Surveillance continue de la S au cours de l'anesthésie pour chirurgie pulmonaire

The multiplicity of potential causes of variations in mixed venous oxygen saturation (Sv2) during one lung ventilation (OLV), including a constant ventilation/perfusion mismatch, explains that it has been suggested as a routine monitoring procedure. To assess its usefulness, 12 adults undergoing OLV were monitored during surgery with an Oximetrix® pulmonary catheter, placed on the side opposite to the surgical field under fluoroscopic control. Seventy two complete sets of haemodynamic measurements were obtained at 6 different times during surgery. We studied the ability of changes in Sv2 to predict changes in arterial oxygen saturation (Sao2), cardiac output (CO), and venous admixture (VA) by calculating sensitivities (Se), specificities (Sp) and predictive values with regard to these variables. There were no complications due to the protocol. However left-sided catheter placement failed in four cases. Correlation between optical and measured Sv2 was very strong (r=0.94; p<0.001). Sv2, oxygen consumption (V˙o2) and the rate of oxygen extraction remained constant throughout the procedure, even when CO, mean arterial pressure, VA, Sao2 and Pao2 varied. Clamping the pulmonary artery returned VA, Sao2 and Pao2 values to those found before OLV, but produced a significant decrease in CO. Sv2 had low Se and Sp for changes in other variables (CO: 76±7, 48±9; Pao2: 79±6, 59±9; VA: 54±7, 48±7 respectively). In this type of surgery, alterations in variables related to oxygen are probably balanced by haemodynamic changes. In fact, according to Fick's formula, Sv2 is almost completely determined by Sao2 and CO, when V˙o2 and haemoglobin remain stable. However, a decrease in Sv2 should alert the clinician to a potential change in haemodynamic indices or to a problem in some component in the oxygen delivery system. Sv2 values do not add to or correlate well with other data, but facilitate their continuous monitoring.