Chapter 50. Therapeutic Hypothermia: History, Data, Translation, and Emergency Department Application

Therapeutic hypothermia (TH) has become standard of care for the management of comatose patients with return of spontaneous circulation (ROSC) after cardiac arrest. The 2010 American Heart Association (AHA) guidelines for management of post–cardiac arrest patients “recommend that comatose (i.e., lack of meaningful response to verbal commands) adult patients with ROSC after out-of-hospital ventricular fibrillation (VF) cardiac arrest should be cooled to 32°C to 34°C (89.6°F to 93.2°F) for 12 to 24 hours (Class I, LOE B). Induced hypothermia also may be considered for comatose adult patients with ROSC after in-hospital cardiac arrest of any initial rhythm or after out-of-hospital cardiac arrest with an initial rhythm of pulseless electric activity or asystole (Class IIb, LOE B).”1 Very few recommendations in the AHA guidelines are assigned a Class I recommendation, and, as practitioners at the portal of entry for the majority of cardiac arrest patients, emergency physicians (EPs) need to be familiar with this therapy and the rationale for the level of recommendation assigned to TH.

Why is TH needed to improve outcomes after cardiac arrest? When a person suffers a cardiac arrest, no effective cardiac contractility occurs, resulting in ischemia, or lack of perfusion. Chest compressions provide some degree of circulation during arrest and can deliver up to 40% of the cardiac output produced by a spontaneously beating heart. The ischemia that occurs during cardiac arrest triggers a number of pathologic processes including the production of reactive oxygen species, initiation of a profound inflammatory cascade, development of metabolic acidosis with accompanying elevated lactate levels, and endothelial and mitochondrial dysfunction, to name some of the dozens of derangements that ensue. When ROSC occurs, ischemic tissues are reperfused with blood, and this reperfusion produces its own injury pattern. The combination of ischemia and reperfusion produces the post–cardiac arrest syndrome (PCAS), a unique disease state requiring specialized care. This was first described as a specific disease entity by Negovsky in 1972 when he stated that postresuscitation syndrome was a unique disease entity with unique pathophysiology that needed to be understood if it was to be treated appropriately.2 Currently, TH is the best-studied and most effective therapy for treating patients with PCAS.3 In this chapter, we will discuss the epidemiology of cardiac arrest, the rationale for using TH, data supporting its use, practical aspects of implementation, and future directions for the therapy.

Although the United States has no standardized, mandatory reporting system, it is estimated that 400,000 cardiac arrests occur each year in the United States, 75% as out-of-hospital cardiac arrests (OHCA) and 25% as in-hospital arrests.4 Viewed from a different perspective, one arrest arrives in an ED in the United States every 2 minutes. Europe also has approximately 400,000 arrests per year.5 In Japan, where a universal cardiac arrest reporting system exists, more accurate numbers are available; in 2007, approximately 78,000 OHCA occurred nationwide.6 Survival from cardiac arrest is abysmal—approximately 7% of OHCA in the United States survive to hospital discharge.4 However, survival varies significantly depending on prehospital, intra-arrest, and postarrest variables including initial arrest rhythm, whether the arrest was witnessed by a bystander, whether bystander cardiopulmonary resuscitation (CPR) was performed, ambulance response times, availability of automated external defibrillators (AED), and quality of postarrest care. In the United States, survival rates vary from 0.2% in Detroit to 17% in the greater Seattle area.7,8 One of the main variables affecting outcome is delivery of quality postarrest care, centered on induced hypothermia. The need for high-quality postarrest care was emphasized in a 2003 publication from the National Registry of Cardiopulmonary Resuscitation (NRCPR), a registry of in-hospital cardiac arrests. Reporting on 14,792 arrests, ROSC was achieved in 39% of the patients, who had a subsequent mortality rate of 68%.9

Physicians have been interested in using induced hypothermia for clinical purposes for millennia. Hippocrates wrote about packing injured patients in ice and questioned whether this simple technique would improve outcomes. In 1814, Baron Larrey, Napoleon's chief battlefield surgeon, made observations on the effect of cold on injured soldiers. While Napoleon's army retreated from Moscow after the Russian campaign, a policy of placing injured officers near the fire while keeping injured foot soldiers in the cold was enforced. Larrey noted that injured foot soldiers appeared to do better with similar injuries than the injured officers and commented, “Cold acts on living parts such that they may remain in a state of asphyxia without losing their lives.” This statement summarizes early ideas of the mechanism of action of induced hypothermia: lowering temperature lowered metabolism, decreasing oxygen and glucose consumption and allowing injured cells time to recover.

Fay, a neurosurgeon at Temple University Hospital in Philadelphia, Pennsylvania, was the first to publish studies of the clinical application of induced hypothermia. In 1940, he reported treating cancer patients with TH.10,11 In 1959, in the journal Anesthesia and Analgesia, Benson et al from Johns Hopkins University Hospital in Baltimore, Maryland, published results of a case series of 27 perioperative cardiac arrests, some of which were treated with TH.12 The stated rationale for using TH was similar to that observed by Baron Larrey almost 150 years earlier: “Hypothermia has been shown to protect the brain against anoxia. There is a reduction in the cerebral oxygen consumption and cerebral blood flow with body cooling.” Nineteen of the arrests were resuscitated with sustained ROSC. Twelve of the 19 were treated with TH. Fifty percent (6/12) of the patients treated with TH survived, all neurologically intact, while only 14% (1/7) of the patients who were not cooled survived. They concluded that, “The improvement in survival rate from 14% to 50% with use of hypothermia is clinically significant and warrants the use of cooling in all patients who have had cardiac arrest with demonstrable neurological injury.”12

In 1964, in an article published in the Journal of the Iowa Medical Society, Safar advocated treating postarrest patients with a comprehensive management strategy centered on induced hypothermia that included attempts to gauge the cause of arrest, support ventilations and circulation, prevent seizures, and monitor closely.13 He recommended early institution of hypothermia, stating, “start within 30 minutes if no signs of CNS recovery.” There are no published case reports documenting clinical experience of Safar and colleagues with TH at the University of Pittsburgh in the 1960s. It is generally believed that they cooled patients to 30°C using ice packs and treated post–cardiac arrest patients as well as patients with other causes of brain injury including traumatic brain injury, ischemic stroke, hepatic encephalopathy, and comatose meningitis. After treating a number of patients, they abandoned the clinical use of TH and began to study its application more closely in animal models of cardiac arrest. Reasons cited for stopping clinical application of TH included coagulopathy, arrhythmias, and hypotension. These potential side effects of TH will be addressed later in this chapter.

Insights from animal experiments including the potential efficacy of hypothermia performed at 33°C instead of lower temperatures, the complexity of ischemia–reperfusion injury, the multiple physiologic processes affected by hypothermia, the need for adequate mean arterial pressure (MAP) to maintain cerebral perfusion, and a focus on patients who remained comatose after ROSC from OHCA led to reinvestigation of the utility of employing TH in humans.14–16 The first prospective study of TH in humans was published in the Annals of Emergency Medicine in 1997.17 The objective of this pilot study was to investigate the effect of induced hypothermia to 33°C begun in the ED and continued for 12 hours in the intensive care unit (ICU) on outcomes in patients with anoxic brain injury after OHCA. Twenty-two comatose OHCA VF patients treated with TH to 33°C for 12 hours were compared with 22 matched historic controls obtained from chart review. Primary endpoints were survival and good neurologic outcomes. Survival was 23% in normothermic controls versus 55% in patients treated with hypothermia (P < .05), and good neurologic outcomes occurred in 14% versus 50%, respectively (P < .05).17

In 2000, in the Journal of the American College of Cardiology, Nagao et al from Tokyo, Japan, published their results treating comatose survivors of cardiac arrest with TH with or without emergency cardiopulmonary bypass (ECPB).18 A convenience sample of 50 patients with an initial rhythm of VF presenting to the ED in ongoing cardiac arrest was treated with standard CPR. If ROSC was achieved and systolic blood pressure (SBP) could be maintained above 90 mm Hg, then hypothermia was induced. If ROSC was not achieved, patients were placed on ECPB as a rescue strategy, followed by treatment with TH if blood pressure was adequate. Of the 23 patients treated with TH, 12 (53%) had good neurologic outcomes. These pilot studies led investigators to pursue true, randomized controlled trials of TH in comatose survivors of OHCA.

In 2001, Hachimi-Idrissi et al published a randomized trial of cooling with a helmet device to achieve a target temperature of 34°C.19 Thirty patients who remained comatose after ROSC from asystole or pulseless electrical activity (PEA) cardiac arrests were randomized to either normothermia (14 patients) or hypothermia (16 patients). The study was designed to test the feasibility of inducing TH using a helmet cooling device, not the efficacy of TH in this patient population. The hypothermia patients reached target core (bladder) temperatures in a mean of 180 minutes from start of therapy. Survival was low in both arms of the study: 3/16 (18.8%) survived in the hypothermia group; 1/14 (7.1%) survived in the normothermia group. Rates of good neurologic outcome were also low: 2/16 (12.5%) patients who were cooled and none of the patients who were not cooled.

Bernard et al followed up their pilot study of TH with a pseudo-randomized trial of TH compared with normothermia, which was published in The New England Journal of Medicine in 2002.20 Seventy-seven patients who remained comatose after resuscitation from OHCA caused by VF were randomly assigned to treatment with hypothermia or normothermia depending on the day of the week. On odd-numbered days patients were treated with induced hypothermia and on even-numbered days with normothermia. Hypothermia therapy was begun by paramedics in the field, who removed the patient's clothing and applied cold packs. The target temperature was 33°C, and hypothermia was maintained for 12 hours. The primary outcome measure was survival to hospital discharge with good neurologic function. Twenty-six percent (9/34) of patients treated with normothermia had good neurologic outcomes versus 49% (21/43) of patients treated with hypothermia (P = .046). When adjusted for potential confounders including age and time to ROSC, the survival benefit with good neurologic outcome associated with TH remained (OR = 5.25; 95% confidence interval 1.47–18.76; P-value = .011) (see Figure 50-1).

In the same issue of The New England Journal of Medicine, the Hypothermia After Cardiac Arrest (HACA) Study Group published the results of a larger, randomized, prospective trial comparing hypothermia to normothermia in patients who remained comatose after being resuscitated from OHCA caused by VF.21 The target temperature range was 32–34°C for 24 hours. The primary outcome measure was good neurologic outcome at 6 months. Secondary endpoints included mortality at 6 months and the rate of complications during the first 7 days. Good outcomes occurred in 55% (75/136) of the TH patients versus 39% (54/137) of the normothermia patients (RR = 1.40; 95% confidence interval 1.08–1.81). Mortality was reduced from 55% in the normothermia group to 41% in the TH group, which was statistically significant. Despite concerns of infection, bleeding, and dysrhythmias, there were no statistically significant differences in the rate of complications between the two groups (see Figure 50-2).

In 2003, the International Liaison Committee on Resuscitation evaluated the results of these randomized trials and concluded that TH should be used to treat comatose patients resuscitated from OHCA caused by VF.22 In 2005, the AHA guidelines for postresuscitation support recommended that “unconscious adult patients with ROSC after out-of-hospital cardiac arrest should be cooled to 32°C to 34°C (89.6°F to 93.2°F) for 12 to 24 hours when the initial rhythm was VF (Class IIa). Similar therapy may be beneficial for patients with non-VF arrest out of hospital or for in-hospital arrest (Class IIb).”23 The central question accompanying these recommendations is the following: When TH is applied in a heterogeneous group of health care settings, will the benefit observed for the therapy in the randomized trials be maintained?

Multiple implementation studies have been published since the randomized controlled trials of TH were published in 2002.24–29 These implementation studies have had varying inclusion criteria including differences in age, the presence of a witness, the presenting rhythm, and duration of downtime. They have used different cooling techniques and maintained hypothermia for varying lengths of time. These diverse implementation studies were summarized in a meta-analysis published by Sagalyn et al in Critical Care Medicine in 2009.30 They examined all nonrandomized studies of adults resuscitated from cardiac arrest, with or without historic controls, published after the Bernard and HACA studies appeared in early 2002. Thirteen studies were included in the analysis with a total of 924 TH patients and 336 normothermic, historic controls. The meta-analysis concluded: “The survival and neurological outcomes benefit from therapeutic hypothermia are robust when compared over a wide range of studies of actual implementation.” The odds ratio for survival when treated with TH was 2.5 (95% confidence interval 1.8–3.3) and for favorable neurologic outcome was also 2.5 (95% confidence interval 1.9–3.4).30

Two large multi-institutional databases have also been published. The first, published by Arrich and The European Resuscitation Council Hypothermia After Cardiac Arrest Registry Study Group in 2007, was an outgrowth of the HACA trial. A total of 587 patients were included, 462 of whom were treated with TH and 123 with normothermia.31 Survival was 57% in patients treated with TH versus 32% in normothermic patients (P < .001); favorable neurologic outcomes were achieved in 45% of TH patients versus 32% of normothermic patients (P = .02). The second, from Nielsen et al, summarized 4 years of data from cases entered into the “Hypothermia Network,” a 34-center, 7-country registry of OHCA patients who remained comatose after resuscitation and were treated with TH.32 Nine hundred and eighty-six patients were included; the median time from collapse to ROSC was 20 minutes (IQR 14–30), the median time from collapse to initiation of TH was 90 minutes (IQR 60–165), and the median time from collapse to target temperature was 260 minutes (IQR 178–400). The presenting rhythm was VF/VT in 686 patients, of whom 412 (61%) survived to 6-month follow-up and 380 (56%) had good neurologic outcomes. Of the 217 patients who presented with asystole, 54 (25%) survived to 6 months and 46 (21%) had good outcomes. For the 66 patients presenting with PEA, 18 (27%) survived to 6 months and 15 (23%) had good outcomes.

These nonrandomized results provide strong, ancillary support to the findings of the randomized trials and suggest that TH should be used to treat the majority of comatose survivors of cardiac arrest regardless of initial rhythm or location of arrest. The totality of the findings from the randomized trials, implementation studies, and databases has informed the changes in the 2010 AHA recommendations about post–cardiac arrest care.1 As the authors acknowledged in the introduction: “There is increasing recognition that systematic post–cardiac arrest care after … ROSC can improve the likelihood of patient survival with good quality of life.”1 They recommend that TH be combined with other interventions to optimize outcomes in PCAS patients. These other interventions include optimization of cardiopulmonary function and vital organ perfusion, early percutaneous coronary intervention, goal-directed critical care, and neurologic support.

Hypothermia therapy can be divided into three distinct phases: induction, maintenance, and rewarming.33 Induction involves bringing the patient from the presenting temperature to target temperature. This can be accomplished by a number of methods: infusion of chilled intravenous fluids, application of ice bags, application of surface cooling equipment, insertion of intravascular cooling catheters, and other, novel equipment. In the Hypothermia Network Registry, the most common methods used to induce hypothermia were chilled fluids, used in 80% of the patients, and ice bags, used in 43%32 (see Figure 50-3). Induction should begin as close to ROSC as possible and occur at as fast a rate as possible. Shivering is the major complication that can occur during induction of TH and can be controlled with a number of different agents including meperidine, magnesium, buspirone, or paralytics. Benzodiazepines and neuromuscular blocking agents were used in all of the patients in both of the TH randomized controlled trials to control shivering and induce sedation. In addition, as the patient's core temperature drops, metabolism slows (decreasing approximately 8%/1°C), oxygen and glucose consumption fall, and ventilator settings need to be adjusted to compensate for decreased carbon dioxide production.33 Continuous core temperature measurement should be established during the induction phase. Place of an esophageal or urinary bladder probe should be considered, particularly if using a device that allows temperature autoregulation via a feedback loop.

During the maintenance phase, patients are kept at the target temperature for a specified period of time. The optimal length of time to maintain hypothermia is unknown. In the Bernard trial, patients were maintained for 12 hours; in the HACA trial, TH was maintained for 24 hours; in the Hypothermia Network Registry, 93% of the patients received TH for 24 hours.20,21 The period of injury after ischemia and reperfusion may last up to 7 days, and longer periods of TH may produce better outcomes.33 In the maintenance phase, several clinical problems need to be addressed: many patients develop postarrest myocardial stunning with accompanying decrease in ejection fraction; TH causes a cold-induced diuresis, and many patients require additional volume infusion to maintain adequate intravascular volume; electrolyte changes include hypokalemia, hypomagnesemia, and hyperglycemia.33

The goal of the rewarming phase is to safely return the patient to normothermia in a controlled fashion. This can be facilitated by using a cooling device with a feedback mechanism and an automatic rewarming program. During the rewarming phase, the patient begins to vasodilate, potentially causing decreased intravascular volume, requiring infusion of additional intravenous fluid. Potassium moves from intracellular compartments to the intravascular compartment and patients can become hyperkalemic; similarly, insulin resistance decreases and patients on continuous insulin infusions can become hypoglycemic. Ventilator settings need to be changed to account for the increased carbon dioxide production as metabolism increases.33

The role EPs play in the management of PCAS patients varies from institution to institution, depending on hospital resources, patient flow, and ED capabilities. Many OHCA patients have ROSC prior to ED arrival; others receive CPR in the ED, and a percentage of them attain ROSC while in the ED. EPs can maximize ROSC by delivering high-quality CPR including early defibrillation,34 quality chest compressions,35 discovery of reversible causes, and prioritization of interventions such as vascular access, intubation, and termination of resuscitation efforts.36 Rapid identification of patients who qualify for TH and other aspects of postarrest care is a central task of EPs.13,20,21 In addition to assessing for persistent coma after ROSC, this may involve, as dictated by the clinical scenario, obtaining basic labs to evaluate coagulation function and head CT to ensure no intracranial hemorrhage. In addition, in the vast majority of cases, induction of TH in patients who remain comatose after resuscitation from OHCA will fall under the auspices of EPs. This may be as simple as identification of qualifying patients, application of ice bags, infusion of chilled saline through peripheral IVs, and rapid transfer to an ICU for definitive management. On the other hand, it may include the first hours of comprehensive postarrest care including induction of TH, placement of arterial and central venous catheters, hemodynamic optimization, initiation of neurologic monitoring, ventilator management, electrolyte management, and bedside echocardiography. Comprehensive PCAS management programs need to be developed with champions from EMS, the ED, cardiology, critical care, neurology, and, potentially, rehabilitation medicine.3,29 It is imperative to delineate the responsibilities of each group of providers, given the potential rapid transition of these critically ill patients and number of interventions that need to be performed in the proximal phase.

The number of PCAS patients qualifying for TH and other aspects of postarrest care who present to different hospitals around the country varies from a few a year to several a month. Studies have demonstrated that outcomes from cardiac arrest vary depending on hospital type with lower survival rates at more rural, nonteaching, and smaller hospitals than at urban, teaching, and larger hospitals.37,38 Therefore, postarrest care may be optimized if regionalization occurs by diverting patients from low-volume hospitals to those treating a larger number of PCAS patients. In Arizona, the state EMS system has instituted a program of comprehensive cardiocerebral resuscitation, including TH delivered at cardiac arrest receiving hospitals. Using this approach, they have increased survival from 3.8% to 9.1%.36

When a cardiac arrest patient does not have ROSC, there are limited options to the physicians attempting to resuscitate the patient. After a period of unsuccessful resuscitation, the EP can halt further resuscitation efforts and pronounce the patient dead. The only alternatives available at this time are continued CPR by conventional means including chest compressions, drug administration, and defibrillation; or placing a patient on ECPB to provide circulation and ventilation for the patient until heart function recovers. During the time the patient is on ECPB, reversible causes of arrest can be addressed. Feasibility studies from Tokyo, Taipei, Seoul, and Los Angeles have demonstrated promising outcomes using ECPB; however, no randomized, prospective studies have been conducted and survivors may reflect selection bias, a Hawthorne effect, or an epiphenomenon of the timing of the intervention.

TH is now recognized as the standard of care for patients who remain comatose after resuscitation from cardiac arrest (see Figure 50-4). An organized program of TH needs to be integrated into a comprehensive plan for management of PCAS patients. EPs must develop these programs in concert with colleagues from EMS, neurology, cardiology, intensive care, and rehabilitation medicine. Clear delineation of who will be cooled, how they will be cooled, and the division of labor among the different care providers is vital to ensure a successful TH program.