The effect of 20 mL of intraarticular bupivacaine (0.25%, with or without 1:200,000 epinephrine), morphine (0.03%, with or without 1:200,000 epinephrine), or normal saline on postoperative analgesia after arthroscopic knee surgery was studied in a randomized, prospective, double-blind trial in ASA I-III outpatients receiving general anesthesia (n = 112) or regional anesthesia (n = 27 [spinal (n = 25) or epidural (n = 2)]). The visual analogue pain scores in the postanesthesia care unit and 3, 6, 12, and 24 h after surgery, time to first analgesic use, and total 24-h analgesic requirements were recorded. In those who received general anesthesia, the visual analogue scores were significantly lower in the bupivacaine group compared with both the morphine- and placebo-treated patients (P less than 0.05). The time to first analgesic use was longer in both the bupivacaine and morphine groups when compared with the control group (P less than 0.05). No significant differences were detected in total 24-h analgesic requirements among the groups. Patients who had received regional anesthesia had lower visual analogue scores compared with patients who had received general anesthesia irrespective of the intraarticular treatment (P less than 0.05). Our results indicate that intraarticular injection of bupivacaine after arthroscopic knee surgery provides prolonged analgesia but that there is no significant prolonged analgesia provided by intraarticular morphine.