BACKGROUND: Although the labour and delivery outcomes of epidural analgesia have been investigated extensively, the effects on breast-feeding success are not clearly identified. AIM: To investigate the effects of intrapartum epidural analgesia on breast-feeding duration. METHODS: Nulliparous women enrolled in a randomised trial investigating labour and delivery outcomes of intrapartum epidural analgesia were asked about breast-feeding outcomes. Breast-feeding duration was ascertained by a self-report at 2 and 6 months post-partum. Breast-feeding outcomes were analysed as a prospective observational study because of high cross-over rates (43.4%) in the original randomised controlled trial. RESULTS: A total of 992 women were recruited to the trial with 690 (69.6%) receiving epidural analgesia in labour. Breast-feeding was initiated by 95% (n = 946). At 2 and 6 months, 625 (63.5%) and 401 (40.7%), respectively, were still breast-feeding. Intrapartum analgesia (trend P-value = 0.036), mode of delivery (P < 0.001), age (P < 0.001), education (P < 0.001), and smoking in pregnancy (P < 0.001) showed univariate associations with breast-feeding duration. In the subgroup of women with spontaneous onset of labour and vaginal deliveries, after controlling for other obstetric and demographic factors, epidural analgesia but not narcotic analgesia was significantly associated with reduced breast-feeding duration (adjusted hazard ratio 1.44, 95% confidence interval 1.04-1.99). CONCLUSIONS: Nulliparous women have a high use of epidural analgesia in labour. Nulliparous women who choose epidural analgesia are more likely to breast-feed for shorter durations. Further exploration of the factors underlying this association should be undertaken.