Knowledge of the actual concentrations of local anesthetic administered by various techniques is essential requisite when undesirable side effects and possible toxicity of a substance are to be evaluated. Therefore, numerous studies of plasma concentrations have been presented, which were carried out with the additional purpose of analyzing the kinetics of different local anesthetics with respect to limiting-value concentrations in the organism. Despite a sufficient degree of precision in the analysis of amide local anesthetics, it is uncertain whether the results of the different studies are comparable, because blood samples have been taken variously from peripheral veins, central veins or arteries. In the present study changes in bupivacaine concentrations were monitored by means of a standardized method consisting in simultaneous sampling of blood in peripheral veins, central veins and arteries. METHODS. Each of 12 patients undergoing orthopedic hip surgery received average 17 ml bupivacaine (0.75%) via peridural lumbar catheter. After the administration of bupivacaine, blood samples were taken simultaneously from peripheral veins, central veins and arteries at 1, 3, 5, 10, 15, 30, 45, 60, and 90 min after injection. Placement of an arterial cannula and central venous catheter was indicated in all patients (hip-joint revision arthroplasty). Quantitative analysis of bupivacaine concentration was carried out by means of high-pressure liquid chromatography (HPLC). All patients had given their informed consent. RESULTS. All patients showed a rapid increase in bupivacaine concentration in the central venous blood within the first few minutes after administration, the maximum being reached between 3 and 10 min after. A similar course was observed with arterial plasma concentrations; absolute values, however, were an average of 10-20% lower at 15 min following administration. Bupivacaine concentrations in peripheral veins rose more slowly and reached a maximum between 15 and 30 min. At 30 min after peridural application the concentration curves in blood from all three sites were similar. DISCUSSION. In earlier studies the influence of the site of blood sampling has often been underestimated. According to our results, central venous and arterial plasma concentrations correspond closely at all times following peridural application. The observed uniform differences in concentrations at the various sites of sampling can be explained by the fact that pulmonary uptake of local anesthetics causes the lower arterial levels. Especially in the early phase of resorption after administration of local anesthetics, the concentration in peripheral blood does not seem to be representative, because an equilibrium is not established between arterial and central venous blood until 30 min after administration at the earliest. In our opinion the peripheral venous concentrations are unreliable, particularly in the early phases, for the evaluation of unwanted effects or toxicity of local anesthetics, because the initial low values and the delayed increase in these could lead to a false sense of security.