An intraoperative decline in regional cerebral oxygen saturation (rSO2) has been associated with postoperative injury to the central nervous system. Wide individual variation in steady-state cerebral oxygen saturation limits the clinical use of rSO2 to monitoring during anesthesia and surgical procedures. Recently, low preoperative rSO2 has been proposed as a predictor of adverse postoperative outcomes in cardiovascular operations. We compared the sensitivity and specificity of preinduction rSO2 as a predictor of adverse operative events and compared this to the widely accepted risk index developed by the Society for Thoracic Surgeons.Methods
2,097 consecutive white patients who underwent cardiac operations from 2010 through 2012 were included. In 1,496 patients (group 1) the preinduction rSO2 was equal to or greater than 60%, whereas in the remaining 601 patients (group 2) it was below 60%. We compared the predictive accuracy of preinduction rSO2 with that of the STS mortality risk score by means of standard statistical techniques, including a receiver operating curve characteristic analysis.Results
Patients with a preinduction rSO2 below 60% had significantly higher STS mortality risk scores than did patients with an rSO2 equal to or greater than 60% (2.0 vs 4.0, p < 0.001). Those with an rSO2 below 60% experienced higher operative mortality (p < 0.001) and after adjustment this determination emerged as an independent predictor of increased mortality (p < 0.001). Receiver operating characteristic curve analysis demonstrated that the rSO2 was slightly less accurate as a mortality predictor (area under the curve: 0.71 vs 0.85).Conclusions
Measurement of rSO2 is considerably less complex than calculation of the STS score and is only slightly less accurate as a predictor of operative mortality. It may be useful when the STS mortality risk score cannot be calculated.