Background: Preimplantation genetic diagnosis is an exciting advance in prenatal diagnosis. However, the safety of embryo biopsy must be determined with respect to both pregnancy rate and cogenital anomalies. Analysis: Too few pregnancies have been reported to allow meaningful inferences to be drawn, for which reason data on pregnancy losses and anomalies after conventional IVF were first reviewed. Loss rates are approximately 25%, and anomaly rates are not increased over that observed in the general population. Unfortunately, considerable methodological problems exist in published surveys: lack of proper controls, failure to take into account potential confounding variables, anomaly surveillance that is inconsistent with respect to the vigor with which anomalies are sought, inclusion or exclusion of minor anomalies, inclusion or exclusion of anomalies evident only on ultrasound, and even inclusion or exclusion of anomalies present in terminated pregnancies. We recommend prospective surveillance for major anomalies, defined as those causing death, major handicap or requiring surgery. Prospective surveillance ideally dictates collection of intake information at the time pregnancy is diagnosed, surveillance during pregnancy to exclude teratogenic influences, and systematic neonatal anomaly surveillance.