Sublingual capnometry versus traditional markers of tissue oxygenation in critically ill patients.

Abstract

The purpose of this study was to determine the prognostic value of sublingual PCO2 (P(SL)CO2), lactate concentration, and mixed venous oxygen saturation (S(MV)O2) in hemodynamically unstable intensive care patients and, additionally, to compare the temporal changes of these variables in response to treatment. Medical/surgical intensive care unit. Fifty-four patients, mean age 58 +/- 8 yrs. Oxyhemodynamic variables, arterial lactate concentration, and P(SL)CO2 were recorded in unselected sequential intensive care patients undergoing pulmonary artery catheterization. A data set was obtained immediately after insertion of the pulmonary artery catheter and repeated 4 and 8 hrs later. Twenty-one patients had severe sepsis or septic shock. Twenty-seven (50%) patients died. The initial P(SL)CO2_PaCO2 gradient (P(SL)CO2-diff) and the initial P(SL)CO2 were highly predictive of outcome (p =.0004 and p =.004, respectively); however, there was no difference in the arterial lactate concentration and S(MV)O2 between the survivors and nonsurvivors. The P(SL)CO2-diff had the best receiver operator characteristic characteristics (area under the curve, 0.75), with a P(SL)CO2-diff >25 mm Hg being the best discriminator of outcome. With treatment, the P(SL)CO2-diff decreased in both survivors and nonsurvivors; however, the lactate and S(MV)O2 remained relatively unchanged during the study period. The baseline P(SL)CO2-diff and P(SL)CO2 were better predictors of outcome than traditional markers of tissue hypoxia and were more responsive to therapeutic interventions. The P(SL)CO2-diff and/or P(SL)CO2 may prove to be a useful marker for goal-directed therapy and for assessing the response to clinical interventions aimed at improving tissue oxygenation.