Central mixed venous oxygen saturation (SvO2) monitoring in critically ill patients to estimate adequacy of peripheral perfusion is gaining increasing popularity. However, a number of unexpected responses, one of which is marked depression of regional (splanchnic) venous oxygen saturation which may coexist with normal or high SvO2, makes interpretation of this parameter difficult. The SvO2 and hepatic venous oxygen saturation levels in seven injured (postoperative) and 15 septic patients were measured. No substantial differences between central and hepatic venous oxygen saturation was noted in nonseptic patients, however, septic subjects exhibited a normal SvO2 of 70.5% +/- 8.7% at a time when the hepatic venous saturation was 55.6% +/- 14.4% which is a significant (p less than 0.05) reduction. This reduced oxygen saturation was noted to arise from an increased regional metabolic rate rather than reduced perfusion. Nevertheless, we conclude that a flow limited regional oxygen consumption may potentially exist despite the presence of a normal SvO2 in certain patient subgroups such as septic subjects. Therefore, a normal SvO2 should not be considered as sole criteria to insure optimal oxygen delivery in critically ill patients.