Is central venous blood an acceptable sample for blood gas analysis?

Blood gas analysis involves measurement of three parameters: pH, partial pressure of carbon dioxide (pCO₂) and partial pressure of oxygen (pO₂). Derived (calculated from these measured values) parameters also generated during blood gas analysis include: bicarbonate (HCO₃) concentration, base excess (BE) and oxygen saturation (sO₂).

Blood gas analysis has two principal purposes: assessment of patient acid-base status and assessment of patient blood oxygenation status. pH, pCO₂ HCO₃ and BE are the parameters used to assess acid-base status, and pO₂, sO₂ are used to assess oxygenation of blood.

The ”gold standard” sample for these measurements is arterial blood collected by needle puncture of an artery or via an indwelling arterial catheter. Although there is an absolute requirement for arterial blood if pO₂ and sO₂ (measures of patient oxygenation) are required, there is accumulating evidence to suggest that venous blood might be an acceptable alternative, if only the acid-base status of patient under investigation is of interest. 

Most intensive care patients who require blood gas analysis have a central venous catheter that allows sampling of venous blood for all laboratory investigations. It would be logistically convenient for clinical staff and more comfortable and safer for the patient if this kind of sample could also be used for blood gas analysis to determine acid-base status. 

The validity of this approach has been tested in several studies, the most recent of which was conducted at the pediatric cardiac surgery intensive care unit of a tertiary care hospital in India. The focus of this study was 30 pediatric patients who required assessment of acid-base status during their intensive care recovery from cardiac surgery. 

For each of these patients blood was sampled simultaneously from an arterial line and via a central venous line. Both samples were submitted for blood gas analysis. A correction factor of +0.05 pH units was applied to measured central venous pH values and a correction factor of –5 mmHg (0.6 kPa) was applied to measured central venous pCO₂ values. 

These adjustment factors are arterio-venous differences determined by previous study. Statistical comparison of adjusted central venous values for pH and pCO₂ with those obtained from arterial samples revealed generally good correlation (pH 0.9544, pCO₂ 0.8738) and acceptable levels of agreement, particularly with regard to pH. 

Bland-Altman plots comparing adjusted central venous pH and arterial pH revealed a small mean bias of 0.008 pH units with limits of agreement of –0.04 to +0.057 pH units. 

Bland-Altman plots comparing adjusted central venous pCO₂ and arterial pCO₂ revealed a mean bias of –3.52 mmHg (–0.47 kPa); limits of agreement were –9.68 to +2.65 mmHg (–1.3 to +0.35 kPa) which the authors judge to be too wide to allow substitution of central venous pCO₂ for arterial pCO₂. The paper includes correlation data and limits of agreement data relating to HCO₃, BE and lactate. 

In summary of their findings the authors suggest that their results indicate that it might be valid to use central venous blood as a substitute sample for trending of acid-base status. 

They concede limitations of their small study and suggest more research is required before a recommendation can be made that central venous blood is an acceptable alternative to arterial blood for assessment of acid-base status in pediatric intensive care patients.