Forty-five ASA physical status I volunteers, divided in three groups of 15 each, received intravenous regional anesthesia (IVRA) of the upper limb with 40 mL meperidine 0.25%, lidocaine 0.5%, or 0.9% sodium chloride (isolated ischemia) by random allocation. Using a double-blind method, the onset and recovery of sensory block was tested at six sites of the forearm and hand. The onset of complete motor block was also assessed. The symptoms after deflation of the tourniquet were recorded. The onset of block, as determined by pin-prick touch, and cold was significantly faster in the meperidine group (P < 0.001) than in the saline group, but also slower (P < 0.001) than in the lidocaine group. After the tourniquet was deflated, recovery occurred in reverse order. A complete motor block was noted in all volunteers from the meperidine and lidocaine groups, but in only 11 cases from the 0.9% sodium chloride group (P < 0.01). In the meperidine group, motor block developed concomitantly or prior to sensory block. There was a significant increase in the incidence of dizziness, nausea, and pain at the injection site in the meperidine group in comparison with the lidocaine group. We conclude that meperidine has local anesthetic action on the peripheral nerve in vivo, but that its single use for IVRA should be a second choice for patients allergic to local anesthetics.