Importance of the sampling site for measurement of mixed venous oxygen saturation in shock

Objectives: To determine if oxyhemoglobin saturation in blood samples taken from the superior vena cava or right atrium can be substituted for oxyhemoglobin saturation in blood taken from the proximal pulmonary artery (<S(V)over bar O-2>)in patients in shock. Design: Prospective clinical investigation. Setting: Mixed surgical/medical intensive care unit in a university hospital. Patients: Thirty consecutive patients in severe circulatory shock who required insertion of a pulmonary artery flotation catheter (PAFC) immediately on intensive care unit admission. All patients fulfilled the criteria described below which were established in advance. Measurements and Main Results: Oxyhemoglobin saturation in the superior vena cava, right atrium, and pulmonary artery (<S(V)over bar O-2>,) was measured by cooximetry in consecutive blood samples from each site during initial insertion of the PAFC. The mean standard deviation of values from these sites was similar: 74 +/- 12.5%, 70.6 +/- 13%, and 71.3 +/- 12.7%, respectively. However, when superior vena cava and right atrial oxyhemoglobin saturations and <S(V)over bar O-2>, were compared, the ranges and 95% confidence limits were found to be clinically unacceptable. The ranges were -19.3 to 23.1% and -19.7 to 16.7%, respectively, and the 95% confidence limits were -18.4 to 24.2% and -18.6 to +17.3%, respectively. Conclusions: These wide range differences and confidence limits would lead to large errors if superior vena cava or right atrial oxyhemoglobin saturations were substituted for true mixed venous blood in oxygen transport or pulmonary venous admixture calculations, or if clinical decision making was based on individual results. In patients in shock in whom clinical decisions may be based on the value of mixed venous oxyhemoglobin, oxyhemoglobin saturation is only reliably measured in samples taken from the pulmonary artery.