Transcutaneous oxygen measurement (PtcO2-measurement) gives by a non-invasive technique continuous information about central PO2. The method has as yet been tested mainly on newborns, and systematic studies of the applicability of the method on adults are still few. The current systematic study reports on consecutive PtcO2 measurements in 15 patients, all with indwelling arterial catheters: 13 of the patients were artificially ventilated; 5 patients had normal body structure and normal skin circulation; 5 patients were obese and had normal skin circulation; and 5 patients were hypothermic with a normal heart frequency or slight bradycardia. The Radiometer equipment TCM1 was used. The electrode surface tempertaure was 45.0 degrees C (within +/- 0.2 degrees C). The oxygen concentration in the inspired air was altered several times for all patients, and arterial samples were drawn at a stable PtcO2 level. On average, seven samples were drawn from each patient. The correlation between PtcO2 and PaO2 was highly significant in 4/5 patients, in both the "obese patient" group and the "normal patient" group. In the "hypothermic patient" group, the correlation between PtcO2 and PaO2 was highly statistically significant in 2/5 patients. The regression line between PtcO2 and PaO2 was calculated for each patient, and a mean regression line was obtained for each group. It was found that the regression line in the "obese patient" group was displaced to the right and parallel to the regression line in the "normal patient" group, while the regression line in the "hypothermic patient" group was less sloping and showed a higher intercept. The PtcO2 monitoring technique reliably reflects changes in the central oxygenation of adult intensive care patients. In some patients with decreased skin circulation, the relationship between PtcO2 and PaO2 seemed to be constant for the same patient, yielding statistically sifnificant correlations. However, the study cannot determine whether this would be true in all clinical situations; the relationship betwee PtcO2 and PaO2 should therefore be checked frequently when measruements are performed in patients with decreased skin circulation. The author's experience with the PtcO2 monitoring method in patients near death has been consistent with the obvious fact that the reliability of the method decreases with decreased skin circulation.