Do PaCO₂ and peripheral venous PCO₂ become comparable when the peripheral venous oxygen saturation is above a certain critical value?

Dear editor,

Peripheral venous blood gas (VBG) analysis is increasingly used as an alternative to arterial sampling in Emergency Departments throughout the world.[1] There are multiple advantages using peripheral venous samples for blood gas analysis - technical ease, reduced pain and fewer complications. The difference in sample site chosen for blood gas analysis between European and Australian centres has been notable for members of our author group, prompting discussion and review of the literature.

The consensus in the literature is that venous and arterial pH, bicarbonate and base excess are relatively interchangeable unless the patient is in a severely shocked state.[2] Peripheral venous partial pressure of CO₂ (PpvCO₂) is reliable as a screen for arterial hypercapnia, considered non-comparable when PpvCO₂ is elevated above the normal physiological arterial range.

In the absence of significant anaerobic metabolism, the amount of CO₂ formed in any tissue is proportional to tissue oxygen delivery via the respiratory quotient. Oxygen delivery to any tissue bed can be measured as [haemoglobin] × (SaO₂ – tissue bed SvO₂). This simple physiology has been used to create a validated method to mathematically “arterialise” peripheral venous blood – reverse modelling the process of oxygen extraction and CO₂ uptake, allowing reliable calculation of arterial CO₂ from a peripheral venous sample.[3] Despite the available evidence base, uptake of mathematical arterialisation of peripheral venous blood gas values remains incomplete.[4]

By the above physiologic rationale, the higher the peripheral venous oxygen saturation (SpvO₂) the closer PpvCO₂ will be to PaCO₂. We use this rationale, the evidence base for mathematical arterialisation and data from a laboratory experiment to generate the hypothesis that above a certain threshold SpvO₂ (approximately 75%) the difference between PaCO₂ and PpvCO₂ will be sufficiently small that it will fall within the limits of clinical agreement.

METHODS

Ethical approval for the experiment was granted by the Ruakura Animal Ethics Committee, Ruakura Agresearch, Hamilton, New Zealand. Interventions - anaesthesia, tracheostomy, ventilation, arterial and venous access were undertaken in rabbits as part of a toxicology protocol which has been previously reported.[5] At the completion of the toxicology protocol a convenience sample of contemporaneous arterial and venous blood gases was taken. Blood gases were analysed using an iSTAT Handheld point of care analyser (Abbott Point of Care in, Princeton, USA). SvO₂ and PvCO₂, were measured from venous samples taken from the internal jugular vein, inferior vena cava and ear vein; PaCO₂ was measured from a central artery cannula. Presence of a relationship between SvO₂ and PvCO₂-PaCO₂ was investigated with linear regression using the vassarstats website (www.vassarstats.net).

RESULTS

There were 23 paired venous and arterial gases for analysis. The difference between arterial and venous PCO₂ decreased with increasing SvO₂. The slope of SvO₂ vs PvCO₂-PaCO₂ was -0.14 mmHg/%saturation (99% confidence interval -0.25 to -0.03). This is displayed graphically in Figure 1.

Figure 1.

SvO₂ vs. PvCO₂-PaCO₂.

DISCUSSION

Intuitively, the degree of difference in PCO₂ between arterial and venous blood is linked to the amount of oxygen extracted from arterial blood. This concept has been extended to a process of mathematical arterialisation of venous blood gas measures, which has a significant evidence base. One of the tenets of this method is that the difference in CO₂ carriage between arterial and venous blood is dependent on oxygen extraction – which will be necessarily greater with lower venous oxygen saturations and lesser with higher venous oxygen saturations. Our laboratory data are consistent with this argument.

The premise upon which this hypothesis rests is that all CO₂ entering venous blood is formed from aerobic metabolism. In the setting of ischaemia tissue bicarbonate buffers the acids formed from anaerobic metabolism, leading to anaerobic CO₂ formation. In the setting of vasodilatory shock with “wasted” perfusion the central venous O₂ saturation may remain high while tissue bicarbonate buffering in poorly perfused tissues increases central venous CO₂. This forms the basis of the “CO₂ gap” concept where an increase arterial to central venous CO₂ partial pressure difference, even in the setting of high central venous oxygen saturations, is considered to be an indicator of inadequate overall perfusion.[6] This cause of increased arterial to venous CO₂ difference seems unlikely to relate to peripheral blood sampling in stable patients.

Taking these arguments and laboratory findings together we hypothesise that, in the absence of shock or local ischaemia, there is a SpvO₂ above which PpvCO₂ provides a clinically acceptable estimation of PaCO₂. Large databases of contemporaneous arterial and venous blood taken in clinical practice exist. It would be possible to pursue this hypothesis further using such data. If this hypothesis were proven it could both decrease use of PpvCO₂ when inappropriate, and potentially reduce unnecessary arterial sampling.