OBJECTIVE It has been argued that venous oxygen saturation from a central venous catheter (ScvO(2)) could be an inexpensive alternative to mixed venous oxygen saturation (SvO(2)). The aim was to evaluate whether ScvO(2) measurements could replace SvO(2) readings after cardiac surgery and to analyze factors influencing any differences found. METHODS A prospective observational study. METHODS A university hospital. METHODS Twenty patients scheduled for elective cardiac surgery. METHODS Patients were followed postoperatively with corresponding ScvO(2)/SvO(2) measurements. RESULTS The overall bias between ScvO(2) and SvO(2) was 1.9. In coronary artery bypass graft (CABG) patients, the bias was 0.6 compared with 6.4 in procedures involving aortic valve replacement. In situations with peripheral saturation (SAT) <92%, the bias was 10.7 compared with 0.8 when SAT was >or=99%. In 25.5% of measurements, the ScvO(2) was more than 10% different from SvO(2), and in only 50% the difference was less than 5%. CONCLUSIONS The ScvO(2) and SvO(2) measurements are not interchangeable, and, especially in patients undergoing aortic valve surgery, this lack of agreement is crucial. However, the present data indicate that ScvO(2) may be used in CABG patients, although it is not completely accurate in terms of absolute venous saturations. A low SAT, low hemoglobin, or low cardiac index increased the venous gap.