex chromosomal anomalies and 15 autosomal anomalies [either numerical (n = 8) or structural (n = 7)], and 22 inherited abnormalities (1.4%; 95% CI 0.9–2.1%) (21 balanced, one unbalanced). In 17/22 inherited cases the chromosomal structural defect was inherited from the father. A significantly higher percentage of 2.1% de-novo prenatal chromosomal anomalies was observed for sperm concentrations of <20×10⁶ sperm per ml, as compared with 0.24% if the sperm concentration was │20×10⁶ sperm per ml (Fisher’s exact test, P = 0.006). No statistical difference in frequency of chromosomal anomalies was observed for lower threshold values of sperm concentration (<1×10⁶, <5×10⁶, <10×10⁶ and <15×10⁶). A statistical difference was observed for motility criteria, but not morphology. Three chromosomal anomalies were found prenatally after use of epididymal or testicular sperm in a total of 94 samples; two (of 83 tested) were from patients with obstructive and one (of nine tested) was from a patient with non-obstructive azoospermia. CONCLUSIONS: A significantly higher rate of de-novo chromosomal anomalies (1.6 versus 0.5% in amniocentesis for a mean maternal age of 33.5 years; P < 0.007) was observed in ICSI offspring, relating mainly to a higher number of sex chromosomal anomalies and partly to a higher number of autosomal structural anomalies. This finding was related to sperm concentration and motility. The significantly higher rate of observed inherited anomalies (1.4 versus 0.3–0.4% in prenatal tests in the general population; P < 0.001) was related to a higher rate of constitutional chromosomal anomalies, mainly in the fathers. The hypothesis of a higher risk of post-zygotic events as a consequence of the ICSI procedure leading to a higher proportion of chromosomal mosaicism was not confirmed in this study. Couples should be informed of the risks of an abnormal result related to sperm quality, and of the risk linked to a prenatal procedure as well as about the relatively benign character of some chromosomal anomalies such as de-novo structural anomalies or sex chromosomal anomalies in order to be able to make a choice for prenatal testing, or not.