Abstract
Acid-base status is frequently assessed in severe sepsis and septic shock. Venous blood gas sampling is proposed as a less-invasive modality but lacks evidence within this population. The objective of this study was to evaluate the correlation and agreement between arterial blood gas (ABG), peripheral venous blood gas (pVBG), and central venous blood gas (cVBG) in severe sepsis and septic shock. We conducted a prospective, observational cohort study in subjects admitted to the medical intensive care unit. Simultaneous blood gas samples, including ABG, pVBG, and cVBG, were analyzed for correlation and agreement. Severity of illness scores revealed a mean (±SD) Sequential Organ Failure Assessment score of 7.9 ± 3.3, Simplified Acute Physiology II score of 49.3 ± 16.5, and a mortality rate of 11.9% in the intensive care unit and 16.4% in the hospital. We found a strong intraclass correlation (>0.85) for pH, partial pressure of carbon dioxide (pCO2), bicarbonate, and base excess for ABG/pVBG, ABG/cVBG, and pVBG/cVBG comparisons. Agreement by the Bland-Altman method was found for pH (bias ± SD, 0.03 ± 0.04, 0.03 ± 0.02, and 0.00 ± 0.03) but not for pCO2, partial pressure of oxygen, bicarbonate, base excess, and oxyhemoglobin saturation. In conclusion, adequate correlation and agreement between ABG/pVBG, ABG/cVBG, and pVBG/cVBG comparisons was found only for pH. The current level of evidence does not support the use of venous blood gas sampling in this setting. © 2018, Copyright © 2018 Baylor University Medical Center.