During isolated limb perfusion (ILP) severe metabolic impairment with a subsequent alteration in oxygen consumption can be observed. The mechanisms responsible for this may be extracorporeal circulation, hyperthermia, and application of cytostatic drugs and cytokines. Thirty-three patients underwent ILP with rhTNF alpha and melphalan for melanoma or soft-tissue sarcoma. Cardiopulmonary monitoring consisted of arterial and mixed venous blood-gas analysis and a Swan-Ganz catheter was inserted after induction of general anesthesia prior to any surgical intervention. Arterial (SaO2) and mixed venous (SvO2) oxygen saturation, serum lactate and end-expiratory CO2 concentration were determined peri- and postoperatively for 72 h. Oxygen supply and consumption rates were measured systemically (DO2I, VO2I) and in the extracorporeal circuit ('DO2I, 'VO2I). For statistical analysis we used the t-test. During extracorporal circulation an increase of DO2I and VO2I was observed. A slight increase of lactate values began during the wash-out phase. Immediately after reperfusion. DO2I, VO2I and lactate increased significantly with normalization until the 2nd postoperative day. SaO2 and SvO2 remained unchanged. A significant correlation between regional toxicity and the postoperative maximum of serum lactate values was found. The increase of DO2I and VO2I in the tissues during ILP and after reperfusion was achieved by a significant increase in cardiac output while the oxygen extraction rate was not altered. Elevation of lactate values after reperfusion and the increase in oxygen utilization might be due to oxygen depletion in the perfused limb. This could contribute to the development of lactacidosis or rhabdomyolysis. Therefore, to minimize toxicity it seems to be mandatory to measure adequate tissue oxygen supply during ILP.