1115

Abstract

Introduction: Severe sepsis is a prevalent disease with approximately 30% mortality. Currently, resuscitative efforts aim to achieve predefined hemodynamic and physiologic goals. The role of CO2 metabolism as reflected by arterial CO2 (PaCO2), central venous CO2 (CvCO2), and end tidal CO2 (ETCO2) is less well understood. Our goal was to determine the relationship between ETCO2 and PaCO2 and CvCO2 in patients undergoing early quantitative resuscitation in the emergency department (ED). Methods: Adult patients meeting criteria for severe sepsis or septic shock (lactate > 4 or hypotension not responsive to 20 mL/kg isotonic fluids) were approached for enrollment. The study protocol began at the start of quantitative resuscitation. For arterial blood gases (ABGs) or central venous blood gases (VBGs) obtained during the 6 hour resuscitative period, ETCO2 was measured within 5 minutes of arterial or venous blood draw. Three ETCO2 measurements were taken during a one-minute timeframe and averaged together. ABGs or VBGs were ordered at the discretion of the treating physicians. Factors that could potentially affect arterial, venous, or end-tidal CO2 levels were documented. Results: There were 42 patients with either ABG or VBG data and comparison ETCO2 measurements. Thirty four patients survived to discharge and 8 expired prior to discharge. There were 26 patients (20 survived, 6 expired) with ABG data and 30 patients (23 survived, 7 expired) with VBG data. The mean PaCO2-ETCO2 difference for surviving patients was 9 mm Hg (95% CI, 6, 13), and 16 mm Hg (95% CI, 10, 22) for patients who expired (p=0.05). When comparing only patients with pulmonary dysfunction, the mean difference was 10 mm Hg for survivors and 17 mm Hg for non-survivors (95% CI difference, -13.4, -1.0, p = 0.03). The mean CvCO2-ETCO2 difference was 15 mm Hg (95% CI, 12, 19) for survivors, and 19 mm Hg (95% CI, 9, 30) for nonsurvivors (p=0.36). There were significant differences in lactate, ventilatory status, hospital length of stay (LOS), shock, and pulmonary dysfunction among those who survived versus expired. Conclusions: We identified a promising relationship between initial PaCO2-ETCO2 difference as a predictor of mortality in patients with severe sepsis or septic shock undergoing early quantitative resuscitation. After accounting for pulmonary dysfunction, this difference was still present. We were unable to find a relationship between CvCO2-ETCO2 difference and mortality.