OBJECTIVE The purpose of this study was to compare the effectiveness of three local anesthetic agents for intravenous regional anesthesia in the upper limb. Side effects and plasma concentrations of the drugs in the doses administered for IVRA were also studied. METHODS Thirty patients in ASA groups I and II received intravenous regional anesthesia for surgery of the upper limb. In a double-blind prospective study, they were randomly allocated to receive one of three local anesthetics: articaine, lidocaine, or prilocaine. Patients received 40 mL of a 0.5% solution of the local anesthetic. The onset time of sensory block was assessed by pinprick and the extent of motor block was was scored as 0-3. Plasma concentrations of local anesthetics were determined in all patients from serial arterial blood samples drawn at predetermined times before and after tourniquet release. RESULTS The onset time of sensory block was significantly shorter (2.5 minutes) in the articaine group than in the lidocaine group (11.1 minutes) or the prilocaine group (10.9 minutes) (Scheffe, P < .05). Development of motor block was equal in all three groups (score 2). Estimation of plasma concentrations by high performance liquid chromatography showed that the peak level in all 30 patients was reached immediately after release of the tourniquet; plasma concentrations thereafter gradually declined. Maximum concentrations of articaine, lidocaine, and prilocaine were, 1.85, 8.5, and 4.4 micrograms/mL, respectively. No signs of local anesthetic toxicity of the cardiovascular or central nervous systems were seen. CONCLUSIONS Articaine had the fastest onset of sensory block and the lowest peak plasma concentration of the three local anesthetics when used for intravenous regional anesthesia.