Objectives To determine the association between fetal position at onset of labor and mode of delivery, specifically left occipito-anterior (LOA) fetal position and spontaneous vaginal delivery (SVD). Methods All nulliparous women who were booked at the Birmingham Women's NHS Foundation Trust over an 18-month period from April 2007 to September 2008 with a singleton live fetus without structural anomalies at term gestation were invited to take part in the study. Women recruited to the study underwent a transabdominal ultrasound scan to determine fetal occiput position at the onset of labor. They were then followed up until birth to determine outcome. The primary outcome measure was mode of delivery, categorized into SVD, instrumental delivery and Cesarean section. Results Of 1647 eligible women, 1250 had valid scans at onset of labor; 155 of the 1250 (12.4%) had fetuses in the LOA position. Analysis showed no evidence of difference in odds ratio (OR) of SVD for fetuses in the LOA position compared with all other positions (OR 0.864 (95% CI, 0.617-1.209); P = 0.394). No difference remained with adjustment for confounding effects of variables known to influence mode of delivery (OR 0.837 (95% CI, 0.551-1.272); P = 0.405). No other occipital position showed significant association with SVD. There was no evidence of the LOA position being associated with Cesarean section, ventouse or forceps delivery. Conclusion There is no evidence of an association between the fetal LOA position at onset of labor and SVD. This finding challenges the conventional theory that LOA is the optimum fetal position at onset of labor, and suggests that antenatal practices encouraging adoption of the LOA position through maternal posturing are unnecessary.