Background Pain during childbirth is arguably the most severe pain some women may experience in their lifetime. Epidural analgesia is an effective form of pain relief during labour. Many women have concerns regarding its safety. Furthermore, epidural services and anaesthetic support may not be available consistently across all centres. Observational data suggest that early initiation of epidural may be associated with an increased risk of caesarean section, but the same findings were not seen in recent randomised controlled trials. More recent guidelines suggest that in the absence of a medical contraindication, maternal request is a sufficient medical indication for pain relief during labour. The choice of analgesic technique, agent, and dosage is based on many factors, including patient preference, medical status, and contraindications. There is no systematically reviewed evidence on the maternal and foetal outcomes and safety of this practice. Objectives This systematic review aimed to summarise the effectiveness and safety of early initiation versus late initiation of epidural analgesia in women. We considered the obstetric and fetal outcomes relevant to women and side effects of the treatments, including risk of caesarean section, instrumental birth and time to birth. Search methods We searched the Cochrane Pregnancy and Childbirth Group's Trials Register (12 February 2014), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2014, Issue 1), MEDLINE (January 1966 to February 2014), Embase (January 1980 to February 2014) and reference lists of retrieved studies. Selection criteria We included all randomised controlled trials involving women undergoing epidural labour analgesia that compared early initiation versus late initiation of epidural labour analgesia. Data collection and analysis Two review authors independently assessed trials for inclusion, extracted the data and assessed the trial quality. Data were checked for accuracy. Main results We included nine studies with a total of 15,752 women.The overall risk of bias of the studies was low, with the exception of performance bias (blinding of participants and personnel). The nine studies showed no clinically meaningful difference in risk of caesarean section with early initiation versus late initiation of epidural analgesia for labour (risk ratio (RR) 1.02; 95% confidence interval (CI) 0.96 to 1.08, nine studies, 15,499 women, high quality evidence). There was no clinically meaningful difference in risk of instrumental birth with early initiation versus late initiation of epidural analgesia for labour (RR 0.93; 95% CI 0.86 to 1.01, eight studies, 15,379 women, high quality evidence). The duration of second stage of labour showed no clinically meaningful difference between early initiation and late initiation of epidural analgesia (mean difference (MD) -3.22 minutes; 95% CI -6.71 to 0.27, eight studies, 14,982 women, high quality evidence). There was significant heterogeneity in the duration of first stage of labour and the data were not pooled. There was no clinically meaningful difference in Apgar scores less than seven at one minute (RR 0.96; 95% CI 0.84 to 1.10, seven studies, 14,924 women, high quality evidence). There was no clinically meaningful difference in Apgar scores less than seven at five minutes (RR 0.96; 95% CI 0.69 to 1.33, seven studies, 14,924 women, high quality evidence). There was no clinically meaningful difference in umbilical arterial pH between early initiation and late initiation (MD 0.01; 95% CI -0.01 to 0.03, four studies, 14,004 women, high quality evidence). There was no clinically meaningful difference in umbilical venous pH favouring early initiation (MD 0.01; 95% CI -0.00 to 0.02, four studies, 14,004 women, moderate quality evidence). Authors' conclusions There is predominantly high-quality evidence that early or late initiation of epidural analgesia for labour have similar effects on all measured outcomes. However, various forms of alternative pain relief were given to women who were allocated to delayed epidurals to cover that period of delay, so that is it hard to assess the outcomes clearly. We conclude that for first time mothers in labour who request epidurals for pain relief, it would appear that the time to initiate epidural analgesia is dependent upon women’s requests.