Measurements of toe temperature and transcutaneous PO2 (PtcO2) have been both suggested for non-invasive assessment of peripheral blood flow in acute circulatory failure. The underlying principle of the two methods is that cutaneous vasoconstriction occurs early when tissue perfusion is altered. In 15 patients, we compared the two measurements during cardiogenic shock (27 measurements) or septic shock (29 measurements). Toe-ambiant temperature gradient and PtcO2 correlated well together (r = 0.66, p less than 0.001) especially in hyperkinetic septic shock (r = 0.79, p less than 0.001). In cardiogenic shock, toe-ambiant temperature correlated well with cardiac index (r = 0.63), stroke index (r = 0.64) and oxygen transport (r = 0.65), and these correlations were stronger than for PtcO2. In septic shock, both techniques were poor indicators of blood flow indexes but PtcO2 rather correlated with arterial pressure (r = 0.66) and left ventricular work (r = 0.66). Trend evaluation of data revealed in cardiogenic shock that the increase in toe temperature usually preceded the increase in PtcO2. Since measurement of PtcO2 is technically more complicated, correlates less well with standard hemodynamic parameters and later reflects cardiovascular improvement, it has no advantage over measurement of toe temperature in circulatory shock. In cardiogenic shock, measurements of toe temperature can reliably track cardiac output changes. In septic states, however, non-invasive assessment of skin perfusion is of limited interest.