Arterial blood gas (ABG) sampling is currently the standard for determining acid-base status and oxygenation of patients on mechanical ventilation in the intensive care unit (ICU). Previous studies have shown that a central venous blood gas (VBG) can substitute for an ABG in patients with COPD, diabetic ketoacidosis and trauma. However, it is unknown whether these measurements correlate in patients with sepsis.
The objective is to determine if there is agreement between ABG and VBG values for pH and PCO2, and whether a VBG can be a surrogate for an ABG in septic patients in the ICU.
A single-center, prospective study was performed on septic medical ICU patients with central venous and arterial catheters. 95 pairs of blood gas samples from 64 patients were drawn within 15 minutes of each other. Agreement between the pH and PCO2 of VBG and ABG samples was evaluated using the Bland-Altman bias method. Validity was assessed using the adjusted VBG rule: Arterial pH=Venous pH+0.05 and Arterial PCO2=Venous PCO2–5.
The mean pH difference between ABG and VBG samples was 0.03, with 95% upper and lower limits of agreement (LOA) of 0.11 and -0.06, respectively. The magnitude of difference increased with decreasing venous pH (p<0.009), decreasing venous O2 saturation (p<0.002), and increasing venous PCO2 (p< 0.03). The mean difference between arterial and venous PCO2 was -6.09, with upper and lower LOA of 7.83 and -20.01, respectively. The magnitude of the difference in PCO2 increased with decreasing venous pH (p<0.013), decreasing venous O2 saturation (p<0.002), and increasing venous PCO2 (p<0.001). After VBG adjustment, the mean VBG pH decreased from 0.03 to -0.02, with LOA between 0.06 and -0.11. The mean VBG PCO2 decreased from -6.09 to -1.09, with LOA between 12.83 and -15.01.
Our study demonstrates a correlation of pH and PCO2 between a VBG and ABG samples in septic patients, although the difference widens at extremes of pH and PCO2. Previously formulated adjusted VBG formulas are also less accurate in these patients. Therefore, neither VBG nor adjusted VBG is a good surrogate for an ABG, and should be used with caution when making clinical decisions in septic patients.