In this prospective meta-analysis, we have evaluated the effect of epidural analgesia with ropivacaine for pain in labour on neonatal outcome and mode of delivery compared with bupivacaine. In six randomized, double-blind studies, 403 labouring women, primigravidae and multiparae, received epidural analgesia with ropivacaine or bupivacaine 2.5 mg ml-1. The drugs were administered as intermittent boluses in four studies and by continuous infusion in two. Apgar scores, neurological and adaptive capacity scores (NACS), degree of motor block and mode of delivery were recorded. The studies were designed prospectively to fit meta-analysis of the pooled results. Results showed similar pain relief and consumption of the two drugs. In the vaginally delivered neonates, NACS scores were approximately equal for both groups at 2 h, but at 24 h there were fewer infants with NACS less than 35 in the ropivacaine compared with the bupivacaine group (2.8% vs 7.6%; P < 0.05). Spontaneous vaginal deliveries occurred more frequently overall with ropivacaine than with bupivacaine (58% vs 49%; P < 0.05) and instrumental deliveries (forceps and vacuum extraction) less frequently (27% vs 40%; P < 0.01), while the frequency of Caesarean section was similar between groups. The intensity of motor block was lower with ropivacaine. There were no significant differences in adverse events.