Is a Central Venous Blood Gas a Useful Surrogate for an Arterial Blood Gas?
Abstract & Commentary
By Saadia R. Akhtar, MD, MSc St. Luke's Idaho Pulmonary Associates, Boise. Dr. Akhtar reports no financial relationship to this field of study. This article originally appeared in the October 2010 issue of Critical Care Alert. It was edited by David J. Pierson, MD, and peer reviewed by William Thompson, MD. Dr. Pierson is Professor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington, Seattle, and Dr. Thompson is Associate Professor of Medicine, University of Washington. Drs. Pierson and Thompson report no financial relationships relevant to this field of study..
Synopsis: This observational study compared simultaneous arterial and central venous blood gases in a mixed patient population and found high agreement between normal gases after the application of a predefined adjustment algorithm.
Source: Walkey AJ, et al. The accuracy of the central venous blood gas for acid-base monitoring. J Intensive Care Med. 2010;25: 104-110.
The authors set out to compare arterial and central venous blood gases (ABG and VBG, respectively), to evaluate the utility and accuracy of a predefined algorithm for adjusting VBG pH and pCO2 to approximate that of ABG, and to define clinical scenarios where a VBG may not be a useful substitute for an ABG. The study took place over 2 months at a tertiary care academic medical center in the United States.
Patients older than age 18 years admitted to any adult ICU or the cardiac catheterization laboratory with a central venous line in place and a need for a blood gas were eligible. Simultaneous ABG (radial or femoral) and VBG (tip of line in distal superior vena cava or right atrium) were drawn and analyzed by a standard point-of-care system. A predefined algorithm (based on prior literature on arteriovenous pH and pCO2 differences) was used to derive an adjusted VBG (aVBG) by adding 0.05 to the VBG pH and subtracting 5 cm H2O from the VBG pCO2. A clinical acid-base diagnosis (e.g., "metabolic acidosis") was given to each blood gas result. Standard statistical methods were used to measure agreement between ABG and VBG or aVBG results (detailed below). Multiple logistic regression modeling was applied to define patient characteristics most highly associated with ABG-aVBG agreement in clinical acid-base diagnosis.
Of 202 consecutive patients, 187 were enrolled in the study. The majority (56%) of patients were admitted to the cardiac catheterization laboratory and more than half (54%) of patients in the study had normal blood gases. Only 32% of the patients were intubated, 8% required vasopressors, and small percentages had admitting diagnoses such as sepsis (8.4%) or the acute respiratory distress syndrome (ARDS; 2.6%). Venous O2 saturation was < 65% in 30% of subjects. As expected, ABG and VBG pH and pCO2 showed fairly high gross correlation (using scatterplots and Pearson's correlation coefficient). The agreement between the values was more robustly evaluated using Bland-Altman plots, which revealed 95% limits of agreement of 0.025 to -0.057 for pH and 7.5 to -6.5 for pCO2. Agreement between clinical acid-base diagnosis based on ABG and VBG values was low at 45%, but increased to 74% between ABG and aVBG (κ 0.61); it was 90% when the aVBG was normal. Finally, presence of an abnormal aVBG had the highest association with discrepancy with ABG-based clinical acid-base diagnosis; other factors such as presence of sepsis or ARDS or venous O2 saturation < 65% showed no association.
Commentary
An ABG is the simplest and most accurate method for assessment of acid-base status and ventilation that is available to the clinician at the bedside. Unfortunately, there are always occasional patients in whom arterial lines are unable to be placed or ABGs are not able to be obtained (or only with great difficulty). Severe peripheral vascular disease, marked soft-tissue edema, or hematomas related to a prior arterial line are some potential reasons for these difficulties. It would be attractive to have an alternative test or surrogate measure for serum pH and pCO2 in these situations. Patients could also then be spared the risks and discomfort of arterial punctures and arterial lines.
The current study is an interesting one. The authors should be commended for trying to go a step beyond what other investigators have done; they are the first to use an adjustment algorithm for the VBG and the first to consider comparing not just the ABG and VBG results but the associated clinical acid-base diagnoses. The decision to obtain central as opposed to peripheral venous blood gases was prudent as the former are not directly impacted by variations in distal tissue perfusion and thus may be more likely to demonstrate a consistent relationship with ABGs. Unfortunately, this study has a number of limitations and inconclusive results.
The majority of the study population consists of patients coming in for cardiac catheterization; "usual" ICU diagnoses and scenarios are underrepresented, thus the results cannot be generalized to most of our patients. The limits of agreement found between ABG and VBG mean, in simplistic terms, that a venous pH of 7.35 could represent an arterial pH of 7.375-7.293; similarly, a venous pCO2 of 50 could represent an arterial pCO2 of 43.5-57.5. Clinical decision-making and response to the results at the extremes of these ranges could differ considerably; therefore, the study results suggest that a single VBG may not be a very useful surrogate for an ABG. The authors describe similarly disappointing findings when comparing the clinical acid-base diagnoses derived from venous vs. arterial blood gases in their study; even using adjusted VBGs, the κ coefficient of 0.61 (where 0 is no agreement and 1 is perfect agreement) shows only fair-to-moderate agreement. The discussion rightly notes that there were only a small number of blood gases in each "abnormal" diagnostic category so that no clear conclusions can be drawn from the relatively low diagnostic agreement; these data at least do not support a VBG being a good surrogate for an ABG. Finally, this work attempts to evaluate the relationship between some clinical characteristics and the likelihood of discrepancy between an ABG and the aVBG; however, these characteristics are not defined a priori and the numbers of patients with each characteristic are relatively small, putting the validity of this analysis into question as well.
The clearest findings seem to be that the authors' adjustment algorithm is reasonable under normal circumstances and that a normal aVBG (and its diagnosis of normal acid-base status) accurately predicts a normal ABG. This information may be valuable in certain specific clinical scenarios when an ABG is unable to be obtained.
It is important to remember that the results of a single abnormal VBG or aVBG must be considered with caution and should not be the primary data guiding significant clinical decision-making. I suggest that trends in these gases may prove more helpful but, again, only in combination with other clinical data.
This observational study compared simultaneous arterial and central venous blood gases in a mixed patient population and found high agreement between normal gases after the application of a predefined adjustment algorithm.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content