Sixty healthy patients scheduled for elective cesarean delivery under epidural anesthesia were randomized to receive either lidocaine or 2-chloroprocaine as the primary local anesthetic agent. When patients first complained of postoperative pain in the recovery room, they were given either fentanyl 50 micrograms or butorphanol 2 mg, epidurally, in a randomized, blinded fashion. Postoperative analgesia, quantitated on a visual analogue scale, as well as time elapsed until first request for supplemental opioid, did not differ for patients receiving butorphanol after either 2-chloroprocaine or lidocaine anesthesia. In contrast, epidural fentanyl produced a shorter and lesser degree of sensory analgesia after 2-chloroprocaine use, whereas epidural fentanyl after lidocaine anesthesia provided pain relief similar to that seen in the butorphanol groups. Side effects were limited to somnolence with butorphanol and pruritus with fentanyl. No evidence of respiratory depression was seen in any patient. We conclude that 2 mg of butorphanol epidurally provides approximately 2 to 3 h of effective analgesia after cesarean delivery with either lidocaine or 2-chloroprocaine anesthesia. Epidural fentanyl seems to be antagonized when 2-chloroprocaine, but not lidocaine, is used as the primary local anesthetic agent. We suggest a possible mu-receptor-specific etiology for this effect.