Rising Lactate & the Art of Venous Blood Gas Interpretation
A 23 year old woman is admitted with severe abdominal pain following 5 days of profound non-bloody diarrhea and 72 hours of recalcitrant non-bloody emesis. She has lost 7 pounds in this time frame and has been unable to maintain oral hydration. Her eyes are sunken and her vital signs are notable for a heart rate of 144, blood pressure 70/55, her respiratory rate is 32 and she is afebrile. A CT of her abdomen reveals significant colitis and ileus and a point of care venous blood gas demonstrates the following: pH 7.19, PvCO2 91 mmHg and venous oxygen saturation of 91%. Her lactate is 8.0 mmol/L; a sepsis alert is called.
A & B Lactates
The distinction between the A-type lactate and B-type lactate in this post will rest on the notion of true oxygen deficiency. Oxygen deficiency will be defined as cessation of oxidative phosphorylation and the Krebs Cycle. When oxygen is unable or unavailable to act as the terminal electron acceptor within the mitochondrion, the electron transport chain ceases to function and protons within the mitochondrial inner membrane diffuse out. Pyruvate levels rise, lactate levels rise and acidosis ensues. This is true anaerobic biochemistry*! This is the A-type lactate elevation [see figure 1A].
Figure 1A: The A-type lactate. Note that oxidative phosphorylation [i.e. the electron transport chain] within the mitochondrion is deficient in oxygen and has stopped. The Krebs Cycle ceases and rogue protons must be buffered. Pyruvate cannot enter the mitochrondrion and it is shunted to lactate. The lactate:pyruvate ratio rises.However, “oxygen deficiency” is often invoked at the bedside when the electron-transport chain is actually fully functional – just slow. What is imperative to understand here, is that the breakdown of glucose to pyruvate is much faster than the Krebs Cycle can handle. When pyruvate levels rise in this situation – as occurs in response to the stress response [e.g. sepsis, exercise] and/or adrenergic tone – pyruvate is converted to lactate in a proton-neutral process [i.e. degrading glucose to pyruvate produces 2 protons and 2 pyruvate molecules and converting 2 molecules of pyruvate to lactate consumes 2 protons]. Lactate levels rise, but the Krebs Cycle slowly lumbers along – without oxygen deficiency! This is a B-type lactate.
Figure 1B: The B-type lactate. Note that oxidative phosphorylation [i.e. the electron transport chain] remains fully functional, so too does the Krebs Cycle. However, the 'stress response' [e.g. sepsis, exercise, exogenous catecholamines] increases glycolytic flux relative to Krebs flux. The lactate:pyruvate ratio remains the same [note there are multiple other mechanisms of B-type elevation e.g. pyruvate dehydrogenase deficiency].B-type lactate elevation may actually be the rule in sepsis; fully aerobic* lactate elevation in sepsis is multifactorial, but is often due to the effect of adrenaline on muscle fibers. Adrenaline activates glycolytic flux and therefore raises lactate without any embarrassment of the slowly moving Krebs Cycle. The provision of fluid in sepsis lowers adrenergic tone such that falling lactate following volume resuscitation physiologically mirrors the fall in heart rate; thus, the differential diagnosis of lactate elevation should be very similar to that of sinus tachycardia.
Tissue Oxygen Consumption & Tissue Carbon Dioxide Production
When cardiac output falls, so too does the delivery of oxygen [DO2] to tissue. Notably, as DO2 falls, the consumption of oxygen [VO2] by the tissues remains strikingly constant until a critical value of oxygen delivery [cDO2]. The reason that VO2 remains constant is because the tissues extract more oxygen from hemoglobin. This flat portion of the VO2/DO2 curve represents aerobic* physiology and the tissues continue to consume the same amount of oxygen. Because oxygen consumption is constant, carbon dioxide production is also constant. The ratio of carbon dioxide production to oxygen consumption [i.e. VCO2/VO2] is 0.7- 1.0 and this is commonly referred to as the respiratory quotient [R]; it depends largely on carbohydrate intake [see figure 2].
Figure 2: The DO2-VO2 relationship. Hypothetical venous carbon dioxide tension and venous oxygen saturation is noted below the graph [assuming an input PaCO2 of ~ 40 mmHg and SaO2 of 100%]. In practice the critical oxygen delivery [cDO2] is variable and may change during disease course. DO2 is oxygen delivery and is the product of cardiac output and blood oxygen content. VO2 is oxygen consumption while VCO2 is carbon dioxide production. Under aerobic* conditions [green portion] the ratio of oxygen consumption to carbon dioxide production remains constant. Below the cDO2 [red portion] the ratio of carbon dioxide production to oxygen consumption rises.When interpreting a venous blood gas, the aforementioned physiology is imperative to understand. While on the aerobic* portion of the curve, as cardiac output and/or oxygen delivery falls, the venous oxygen saturation [SvO2] should fall as more oxygen is extracted to maintain constant VO2. Additionally, as blood flow falls, the venous effluent will have a higher partial pressure of carbon dioxide [PvCO2]. The reason is that, because oxygen consumption is constant, CO2 production also remains constant, however, the lower blood flow results in a comparatively greater amount of carbon dioxide added per unit of blood volume. The normal rise in arterial to venous carbon dioxide partial pressure is less than 6 mmHg.
Another, more advanced way, of thinking about the flat [i.e. aerobic*] portion of this curve is that the ratio of VCO2/VO2 will remain < 1.0, that is, it will recapitulate the respiratory quotient. One could make this calculation by determining the both the VCO2 and VO2 from the Fick Equation. The difference between venous carbon dioxide content [calculated from a VBG] and arterial carbon dioxide content [calculated from an ABG] is directly related to the production of carbon dioxide by tissue [i.e. VCO2]. Similarly, the difference in arterial oxygen content [from an ABG] and venous oxygen content [from a VBG] is directly related to the tissue oxygen consumption [i.e. the VO2]. Accordingly, dividing the venous-arterial carbon dioxide content difference [i.e. VCO2] by the arterial-venous oxygen content difference [i.e. the VO2], yields the respiratory quotient, which is less than 1.0 under aerobic* conditions. A lactate elevation in this situation, would reflect B-type physiology as described above.
The Critical Oxygen Delivery [cDO2]
But what happens when the tissue oxygen consumption [VO2] begins to fall? That is, when tissue becomes truly anaerobic*? Firstly, as oxygen consumption falls, the difference between arterial oxygen content and venous oxygen content narrows. On the venous blood gas, this is suggested by a rising venous oxygen saturation. Crucially, one might expect tissue carbon dioxide production [VCO2] to fall in proportion to the fall in oxygen consumption [VO2] such that the ratio of carbon dioxide production to oxygen consumption remains unchanged. This does not occur, however, because anaerobic* carbon dioxide production rises. Why? As described above, the rogue protons that result when the electron transport chain slows, are buffered by bicarbonate and this increases tissue carbon dioxide. Thus, on the venous blood gas, the PvCO2 continues to rise disproportionately. Therefore, under anaerobic* conditions, the VCO2/VO2 ratio starts to rise. Indeed, this has been recently, convincingly demonstrated.
In a must-read study, Mallat et al. used a number of markers to predict VO2/DO2 dependence. VO2/DO2 dependency is a term used to describe patients who have fallen below the cDO2 and are on the ‘descending limb’ [or anaerobic*] portion of the VO2-DO2 curve. These patients were defined as having an increase in VO2 by at least 15% following a volume challenge. It is assumed, therefore, that only patients who are on the descending limb of the curve will raise their VO2 in response to an increase in DO2. Trying to predict patients on the descending limb failed with respect to central venous oxygen saturation [AUC 0.624] and with lactate [AUC 0.745], but was astonishingly excellent when using a VCO2/VO2 ratio above 1.02 [AUC 0.965]. The likely reason for the poor performance of lactate is the common entity of B-type physiology in sepsis.
Importantly, at the bedside, because the relationship between carbon dioxide content and the partial pressure of carbon dioxide is largely linear, the VCO2 in the numerator can be simply replaced by the absolute difference between the venous and arterial partial pressures of carbon dioxide. Even when this short-hand is used, the predictive power is excellent [AUC 0.962] for a ratio of 1.7 mmHg/mL.
Return to the case
At the outset, we do not have an arterial blood gas for comparison. Nevertheless, there are some clues from the patient’s venous blood gas. Firstly, her venous oxygen saturation is quite high which suggests that oxygen consumption [VO2] is low. Strikingly, her PvCO2 is very high. We do not know the input partial pressure of arterial carbon dioxide [PaCO2], but it is fair to assume that it is much, much lower than 91 mmHg given her young age and tachypnea. Together, these numbers suggest, but do not prove, a high VCO2/VO2 ratio which points towards a patient below her cDO2 with truly anaerobic* tissue. Her lactate almost certainly has some A-type physiology and she should be considered for admission to the ICU.
Lastly, the venous blood gas analysis reflects only the tissue metabolism distal to the point of venipuncture. In clinical practice, this is typically the arm. In the Mallat study, it was central venous blood; if it were venous blood from the jugular vein, it would represent cerebral oxygen kinetics, etc. Clearly, the site of analysis must be considered when interpreting the venous blood gas.
The terms 'anaerobic' and 'aerobic' are used above in the clinical, colloquial [and anthropocentric] sense; that is, whether or not there is tissue oxygen deficiency; however, from an evolutionary-biochemical perspective, anaerobic simply means that oxygen is not necessary for a biochemical reaction - glycolysis, for example, is always anaerobic. See this 2018 post for further reference.