Shock is managed (1) at an urgent tempo; and by (2) identifying and treating acute, reversible causes; (3) restoring intravascular volume; (4) infusing vasoactive drugs; (5) using mechanical adjuncts, when applicable; and (6) supporting vital functions until recovery.
Minutes matter when recognizing and resuscitating shock. This principle is emphasized in speaking of the “golden hour” in which the circulation and tissue perfusion are restored before progressive tissue injury and organ failures that rapidly become refractory to subsequent attempts at resuscitation. This golden hour is a time-honored tenet in trauma, and more recently recognized to be also pertinent in septic shock. The initial Early Goal Directed Therapy (EGDT) trial demonstrated that rapid resuscitation to endpoints of central venous pressure, mean arterial pressure, and central venous oxyhemoglobin saturation within 6 hours of presentation improved outcomes compared to a less-aggressive resuscitation.11 Three subsequent trials of septic shock did not confirm the utility of the original EGDT protocol.19,20,21 However, the mortality rates reported in these later trials, performed in an era when aggressive early resuscitation was the norm, were significantly lower than in previous reports. One interpretation of these results is that, while the appropriate endpoints of resuscitation remain unclear, an environment that favors early, aggressive resuscitation improves outcomes in septic shock.
The appropriate endpoints of shock resuscitation remain elusive. Importantly, resuscitation to an arbitrarily set MAP of at least 65 mm Hg is not sufficient and possibly not necessary. As an example, treatment with a nitric oxide synthase inhibitor leads to increased blood pressures and lower catecholamine use, but also increased mortality.22 These seemingly contradictory effects of therapy may be explained by discrepancies between systemic hemodynamics and the microcirculation; lack of validity of the targets of macrovascular resuscitation (perhaps MAP is less important than believed); or to unrecognized adverse effects of the drug. Several studies have compared specific blood pressure targets finding that achieving a MAP of higher than 65 mm Hg (such as 75 or 85 mm Hg) does not improve outcomes. Targeting microvascular resuscitation is attractive in theory, but real-time assessments of microvascular function are not readily available for clinical use, and effective methods to safely and reliably increase microvascular function have not been found. We advocate a comprehensive assessment of the adequacy of perfusion to guide resuscitation, rather than merely aiming for an arbitrary mean arterial pressure. Serial assessments are likely to be valuable since shock and its resuscitation can produce dramatic changes within hours.
Identify and Treat Reversible Causes
Several causes of shock require specific identification and treatment because general supportive measures will surely fail. Good examples include tension pneumothorax, cardiac tamponade, and ruptured abdominal aortic aneurysm (Table 21–3). These can be subtle at times, requiring a careful, systematic approach to shock. Intensivist-conducted ultrasound has changed fundamentally the initial examination of the shock patient. Its ability to quickly signal cardiac dysfunction, pericardial effusion, hypovolemia, deep vein thrombosis, pulmonary embolism-in-transit, pneumothorax, aortic rupture, free peritoneal blood, traumatic injuries, sources of sepsis, and other crucial findings makes ultrasound an essential skill for early diagnosis.23 Moreover, the intensivist can repeat the ultrasound examination at will to judge the response to interventions or identify complications.
Table 21–3Differential diagnosis of shock. ||Download (.pdf) Table 21–3 Differential diagnosis of shock.
Hemorrhagic shock (GI, trauma, aortic rupture, etc)
Renal loss (diuresis, osmotic loss, diabetes insipidus)
Gastrointestinal loss (diarrhea, vomiting)
“Third space” loss (pancreatitis, postsurgical, sepsis, anaphylaxis, trauma, toxin, idiopathic systemic capillary leak syndrome)
Drugs (sedatives, analgesics, nitrates, Ca-channel blockers)
Impaired cardiac filling
High pleural pressure (PEEP, autoPEEP, tension pneumothorax, massive effusion, abdominal compartment syndrome)
Other obstruction to cardiac filling (tumor, thrombus)
Arrhythmia (fast or slow)
Systolic LV or RV dysfunction
Drugs (Ca-channel blockers, beta-blockers, other)
Tamponade (as above)
Aortic stenosis; hypertrophic obstructive cardiomyopathy
ARDS and its treatment (elevated alveolar pressure)
Papillary muscle rupture
Device malfunction (ECMO, VAD, IABP)
Distributive (high cardiac output) shock
Hepatic failure (acute and chronic)
Toxins (salicylate, cyanide, carbon monoxide)
Cellular dysoxia (prolonged shock states)
The timing of antibiotics in confirmed or suspected septic shock deserves specific mention in relation to the tempo of shock resuscitation. Appropriate antibiotics must be given within the first hour following the recognition of septic shock. Antibiotic therapy is frequently delayed and often ineffective for the final microbiologic diagnosis. Orders may be delayed due to diagnostic confusion and caregiver attention toward invasive procedures and hemodynamic resuscitation. Systems issues between ordering and administering antibiotics also contribute to these delays. Regardless of cause, delays in appropriate antibiotic administration worsen mortality by approximately 8% per hour of delay.24 For these reasons, broad-spectrum antibiotics should be ordered and administered promptly after a diagnosis of shock when sepsis is in the differential, preferably guided by preplanned order sets. Antibiotics should then be tailored to microbial susceptibilities, as these data are available, or discontinued promptly if an alternative etiology of shock is identified.
Restoring Intravascular Volume
Rapid restoration of intravascular volume is an essential principle of shock resuscitation since fluids may promptly restore perfusion and prevent organ failures. Fluids should be infused at a rapid pace (usually much faster than typical ICU infusion pumps will allow), and in sufficient volume (which can be many liters). This practice allows for periodic reevaluation for clinical response: slower infusions of small volumes may confound the perception of response. Although colloid-containing fluids have some theoretical advantages over crystalloids, clinical trials generally show equivalence. Some colloids, especially synthetic starches, are clearly detrimental and should not be used.25 Because crystalloids are more widely available, cheaper, and at least as effective, they are preferred for shock resuscitation. Although vasomotor function and vasopressors are both less active in acidemic environments, attempts to correct a metabolic acidosis with bicarbonate infusions do not speed resuscitation nor reduce vasopressor requirements, and may lead to worsening intracellular acidosis. Accordingly, bicarbonate infusions should be avoided.
In the setting of acute traumatic shock, questions have been raised about the targets of early fluid resuscitation. Potential downsides of restoring blood pressure to normal before surgical exploration include dilution of clotting factors, hypothermia, and an increased rate of hemorrhage as arterial pressure rises. Several studies suggest that delayed fluid resuscitation for victims of penetrating trauma (aiming for a systolic blood pressure of 70 mm Hg) might improve outcomes.26 Concerns about the risks of persisting hypotension and doubts about whether these data can be generalized to the broad group of patients with traumatic shock have limited its appeal.
While the importance of urgent fluid resuscitation is undeniable, many patients with shock fail to respond, especially following initial resuscitation when shock is due to sepsis.27 Indiscriminant fluid may produce harm by causing pulmonary edema or other organ failures so, especially in the sickest patients with lung and renal failure, an effort should be made to predict fluid responsiveness. In this regard, static measures of the central venous and pulmonary artery occlusion pressures have been shown invalid28 and have been supplanted by dynamic indicators. High-volume positive-pressure ventilation produces pleural pressure changes that affect stroke volume in a cyclical fashion (largely by varying right atrial filling), giving rise to larger fluctuations in stroke volume; vascular flow; and vena caval diameter in preload-dependent individuals. A 13% variation in pulse pressure with breathing is highly sensitive and specific for predicting fluid responders.29 Similarly, useful cutoff values have been determined for variations in vena caval diameter (superior and inferior); aortic and brachial artery flow velocity; left ventricular outflow tract velocity-time integral; and cardiac volumes derived from bioimpedance and bioreactance. Prerequisites for validity include tidal volume of 8 to 12 mL/kg; a fully passive patient; regular cardiac rhythm; and the absence of acute cor pulmonale. Since this tidal volume is larger than generally appropriate, the ventilator should be adjusted before the measurement of variation in order to produce the conditions for validity but then returned to lung-protective volumes. Passive leg raising is a reliable indicator of fluid-responsiveness irrespective of ventilation mode and cardiac rhythm,30 but is not reliable when there is severe intra-abdominal hypertension. By returning blood held in the capacitance veins to the circulation and thereby raising stroke volume in patients who are on the ascending limb of the Starling curve, it avoids potentially harmful fluid boluses in patients who will not benefit. This method is particularly useful for patients who cannot easily be made passive on the ventilator.
Infusing Vasoactive Drugs
In addition to careful assessment and restoration of circulatory volume, many patients in shock require vasoactive infusions. Norepinephrine is the preferred initial agent given its potency, relatively low propensity to induce arrhythmias, and association with improved mortality compared to dopamine.31 To avoid injury caused by accidental infiltration of vasoconstrictors into peripheral tissues, norepinephrine and other vasopressors should be infused through central venous or intraosseous catheters. However, in keeping with the urgent tempo of shock resuscitation, vasoactive infusions into severely ill patients should not be delayed merely because central access is not yet available. Similarly, vasopressors should not be initiated without attempts to restore circulating volume, yet severely ill patients should be resuscitated simultaneously with vasopressors and fluids, with titration of the vasopressors as circulating volume is restored.
The initiation of vasoactive infusions may also provide additional important clues to the underlying physiology. For example, norepinephrine consistently increases blood pressure. However, a concomitant increase in lactic acidosis and fall in venous oxyhemoglobin should prompt reevaluation for inadequate fluid loading or for cardiogenic shock.
When cardiogenic shock is identified or suspected, inotropic agents, such as dobutamine are useful. Like norepinephrine, careful examination during dobutamine initiation may identify additional physiologic perturbations. A rise in arterial pressure (or decrease in norepinephrine requirements) after initiating inotropic agents supports cardiogenic shock physiology. Because dobutamine also causes some arteriolar dilation, if arterial pressure falls with dobutamine one may suspect inadequate preload or, alternatively, a severely dysfunctional myocardium.
Mechanical Adjuncts for Circulatory Support
When ventricular dysfunction is so extensive that it is refractory to vasoactive infusions (or when valvular incompetence contributes to cardiogenic shock), mechanical adjuncts to aid circulation may be employed. Ideally, these devices are employed as a bridge to definitive correction of the cardiac dysfunction. Intra-aortic balloon counterpulsation has been used extensively in this fashion for decades, although more recent evidence show that it does not improve outcomes in patients with acute myocardial infarction or cardiogenic shock.32,33 Venoarterial extracorporeal membrane oxygenation (ECMO) also provides circulatory support and can be employed rapidly, even during cardiopulmonary resuscitation. The practice of E-CPR (initiation of ECMO within 15 minutes of cardiac arrest) may lead to improved patient outcomes. Ventricular assist devices (VADs) also provide circulatory support: while slower to employ than ECMO, VADs may be more appropriate for medium- and even long-term support.
Patients in shock often have deranged cerebral perfusion and metabolic encephalopathy that would benefit from endotracheal intubation and mechanical ventilation. In severe shock states, lactic acidosis leads to increased respiratory effort, which subsequently increases lactate production, diverts blood flow to respiratory muscles, and draws CO from other vital organs. Mechanical ventilation (either invasive or noninvasive) may decrease oxygen consumption and increase vital organ blood flow and should be considered even in the absence of encephalopathy.
Because shock is usually characterized by inadequate oxygen delivery, increasing arterial oxygen content through transfused red blood cells may be helpful, particularly when shock results from massive hemorrhage. However, this theoretical benefit of transfusions must be balanced with its negative effects, including circulatory overload, inflammatory effects, and immune suppression as well as recognition that transfused red cells may exhibit impaired oxygen carrying capacity. Studies in patients with acute gastrointestinal hemorrhage and in those with septic shock show that liberal transfusion strategies (to keep the [Hgb] > 9 g/dL) are no better and may be inferior to restrictive targets ([Hgb] > 7 g/dL).34,35 While there may be some patients in whom higher targets for hemoglobin may be appropriate (eg, acute coronary syndrome, low central venous oxyhemoglobin saturation despite resuscitation, and other overt manifestation of anemia), we recommend that transfusion be avoided in most patients with shock until the hemoglobin falls lower than 7 g/dL. It is even possible that lower targets could be beneficial for some patients, but these have not been tested.
Critically ill patients with shock have historically been considered fragile, leading to orders for strict bed rest and minimization of physical interventions. More recently, early physical and occupational therapy has been shown safe for critically ill patients and effective in preserving functional independence.36 Many of the subjects in this and similar studies have been in shock, on vasoactive infusions, suggesting that shock is not a contraindication to mobilization. Some patients experience a decrease in vasoactive infusion requirements following mobility sessions suggesting a previously unrecognized, but important intervention for the management of shock.