The transcutaneous measurement of oxygen (tcP(O2)) and carbon dioxide (tcP(CO2)) tensions may serve as a surrogate of arterial oxygen (Pa(O2)) and carbon dioxide (Pa(CO2)) tensions, respectively. We investigated the effects of the electrode temperature of a new device, TCM4, on the measurement of tcP(O2) and tcP(CO2). Twenty-five patients scheduled for major lower abdominal surgery were enrolled. The electrode of the TCM4 was attached to the chest, with its temperature set to 37 degrees C, 40 degrees C, 42 degrees C, 43 degrees C, or 44 degrees C. tcP(O2), tcP(CO2), end-tidal carbon dioxide tension (Et(CO2)), Pa(O2), and Pa(CO2) were simultaneously measured at various Et(CO2) levels and inhaled oxygen concentrations. The times required for stabilization of the tcP(O2) and tcP(CO2) values were measured. A Bland-Altman plot was used to compare the two measurements. The time required for stabilization was shorter with a higher electrode temperature, but the shortest time was still more than 150 s. TcP(O2) correlated well with Pa(O2) at 43 degrees C and 44 degrees C. TcP(CO2) correlated well with Pa(CO2) and Et(CO2) at 43 degrees C. The bias and limits of agreement were larger with lower electrode temperature for TcP(O2)--Pa(O2), tcP(CO2)--Pa(CO2), and tcP(CO2)-Et(CO2). We concluded that the electrode of the TCM4 should be heated to at least 43 degrees C to measure tcP(O2) and tcP(CO2). However, the absolute values of tcP(O2) and tcP(CO2) could not be used as surrogate measurements of Pa(O2) and Pa(CO2), respectively.