A randomized, double-blind study was designed to determine the effects on maternal intraoperative analgesia of adding one of the following opioids to the local anesthetic at the onset of epidural block, before surgery and neonatal delivery: morphine (3 mg), fentanyl (75 micrograms), sufentanil (50 micrograms), buprenorphine (0.3 mg) and oxymorphone (1 mg). The duration of postoperative analgesia, the presence of side effects and the neonatal outcome were also studied. Ninety healthy multiparas, at term, undergoing elective cesarean delivery using lumbar epidural anesthesia with 2% lidocaine were randomized in six equal groups to receive one of the opioids or saline. The predelivery administration of morphine, fentanyl and sufentanil significantly improved the intraoperative analgesia. Patients who received fentanyl, sufentanil, buprenorphine or oxymorphone had more somnolence than the others (p less than 0.01), but this did not interfere with the first mother-infant relationship during surgery. Patients in the buprenorphine group had more vomiting during surgery when compared with the others (p less than 0.01). Morphine provided the longest pain-free interval, followed by oxymorphone, buprenorphine, sufentanil and fentanyl. Postoperatively, the number of patients having pruritus and vomiting was significantly higher in the morphine and buprenorphine groups, respectively (p less than 0.01 versus others). No adverse neonatal effects were noted in any group.