Following arthroscopic surgery of the knee, 60 ASA I-II patients were randomly allocated to three different groups (n = 20 each) in an attempt to establish the best postoperative analgesic protocol: 20 ml of bupivacaine (0.25%; 50 mg) intra-articularly (IA) (Group 1); continuous three-in-one lumbar plexus block using a catheter to administer 0.25 ml/kg of bupivacaine (0.25%; single-shot) plus continuous pump infusion of the same drug (0.03 ml.kg-1 x h-1 of 0.25% bupivacaine for 24 h) (Group 2); or 1 mg of morphine with 20 ml of saline IA (Group 3). The results were appraised in a double-blind manner, and the degree of postoperative pain was evaluated by visual analog scale (VAS) at constant hourly intervals for the first 24 h, by the incidence of secondary effects, and by the need for complementary analgesia. Group 2 (three-in-one) VAS values were lower 16 and 24 h after surgery and also globally (P < 0.05), with respect to the other two groups. Five patients in Groups 1 and 3 required complementary analgesia, while three patients in Group 1 suffered nausea. No other secondary effects were observed. We conclude that all three analgesic methods proved efficient, as reflected by the scant requirements for additional analgesics and the degree of expressed patient satisfaction. However, lumbar plexus block (three-in-one) with continuous bupivacaine infusion was most effective, and, although it constitutes an added technique, its scant complications and easy performance make its use advisable.