Epidural administration of hydromorphone was evaluated using a patient-controlled analgesia (PCA) delivery system in 170 healthy women undergoing elective cesarean delivery with epidural bupivacaine who were randomly assigned to one of four epidural PCA treatment groups: group I, hydromorphone alone by bolus administration; group II, hydromorphone, with a continuous (basal) infusion; group III, hydromorphone in combination with 0.08% bupivacaine by bolus administration; or group IV, hydromorphone and bupivacaine, with a concurrent infusion of both drugs. Patients in group I required significantly less opioid medication (2.1 +/- 1.1 mg [mean +/- SD]) during the first 24 h than patients in group II (3.3 +/- 1.3 mg). Similarly, patients in group III self-administered significantly less hydromorphone (2.0 +/- 1.0 mg) and bupivacaine (23.3 +/- 11.4 mg) during the first 24 h of PCA therapy, compared with patients in group IV (hydromorphone [2.7 +/- 1.1 mg] and bupivacaine [31.5 +/- 11.6 mg]). The concomitant use of a local anesthetic or basal opioid infusion with hydromorphone via epidural PCA did not decrease the number of PCA demands or delivered doses. In addition, patients in all four groups had similar pain, sedation, discomfort, fatigue, and anxiety scores. The frequency of awakening at night to self-administer analgesic medication was not decreased when a basal infusion was used.(ABSTRACT TRUNCATED AT 250 WORDS)