Chorionic villus sampling is being considered as an alternative to amniocentesis for genetic evaluation. The risk of subsequent spontaneous abortion from this invasive procedure, however, remains unclear. A prospective analysis of pregnancy outcome after ultrasound documentation of fetal viability at eight to 12 weeks' gestation was carried out in an uninstrumented population of patients to establish a background loss rate against which to evaluate the safety of chorionic villus sampling. All patients presenting to the authors' private practice from December 1983 to January 1986 were enrolled in the study. An ultrasound examination was performed at the first visit to confirm intrauterine pregnancy and assess gestational age; in gestations less than eight weeks, the sonogram was repeated four weeks later. Four hundred eighty-nine delivered patients are the subject of this study. We evaluated the differences between this group and those patients presenting either with a blighted ovum or beyond the first trimester, as well as the outcome of those patients with spotting early in gestation. This study suggests that if a live fetus is documented by ultrasonography at eight to 12 weeks' gestation, the risk of spontaneous abortion before 20 weeks' gestation in an uninstrumented population is 2.0%.