Shedding light on venoarterial PCO₂ gradient

As an expression on Fick’s principle, the reduction in cardiac output is associated with a parallel increase in both mixed venoarterial CO₂ content difference (C_mv-aCO₂) and arterial-mixed venous oxygen content difference (C_a-mvO₂). Nevertheless, disproportioned elevations in C_mv-aCO₂ compared to those of C_a-mvO₂ ensue when the anaerobic threshold is reached. This results from anaerobic CO₂ production, secondary to the buffering of anaerobically generated protons by bicarbonate.

The clinical approach to venoarterial CO₂ differences usually relies on partial pressure rather than content difference. Unfortunately, the attempts to track C_mv-aCO₂ through mixed venoarterial PCO₂ difference (P_mv-aCO₂) might be misleading. The relationship between CO₂ content and partial pressure is intricate. Moreover, the estimation of CO₂ content from PCO₂ is troublesome, and calculation algorithms frequently produce unreliable results. Since several factors can modify the dissociation of CO₂ from Hb, P_mv-aCO₂ can fail to reflect C_mv-aCO₂ changes. For example, hemodilution induces opposite changes in P_mv-aCO₂ and C_mv-aCO₂. The high cardiac output that develops in such situation increases P_mv-aCO₂ and reduces C_mv-aCO₂ [1]. Other factors, such as metabolic acidosis and Haldane effect, can also play a major role in this relationship and have strong effects on P_mv-aCO₂, regardless of cardiac output changes [2].

Another focus of confusion might reside in the actual meaning of venoarterial PCO₂ difference. Differently to tissue-arterial PCO₂ difference, P_mv-aCO₂ primarily reflects the changes in systemic blood flow and not in microcirculatory perfusion. Physiologic research helps to understand this question. In an experimental model of endotoxemia, all the PCO₂ differences—P_mv-aCO₂, mesenteric venoarterial and mucosal villi-arterial—increased during the phase of hypodynamic shock [3]. After fluid resuscitation, P_mv-aCO₂ and mesenteric venoarterial PCO₂ difference normalized following the improvement in cardiac output and superior mesenteric artery blood flow. Tissue hypercarbia, however, remained present as an expression of villi microcirculatory hypoperfusion.

Although mixed venous and central venous gases are not interchangeable [4], central venoarterial PCO₂ difference (P_vc-aCO₂) has been used a surrogate for P_mv-aCO₂. It might thus be a good marker of cardiac output, even more sensitive than central venous oxygen saturation [5]. Nevertheless, an observational study found that P_vc-aCO₂ did not correlate with cardiac output but with sublingual microvascular perfusion [6]. It was therefore claimed by some authors that P_vc-aCO₂ might reflect tissue perfusion. This speculation is supported neither by physiology [3] nor by relevant clinical studies. In septic shock, patients with a hyperdynamic profile showed lower P_vc-aCO₂ than those with normal systemic hemodynamics, even though the microcirculatory alterations were similar in both groups [7]. So, the lack of correlation between cardiac output and P_vc-aCO₂ found in septic patients [6] should be explained by modifications in the dissociation of CO₂ from Hb. Disorders such as hemodilution and lactic acidosis are commonly present in septic shock and frequently display microvascular abnormalities. Certainly, the relationship between P_vc-aCO₂ and microcirculation should not be interpreted as a causal phenomenon.

In this issue of Annals of Intensive Care, Mallat et al. [8] report that an acute reduction in arterial PCO₂ from 44 to 34 mm Hg was associated with an increase of 2 mm Hg in P_vc-aCO₂. The authors attributed this finding to the concomitant increase in oxygen consumption (VO₂). Unfortunately, methodological issues might limit the relevance of the conclusions: First, the increase in P_cv-aCO₂ not only was quantitatively minor and insignificant from a clinical point of view, but mainly stayed within the error of the method of PCO₂ measurement. This is especially true when taking into account the error propagation produced during the calculation of the PCO₂ difference. Furthermore, the use of central venous instead of mixed venous gases for computation of VO₂ is questionable [4]. In addition, the subtle change in base excess that appeared during hyperventilation might also explain part of the change in P_cv-aCO₂ [2]. Modifications in Hb levels before and after hyperventilation, which might have affected P_cv-aCO₂ [1], were not reported. A comprehensive discussion about any P_cv-aCO₂ change should consider all its determinants.

The effects of hypocapnia on P_vc-aCO₂ have been previously reported in stable cardiac surgery patients [9]. An experimental study also showed that severe hypocapnia increased gut intramucosal-arterial PCO₂ as a probable consequence of regional and tissue hypoperfusion. In contrast, systemic and regional venoarterial PCO₂ gradients did not change [10]. In this way, the effects of hypocapnia on P_vc-aCO₂ are uncertain.

Although the study from Mallat et al. [8] does not add new physiologic information and has major limitations, it emphasizes that P_vc-aCO₂ is not a straightforward surrogate for blood flow. The messages for physiologists and practitioners should be that P_vc-aCO₂ monitoring might contribute to the assessment of systemic hemodynamics but requires a comprehensive interpretation.