Chapter 73: Controversies: ScvO₂ Versus Lactate Clearance to Guide Resuscitation in Septic Shock?

Septic shock has been defined as a state in which hypotension persists despite adequate fluid resuscitation or the presence of a lactate level more than 4 mmol/L (so-called sepsis-induced tissue hypoperfusion) in patients with confirmed or suspected infection.¹ In these conditions, the Surviving Sepsis Campaign Guidelines suggest to restore mean arterial pressure (MAP) more than 65 mm Hg and normalize lactate levels and central or mixed venous hemoglobin saturation (ScvO₂/SvO₂).¹ Basically this reflects restoring perfusion pressure, improving tissue perfusion and restoring the balance between oxygen delivery and oxygen demand.

A recent study showed a 46% mortality rate in septic patients with increased lactate levels and hypotension.² In addition, a study using ScvO₂ as an endpoint of resuscitation³ and a study using both ScvO₂ and lactate clearance as endpoints of resuscitation⁴ showed significant improvements in mortality. However, a study (mostly septic-shock patients) comparing the use of either ScvO₂ or lactate clearance in early resuscitation with common goals for central venous pressure and MAP did not show differences in mortality.⁵ In addition, a recent study in septic-shock patients comparing early goal-directed therapy based on ScvO₂ measurements, protocol-based standard therapy based on systolic blood pressure, and heart-rate measurement without mandatory central venous catheter placement and usual-care based on the bedside physician similarly did not show differences in mortality.⁶

Therefore, controversy exists on how to use ScvO₂ or lactate levels in the resuscitation of septic-shock patients. In this chapter, both the background physiology of both parameters and their clinical use in the treatment of septic-shock patients is reviewed.

Many aspects of venous oximetry have recently been reviewed.^7,8 There is an abundance of oxygen available in the circulation. The red blood cells leave the left ventricle with almost 100% saturated hemoglobin to return, after exchanging oxygen in the microcirculation, to the right ventricle with around 75% saturated hemoglobin. When oxygen delivery to the tissues is decreased, stepwise oxygen consumption is maintained by using this surplus of oxygen resulting in a gradual decrease of venous oxygenation.⁹ When the decrease in oxygen delivery reaches a critical level, lactate levels generally start to increase.^10,11 In these models, venous oxygenation is proportional to cardiac output (CO) as arterial oxygen saturation and hemoglobin levels are usually stable. Venous oxygenation in these conditions thus reflects the balance between whole body oxygen demand and CO. It is important to realize that SvO₂ does not reflects regional venous oxygenation levels.¹² When cardiac function is normal, decreases in oxygen content are met by increases in CO thus minimally affecting venous oxygenation.¹³ However, when cardiac function is limited or the decrease in oxygen content is excessive, the surplus of oxygen in the system is mainly used to compensate for changes in oxygen demand so that venous oxygenation decreases.¹⁴ Although in clinical conditions, venous oxygenation is dependent on many factors (Figure 73–1), the response of CO to changes in oxygen content and oxygen demand is probably the main factor.

In hyperdynamic septic shock, patients seldom exhibit low levels of venous oxygenation,¹⁵ whereas it is more frequently found in the early unresuscitated patients with severe sepsis.³

The true venous oxygenation reflecting total body oxygen exchange is the pulmonary artery hemoglobin saturation. However, as pulmonary artery catheters are hardly used anymore and central venous catheters are recommended in the complex septic-shock patients, ScvO₂ is frequently used as a surrogate for the SvO₂.

In controlled experimental conditions, where usually oxygen demand is stable, ScvO₂ and SvO₂ highly correlate in various pathologic conditions.¹⁶ However in clinical conditions, significant differences between ScvO₂ and SvO₂ may exist.¹⁷ This could be related to the interindividual variation in the balance between oxygen demand and oxygen consumption as in low CO states, the ScvO₂ adequately reflects SvO₂.¹⁸ When conditions are stable and oxygen demand is acutely changed, ScvO₂ reflects the decrease in oxygen demand just like SvO₂.^19,20

Other sites for venous sampling (eg, femoral vein) or peripheral tissue oxygen saturation (eg, by near infrared spectroscopy) should not be used to replace ScvO₂.^21,22

Given the relationship of ScvO₂ with CO, a normal to high ScvO₂ might suggest adequate perfusion. However, normal or high venous oxygenation levels may not reflect adequate tissue oxygenation as microcirculatory shunting may result in increased venous oxygenation levels.²³ In addition, patients with high venous oxygenation levels, independently associated with increased mortality,²⁴ may have low CO responsive to fluid resuscitation that might contribute to this high mortality.²⁵ The venous-arterial carbon dioxide difference (Pv-aCO₂) is dependent on CO in septic-shock patients.²⁶ In the presence of a normal or even increased ScvO₂, an increased Pv-aCO₂ may reflect low CO²⁷ that has been associated with a worse outcome.²⁸

Lactate, as a normal end product of glucose metabolism, is produced from pyruvate. When pyruvate concentration exceeds the capacity of the slow metabolic Krebs cycle pathway, the fast metabolic pathway to lactate production is preferred resulting in the net production of 2 mol ATP per mol glucose metabolized. Although this amount of ATP is small compared to the additional 34 mol ATP produced in the Krebs cycle, the lactate pathway is fast and can thus produce large amounts of energy.²⁹ The increased levels of lactate in these conditions thus do not reflect the pathologic condition where pyruvate accumulates due to hypoxia. Lactate at the level of the organism also acts as an intermediate fuel that can be exchanged between various tissues.³⁰ In clinical practice, it is frequently difficult to separate these 2 states that clearly have significantly different impact on survival.

Lactate and Tissue Hypoxia

Decreasing oxygen delivery to the cells will ultimately result in decreasing oxygen consumption and increasing lactate levels.^31,32 It is important to realize that lactate levels do not necessarily represent the adequacy of oxygenation in different regional circulations as these have different levels of critical oxygen delivery.^33,34

When oxygen delivery decreases below a critical level, oxygen consumption starts to decrease and lactate levels increase. This supply-dependent oxygen consumption in relation to increased lactate levels is a key characteristic of shock. Whereas this relationship is easily shown in experimental conditions, clinical circumstances, and ethical considerations limit the possibility to study this in humans. However, a few studies have identified the presence of supply-dependent oxygen consumption in relation to increased lactate levels also in patients with septic shock.^35,36,37

In a model of tamponade, Zhang et al³² recently showed that decreasing oxygen delivery beyond the critical value resulted in a rapid increase in lactate levels. Subsequent resolution of the tamponade resulted in supply-independent oxygen consumption and normalization of lactate levels (Figure 73–2). In the same model, van Genderen et al³⁸ showed that increasing tamponade was associated with decreased microvascular perfusion and increased lactate levels. In this model, resolving tamponade was associated with a normalization of microvascular perfusion and rapidly decreasing lactate levels. In patients with septic shock, Friedman et al showed similar findings.³⁵ Supply-dependent oxygen consumption was a characteristic of patients early in septic shock with hyperlactatemia, whereas supply-independent oxygen consumption was found following resuscitation and normalization of lactate levels. In addition, improving microvascular perfusion in septic-shock patients with increased lactate levels has been associated with decreasing lactate levels.³⁹

Lactate and Aerobic Metabolism

These important observations as mentioned earlier underscore the importance of evaluating increased lactate levels to diagnose and treat supply-dependent oxygen consumption in critically ill patients. However, as lactate is a normal end product of glucose metabolism, increased aerobic glucose metabolism has been shown to increase lactate levels in the presence of adequate tissue oxygenation. Increases in pyruvate either by increasing glycolysis by means of hyperventilation or by infusing pyruvate result in increased lactate levels.^40,41 Similarly, the administration of epinephrine and corticosteroids, known to increase glycolysis, results in dose-dependent increase in lactate.^42,43 Very high lactate levels found in patients suspected of malignant lymphoma may have aerobic lactate production due to what is called the Warburg effect.⁴⁴ Treatment of the lymphoma is associated with a decrease in lactate levels, whereas subsequent tumor load leads to rapid increases in lactate levels.²⁹ In patients with septic shock, the increased activity of the cellular sodium/potassium pump has been associated with increased lactate levels not related to tissue hypoxia.⁴⁵ In addition, sepsis is also characterized by increased glycolysis due to cytokine-mediated glucose uptake.⁴⁶ Finally, mitochondrial dysfunction, not related to tissue hypoxia, can increase pyruvate levels and thus lactate levels.^47,48 In clinical practice, infusion of Ringer’s lactate does not interfere with lactate measurements,⁴⁹ whereas high-volume hemofiltration using lactate-buffered solutions do result in transiently increased lactate levels.^50,51 Many other causes of increased lactate levels, not related to the presence of tissue hypoxia, have been identified.^{52,53,54,55,56} In the presence of ethylene glycol intoxication, a high lactate level measured with a point-of-care device may be confounded by the adverse reaction to the lactate electrode in the machine.⁵⁷ A subsequent normal lactate in the central laboratory can even be used as a diagnostic tool in these circumstances.⁵⁸

Clearance of Lactate

The body is able to clear large lactate loads rapidly as demonstrated by the decrease in lactate levels following exercise, cessation of seizures, or return of circulation in cardiac arrest. However, in specific circumstances, the clearance of lactate maybe limited that could result in prolonged increased lactate levels following resuscitation or in limited increased lactate levels in the case of normal production. Impaired liver function is known to decrease the clearance of lactate by the liver.^36,59 Also in patients following cardiac surgery, clearance is impaired following a bolus infusion of lactate.⁶⁰ The dysfunction of pyruvate dehydrogenase in septic conditions, which can be alleviated by the administration of dichloroacetate, also limits lactate metabolism and results in increased lactate levels, where adequate oxygenation maybe present.^61,62

Various devices are available to rapidly measure lactate levels. A point-of-care blood gas analyzer has to advantage of delivering both the measurement of ScvO₂ and the lactate level. This can be used in clinical practice as the sampling site for lactate measurement (arterial, venous, capillary, etc) does not have a significant impact on the results.^63,64,65 In addition, hand-held devices to measure lactate have acceptable limits of agreement and clinical value when not used interchangeably with laboratory devices.^66,67

Physiologically, it is clear that low ScvO₂ levels represent a state where the oxygen delivery is not matched to oxygen demand. When this is coincided with an increased lactate level, supply-dependent oxygen consumption might be present. As lactate is not specific for tissue hypoxia and ScvO₂ is only specific when very low, the clinical identification of tissue hypoxia represents a challenge. Therefore, clinical context and other signs of abnormal perfusion or limited oxygen transport should be included. Especially in sepsis, no critical values for venous oxygenation exist.²⁴ This is probably related to the interindividual range of oxygen demand⁶⁸ and the microcirculatory shunting that is frequently present.²³ However, other signs also relate to abnormal perfusion and abnormal organ function.^{69,70,71,72,73} In the presence of hypotension, patients with normal lactate levels have less organ failure and better preserved microcirculatory perfusion than hypotensive patients with increased lactate levels.⁷⁴ Some of these are even strongly related to a worse outcome.⁷⁵ Next to other signs of tissue hypoperfusion the context/history of the patient is important. We usually accept low ScvO₂ levels in patients with chronic cardiac dysfunction when no signs of impaired organ perfusion exist. In addition, a diabetic patient on metformin presenting with high lactate levels without a clear history that might result in hypovolemia or myocardial dysfunction is likely to have metformin-associated hyperlactatemia even in the presence of acidosis.⁷⁶ This is clearly different from a patient admitted with severe trauma and increased lactate levels.^77,78

As ScvO₂ responds immediately to changes in blood flow⁷⁹ and lactate clearance takes at least 15 to 20 minutes,^40,80 there is no clear relationship between the change in lactate levels and venous oxygenation levels.⁸¹

Given the increased risk of the presence of tissue hypoperfusion and concomitant hypoxia, the Surviving Sepsis Campaign Guidelines and the Task Force of the European Society of Intensive Care on circulatory shock and hemodynamic monitoring recommend targeting a ScvO₂ of 70% and normalization of lactate levels in the resuscitation of septic-shock patients.^82,83

Given the physiologic coupling of abnormal ScvO₂ and increased lactate levels, the Task Force maintained this advice despite the study which showed that either targeting lactate or ScvO₂ had no difference in outcome⁵ and the study that showed that excluding ScvO₂ and lactate measurements from a protocol-based therapy did not affect outcome.⁶

The only study available today showing a significant impact on mortality when combining both ScvO₂ and normalization of lactate in the resuscitation of patients with circulatory failure (40% of whom had severe sepsis/septic shock) had a limited time frame of interventions.⁸⁴ The aggressive therapy lasted for not more than 8 hours. Given the current possibilities to rapidly improve tissue perfusion and oxygenation by different interventions,^39,85,86 the changes of persisting tissue hypoxia that responds to optimizing perfusion are small. Recently, Hernandez et al suggested the presence of flow-dependent phase in lactate clearance in survivors of septic shock.⁸⁷ This so-called “inflection point” in the lactate versus time graph could identify a point where increased lactate levels are unlikely to be caused by persistent tissue hypoxia. In the study by Jansen et al, where the target was to decrease lactate by 20% every 2 hours (mainly by optimization of the balance between tissue oxygenation and tissue oxygen demand), this goal was met only in the first 2 hours of the study.⁸⁴

There is a physiologic coupling between decreases in venous oxygenation and increasing lactate levels. However, this relationship is dependent on many factors that are frequently present and can rapidly change in septic-shock patients. Thus, both venous oxygenation and increased lactate levels can have limited specificity and sensitivity to detect tissue hypoperfusion and hypoxia. Other parameters of tissue hypoperfusion are available and should be used to create context to venous oxygenation and increased lactate levels. Although ScvO₂ can be used as a surrogate of the true venous compartment, caution should be used in specific circumstances. The venoarterial PCO₂ difference, easily available in these circumstances, can help to disguise a low flow state in the presence of normal to high ScvO₂ levels. As the persistence of low ScvO₂ values following initial resuscitation is very low¹⁵ and the baseline values for ScvO₂ in recent randomized controlled trials were normal,^5,6 lactate monitoring in circumstances where resources and placement of a central venous catheter is limited, represents an easy, cheap, and fast way to assess risk of morbidity and mortality of septic-shock patients.^67,88 The use of both parameters, however, requires adequate understanding of their physiology.