Venous–arterial CO₂ to arterial–venous O₂ difference ratio as a resuscitation target in shock states?

In shock states, aerobically generated carbon dioxide (CO₂) decreases together with oxygen (O₂) consumption (VO₂). When VO₂ becomes dependent on O₂ delivery (DO₂), CO₂ can be produced anaerobically mostly due to bicarbonate buffering of protons produced in excess secondary to the hydrolysis of adenosine triphosphate, so that VCO₂ can exceed VO₂. It has been estimated that anaerobic ATP compensates approximately 6 % of the accumulated O₂ depth [1]. Therefore, a respiratory quotient (RQ) >1 may be interpreted as a sign of anaerobic metabolism, since it may indicate that more CO₂ is produced than O₂ is consumed (Fig. 1), although both are decreased. According to the Fick equation, VCO₂ equals the product of cardiac output by the difference between mixed venous and arterial CO₂ contents (C_mv–aCO₂) whereas VO₂ equals the product of cardiac output by the difference between arterial and mixed venous O₂ contents (C_a–mvO₂) (Fig. 1). By eliminating the cardiac output value, which is common to the numerator and denominator of the RQ (Fig. 1) and taking PCO₂ as a surrogate of CO₂ content, an increased ratio between mixed venous–arterial PCO₂ difference and C_mv–aCO₂ (P _mv–aCO₂/C_a–mvO₂) was shown to be a good indicator of anerobiosis assessed by hyperlactatemia [2]. However, the relationship between CO₂ content and PCO₂ is curvilinear rather than linear and is influenced by the degree of metabolic acidosis, the hematocrit and the O₂ saturation. In this issue of ICM, Ospina-Tascon et al. [3] demonstrate that a C_mv–aCO₂/C_a–mvO₂ >1 predicts outcome early in septic shock, similar to increased arterial lactate concentration. Patients with both an increased C_mv–aCO₂/C_a–mvO₂ and lactate concentration 6 h after the study start had the highest mortality [3]. The authors propose that the C_mv–aCO₂/C_a–mvO₂ could become a resuscitation target [3]. The authors should be congratulated on having conducted this study and having performed the relatively complex calculations of the CO₂ contents. However, before targeting to normalize the C_mv–aCO₂/C_a–mvO₂, several issues should be considered.

Fig. 1

In cases of shock states, tissue hypoxia results in decreased oxygen consumption (VO₂) and aerobically generated carbon dioxide (CO₂) production (VCO₂). However, the global VCO₂ decreased to a lesser extent than VO₂ due to production of anaerobically generated CO₂. Consequently, the VCO₂ over VO₂ ratio increases. Therefore, after elimination of cardiac output (present in both numerator and denominator), the difference between mixed venous and arterial CO₂ contents (C_mv–aCO₂) over the difference between arterial and mixed venous O₂ contents (C_a–mvO₂) should increase in such hypoxic conditions

• 1. Calculating VCO₂ by multiplying C_mv–aCO₂ with blood flow is only valid under steady-state conditions. If poorly perfused tissues regain flow, CO₂ stores are washed out and calculated CO₂ production is likely to be overestimated. This may have happened in patients in the study of Ospina-Tascon et al. [3] who were in the resuscitation phase of septic shock.

• 2. The amount of anaerobically produced CO₂ is low compared to CO₂ produced under aerobic conditions. It is therefore be questioned whether such small amounts can increase the VCO₂ above VO₂. For instance, when DO₂ was stepwise reduced to 16 % of baseline values in an in situ, vascularly isolated, innervated dog limb, VO₂ remained above VCO₂ despite continually increasing RQ [4]. Admittedly, the hindlimb VCO₂/VO₂ relationship may not represent global RQ well.

• 3. The treatment may have influenced the findings of Ospina-Tascon et al. [3]. Patients with high lactate values at 6 h received more norepinephrine than those with normal values, despite a greater number of the former (around 50 %) being treated with vasopressin. Vasopressin may constrict the mesenteric vascular bed [5, 6]. If global blood flow is low, increasing mesenteric lactate production may not be cleared by the liver and arterial lactate may increase. Conversely, if the liver is able to metabolize the excess of lactate, systemic RQ may rise as a consequence of mesenteric dysoxia. Since the groups at study baseline (T ₀) and after 6 h (T ₆) do not represent the same population, it would be interesting to know the treatment in patients who increased versus decreased their C_mv–a CO₂/C_a–mv O₂ and lactate values between T ₀ and T ₆.

• 4. Computation of CO₂ content seems to be cumbersome and subject to errors due to the number of variables included in the formula. This raises the question of its practical use in routine. More than a decade ago, Mekontso-Dessap et al. [2] proposed to calculate the P _mv–aCO₂/C_a–mvO₂ ratio to detect the presence of global anaerobiosis and showed that a value of P _mv–aCO₂/C_a–mvO₂ >1.4 can reliably predict the presence of hyperlactatemia in the general population of critically ill patients. In spite of this interesting finding, the use of this ratio did not become popular for managing critically ill patients, maybe because measurements of blood lactate concentration are easier to obtain. In addition, determination of the P _mv–aCO₂/C_a–mvO₂ ratio as well as the C_mv–aCO₂/C_a–mvO₂ ratio require combined samplings of arterial blood and mixed venous blood and thus require pulmonary artery catheterization, which is less and less performed in the intensive care unit. Central venous blood variables are becoming more popular than mixed venous blood variables. Central venous blood O₂ saturation and the difference between central venous PCO₂ and arterial PCO₂ (P _cv–aCO₂) are recommended to be used to assess the adequacy of cardiac output to the global metabolic conditions, although the quality of evidence is only moderate [7]. During sepsis, where central venous O₂ content (C_cvO₂) and ScvO₂ can be in the normal range despite global tissue hypoxia owing to low O₂ extraction, it has recently been shown that hyperlactatemia and increased P _cv–aCO₂/C_a–cvO₂ ratio can predict the presence of VO₂/DO₂ dependence, whereas ScvO₂ cannot [8]. This may suggest that the P_cv–aCO₂/C_a–cvO₂ ratio could be used as a surrogate for the P _mv–aCO₂/C_a–mvO₂ ratio to assess global tissue hypoxia. It must be further shown that taking central venous CO₂ content instead of pressure, which would be more physiological, can be easily done at the bedside by using simple software able to avoid cumbersome calculations of CO₂ content. For a routine use of these surrogates of RQ, it must also be shown that they respond to changes in global tissue oxygenation faster than blood lactate concentration. Finally, one must keep in mind that all these parameters allow the assessing of global but not regional or local tissue oxygenation, knowing that dissociation between systemic and local blood flows may exist in patients with septic shock [9].

In conclusion, the paper by Ospina-Tascon et al. [3] adds interesting information to metabolic consequences of septic shock in the phase when treatment is administered with the aim to improve DO₂ and VO₂. Whether a C_mv–aCO₂/C_a–mvO₂ ratio above 1 indicates anaerobic metabolism in unstable patients in septic shock and, if yes, why it can be associated with and without arterial hyperlactatemia, should be evaluated in more detail.