Airway and Ventilation

Obstruction of the airway can occur in any patient with impaired consciousness including cardiac arrest.²⁵ If uncorrected, this obstruction prevents oxygenation and ventilation, leading to or perpetuating cardiopulmonary collapse. In patients who are comatose because of primary cardiac arrest, the airway usually is not patent. Animal models typically do not mimic ventilation through the human airway, because the most common research animals (dogs and swine) have straight oral-tracheal passages.

Agonal respirations occur after acute cardiac arrest for an additional 1 to 2 minutes.²⁶ These respirations may confuse lay people, delaying recognition of cardiac arrest. It is unclear whether agonal respirations can generate sufficient ventilation to support life. The presence of gasping is associated with survival, but it also may be a surrogate marker for brief collapse-to-resuscitation intervals.²⁷ Regardless, the amplitude and frequency of agonal respirations declines over 1 to 2 minutes, necessitating artificial ventilation for all patients requiring more than momentary resuscitation efforts.

Simple maneuvers can open the human airway. Extension of the neck (head tilt) and forward displacement of the mandible (chin lift) straightens and opens the pharynx. The tongue can be displaced from the posterior pharynx by insertion of an oropharyngeal airway. With these steps, positive-pressure ventilation can be provided using mouth-to-mouth or bag-valve-mask ventilation. A positive-pressure breath of as little as 400 mL in adults (6-7 mL/kg) delivered over 2 to 3 seconds will cause the chest to rise.²⁸ Recent studies have indicated that hyperventilation or hyperexpansion of the chest can impair venous return and decrease circulation during resuscitation.²⁹ In addition, minute ventilations that are smaller than those required for long-term support probably provide adequate gas exchange during cardiac arrest.

The optimal rate and timing for positive-pressure breathes during resuscitation in humans has not been established. The need for gas exchange must be balanced against the fact that interrupting chest compressions even for a few seconds can reduce coronary perfusion pressure (Figure 33-3).³⁰ In swine, comparison of different ratios of chest compressions to ventilation suggests that 2 breaths per 50 chest compressions or more may be optimal for resuscitation.³¹ An extreme point of view is that chest compressions without any artificial ventilation may be sufficient to accomplish resuscitation in certain individuals.³² This position is contrary to early work demonstrating that the human airway collapses in most unconscious subjects, and that compressions alone are unable to provide ventilation.²⁵ Recent studies in humans indicate that chest compressions generate tidal volumes that are likely less than the physiologic dead space.³³ Nevertheless, passive insufflation of oxygen without positive-pressure ventilation has been reported to be beneficial in humans.³⁴ As a compromise, some medical systems employ continuous chest compressions without pauses for positive-pressure ventilation. In these systems, positive-pressure breathes are delivered asynchronously by bag-mask during chest compressions, and the actual minute ventilation achieved is unknown. Therefore, the most recent guidelines recommend a ratio of 30 chest compressions to 2 ventilations, but even higher ratios or asynchronous ventilation may be optimal. Certainly, the duration of any pauses to deliver breaths must be minimized.

Figure 33-3 Chest compressions provide coronary perfusion by creating a pressure gradient between aorta (Ao) and inside of ventricles (approximated by right atrium [RA]). Gradient between sites is the coronary perfusion pressure (CPP). During chest compression, pressure increases in both Ao and RA. During relaxation, pressure persists in Ao more than RA. Thus, myocardial blood flow is most related to CPP during relaxation phase of chest compressions. Note that CPP declines within 1 to 2 seconds when compressions pause for ventilation. (Unpublished laboratory data.)

Waveform capnography is an extremely useful monitor during resuscitation, both for confirming ventilation and for monitoring adequacy of circulation. During cardiac arrest, end-tidal CO₂ measurement is related to cardiac output and pulmonary blood flow.³⁵ Therefore, CO₂ levels may be very low (<10 mm Hg) at the onset of resuscitation. Adequate artificial circulation will cause CO₂ levels to increase, and these levels may be used as a feedback to improve or modify chest compressions. Data from emergency department patients suggest that an end-tidal CO₂ level greater than 15 to 16 mm Hg is associated with successful cardiac resuscitation.^36,37 Conversely, end-tidal CO₂ less than 10 mm Hg after 20 minutes of resuscitative efforts appears to confirm failure of resuscitation.³⁸ However, drugs commonly used during resuscitation can disrupt the association between capnography readings and pulmonary blood flow. For example, epinephrine infusion reduces CO₂ levels, and sodium bicarbonate infusion produces a transient but profound elevation of CO₂ levels. An abrupt increase in end-tidal CO₂ levels, usually to levels over 35 mm Hg, accompanies the return of spontaneous circulation. This finding may be useful for recognizing return of circulation, thereby minimizing any interruptions of chest compressions for pulse checks (Figure 33-4).

Figure 33-4 End-tidal CO₂ changes during resuscitation. Tracings from monitor-defibrillator with an accelerometer to measure chest compression depth (top trace) and waveform capnography to measure exhaled CO₂ (bottom trace). Exhaled CO₂ confirms correct placement of an advanced airway. Note frequent interruptions in chest compressions prior to placing advanced airway. More continuous chest compressions are followed by an abrupt rise in exhaled CO₂, corresponding to return of circulation, including pulmonary circulation. Chest compressions continue for 2 to 3 minutes until providers detect a palpable pulse.

Artificial Circulation

In the patient without pulses, circulation of blood can be accomplished by mechanical compression of the heart and chest. The critical parameter for restoring spontaneous circulation is the development of adequate coronary perfusion pressure (CPP). CPP is quantified by the pressure gradient between the aorta and the inside of the ventricles (usually approximated by the pressure in the right atrium or the central venous pressure [CVP]). Measurement of CPP in clinical practice is difficult unless the patient has invasive monitoring prior to cardiopulmonary collapse. CVP can be estimated from a central line, and peripheral arterial pressures developed approximate aortic pressures. In the spontaneously beating heart, most blood flows through the ventricular walls during diastole, when the ventricular pressure is lowest. With mechanical compression of the heart and chest during resuscitation, the primary perfusion of the heart occurs during the relaxation phase (see Figure 33-3). Therefore, CPP is usually measured at the end of the relaxation phase. CPP is highly correlated with myocardial perfusion, and consequently with the likelihood of resuscitation.⁴⁶ In humans, return of circulation requires that the developed CPP exceed 15 to 20 mm Hg. It is likely that with CPP less than 15 mm Hg, perfusion is inadequate to replete the energy state of the myocardium during cardiac arrest.

Even brief interruptions in chest compressions can result in decreased CPP. This fact has clinical importance, as evidenced by the fact that more uninterrupted chest compression are highly associated with restoring circulation and survival.²³ When chest compressions are measured during actual resuscitations by paramedics or hospital providers, interruptions and pauses are frequent.^47,48 These observations have prompted the development of monitor-defibrillators with features to measure and record chest compressions, and even to provide real-time feedback to providers.⁴⁹ These devices often employ an accelerometer in the defibrillation pads or in a separate attachment that is placed on the sternum of the patient. It is unknown if real-time feedback can improve resuscitation success, but it is clear that the continuity and quality of chest compressions are important parameters to maximize for resuscitation to succeed. Recent literature has used the chest-compression fraction to quantify this parameter (Figure 33-5).

Figure 33-5 Chest compression fraction (CCF) describes the continuity of chest compressions during resuscitation. Tracings of chest compressions detected by the accelerometer of a monitor-defibrillator can be used to calculate the proportion of time when chest compressions are occurring. In top tracing, there are few pauses, and CCF is 0.90. In lower tracing, frequent interruptions for breaths or procedures results in CCF of 0.60.

It is important to recognize that peak arterial pressure or palpable pulses measured during chest compressions do not necessarily represent CPP, because ventricular pressures are simultaneously elevated. Consequently, palpation of pulses developed by chest compressions and systolic pressures developed by chest compressions may be misleading. It is most useful to follow the “diastolic” or relaxation-phase arterial pressure. If unable to follow any of these pressures, the clinician must rely on indirect evidence of myocardial perfusion, such as improved electrical and mechanical activity or increased pulmonary CO₂ excretion.

Direct cardiac compression via a thoracotomy is more effective than external chest compressions, producing roughly threefold increases in CPP.^50,51 This approach also allows recognition of cardiac tamponade and treatment by pericardiotomy. Mechanical activity and fibrillation are immediately visible, and electrical rescue shocks or pacing can be applied directly to the heart. In the setting of cardiopulmonary collapse due to exsanguination, thoracotomy also allows aortic compression to shunt blood to heart and brain, as well as direct control of intrathoracic bleeding. Until the 1960s, thoracotomy was the standard approach for treatment of sudden cardiac arrest, but this procedure has now been supplanted by closed-chest compressions. Cases series describe how this technique continues to be successful, and its use should be considered when closed-chest compressions are ineffective.⁵⁰ Open-chest cardiac massage is most likely to succeed if initiated early during resuscitation.⁵²

Delivery of chest compressions is often inadequate, and uninterrupted chest compressions are critical for restoration of circulation.^23,30,53,54 A variety of mechanical devices have been developed to provide more consistent and continuous chest compressions.⁵⁵ Some of these devices exploit circumferential compression or active compression/decompression of the chest to increase the efficiency of artificial circulation. Clinical trials comparing these devices have not demonstrated any benefit versus manual chest compressions, and one trial found a trend towards worse neurologic outcome.⁵⁶ While no current device is poised to solve this challenge, the constant attention of industry to this area illustrates the need for strategies to improve delivery of chest compressions.

Extracorporeal perfusion for restoration of circulation is highly effective and can be used to resuscitate subjects for whom chest compressions have failed.^57,58,59 With extracorporeal support, more time becomes available to address the primary cause for cardiac arrest. However, this approach requires specialized technical skill and has increased cost and risk. Logistical issues include limited availability of perfusion equipment, setup time for circuit priming, and delays in establishing adequate venous and arterial access. Development of portable cardiopulmonary bypass devices that can be primed quickly, along with improved techniques for rapid vascular access, could broaden the use of this technology. At present, this approach is used only in specialized centers that have devoted institutional resources to set up a dedicated program.

ECG Monitoring

Continuous three-lead ECG monitoring is essential for guiding resuscitation. A practical division of the ECG is to divide rhythms into organized and not organized. Organized rhythms include supraventricular rhythms or ventricular tachycardia (VT). Not-organized rhythms include ventricular fibrillation (VF) and asystole. Not-organized rhythms cannot support the pumping of blood, regardless of volume status, cardiac muscle state, and vascular integrity. Therefore, restoring cardiac electrical activity to an organized rhythm is an essential step in resuscitation. Organized rhythms can support pumping of blood unless they are too slow (<30-40 complexes/min) or too fast (>170-180 complexes/min). An organized rhythm in the absence of pulses is termed pulseless electrical activity (PEA).

Any organized complex that is not associated with perfusion should be considered PEA. The absence of perfusion in the presence of organized electrical activity may result from damage to heart muscle (as in massive MI) or from uncoupling of electrical and mechanical activity (as in prolonged circulatory arrest). Perfusion may be so poor that pulses are absent in VT, supraventricular tachycardia, and atrial fibrillation with rapid ventricular response, which are unresponsive to the filling of the heart. These tachyarrhythmias should be corrected by rescue shock. Outside of these tachyarrhythmias, the rate of complexes in PEA is related to the ischemic state of the heart and may be used to monitor resuscitation efforts. With increasing ischemia, energy depletion will occur in the electrical system, and the rate of PEA will slow. If resuscitation is improving the energy state of the heart, the rate of PEA will accelerate. Anecdotally, narrow complexes reaching rates of 80 to 100 beats per minute often herald the return of pulses. Falling rates of complexes in PEA reflect unsuccessful resuscitation efforts, probably because of inadequate perfusion of the cardiac conduction system.

VF and asystole lie along a continuum of not-organized ECG. Arbitrary peak-to-peak amplitude of the ECG is usually used to distinguish asystole (amplitude < 0.1-0.2 mV) from VF (amplitude > 0.2 mV).⁶⁰ However, VF also exhibits temporal structure that may be absent in asystole.⁶¹ VF is a chaotic electrical activity formed by multiple interacting waves of activation within the heart.⁶² VF emerges from broken wavefronts that result from an area of ischemia (as in MI), an area of prolonged refractoriness (as in drug-induced or inherited prolonged QT intervals), or too-rapid succession of activation potentials (as in tachycardia or an “R on T” premature beat). As the organization and amplitude of these waves decline, because of ischemia or hypoxemia, the amplitude of the ECG also declines. Reperfusion of the heart in asystole may restore VF. Furthermore, the amplitude and organization of the VF increase with reperfusion, providing a marker of adequate artificial perfusion.

Rational Use of Rescue Shocks for Defibrillation

Delivery of immediate transthoracic electric (rescue) shocks to patients in VF can convert VF into an organized cardiac rhythm. Rescue shocks are highly effective when VF is of very brief duration (<1-2 minutes). These shocks may work by depolarizing the heart, canceling the original wavefronts, or by prolonging the refractory periods.⁶² Although rescue shocks can successfully restore an organized rhythm, repeated shocks may directly damage the myocardium. The precise magnitude of this damage is still unclear.⁶³ Nevertheless, optimal therapy should provide rescue shocks at the lowest effective energy while minimizing the number of unsuccessful rescue shocks.

Rescue shocks are more likely to fail when cardiac arrest has lasted more than a few minutes. In the out-of-hospital setting when the collapse is not witnessed by the paramedic, only 9% to 12% of rescue shocks restore an organized ECG.^64,65 Furthermore, resuscitation is less likely after rescue shocks that convert VF into asystole.⁶⁶ In one model for defibrillation, a “critical mass” of the heart must be depolarized by a rescue shock to ensure that VF activation potentials are extinguished.⁶² If a critical mass is not defibrillated, chaotic activity in the remaining regions will spread throughout the heart, rekindling VF. However, even when the entire heart is depolarized by the shock, VF may recur, perhaps because of heterogeneous areas of refractoriness or persistent foci.⁶⁷ Regardless, shocks must be of sufficient intensity to deliver depolarizing current to the majority of the heart.

Several maneuvers can facilitate electrical defibrillation. Multiphasic shock waveforms that produce more effective depolarization of individual myocytes (biphasic waveforms, for example) tend to accomplish defibrillation with less energy than monophasic waveforms.⁶² Consequently, most defibrillators available now deliver biphasic waveforms. Increased pressure of paddles from 0.5 to 8 kg on the chest will decrease transthoracic impedance by as much as 14%, increasing delivery of current to the heart.^68,69 This advantage of paddles must be weighed against the increased safety and convenience afforded by hands-free self-adhesive defibrillation pads, and most settings now have adopted the hands-free pads. In the past, multiple shocks would be delivered in rapid succession to decrease chest impedance. However, repetitive shocks decrease chest impedance only about 8% or less in actual patients,^68,70,71 which does not justify the interruption of artificial circulation to deliver “stacked” shocks. Minimal interruption of chest compressions to deliver a single shock, followed by immediate resumption of chest compressions, is now recommended. Reducing the delay from the last chest compression to the delivery of the rescue shock has also been associated with greater resuscitation success.⁷² Therefore, chest compressions should continue while the defibrillator is charging and only stop at the last moment prior to shock. Rescue shocks of sufficient energy with multiphasic waveforms should be delivered singly to the patient in VF using firm paddle pressure on the chest or hands-free self-adherent pads.

For VF that has lasted more than a 3 to 4 minutes, preclinical data suggest that delaying rescue shocks until after a few minutes of chest compressions will improve rescue shock success. In animals, reperfusion prior to rescue shocks appears to be preferable to immediate rescue shock after more than 5 minutes of untreated VF.^73–76 To date, two clinical studies in out-of-hospital cardiac arrest have found that either 90 seconds or 3 minutes of chest compressions prior to delivery of the initial rescue shock improved resuscitation rates for subjects with VF outside the hospital, particularly when rescuer response intervals are longer than 4 minutes.^2,77 However, a third study found no difference in outcomes with 5 minutes of chest compressions prior to shock,⁷⁸ while a fourth study found no difference in outcomes with 3 minutes of chest compressions prior to shock.⁷⁹ Finally, a large multicenter trial comparing immediate rescue shock to 3 minutes of chest compressions prior to rescue shock recently stopped enrollment, reportedly finding no difference between groups.⁸⁰ Taken together, the clinical data suggest that the first rescue shock for VF should be delivered as soon as possible within 3 to 5 minutes as long as chest compressions are started immediately, but that there is no reason to intentionally delay the rescue shock.

Quantitative analysis of the VF waveform can distinguish early VF from late VF and may be useful for estimating the likelihood of rescue shock success.⁸¹ Larger amplitude of VF suggests early VF and is associated with more successful resuscitation.⁸² However, amplitude can be affected by body habitus and other recording conditions. Frequency-based measures, as well as nonlinear dynamical measures, also can be used to quantify VF and estimate the probability of rescue shock success and are less dependent on recording conditions.^83–86 All these measures, or some combination of these measures, are likely to be implemented in future generations of defibrillators. These devices may provide real-time, semiquantitative estimates of the probability that a rescue shock will succeed in restoring an organized rhythm. Using this information, the clinician will be able to choose to shock VF when the probability of shock success is high, or to concentrate on improving the situation with artificial perfusion when the probability of shock success is low.

Drug Therapy

All drug therapy in cardiac arrest can be divided into three categories: pressors, antidysrhythmics, and metabolic drugs. There is good evidence that pressors improve artificial circulation, making restoration of pulses more likely. However, recent data in out-of-hospital cardiac arrest accentuate the point that no drug therapy has been demonstrated to improve long-term survival.⁸⁷ Antidysrhythmic drugs are effective for preventing dysrhythmias and therefore have a role in stabilizing the heart once circulation is restored. The value of antidysrhythmic drugs for terminating VF or reversing asystole is less clear. Metabolic drugs, primarily bicarbonate, can be used to reverse acidosis or other electrolyte problems when they are recognized. However, there are no data to support routine use of these drugs for all patients.

Pressors used during resuscitation include epinephrine and vasopressin. Both of these drugs can increase CPP via actions on α-adrenergic (epinephrine) or vasopressin receptors (Figure 33-6).^88,89 Epinephrine is usually administered in 1-mg (~0.015 mg/kg) increments. In laboratory studies, the pressor effects of epinephrine during cardiac arrest are brief (~5 minute). Vasopressin has been administered as 40-unit boluses (~0.5 units/kg) and produces a longer-lasting increase in CPP (~10 minutes). Both drugs should be titrated to improvement in clinical indicators (ECG waveform, mechanical activity, changes in end-tidal CO₂ or CPP). Direct comparisons between vasopressin and epinephrine have failed to demonstrate a clear superiority of one drug over the other or of the combination of the two over either alone.^90–93

Figure 33-6 Administration of a vasoactive drug, in this case vasopressin, can increase coronary perfusion pressure (CPP) produced by chest compressions. Note that CPP generated by chest compressions alone is below the 15 to 20 mm Hg believed necessary for restoration of circulation. However, aortic pressure (Ao) and thus CPP increases above this threshold 40 to 60 seconds after drug administration (arrow), while right atrial (RA) pressure remains unchanged. When treating cardiac arrest, it is reasonable to expect vasoactive drug will act after 60 more seconds of chest compressions. (Unpublished laboratory data.)

The 1-mg dose of epinephrine may be too small to restore pulses after circulatory arrest lasting more than a few minutes. However, trials in out-of-hospital patients comparing higher initial boluses of epinephrine (15 mg versus 1 mg) found a higher rate of restoration of pulses (13% versus 8%) and admission to the hospital (18% versus 10%),⁹⁴ but overall survival was not different. Comparison of a lower dose (7 mg versus 1 mg) of epinephrine in both in-hospital and out-of-hospital cardiac arrest found no change in restoration of pulses or survival.⁹⁵ Likewise, comparison of 0.02 mg/kg versus 0.2 mg/kg epinephrine found no change in restoration of pulses, or survival.⁹⁶

It is possible that the β-adrenergic effect of these higher doses of epinephrine produces toxicity that limits long-term survival. Postarrest impairment of cardiac index and oxygen delivery has been related to epinephrine dose.⁹⁷ Likewise, neurologic impairment has been related to epinephrine dose.⁹⁸ No trial has completed a direct comparison of epinephrine with more selective α-adrenergic agents such as phenylephrine. However, one trial found no advantage from administration of 11 mg of norepinephrine.⁹⁴

Vasopressin can increase CPP without complicating β-adrenergic effects. Resuscitation rates and survival are identical for patients resuscitated with vasopressin and standard doses of epinephrine after in-hospital⁹⁰ or out-of-hospital cardiac arrest.⁹¹ However, post hoc analyses suggest vasopressin may be superior for resuscitation and survival of patients whose first ECG rhythm is asystole and for those subjects requiring multiple doses of vasopressors.⁹¹ Subsequent studies examined the combination of epinephrine with vasopressin versus epinephrine alone for treatment of cardiac arrest.^92,93 These studies found no difference in outcomes with the combination of drugs. Therefore, use of either vasopressin or epinephrine is justified in the setting of cardiac arrest. None of the studies on these pressors standardized post–cardiac arrest care, and therefore all are limited for assessing drug effects on neurologic recovery.

Although preclinical and existing clinical data support the conclusion that vasoactive drugs during CPR can increase the probability of restoring spontaneous circulation, it is unclear whether these drugs actually improve overall survival. A recent trial in out-of-hospital cardiac arrest compared resuscitation without intravenous (IV) drugs to resuscitation with IV drugs.⁸⁷ In this study, which accounted for many aspects of post–cardiac arrest care, IV drugs clearly increased the rate of restoration of pulses (32% versus 21%), but the rate of neurologically good survival did not differ (9.8% versus 8.1%). It is possible the study was underpowered to detect differences in long-term outcomes. However, the data raise the worrisome possibilities that when IV drugs are required to restore cardiac activity, severe brain injury has already occurred, or even that the drugs used add to brain injury.⁹⁸

The role of antidysrhythmic drugs during cardiac arrest is equivocal.^99,100 Atropine may relieve bradycardia when it is vagally mediated. However, nervous system influences on the heart are largely eliminated after more than 1 to 2 minutes of circulatory arrest. Therefore, there is little expectation that atropine will improve resuscitation from asystole or PEA. Lidocaine, procainamide, and bretylium have a long history of use in the treatment of VF. The basis for this use is principally the observation that these drugs can suppress dysrhythmias prior to cardiac arrest. Once VF is established, lidocaine can actually increase the electrical energy required to defibrillate by more than 50%.¹⁰¹ This effect is not true for agents that are less potent, as sodium channel antagonists. For example, administration of amiodarone (5 mg/kg) is superior to placebo¹⁰² and to lidocaine¹⁰³ for restoration of pulses to out-of-hospital patients with VF that is not terminated by three rescue shocks. These studies did not control subsequent critical care and are thus not designed to determine any effect on long-term survival. In summary, antidysrhythmic drugs are commonly used during resuscitation, but only amiodarone use has supporting human data.

Empirical treatments of metabolic disturbances during cardiac arrest are not supported by prospective human data. Bicarbonate or other buffers may improve acidemia resulting from ischemia, and systems with increased use of bicarbonate report higher rates of successful resuscitation.¹⁰⁴ However, no trial has demonstrated that sodium bicarbonate administration improves outcome.^105,106 Aminophylline has been proposed as an antagonist of adenosine released during ischemia. Adenosine is hypothesized to suppress cardiac electrical activity. Two prospective studies of aminophylline administration to subjects with PEA or asystole failed to demonstrate any improvement in resuscitation.^107,108 Use of dextrose-containing fluids versus dextrose-free fluids did not alter outcome for out-of-hospital cardiac arrest.¹⁰⁹ Other metabolic therapies including calcium and magnesium also lack supporting data.^110,111 However, it is appropriate to consider specific use of these agents to correct known abnormalities that are contributing to cardiac arrest, such as known hyperkalemia, calcium channel blocker overdose, torsades, or hypomagnesemia.

Taken together, data support a simple pharmacologic approach to treatment of cardiac arrest. First, the vasopressors epinephrine and vasopressin are useful for augmenting CPP generated during chest compressions. Other vasopressors should also be useful but lack prospective data. Second, antidysrhythmic drugs may be useful for maintaining organized rhythms but not for terminating VF. Only amiodarone has clinical data supporting its use during VF that persists after rescue shocks. All other drug therapy should be based on the clinical situation and the response of the patient.

Primary Cardiac Events

Cardiac arrest is most commonly attributable to cardiac disease. A primary dysrhythmia or cardiogenic shock is the most common proximate cause of cardiac arrest.^112,113 Patients undergoing angioplasty have 1.3% incidence of cardiac arrest, and survival in these patients resembles survival in other populations.¹¹⁴ Among patients admitted to the hospital with acute MI, cardiac arrest occurs in 4.8%.¹⁷ Dysrhythmias are common during the hours after reperfusion therapy,¹¹⁴ although reperfusion therapy reduces the overall risk of cardiac arrest.¹¹⁵ During acute MI, cardiac arrest is most likely in patients with lower serum potassium levels, more than 20 mm of total ST elevation, and a prolonged QTc interval during the first 2 hours of their event.¹¹⁵ With a mean follow up of 43 months, 3.3% of subjects surviving acute MI suffered sudden cardiac death.¹¹⁶ Abnormalities of the heart are present in most cases of cardiac arrest, with coronary artery disease present in at least 65% of autopsies.¹¹⁷ Taken together, these data suggest that most patients with cardiac arrest will have contributing cardiovascular disease.

An acute coronary syndrome is present in more than half of patients presenting with primary cardiac arrest outside of the hospital. When angiography was performed on consecutive patients resuscitated from cardiac arrest, coronary artery occlusion was identified in 48%.¹¹⁸ Similarly, 51% of initially resuscitated outpatients exhibited cardiac enzyme elevation or ECG evidence of acute MI.¹¹⁹ In one series, troponin T was elevated in 40% of out-of-hospital patients undergoing CPR, regardless of whether circulation was restored.¹²⁰ The direct myocardial injury from defibrillation and CPR may cause spurious elevations of creatine kinase that are unrelated to cardiovascular disease.¹²¹ However, cardiac troponin elevations are believed to reflect acute MI rather than injury from electric shocks.¹²² Thus the 40% of subjects undergoing CPR with elevated troponin probably suffered myocardial injury prior to collapse.

The high likelihood of an acute coronary syndrome in the patient suffering cardiac arrest should prompt consideration of antiplatelet therapy, anticoagulation, beta-blockade, and nitrates during the post–cardiac arrest care. Unless a clearly noncardiac etiology for cardiac arrest is evident, acute coronary angiography may reveal an indication for angioplasty, thrombolysis, or other reperfusion therapy. Early angioplasty or reperfusion therapy is associated with improved survival and outcome.^{17,118,123,124} Primary angioplasty is safe in comatose patients undergoing hypothermia treatment, and good outcomes have been reported.^125,126 Therefore, coma and its treatment should not delay emergent treatment of acute coronary syndromes if they are suspected.

Primary ventricular tachyarrhythmias are rapidly reversible and may be more likely than PEA or asystole for patients with a primary cardiac cause of collapse. Ventricular tachyarrhythmias are the initially recorded rhythm in 23% to 41% of out-of-hospital cardiac arrests^10,11,127 and in 25% of in-hospital cardiac arrests.¹⁴ Because VF is rapidly reversible, patients with this rhythm comprise the majority of survivors of cardiac arrest. Data collected over 3 decades in one city noted that the prevalence of VF in out-of-hospital cardiac arrest has declined since 1978.¹⁰ This trend may reflect a change in preventive medicine or in the epidemiology of cardiovascular disease over time. Once defibrillation is accomplished, prophylactic treatment of ventricular dysrhythmias is usually not required and may be harmful, given the side effects of most drugs. However, recurrent malignant dysrhythmias may be treated with infusions of lidocaine, amiodarone, procainamide, or other antidysrhythmics while searching for and treating the underlying etiology. Current practice most often employs amiodarone or lidocaine.

Long-term antidysrhythmia treatment should be considered for patients who survive sudden cardiac arrest. At a minimum, treatment should be considered for patients with decreased left ventricular function or primary dysrhythmia without a reversible cause.¹²⁸ Importantly, subjects surviving a life-threatening ventricular dysrhythmia had a 15% to 20% risk of death during a mean of 16 months of follow-up, even when a reversible cause of the dysrhythmia such as electrolyte disturbance or hypoxia can be identified.¹²⁹ Implantable defibrillators have been found to be superior to antidysrhythmic drugs for reducing the risk of subsequent death.¹³⁰ This benefit is primarily in subjects with a left ventricular ejection fraction (LVEF) less than 0.35.¹³¹ Implantable defibrillators were not better than antidysrhythmic drugs in a European trial that enrolled subjects resuscitated from cardiac arrest secondary to ventricular dysrhythmia without regard to LVEF.¹³² Nevertheless, these devices offer significant hope of preventing sudden cardiac death, and identification of patients they may benefit is an active area of research. At present, implantable defibrillators should be discussed for patients who recover from coma with LVEF less than 0.35 or survive a ventricular arrhythmia in the absence of clearly reversible causes.

Asphyxia

Asphyxia-induced cardiac arrest can result from drowning, choking, asthma, progressive respiratory failure with hypoxemia, or traumatic coma with hypoventilation. Acute asphyxia causes transient tachycardia and hypertension, followed by bradycardia and hypotension, progressing to PEA or asystole. This period of blood flow with severe hypoxemia prior to cardiac arrest may make asphyxiation a more severe injury than VF or other rapid causes of circulatory arrest.¹³³ Brain edema is more common on CT scans after resuscitation when cardiac arrest was caused by pulmonary rather than cardiac etiologies.¹³⁴

During cardiac arrest, pulmonary edema develops from redistribution of blood into the pulmonary vasculature.¹³⁵ Thus, oxygenation is only worsened in the asphyxiated patient. Attention to the primary cause of asphyxia, as well as to maneuvers that will increase oxygenation (e.g., increased end-expiratory pressure or increased inspiration to expiration time ratios), may be necessary.

Pulmonary Embolism

Pulmonary emboli may occur in the postsurgical patient, as well as in medical patients with impaired mobility.¹³⁶ In one series, pulmonary emboli were present in 10% of in-hospital deaths,¹³⁷ and the prevalence among out-of-hospital deaths was similar.¹³⁸ Pulmonary emboli can result in rapid cardiopulmonary collapse and should be considered as a possible etiology of cardiac arrest in the proper clinical setting or when collapse is preceded by sudden shortness of breath, hypoxemia and/or pleuritic chest pain.

Physiologically, pulmonary emboli can result in cardiac arrest if a large thrombus obstructs right ventricular outflow into the pulmonary arteries. This situation results in a dilated, distended right ventricle and an empty left ventricle. Right ventricular dilation is sufficiently profound that it can be seen on transthoracic echocardiogram. Circulation cannot be restored unless this obstruction is relieved. Because the primary disturbance is hypoxemia and decreased cardiac output, cardiac arrest from pulmonary embolism should present with an initial rhythm of PEA or asystole.

Administration of bolus fibrinolytic drugs (tissue plasminogen activator, streptokinase, or urokinase) has been speculated to help acutely during resuscitation of a patient with a suspected pulmonary embolism. Smaller pulmonary emboli can lead to cardiac arrest due to hypoxemia, and resuscitation may be possible prior to fibrinolysis if adequate oxygen exchange can be restored. Tenecteplase has been used with reported success in non-randomized trials during resuscitation of undifferentiated patients,¹³⁹ but it failed to demonstrate benefit in a larger randomized trial of undifferentiated patients in cardiac arrest.¹⁴⁰ Likewise, a randomized trial of tissue plasminogen activator to patients with out-of-hospital cardiac arrest and an initial rhythm of PEA failed to demonstrate any benefit, although drug administration was late during resuscitation.¹⁴¹ There are few data about treatment of cardiac arrest resulting from massive pulmonary emboli, and certainly the overall prognosis in this setting is dismal. Individual decisions to use fibrinolytics in this setting must be tempered by the low probability of success.

Electrolyte Disturbances

Potassium disturbances are the most likely electrolyte disturbance to result in cardiac arrest. In cardiac patients, hypokalemia has been linked to the incidence of VF after MI.^115,142 Hypokalemia also may account for the increased incidence of sudden death in patients taking large doses of diuretics. VF is rare in patients where serum potassium is maintained over 4.5 mEq/L. Conversely, hyperkalemia can prolong repolarization increasing the likelihood of VF initiation. Hyperkalemia may also suppress automaticity in the myocardial electrical system, leading to bradycardic PEA or asystole. Interestingly, cardiac arrest occurring during hemodialysis is not associated with high or low potassium levels but is more common when patients are dialyzed against a low (0 or 1 mEq/L) potassium dialysate.¹⁴³ These data suggest that rapid changes in potassium rather than the absolute value are important triggers of cardiac arrest in this population. Derangements of calcium and magnesium may produce similar or synergistic changes in cardiac conduction.

The clinical setting of cardiac arrest may suggest a primary electrolyte disturbance. Heavy diuretic use or intestinal fluid loss, for example, suggests potassium depletion. Suspected hypokalemia will not change acute resuscitation, but it must be addressed promptly in the post–cardiac arrest stabilization of the patient. Cardiac arrest in a patient with renal failure or during potassium infusion suggests hyperkalemia. Widened ventricular complexes with repolarization abnormalities on ECG would heighten this suspicion. If hyperkalemia is suspected, the usual acute resuscitation maneuvers can be supplemented by bolus injection of calcium carbonate (1 mg), bicarbonate (1 mEq/kg), and perhaps insulin (0.1 units/kg) with glucose (0.5-1 gm/kg). These drugs may improve cardiac electrical stability, facilitating restoration of circulation.

Poisoning

Cardiac arrest can result from drug overdose. Therapy does not change except when specific antidotes or countermeasures to the poison are available. For example, calcium channel blocker overdose may be countered by administration of IV calcium.¹⁴⁴ Beta-blocker toxicity may require large doses of inotropic agents¹⁴⁵ or may respond to glucagon.¹⁴⁶ Digoxin overdose may respond to digoxin-binding antibodies.¹⁴⁷ In the case of narcotic-induced respiratory depression, subsequent cardiac arrest usually is a specific case of asphyxia rather than specific cardiotoxic effects. Case reports have suggested treating local anesthetic toxicity with 1 to 3 mL/kg of Intralipid.¹⁴⁸ This intervention is likely to be explored for other lipid-soluble poisonings.¹⁴⁹ One principle of poisonings is that the patient was often healthy prior to the event, and may recover well once the poison is eliminated. This potential for a better outcome may justify longer and more aggressive efforts at resuscitation.

Sepsis

Cardiac arrest can develop from sepsis for several reasons. Direct myocardial depression occurs, probably due to humoral factors.^150,151 Vasodilation results in apparent hypovolemia. Finally, impaired oxygen extraction, shunting, and mitochondrial depression can produce cellular hypoxia. Because pump and vascular failure are the principle physiologic derangements, the most common initial ECG rhythm would be expected to be a rapid PEA that slows to asystole with ischemia. When these processes have progressed to cardiac arrest, large doses of inotropes, vasoconstrictors, and volume may be needed to restore circulation. Acute volume resuscitation may require 100 mL/kg of isotonic fluids or more and must be titrated to physiologic endpoints (for example, central venous oxygen saturation, CVP, or urine output) rather than according to recipe. Because the underlying sepsis physiology will still be present if pulses are restored, these patients may prove exceedingly unstable during the subacute recovery period and have a reduced chance for survival.^27,152–154

Trauma/Hemorrhage

Hypovolemic cardiac arrest occurs after severe trauma, gastrointestinal hemorrhage, or other blood loss. Absence of venous return results in an empty heart that cannot produce cardiac output despite normal inotropic state and normal vascular tone. As with sepsis, this situation would most likely present with a rapid PEA that slows to asystole, but VF can develop in response to the global ischemia. Because cardiac function and vascular function are initially normal, inotropes and vasoconstrictors are unlikely to benefit hypovolemic cardiac arrest. Rapid replacement of volume with crystalloid infusion is indicated. Colloid or blood infusion should correct the situation more rapidly.¹⁵⁵ After restoration of circulation, patients with hemorrhagic cardiac arrest are likely to develop multisystem organ failure.¹⁵⁵

During hypovolemic cardiac arrest, the empty cardiac ventricles render external chest compressions ineffective. If blood loss is ongoing or if massive volume replacement cannot be instituted rapidly, thoracotomy allows clamping or compression of the aorta, perhaps retaining sufficient blood in the proximal aorta to perfuse the coronary and cerebral arteries. This procedure has reported success in the treatment of penetrating traumatic injuries,¹⁵⁶ but not in blunt trauma.¹⁵⁷ Survival is better if thoracotomy occurs in the operating room after brief loss of pulses, and best if the penetrating injury has created cardiac tamponade that is directly relieved by pericardotomy. Clearly, restoration of circulation must be accompanied by repair of the site of hemorrhage.

Hypothermia

Hypothermia represents an important situation where prolonged resuscitative efforts are justified. If hypothermia develops prior to circulatory arrest, the tolerance of the heart and brain to ischemia is greatly prolonged. Survival with favorable neurologic recovery has been reported after cold-water submersion or exposure with cardiac arrest and resuscitation efforts lasting several hours.^158,159 Although all data are retrospective, subjects in whom circulatory arrest occurs because of hypothermia appear to be more salvageable than subject who asphyxiate or have circulatory arrest prior to becoming cold.¹⁶⁰

Treatment should be based upon the initial temperature of the patient. Between 32°C and 37°C, no change in drug or electrical treatment is required, and this level of hypothermia may be beneficial for resuscitation of both brain and heart.^161,162 Between 29°C and 32°C, cardiac activity may be preserved, and external warming (warm air, heating lights, warm blankets) and warm IV fluids should accompany usual resuscitation efforts. The likelihood of generating sufficient perfusion to rewarm the body declines as temperature decreases from 32°C to 29°C, and more invasive warming should be considered if there is not a rapid response with external warming. More invasive and aggressive treatment will almost certainly be required for cooler patients, because both mechanical and electrical activity of the heart is disrupted at temperatures below 28°C. Patients below this temperature may exhibit PEA, VF that is refractory to defibrillation attempts, or asystole. Repetitive rescue shocks in such patients are not justified and may be detrimental. Efficacy of most resuscitation drugs may be impaired.

Several techniques for active rewarming during resuscitation of victims of severe hypothermia are available. Given the potentially prolonged tolerance of the cold patient to ischemia, there may be sufficient time to establish arterial and venous access for partial or complete cardiopulmonary bypass. Extracorporeal circulation is particularly useful in these subjects because it can provide artificial circulation at the same time as rewarming.^160,163,164 In the absence of extracorporeal circulation, placement of thoracostomy tubes and lavage of the chest with warm fluids is an option.¹⁶⁵ Thoracostomy is intuitively preferable to peritoneal lavage because the heart will be directly warmed. Warm air forced over the body surface can rewarm a patient, although this technique may provide the least heat exchange.¹⁶⁶ In any case, it is difficult to determine whether or not circulation can be reestablished in the profoundly hypothermic patient until near physiologic core temperatures (33°C-37°C) are restored.

Other Medical Conditions

Comorbidities have a tremendous influence on the outcome from cardiac arrest.^113,167,168 In some cases, cardiac arrest may be an expected progression of the patient’s disease, but guidelines about limiting resuscitation were not set out. For example, no survivors were reported among cancer patients with expected cardiac arrest.¹⁶⁸ Therefore, it may be appropriate to set limits on resuscitation efforts in certain medical conditions prior to cardiopulmonary collapse. Ideally, discussion about the expectations for resuscitative efforts should be held with the patient, their family, or the patient’s representatives prior to cardiac arrest. If those discussions did not occur prior to the first cardiac arrest, they should follow promptly any initially successful resuscitation.

Temperature Control

Meticulous avoidance of fever is important during the first 24 to 48 hours after ischemic brain injury. Temperature control after cardiac arrest may be confounded by the fact that bacteremia and spontaneous fever is common in the resuscitated patient.^182,183 The benefit of lower temperatures for injured brain has been demonstrated after traumatic brain injury, stroke, and cardiac arrest.^179,184,185 Mechanistically, temperature probably affects more than brain metabolic rate. For example, manipulations of temperature that improve neurologic recovery in laboratory studies produce no effect on jugular venous lactate or oxygen uptake.¹⁸⁶ Recent laboratory investigations suggest that a variety of signaling pathways and cellular responses are sensitive to relatively small (1°C-2°C) changes in brain temperature.^172,173

Induction of mild hypothermia for resuscitated patients produces a 24% to 30% relative risk reduction for death or poor neurologic outcome.¹⁸⁷ Mild hypothermia (32°C-34°C) maintained for 12 or 24 hours significantly improved the odds of survival and good neurologic outcome for subjects resuscitated from VF cardiac arrest.^161,178 There is no biological basis to believe that this neurologic benefit of induced hypothermia is specific to patients with one type of cardiac rhythm, and use in all patients who remain comatose after cardiac arrest seems reasonable. Multiple case series report successful application of therapeutic hypothermia for patients after out-of-hospital and in-hospital cardiac arrest with all initial rhythms.^3,188,189 At the time of resuscitation, many patients are already mildly hypothermic, with core temperatures between 35°C and 35.5°C.^161,178,190 This spontaneous cooling may result from equilibration of core and peripheral blood compartments during circulatory arrest. Subsequent to restoration of circulation, patients will rewarm within a few hours unless specific interventions are instituted.^179,191

The optimal duration of cooling, the maximum delay in achieving target temperatures, the optimal target temperature, and the preferred rate of rewarming are unknown. Laboratory studies suggest that cooling to between 32°C and 35°C for 12 to 24 hours is beneficial, particularly if cooling is achieved within 6 hours after resuscitation. These studies also suggest that temperature is less important more than 48 hours after resuscitation, but that rewarming should be performed slowly (<0.25°C/hour). Clinical data to answer these practical questions are likely to become available over the next few years as therapeutic hypothermia becomes widespread.

After cardiac arrest, mild hypothermia can be induced and maintained by a variety of techniques including surface cooling and endovascular devices.¹⁶² Surface cooling with ice packs and cooling blankets is tolerated by the comatose patient.^178,192 Initial studies using surface cooling alone suggested that it is slow and may require 4 to 6 hours to reach 34°C.^161,184,193 However, neuromuscular blockade and sedation can help prevent shivering or other compensatory responses and thereby greatly facilitate surface cooling. Recent reports indicated that core temperature can be reduced below 34°C for the majority of patients within 4 hours with surface cooling.¹⁹⁴ There are few direct comparisons of surface cooling and endovascular cooling. What data exist suggest that endovascular devices are no faster than water-circulating surface devices, but that endovascular catheters provide more stable control of temperature over time.^195,196 Local cooling of the head is unlikely to produce brain hypothermia when there is adequate perfusion by warm core blood,¹⁹⁰ although the head can be an effective site for removing heat from the body.¹⁹⁷

Additional maneuvers are available to accelerate cooling after cardiac arrest. For example, rapid infusion of 30 mL/kg cold (4°C) crystalloid produces a rapid decrease in core temperature and is tolerated by the post–cardiac arrest patient.^{191,198–200} Effective use of cold fluid boluses requires that they are administered quickly into the central circulation (via central line or under pressure infusion via peripheral line). The volume required may limit this intervention to those patients without renal failure or pulmonary edema. Gastric or bladder lavage are labor intensive and less effective than cold fluid administration. It is also critical to appreciate that cold IV fluids only produce a transient decrease in core temperature, requiring that a maintenance technique (endovascular or surface cooling device) be in place after the infusion.^191,200 Taken together, these data support the rapid induction of mild hypothermia by bolus infusion of cold IV fluids (unless contraindicated), followed by maintenance of hypothermia by surface or endovascular cooling, neuromuscular paralysis, and sedation. One additional precaution that must be considered if neuromuscular blockade is required is the high incidence of seizures after cardiac arrest. Neurophysiologic monitoring should be in place if the patient is paralyzed.

Fluid and electrolyte shifts are the primary management concerns during induction of hypothermia. Induction of cooling can result in peripheral vasoconstriction, with an apparent reduction in vascular volume.²⁰¹ CVP will rise, followed by diuresis. Conversely, at the time of rewarming, vessels will dilate, CVP will fall, and the patient may appear relatively hypovolemic. Volume status should be followed closely, and a need for additional volume infusion to maintain blood pressure and urine output should be anticipated at the time of rewarming. Inattention to this fluid shift was cited as a pitfall in trials of therapeutic hypothermia for traumatic brain injury.²⁰¹ The initial diuresis, along with shifts between intracellular and extracellular compartments, can result in hypokalemia, hypophosphatemia, and hypomagnesemia at cooling, followed by hyperkalemia at rewarming.^202,203 Frequent monitoring and correction of electrolytes during these transitions is warranted.

When indicated, primary angioplasty can be conducted in patients undergoing hypothermia treatement.^125,126 Conscious patients undergoing angioplasty for acute MI tolerate mild hypothermia.¹⁶² In these patients, cooling did not interfere with defibrillation when it was required. Cardiovascular complications of hypothermia are rare with temperatures greater than 30°C. Cooling from 37°C to 31°C actually has a positive inotropic effect, increasing stroke volume to a greater extent than it decreases heart rate.²⁰⁴ Systemic vascular resistance does not appear to change greatly. Clinical data report a transient 18% decline in cardiac index with cooling to 33°C.²⁰⁵

Other complications of mild hypothermia are few when the cooling period lasts less than 24 hours. Infections become more common if cooling is prolonged for more than 24 hours. There is a suggestion that infections were slightly more common in post–cardiac arrest patients cooled for 24 hours,¹⁶¹ but not in those cooled for 12 hours.¹⁷⁸ Although mild hypothermia can inhibit platelet function and coagulation,²⁰⁶ these changes are of small magnitude, leading to few bleeding complications in studies to date. These studies included subjects with concurrent trauma or administration of heparinoids and glycoprotein IIb/IIIa inhibitors.^162,184 There are reports of infrequent but increased bleeding in post–cardiac arrest patients treated with hypothermia and cardiac catheterization (6.2% of patients).²⁰⁷ Elevations of pancreatic enzymes have been reported in cooled patients, but these changes resolve with rewarming.^161,205 Creatinine clearance and platelet count may fall during cooling, but both parameters normalize with rewarming.²⁰⁵

Blood Pressure and Cerebral Blood Flow

After cardiac arrest, the heart experiences a reversible period of decreased mechanical function.²⁰⁸ The biochemical basis for this dysfunction is an active area of study. Moreover, reperfusion at the time of resuscitation includes oxidative stress or other triggers that can lead to myocyte death over time.²⁰⁹ From a clinical standpoint, some vasoactive drug support is necessary in a large proportion of patients resuscitated from cardiac arrest lasting more than 3 or 4 minutes. This dependence on vasoactive drugs should decline over the subsequent 24 to 48 hours.

Autoregulation of cerebral blood flow is disturbed after cardiac arrest. Although rarely used clinically, measurement of oxygen saturation in the jugular bulb venous blood allows calculation of brain oxygen extraction. Furthermore, cerebral blood flow can be estimated using transcranial Doppler ultrasound or nuclear imaging. During the first day after resuscitation from cardiac arrest, patients exhibit increased cerebral vascular resistance²¹⁰ and impaired cerebral autoregulation.^211,212 When autoregulation is present, it is right shifted such that brain perfusion declines when mean arterial pressure (MAP) declines below 80 to 120 mm Hg. When blood pressure is maintained, clinical positron emission tomography (PET) studies suggest that regional perfusion remains matched to metabolic activity after cardiac arrest.^213,214 Thus, available data indicate that normal perfusion of the brain requires a higher MAP than normal during the first few hours after cardiac arrest. Periods of hypotension after restoration of pulses may add significant secondary ischemic brain injury.

If tolerated by the heart, relative hypertension (MAP of 80–100 mm Hg) should be maintained to prevent brain hypoperfusion. Maintaining this level of hypertension often requires infusions of inotropes and/or pressors. In support of this recommendation, hypotension during the first 2 hours after resuscitation is associated with poor neurologic recovery in patients admitted to the hospital after cardiac arrest.¹⁸⁰ No specific choice of pressors has been demonstrated to be superior. Dopamine 5 to 20 µg/kg/min, norepinephrine (0.01-1 µg/kg/min) and/or epinephrine (0.01-1 µg/min) are all potential agents. In addition, the doses of these agents must be titrated, and patient requirements may be very dynamic in the early period after cardiac arrest.

Glucose Control

Elevated serum glucose is associated with poor outcome after cardiac arrest¹⁰⁹ and may be a marker of prolonged or difficult resuscitation. Both epinephrine and physiologic stress can elevate serum glucose. However, multivariate models that accounted for resuscitation time and medication usage still show an effect of serum glucose on admission and during the first 48 hours of intensive care on long-term outcome.^5,215 Despite this association, both studies noted that monitoring of glucose in nondiabetic patients was infrequent.

It has not been established whether any specific strategy to control hyperglycemia will improve outcome after cardiac arrest. Intensive glycemic control with a low target range (72-108 mg/dL, 4-6 mmol/L) is quite controversial outside of the surgical population where it was first studied²¹⁶ and has not proven beneficial and may be harmful in medical intensive care.^217,218 After cardiac arrest, there was no difference in outcome when a moderate glucose was targeted (108-144 mg/dL; 6-8 mmol/L) versus a strict lower range (72-108 mg/dL; 4-6 mmol/L).²¹⁹ However, the incidence of hypoglycemic events was higher in the strict versus moderate control group (18% versus 2% of patients). Given the available data, control of glucose levels above 180 mg/dL (10 mmol/L) seems reasonable, and 144 to 180 mg/dL (8-10 mmol/L) would be a safe target range. However, lower ranges (72-108 mg/dL; 4-6 mmol/L) have risk of complication without benefit and should be avoided.

Hematologic Changes

Cardiac arrest is associated with activation of coagulation that is not balanced by fibrinolysis. This hematologic profile is reminiscent of disseminated intravascular coagulation and may contribute to subsequent end-organ dysfunction. Markers of thrombogenesis that have been reported include increased thrombin-antithrombin complexes and fibrinopeptide A.^220,221 These increases are not balanced by fibrinolytic factors for at least 24 hours. The etiology of these changes is unknown and may be related to ischemic injury to the endothelium.

At present, use of anticoagulation is variable, and there are no prospective trials evaluating the effect of anticoagulation after resuscitation. Anticoagulation and even fibrinolytic drugs are safe after cardiopulmonary resuscitation.^{140,221–223} A retrospective series noted a univariate relationship between anticoagulation and 6-month survival that was not significant in a multivariate model.²¹⁵ Given the hematologic evidence of active thrombogenesis, these data suggest that at least anticoagulation should be considered immediately after resuscitation.

Infection

The physiology of the post–cardiac arrest patient resembles the systemic inflammatory response syndrome, and post–cardiac arrest infections are common. Bacteremia has been noted in 39% of patients during the first 12 hours after resuscitation.¹⁸² Fever is common in patients within 48 hours after resuscitation from cardiac arrest if active temperature control is not in place. Potential causes include contamination during emergent line placement, aspiration or transient bacteremia during airway management, and mesenteric ischemia contributing to bacterial translocation from the gut. Endotoxin and various cytokines are elevated in serum after resuscitation.²²⁴ While an intestinal origin of endotoxin was suspected, pulmonary infections were more common than bacteremia. Pneumonitis is evident in 33% to 41% of patients after cardiac arrest,^161,207 although microbiological confirmation of pneumonia is often lacking. Severe infections may contribute to overall mortality.¹⁸³ Despite these observations, the role of routine antibiotics and antipyretics has not been examined.

Neurophysiology

It is common for an electroencephalogram (EEG) to be obtained for prognostic purposes. However, the prognostic value of EEG after resuscitation from cardiac arrest is limited by its nonspecific nature and by the dynamic changes in EEG over time.²³⁸ Perhaps the greatest utility of the EEG is to diagnose seizures and exclude nonconvulsive seizures as an etiology of unresponsiveness. Seizures are diagnosed clinically in 5% to 20% of comatose patients after cardiac arrest,^161,239 and the true incidence of nonconvulsive electrographic seizures may be higher. Termination of seizures, if possible, is essential to allow untainted assessment of the neurologic examination. Aside from seizures, certain malignant EEG patterns have a strong albeit imperfect association with poor outcome. Specifically, generalized suppression (<20 µV), burst-suppression pattern associated with generalized epileptic activity, or diffuse periodic complexes on a flat background during the first week after resuscitation are associated with poor neurologic outcome.²³¹ Thus EEG cannot be used by itself to determine prognosis, but the information provided by EEG can exclude confounders (seizures) and can be integrated into the total clinical picture used to assess prognosis.

Electrophysiologic response to stimuli also can be used to assess whether the patient is neurologically intact. Recovery of longer-latency event-related potentials is associated with awakening.^238,240,241 Conversely, absence of a cortical response to somatosensory evoked potentials (SSEPs) is very specific for poor neurologic outcome.^230,231 Cortical response is usually assessed as the N20 response to electrical stimulation of the median nerve. Like EEG, SSEP responses vary with the elapsed time since resuscitation.²³⁸ Recent data suggest that the use of therapeutic hypothermia may increase the time-dependent changes in SSEP. One case series reported two patients treated with hypothermia who had absent N20 responses at 3 days after cardiac arrest but recovered cognition.²³⁵ Therefore, it may be reasonable to repeat SSEPs that show absent N20s several days apart in order to avoid false-negative tests.

Blood Markers

Several neuronal peptides appear in the blood after injury to the brain, including neuron-specific enolase (NSE) and the glia-derived protein, S-100B. After cardiac arrest, NSE reaches a maximum level in serum at 72 hours. High NSE levels at 48 to 72 hours after resuscitation are associated with poor outcome.^241–243 Serial NSE levels that continue to rise over the first 72 hours also are associated with poor outcome.²⁴³ In contrast to NSE, peak levels of S-100B in serum occur during the first 24 hours after resuscitation.²⁴² Higher S-100B levels are also associated with poor neurologic outcome.^242,244 Hypothermia treatment appears to alter serum NSE levels.²⁴⁵ Use of NSE or S-100B to determine prognosis is limited by the absence of a clear cutoff value that is unsurvivable. In addition, laboratory assays are not sufficiently standardized to be certain that cutoff values derived in one location are relevant to another, and at least NSE can be released by non-brain injury.²⁴⁶ These neuronal markers might be considered a tool for following brain injury, analogous to troponin levels for following myocardial injury, but lack the specificity or clarity required to guide therapeutic decisions.

Imaging Studies

Imaging of the brain is important to exclude injury incurred at the time of collapse and to exclude intracranial causes of collapse. A noncontrast cranial computed tomography (CT) scan to exclude hemorrhage may be prudent in the comatose patient after cardiac arrest prior to anticoagulation or fibrinolytic therapy. Several series have reported a reasonable yield of clinical information from CT scans after cardiac arrest.^247,248 In general, noncontrast CT scan is insufficiently sensitive to determine prognosis after cardiac arrest unless severe changes are present. For example, severe generalized edema is often associated with loss of brainstem reflexes and may progress to herniation and brain death (Figure 33-7).

Figure 33-7 Imaging of the brain after cardiac arrest.

A, Severe cerebral ischemia appears as sulcal effacement, with loss of contrast between gray matter and white matter. Congestion of blood in meninges (pseudo-subarachnoid hemorrhage [arrow]) sometimes is evident. This pattern on early computed tomography (CT) scan often progresses to herniation and brain death. B, Less severe early changes show edema (hypodensity) restricted to basal ganglia (arrow), with sparing of cortex. C, Increased magnetic resonance imaging (MRI) signal from extensive areas of cortex on diffusion-weighted images (DWI) of this patient correspond to devastating brain injury with persistent coma. This patient showed no improvement in coma. D, DWI for same patient in B illustrates high-intensity signal from damaged subcortical areas. In this case, cortex and other structures are normal. After 5 days of coma, this patient awoke, completed rehabilitation, and recovered completely.

Magnetic resonance imaging (MRI) has a capacity to visualize more subtle changes in brain after cardiac arrest. For example, increased cortical signals on diffusion-weighted images (DWI) or fluid-attenuated inversion recovery (FLAIR) are associated with poor neurologic outcome.²⁴⁹ For patients who remain comatose for several days and in whom clinical or electrophysiologic testing is indeterminate, MRI may be considered as an adjunct to assess the extent of brain injury. Expectations and enthusiasm for long-term support may be reduced if extensive cortical lesions are present, whereas persistence may be justified if injury is limited. In the long term, cognitive deficits are associated with global brain volume loss after cardiac arrest.²⁵⁰ An important caveat with interpretation of all brain imaging after global ischemia is the differing clinical impact of lesions in different brain regions. The anatomic complexity of the brain precludes any simple quantitative relationship between the number or size of lesions and outcome. At present, the role of MRI as an adjunct for assessing post–cardiac arrest brain damage is evolving, and its interpretation should be considered in combination with neurologic or neuroradiologic expertise.

In summary, the determination of neurologic prognosis after cardiac arrest varies from patient to patient. Changes in the clinical examination are the cornerstone of prognostication. While the patient is recovering, hypothermia and supportive care may increase the likelihood of recovery. However, electrophysiologic and imaging techniques can add additional useful information to help guide clinicians and families. The approximate timing for each of these studies is depicted in Figure 33-8.

Figure 33-8 Rational approach to neurologic treatment, monitoring, and testing after cardiac arrest. For patients remaining in coma, hypothermia treatment followed by therapy and stimulation may improve recovery. Early electroencephalogram (EEG) monitoring is recommended not for prognosis but to allow detection and treatment of seizures. Absence of cortical response on somatosensory evoked potential (SSEP) or malignant EEG patterns after hypothermia treatment may help identify subsets of patients unlikely to ever show improvement. When those tests are indeterminate, and there is no sign of clinical improvement, magnetic resonance imaging (MRI) of the brain may help quantify extent and location of injury, which in turn assists in decisions about continuing long-term support. Most tests are not required when patients exhibit clear clinical improvement. Conversely, patients who progress to brain death should undergo brain death testing once toxicologic confounds, metabolic abnormalities, and shock have been corrected.