Objective: To investigate the morbidity associated with assisted vaginal birth (AVB) and an occipito-posterior (OP) fetal head position. Design: Observational study based on secondary analyses of a randomised controlled trial and cohort study. Setting: Two university affiliated maternity hospitals in Ireland. Population: A cohort of 1081 nulliparous women at term who experienced AVB. Methods: Univariable and multivariable logistic regression analyses were performed. Main Outcome Measures: Postpartum haemorrhage (PPH), obstetric anal sphincter injury (OASI), fetal acidosis, neonatal trauma, failed AVB leading to caesarean section (CS). Results: Of the 1081 AVBs, a total of 192 (17.8%) had an OP position with 103 of these (53.6%) rotated to occiptio-anterior (OA) prior to AVB and 89 (46.4%) delivered direct OP. AVB completed in a direct OP position compared with OP rotated to OA was associated with an increased risk of PPH (34% vs 15%, adjusted Odds Ratio 2.98; 95% Confidence Interval 1.48 to 6.02), OASI (16% vs 2%, adj OR 19.56;95% CI 4.09 to 93.62) and CS (27% versus 5%, adj OR 7.24; 95% CI 2.63 to 19.93). There were no significant differences in neonatal outcomes. The diagnosis of OP was incorrectly assigned in 35 of the 192 cases (18.2%) but did not impact on morbidity outcomes compared to known OP. Conclusions: This study highlights the increased risk of PPH, OASI, and failed AVB when attempting to deliver the baby in a direct OP position rather than rotating to OA. This has important implications for training and guidelines.