In recent years external cephalic version (ECV) has been increasingly recommended as a safe alternative to either vaginal breech delivery or cesarean delivery. Reports suggest that both the latter outcomes are in fact less frequent with ECV, but fetal risk has not been reliably assessed. This study is a 10-year review of ECV performed at three hospitals from 1988 through 1997. Women having a single fetus in breech presentation, confirmed ultrasonically, were entered into the study when they agreed to attempted version. Version was done using a tocolytic agent, ritodrine, to relax the uterus, but without maternal analgesia. Electronic fetal monitoring was used to detect fetal cardiac dysfunction. The “forward roll” version technique was used; a “back flap” technique was occasionally tried if the fetus did not cross the midline. A second attempt was allowed if the first failed or if the fetus resumed a breech position.Among 923 women enrolled in the study, ECV succeeded in 62.7% of cases, with comparable rates at all three participating hospitals. Success rates were higher in multiparas than in nulliparas (76.6% vs. 55.3%), with a normal amniotic fluid volume (93% with polyhydramnios vs. 65% with a normal fluid volume and 29.6% with oligohydramnios), and incomplete compared with frank breeches (77.3% vs. 56.6%). Version succeeded in converting 72% of 30 transverse-lie fetuses. Two-thirds of 19 women having previous cesarean delivery and three-fourths of those undergoing myomectomy had a successful outcome. Logistic regression analysis disclosed that polyhydramnios was the strongest predictor of success, with a relative risk of 5.8. Version attempted on a multipara or for incomplete breech was 2.5-fold likelier to succeed than when done on a nullipara or for a frank breech. A posterior rather than anterior placenta also was a favorable factor. A second attempt succeeded in 10 of 24 women when the initial attempt failed and in all 8 women when an initially successful result reverted to breech. The risk of vaginal bleeding after version was 1.5%, and that of transient fetal bradycardia, 5.8%. The rate of cesarean delivery was 43% at these hospitals, where it is standard practice to perform cesarean delivery in cases of breech presentation. With one exception (femoral fracture when cesarean delivery followed a failed version), neonatal outcomes were favorable.The investigators believe that ECV is effective in lowering the need for cesarean delivery in term infants in breech presentation. No major adverse infant outcomes have been noted.