To evaluate whether epidural analgesia is associated with a higher rate of abnormal fetal head position at delivery. We conducted a prospective cohort study of 1,562 women to evaluate changes in fetal position during labor by using serial ultrasound examinations. Ultrasound examinations were performed at enrollment, epidural administration, 4 hours after the initial ultrasonography if epidural had not been administered, and late in labor (> 8 cm). Information about fetal head position at delivery was obtained from the provider. Regardless of fetal head position at enrollment (occiput transverse, occiput posterior, or occiput anterior), most fetuses were occiput anterior at delivery (enrollment position: occiput transverse 78%, occiput posterior 80%, occiput anterior 83%, P = .1). Final fetal position was established close to delivery. Of fetuses that were occiput posterior late in labor, only 20.7% were occiput posterior at delivery. Changes in fetal head position were common, and 36% of women had an occiput posterior fetus on at least one ultrasound examination. Women receiving epidural did not have more occiput posterior fetuses at the enrollment (23.4% epidural versus 26.0 no epidural, P = .9) or the epidural/4-hour ultrasound examination (24.9% epidural, 28.3% no epidural), but did have more occiput posterior fetuses at delivery (12.9% epidural versus 3.3% no epidural, P = .002); the association remained in a multivariate model (adjusted odds ratio 4.0, 95% confidence interval 1.4-11.1). Fetal position changes are common during labor, with the final fetal position established close to delivery. Our demonstration of a strong association of epidural with fetal occiput posterior position at delivery represents a mechanism that may contribute to the lower rate of spontaneous vaginal delivery consistently observed with epidural.