Addressing Global Adequacy of Tissue Oxygenation

Adequacy of tissue oxygenation is defined as an adapted oxygen supply (or DO₂) to oxygen demand.¹ Oxygen demand varies according to tissue type and according to time. Although oxygen demand cannot be measured or calculated, oxygen uptake or consumption () and DO₂ both can be quantified; they are linked by a simple relationship:

where ERO₂ represents oxygen extraction ratio (ERO₂ in %; and DO₂ in mL O₂/kg/min). DO₂ represents the total flow of oxygen in the arterial blood and is given as the product of cardiac output () by arterial oxygen content (CaO₂):

with CaO₂ being the product of hemoglobin (Hb, g/100 mL) by arterial oxygen saturation (SaO₂, %) and Hb O₂ capacity (1.39 mL O₂/g Hb): CaO₂ = Hb × SaO₂ × 1.39.

Under physiologic control, oxygen demand equals (≈2.4 mL O₂/kg/min for a 12 mL O₂/kg/min DO₂, which corresponds to a 20% ERO₂). The rate of oxygen delivered by blood is physiologically larger than the rate of : DO₂ is adapted to oxygen demand. When oxygen demand increases (e.g., during exercise), DO₂ has to adapt and increase.

During circulatory shock and/or severe hypoxemia, as DO₂ declines secondary to a decrease in and/or a decrease in CaO₂, can be maintained by a compensatory increase in ERO₂, and DO₂ remaining therefore independent. But as DO₂ falls further, a critical point (DO₂ crit) is reached; ERO₂ can no longer compensate for this fall in DO₂, and at this critical level, becomes DO₂ dependent (Figure 91-1). At this DO₂ crit (4 mL/kg/min), for a of about 2.4 mL/kg/min, ERO₂ reaches its critical point (ERO₂ crit) of 60%. When is higher, DO₂ crit is higher as well. Increase in oxygen extraction occurs via two fundamental adaptive mechanisms²: (1) redistribution of blood flow among organs via an increase in sympathetic adrenergic tone and central vascular contraction (this is responsible for a decreased perfusion in organs with low ERO₂, such as the skin and splanchnic area, and a maintained perfusion in organs with high ERO₂, such as heart and brain); and (2) capillary recruitment within organs responsible for peripheral vasodilation (opposite to central vasoconstriction).

Figure 91-1 O₂ uptake ()-to-O₂ supply (DO₂) relationship.

When is supply independent (“independency”) following the relation = y, whole body O₂ needs are met. When becomes DO₂ dependent (“dependency”) according to the relation = x DO₂, starts to be linearly dependent on DO₂ at the critical DO₂ value (DO₂ crit), which corresponds to dysoxia (insufficient ATP synthesis as related to needs) and shock state. DO₂ crit is influenced by global organism O₂ needs: when is decreased (e.g., by rest, sedation, hypothermia), the DO₂ crit is decreased as well (lower dotted line; DO₂ crit[1]); conversely, increased (e.g., by increased muscle activity, awakening, hyperthermia, sepsis) is associated with increased DO₂ crit (upper dotted line; DO₂ crit[2]).

Using Mixed Venous Oxygen Saturation to Assess Adequacy of Global Tissue Oxygenation

In the clinical setting, mixed venous oxygen saturation (SvO₂) can be used for assessing whole-body -to-DO₂ relationships. Indeed, according to the Fick equation, tissue is proportional to cardiac output:

where CvO₂ is mixed venous blood oxygen content. To some extent, is approximately equal to cardiac output × (SaO₂ − SvO₂) × Hb × 1.39, and SvO₂ is approximately equal to SaO₂ − /( × Hb × 1.39).

Four situations can be responsible for a decrease in SvO₂: a decrease in SaO₂ (hypoxemia), in Hb (anemia) or in cardiac output, or an increase in (like in exercise). At DO₂ crit, SvO₂ is about 40% (SvO₂ crit) with an ERO₂ of 60% and a SaO₂ of 100%. This SvO₂ crit has been identified in humans.³ It is important to emphasize that for the same decrease in CaO₂ (induced by a decrease of Hb or SaO₂), the decrease in SvO₂ will be more pronounced if cardiac output cannot adapt. Hence, SvO₂ represents adequacy of global flow to CaO₂ decrease. A 40% SvO₂ can be taken as an imbalance between arterial blood oxygen supply and tissue oxygen demand with evident risk of dysoxia. In the clinical setting, a decrease of SvO₂ of 5% from its normal value (77%-65%) is representative of a significant fall in DO₂ and/or an increase in oxygen demand (Figure 91-2). If initial probabilistic treatment (fluid resuscitation and/or low-dose inotropes and/or red blood cell transfusion) does not allow SvO₂ to be restored to a minimal 65%, Hb, SaO₂, and cardiac output should then be individually measured to introduce the appropriate treatment.

Figure 91-2 Venous O₂ saturation (SvO₂)-to-cardiac index (CI) relationship.

According to the modified Fick equation, the relationship SvO₂/CI is curvilinear. Subsequently, when O₂ uptake () is constant, CI variations lead to large variations in SvO₂ when the initial CI value is low. In contrast, when initial CI values are already high, CI variations do not influence SvO₂ very much. These relationships are modified when CI variations are associated with large modifications in .

Assessing Global Flow

Global flow (i.e., cardiac output) is dependent on preload, myocardial contractility, afterload, and heart rate. Regional flow distribution is not homogeneous and is dependent on central and peripheral vascular tone, which ultimately results in the composite systemic vascular resistances (SVR). As an oversimplification, mean arterial pressure (MAP) can be estimated as the product of cardiac output by SVR. When flow decreases, MAP remains stable when SVR increases; this corresponds to increased sympathetic adrenergic tone and central vascular contraction in low ERO₂ organs, and preserved peripheral vasodilation in high ERO₂ organs. Overall, ERO₂ increases and SvO₂ decreases.

Minimal data exist to guide selection of the threshold for blood pressure maintenance. Arbitrary values of a systolic blood pressure of 90 mm Hg or a MAP of 60 to 65 mm Hg have traditionally been chosen. Observation of an inappropriate tissue perfusion (e.g., raised blood lactate level, metabolic acidosis, SvO₂ lower than 65%, decreased urinary flow) and its persistence despite probabilistic therapy (fluid, low-dose inotropes, red blood cells) should lead to optimizing flow according to the Frank-Starling curve. This can be assessed by invasive and noninvasive investigative procedures (see later).

During circulatory shock, -to-DO₂ dependency with a rise in blood lactate levels implies oxygen debt. Several authors have reported that oxygen debt is related to the likelihood of multiple organ failure and mortality in postoperative or polytrauma patients.^4,5 Patients who survive multiple organ failure have been shown to have higher cardiac index, lower SVR, higher , and higher SvO₂ than nonsurvivors.^6,7 Rixen and Siegel⁵ demonstrated that the degree of tissue oxygen debt is related to an enhanced inflammatory response, associated with an increased risk of acute respiratory distress syndrome and higher mortality rates.

Recent research has emphasized the potential interest of central venous oxygen saturation (ScvO₂) for detecting global oxygenation impairment.⁷ Experimental studies reported that changes in SvO₂ and ScvO₂ closely reflect circulatory disturbances during periods of hypoxia, hemorrhage, and subsequent resuscitation (ScvO₂ being approximately 5% higher than SvO₂ in the critically ill). Fluctuations in these two parameters correlated relatively well, although absolute values differed.⁸ Finally, observational data found ScvO₂ to be a useful parameter in detecting occult tissue hypoperfusion in both sepsis and cardiac failure.^9,10 An important feature with ScvO₂ monitoring is that ScvO₂ can be continuously provided by central venous catheters equipped with optic fibers (e.g., PreSep oximetry catheter [Edwards Lifesciences, Irvine, California]). In initial resuscitation of circulatory shock, insertion of a central venous catheter is a standard, rapid, and easy approach, much easier than any other invasive hemodynamic monitoring, especially in patients who are not yet sedated, intubated, and ventilated.

In a landmark trial by Rivers et al., patients with severe sepsis and septic shock admitted to the emergency department were randomized to standard therapy (n = 133) or to early goal-directed therapy (n = 130) targeted to achieve a central ScvO₂ of greater than 70%.¹¹ Standard therapy included antibiotics, fluid resuscitation, and vasoactive drugs to achieve a central venous pressure between 8 and 12 mm Hg, MAP greater than 65 mm Hg, and urine output greater than 0.5 mL/kg/h. Patients in the early goal-directed therapy group, in addition to the standard goals, had to reach an ScvO₂ of greater than 70% by optimizing fluid administration, hematocrit above 30%, and/or prescription of an inotrope (dobutamine < 20 µg/kg/min). Initial ScvO₂ in both groups was quite low (49 ± 12%), confirming that severe sepsis is hypodynamic before any fluid resuscitation has started. This study demonstrated a significant reduction in hospital mortality: 30.5% in the early goal-directed therapy group compared with 46.5% in the standard therapy group (P = .009). An important point in this study is that 99.2% of patients receiving early goal-directed therapy achieved their hemodynamic goals within the first 6 hours, compared with 86% of those receiving standard therapy. From the first to the 72nd hour, total fluid loading was not different between the two groups (approximately 13,400 mL); in contrast, from the first to the seventh hour, the amount of fluid received was significantly larger in the early goal-directed therapy patients (approximately 5000 mL versus 3500 mL). In the follow-up period between the seventh and the 72nd hour, in patients receiving early goal-directed therapy, mean ScvO₂ was higher (70.6 ± 10.7% versus 65.3 ± 11.4%; P = .02), mean arterial pH was higher (7.40 ± 0.12 versus 7.36 ± 0.12; P = .02), and lactate plasma levels were lower (3.0 ± 4.4 mmol/L versus 3.9 ± 4.4 mmol/L; P = .02), as was base excess (2.0 ± 6.6 mmol/L versus 5.1 ± 6.7 mmol/L; P = .02). The multiple organ failure score was significantly altered in patients receiving standard therapy when compared with early goal-directed therapy patients. This was the first study demonstrating that early identification of patients with sepsis, associated with early initiation of goal-directed therapy to achieve adequate tissue oxygenation by O₂ delivery (ScvO₂ monitoring), significantly improves mortality rates.¹¹ This study was then supported by more than 10 following trials,¹² and further multicentric prospective studies are under way.

Diagnosing Shock Type

Deciding When to Admit Patient to Intensive Care Unit

Admission to the intensive care unit (ICU) is requested when hemodynamic instability is present and requires use of inotropes (inoconstrictors or inodilators). This occurs when shock does not readily respond to initial fluid therapy, requires ventilatory support (with a noninvasive interface or after intubation) or imposes hemofiltration (severe electrolyte disorder, fluid overflow, poisoning), and more generally when invasive procedures become necessary (invasive blood pressure monitoring). A patient becomes eligible for an ICU bed at the time failure of one or more organs develops.

Choosing Appropriate Monitoring

The discussion on types of monitoring has no meaning until the cardiorespiratory emergency has been treated. The minimal monitoring device consists of electrocardiography, pulse oximetry, and rapid arterial pressure recordings (every 5 minutes and, at best, continuous and invasive). A central venous catheter allows measurement of central venous pressure, which often cannot help much in deciding fluid administration (except when it remains lower than 5-8 mm Hg), but which facilitates infusion of drugs, crystalloids, or colloids. The central venous line also allows for monitoring and/or sampling of ScvO₂ (surrogate for mixed SvO₂) if the catheter is not equipped with optic fibers. Central venous catheters are easier, should be cheaper, and carry less iatrogenic risk than Swan-Ganz catheters.

A Swan-Ganz catheter (with continuous cardiac output and SvO₂ monitoring) and/or any noninvasive flow assessment (transesophageal echography, esophageal Doppler echography) is recommended when optimized cardiac output is doubtful. This requires that some preliminary cardiorespiratory stability has been obtained. In that context, fluid administration should be continued (the heart is preload dependent) until cardiac output increases no further (becomes preload independent). When cardiac output is insufficient to maintain MAP or urine output, when SvO₂ remains low, or when lactate concentration remains elevated, an inotrope should be given. Cardiac echography must be performed in the context of congestive heart failure and/or myocardial ischemia to diagnose ventricular or valvular dysfunction. In the sedated, intubated, and ventilated patient, recordings of systolic pressure variation or pulse pressure variation can be helpful: the heart remains preload dependent until systolic pressure variation is smaller than 10 mm Hg or pulse pressure variation is less than 10%, or both.¹⁵ Arrhythmia and tidal volume below 7 mL/kg limit this type of evaluation.

Iterative blood gas analysis (another approach justifying insertion of an arterial line), metabolic acidosis and lactate concentration evaluation, is a way to assess global tissue oxygenation and completes ScvO₂ or SvO₂ information.

Symptomatic Treatment

Emergency therapeutic principles of care need to be decided at the time the initial diagnostic strategy is considered. It is necessary to give supplemental O₂ and ventilatory support in response to acute respiratory failure (acute lung injury, mechanical failure, respiratory distress) either through a face mask or by endotracheal intubation and ventilation. Acute circulatory failure is treated by initial fluid loading in the absence of left ventricular failure (see later). If decreased global contractility is present, inotropic support is considered with either dobutamine or dopamine. In case of anaphylactic shock, emergency treatment is to give intravenous epinephrine to treat allergy-induced vasodilation.

Fluid loading is the first step in treatment, and its first goal is to optimize left ventricular preload to improve DO₂ by increasing cardiac output.¹⁶ There is, however, an associated risk of interstitial edema, in particular pulmonary edema. Unless the patient has an acute lung injury, fluid loading aims at maximizing cardiac output¹⁶ according to the Frank-Starling relationship, decreased lung gas exchange being detected by a decrease in SaO₂ (or by a decrease in its surrogate, pulse oximetry).

Swan-Ganz catheter–derived pulmonary artery occlusion pressure has long been the most used static clinical variable for guiding fluid infusion. In septic shock, it was accepted that maximal cardiac output was obtained for values between 12 and 15 mm Hg.¹⁷ To better estimate left ventricular preload, left ventricular end-diastolic surface has now been proposed. In fact, in the sedated, intubated, and ventilated patient, ventilatory-induced systolic pressure variation predicts increased systolic ejection volume to fluid loading much better than pulmonary artery occlusion pressure.¹⁸

Synthetic colloids are first-line agents. They may induce less pulmonary edema than crystalloids, especially in patients in septic shock. Crystalloids are recommended as first-line agents during anaphylactic shock. Normalization of hemoglobin concentration, [Hb], by red blood cell transfusion is not required. However, a [Hb] between 8 and 10 g/dL¹⁶ might be preferred in patients with severe sepsis and/or coronary disease and/or decreased cardiac contractility. In those latter cases, decreased [Hb] is not compensated by increased cardiac output, and DO₂ crit is reached more rapidly. In each case targeting a ScvO₂ larger than 70% may be a helpful guide for transfusion.¹⁹

Catecholamines help in restoring perfusion pressure and maintaining cardiac output, thus allowing sufficient DO₂; this should allow regional flow distribution and improved ERO₂. All catecholamines are inotropes; they can be divided into (1) inodilators when they combine inotropic and vasodilatory properties (low-dose dopamine, any dose of dobutamine or dopexamine); or (2) inoconstrictors when they combine inotropic and vasoconstricting properties (high-dose dopamine, any dose of epinephrine or norepinephrine). Inodilators increase flow; inoconstrictors increase perfusion pressure. Because of variable individual sensitivity to catecholamines, dose titration is strongly recommended.¹⁷ More potent vasopressors such as vasopressin have been tested with conflicting results, in particular as regards regional circulation. More recently, in a large multicenter, randomized, double-blind trial, vasopressin showed no benefit as a first-line vasopressor in comparison to norepinephrine in septic shock.²⁰ It is important to emphasize that a rise in blood pressure may not be a surrogate of clinical benefit. Indeed, in a large placebo-controlled clinical trial, administration of the nonselective nitric oxide inhibitor, N^G-methyl-L-arginine, in septic shock produced both significant increases in blood pressure and significant increases in mortality.²¹

In septic shock, several studies demonstrated that increasing MAP from 65 to 85 mm Hg was associated with no difference in organ perfusion variables.¹⁶ Because increasing blood pressure through vasoconstriction may be associated with a decrease in flow, a tradeoff may exist between raising blood pressure and decreasing cardiac index that will vary depending on the specific vasopressor or combined inotrope/vasopressor.¹⁷ Applying such principles for symptomatic treatment in septic shock patients has resulted in decreasing unadjusted hospital mortality from 37% to 31% over 2 years (P = .001). The adjusted odds ratio for mortality improved the longer a site was involved in the Surviving Sepsis Campaign, resulting in an adjusted absolute drop of 0.8% per quarter and 5.4% over 2 years (95% CI, 2.5%-8.4%).²²

Other Therapeutic Principles

The importance of correction of metabolic acidosis and the use of intravenous bicarbonate for shock-induced anion gap acidosis have been overemphasized in the past. Indeed, clinical studies, including one randomized, prospective trial, failed to show any hemodynamic benefit from bicarbonate therapy either to increase cardiac output or to decrease vasopressor requirements, regardless of the degree of acidemia. Cardiac function does not appear to be significantly decreased when the arterial pH remains higher than 7.00. Bicarbonate infusion, apart from renal or digestive losses, is therefore not recommended unless the patient has hyperkalemia.²³

In patients with septic shock, stress-dose (low-dose) steroid therapy (hydrocortisone 200 mg/day) needs to be considered, especially if the decrease in blood pressure requires high or increasing concentrations of vasopressors, once appropriate antibiotics are being given or the infectious site is controlled.¹⁶ Steroid therapy may be weaned once vasopressors are no longer required. Beyond 72 hours, absence of any hemodynamic improvement suggests the hydrocortisone treatment is futile.

Although not oriented toward better circulatory efficacy, a number of treatments are essential in septic shock.¹⁶ Control of the infectious source is essential. Empirical or probabilistic antibiotics must be directed against gram-negative microorganisms but also against potentially resistant pathogens. This justifies double or sometimes triple antibiotherapy. It theoretically offers the following advantages: widening of the spectrum of activity, antibacterial synergy, increased bactericidal speed, and decreased risk for emergent resistant germs.