Tissue oxygenation monitoring using resonance Raman spectroscopy during hemorrhage.

BACKGROUND The ability to monitor the patient of hemorrhage noninvasively remains a challenge. We examined the ability of resonance Raman spectroscopy to monitor tissue hemoglobin oxygenation (RRS-StO2) during hemorrhage and compared its performance with conventional invasive mixed venous (SmvO2) and central venous (ScvO2) hemoglobin oxygen saturation as well as with near-infrared spectroscopy tissue hemoglobin oxygenation (NIRS-StO2). METHODS Five male swine were anesthetized and instrumented followed by hemorrhage at a rate of 30 mL/min for 60 minutes. RRS-StO2 was continuously measured from the buccal mucosa, and NIRS-StO2 was continuously measured from the forelimb. Paired interval measures of SmvO2, ScvO2, and lactate were made. Pearson correlation was used to quantify the degree to which any two variables are related. Receiver operating characteristic (ROC) area under the curve values were used for pooled data for RRS-StO2, NIRS-StO2, SmvO2, and ScvO2 to compare performance in the ability of tissue oxygenation methods to predict the presence of an elevated arterial blood lactate level. RESULTS Sequential RRS-StO2 changes tracked changes in SmvO2 (r = 0.917; 95% confidence interval [CI], 0.867-0.949) and ScvO2 (r = 0.901; 95% CI, 0.828-0.944) during hemorrhage, while NIRS-StO2 failed to do so for SmvO2 (r = 0.283; 95% CI, 0.04919-0.4984) and ScvO2 (r = 0.142; 95% CI, -0.151 to 0.412). ROC curve performance of oxygenation measured to indicate lactate less than or greater than 3 mM yielded the following ROC area under the curve values: SmvO2 (1.0), ScvO2 (0.994), RRS-StO2 (0.972), and NIRS-StO2 (0.611). CONCLUSION RRS-StO2 seems to have significantly better ability to track central oxygenation measures during hemorrhage as well as to predict shock based on elevated lactate levels when compared with NIRS-StO2.