In a double-blind, randomheed, prospective study 150 women in labour received intermittent epidural injections of 10 mlO.l25% bupivacaine with adrenaline (1 :800000) with 5, 7.5 or 10 pg of sufentanil added. The onset, duration, and quality of analgesia were compared. Motor block, type of delivery and neonatal Apgar scores were noted. The onset, duration, and quality of analgesia were generally similar in the three groups, except following the second injection when the quality of analgesia was signijcantly superior in the sufentanil 7.5 and 10 pg groups. Motor blockade and type of delivery did not difer between the groups and there were no diferences in neonatal Apgar scores. No patient required more than three injections. We conclude that 7.5 pg sufentanil is the optimal dose to add to intermittent epidural injections of 10 mi 0.125% bupivacaine with adrenaline ( 1 : 800 000) for pain relief in labour.