Objectives: To assess in a prospective randomised study the association between motor block resulting from high and low dose epidural infusions of bupivacaine in labour and the incidence of long term backache after childbirth, and to compare the incidence of backache in women not receiving epidural analgesia. Design: Women requesting epidural analgesia in labour between October 1991 and March 1994 were randomised to receive infusions of either bupivacaine alone or low dose bupivacaine with opioid. Data were collected during labour and the immediate postpartum period from these women and from women recruited at random over the same time from those who had laboured without epidural analgesia. A postal questionnaire about symptoms was sent three months after childbirth to all women. Further data were collected one year after childbirth from those who had reported new backache at three months. Setting: St Thomas's Hospital, London. Subjects: 599 women were recruited, of whom 450 (75%) replied to a follow up questionnaire. Results: 152 women (33.8% of responders) reported backache lasting three months after delivery and, of these, 33 (7.3%) had not previously suffered with backache. There were no significant differences between the treatment groups in the incidence of postnatal backache overall or of new backache or any symptoms after childbirth. Among all demographic, obstetric, and epidural variables examined the only factors significantly associated with backache after childbirth were backache before and during pregnancy. Conclusions: The incidence of new long term backache was not significantly increased in women who received epidural analgesia in labour. Motor block resulting from epidural local anaesthetic administration was not a significant factor in the development of backache. Key messages About half of all women suffer backache during pregnancy, but many forget this when questioned retrospectively A prospective study showed that the incidence of new postpartum backache is 7.3 The use of epidural analgesia in labour had no effect on the incidence of postpartum backache In a randomised trial motor block in labour was not associated with an increase the incidence of backache