Mixed venous oxygen saturation in critically ill septic shock patients. The role of defined events.

STUDY OBJECTIVE To investigate the frequency and extent of spontaneous changes ("events") in continuously measured mixed venous oxygen saturation (SvO2) in septic patients and to determine whether attention to individual event-frequency offers additional information for patient management. DESIGN Nonrandomized prospective study. SETTING General intensive care unit at a university hospital. PATIENTS Fifteen patients suffering from septic shock and multiple organ dysfunction syndrome. MEASUREMENTS For the continuous assessment of SvO2 a fiberoptic pulmonary artery catheter (Baxter Edwards) was inserted in all patients. A certain event was defined as a sudden change in SvO2 of > or = 5 percent lasting for > 10 min. All events were grouped as either moderate (< or = 10 percent changes in SvO2) or severe events (> 10 percent changes). Hemodynamics and inotropic support, oxygenation and ventilatory support, hemoglobin levels and body temperature were determined at the event and compared with the ultimate values registered before the event. RESULTS We evaluated 377 events during an observation period of 1,575 h. Patients' mean SvO2 levels ranged between 72 +/- 7 and 82 +/- 4. Desaturations below 65 percent (39 out of 377 events) occurred in 11 patients. Overall, 74 percent of all events were moderate and 26 percent were severe. The incidence of events was 5.6 +/- 1.5 during 24 h in survivors (n = 10) and 6.3 +/- 1.6 during 24 h in nonsurvivors (n = 5). While in survivors only 20 percent of all events were severe events, this portion was significantly higher in nonsurvivors (34 percent; p = 0.03). In 67 percent of all events we observed changes in the registered physiologic parameters or therapeutic interventions probably causing the event. The cause of the remaining 33 percent of all events could not be elucidated. CONCLUSIONS The SvO2 of septic shock patients is mainly normal or even supranormal. However, short-term changes in SvO2 do occur frequently in these patients. Nonsurvivors exhibit a higher frequency as well as a significantly greater severity of events, which may point toward a concealed mismatch of oxygen supply and demand. A high incidence of short-term SvO2 changes in a septic shock patient may be of diagnostic and prognostic significance. Therefore, we recommend the installation of a computerized alarm-function for the automatic detection and indication of frequent events.