We assessed the efficacy of intravenous patient-controlled analgesia (PCA) and femoral nerve PCA after trochanteric fracture repair. A total of 40 patient were randomised to receive either femoral nerve PCA (Group I) or intravenous (IV) PCA (Group II) in the postoperative period. Group I received a continuous infusion of 0.125 bupivacaine at rate of 4 ml h-1 plus PCA boluses of 3 ml with lockout time of 20 min. Morphine by IV PCA was applied as 1 mg bolus dose and 7 min lockout time in Group II. Pain scores (active-passive movement and during the resting), side effects, and patient satisfaction were recorded. Pain scores at rest were similar in both groups. Significantly better pain control at movement was observed in the Group I. Side effects were observed significantly more frequently in the Group II. Patient satisfaction was greater in Group I. Of the two PCA techniques tested, femoral nerve PCA with continuous infusion provides greater patient comfort both at rest and with ambulation than does IV PCA. © 2003 Elsevier Science B.V. All rights reserved.