Fetal blood sampling is now performed in many centres through different approaches (fetoscopy, placentacentesis, cardiac puncture, umbilical cord needling, intrahepatic vein puncture) for prenatal diagnosis of congenital defects, management of intrauterine growth retardation and fetal therapy. One hundred and thirty-nine fetal blood samples have been performed during a 10 month period at Queen Charlotte's Maternity Hospital, London, using an individualized approach. One failure to obtain fetal blood occurred and there were four fetal losses, three of which followed an intrauterine transfusion in very severely affected fetuses. Two of these losses were associated with peculiar circumstances (see above). The procedure-related risk is nowadays more difficult to evaluate than in the past, when most fetal blood samplings were carried out in the second trimester for prenatal diagnosis. Most case studies, and ours as well, are not homogeneous and high-risk patients such as those with Rhesus disease or intrauterine growth retardation are also included. It seems, however, that transabdominal needling of the cord, at either placental or fetal insertion, is a low-risk procedure although a larger number of cases should be collected to draw definite conclusions about sampling from the intrahepatic vein. Fetoscopy also has a low risk in experienced hands, but the training period is certainly longer and the application in the second half of the pregnancy has been limited to a few cases. It is likely to be used only very little in the future. A flexible approach to fetal blood sampling allows the best choice of technique and utilizes the advantages of each technique.