Special Feature

Use of Therapeutic Hypothermia After Cardiac Arrest

By Richard J. Wall, MD, MPH, Pulmonary Critical Care & Sleep Disorders Medicine, Southlake Clinic, Valley Medical Center, Renton, WA, is Associate Editor for Critical Care Alert.

Dr. Wall reports no financial relationship to this field of study.

When patients sustain a sudden cardiac arrest, they require immediate life-saving therapies to restart cardiac function and prevent secondary anoxic brain injury. Unfortunately, resuscitative therapies in these patients are often delayed. Furthermore, even if successfully resuscitated, only 10%-30% of such patients are able to leave the hospital alive and resume an independent life. One of the key reasons many patients do not recover from such an ordeal is that they sustain severe anoxic brain injury as a result of their arrest.

The care delivered immediately after return of spontaneous circulation (ROSC) is crucial for survival and should be considered as a continuum of advanced life support. Indeed, most deaths among patients initially resuscitated after cardiac arrest occur within the first 24 hours. Thus, while the "best" care for patients after ROSC is not yet known, there is increasing interest in optimizing practices that have an increased chance of improving outcomes. In this review, I will discuss one of these interventions for adults, therapeutic hypothermia (TH).


TH is a technique for preserving cerebral function in patients who are resuscitated after cardiac arrest. After patients have been stabilized from a cardiovascular standpoint, their body temperature is lowered to 32-34° C for 12-24 hours. TH benefits patients in multiple ways. During reperfusion after arrest, there is a post-resuscitation syndrome wherein free radicals, neurotransmitters, and other mediators further damage the brain. TH mitigates these effects. With hypothermia, there is also better preservation of the blood-brain barrier, cell death is lessened, and brain energy stores of adenosine triphosphate are better preserved.


TH has been studied since the 1950s, but the first randomized controlled trial (RCT) of its use was not published until 2001.1 This was a small feasibility study of 30 patients who sustained an out-of-hospital arrest due to asystole or pulseless electrical activity (PEA). A helmet device was used to induce mild hypothermia. No complications were noted. Since it was a feasibility study, no neurologic or clinical outcomes were reported. Shortly thereafter, however, two landmark RCTs were simultaneously published showing that induced hypothermia has a neuroprotective and mortality effect in patients resuscitated after cardiac arrest due to ventricular fibrillation (VF).2,3

In the first study, Bernard and colleagues randomized 77 adult patients to either hypothermia at 33° C for 12 hours within 2 hours after ROSC (n = 43) or normothermia (n = 34).2 The initial rhythm was VF. All patients had persistent coma after ROSC. Patients assigned to hypothermia underwent basic cooling in the ambulance, followed by vigorous cooling using external packs as soon as they arrived at the hospital. Patients were sedated, mechanically ventilated, and paralyzed if necessary (to prevent shivering). At 18 hours, they were actively rewarmed using an external heated-air blanket. Overall, 49% of the hypothermia patients survived, compared with 26% in the normothermia arm (P = 0.046). Among those discharged from the hospital, the adjusted odds ratio for a good neurologic outcome in the hypothermia arm was 5.25 (P = 0.011). There was no significant difference in adverse events between the groups.

The Hypothermia After Cardiac Arrest group conducted another multicenter RCT with a similar design.3 After screening 3551 patients with persistent coma after ROSC from VF arrest, 275 were randomized. Hypothermia patients were cooled for 24 hours using a specialized cooling mattress and, if necessary, ice packs were also added. Patients were sedated, mechanically ventilated, and paralyzed as needed (to prevent shivering). They were passively rewarmed over an 8-hour period. At 6 months follow up, 59% of the hypothermia patients were alive, compared with 45% in the normothermia arm (P = 0.02). The hypothermia group were 40% more likely to have a cerebral performance category (CPC) score of 1 or 2 (5-point scale; 1 = good cerebral performance, 5 = brain death), as compared with the normothermia group (P = 0.009). A caveat is that this study only enrolled 8% of the patients initially assessed, thus raising questions about its generalizability.

Not surprisingly, a subsequent meta-analysis of these three RCTs showed a benefit for hypothermia in terms of neurologic outcome and survival after cardiac arrest.4 Likewise, the Cochrane investigators came to a similar conclusion when they examined the topic in 2009.5 The Cochrane group performed a systematic review of five RCTs (4 trials and 1 abstract). They deemed study quality as "good" in the three studies described above.1-3 Overall, they found that patients in the hypothermia group were 35% more likely to survive to hospital discharge compared to standard care. Patients in the hypothermia group were 55% more likely to achieve a favorable CPC score of 1 or 2 by the end of their hospital stay.

A recent study in 59 intensive care units (ICUs) across the Netherlands is the largest observational study ever published on TH.6 Using a nationwide registry, the authors examined mortality rates over a 10-year period. They included all patients admitted to a Dutch ICU after any type of cardiac arrest. A key feature of this study is that more than 90% of Dutch ICUs currently use TH. Among the 5317 patients analyzed, 1547 were treated before and 3770 were treated after implementation of TH. Overall, the adjusted odds ratio of hospital mortality for patients treated with TH was 0.80 (95% confidence interval [CI] 0.65-0.98; P = 0.029). Moreover, the various ICUs used different cooling techniques and protocols, suggesting that TH can be effectively implemented on a widespread basis.


Patients surviving VF arrest often require additional therapies for treatment of their coronary disease. Case series have reported the feasibility of using TH in combination with emergent percutaneous coronary intervention. Others have described the safe use of TH in patients with cardiogenic shock after ROSC. Although there are descriptions of using fibrinolytic therapy for acute myocardial infarction after ROSC, safety data are lacking in this situation. Given the inherent risk of hypercoaguability and bleeding in TH (see below), fibrinolytics should be given with extreme caution.


No prospective RCTs have compared outcomes for patients with an initial non-VF rhythm (i.e., asystole or PEA). However, several observational studies have shown a benefit with use of TH in comatose survivors after both in-hospital and out-of-hospital cardiac arrest associated with any arrest rhythm. Curiously, a few of these studies suggested the benefit of TH is better for out-of-hospital arrest patients, even though code response times are presumably faster in the hospital. The reason for this observation is unclear, but it is probably due to the comorbidities of such patients.

Given the low survival rate in non-VF patients, it is unlikely that a clinical trial will ever be undertaken to test the efficacy of TH in this setting. Such a study would require an enormous sample size. In other words, clinicians need to rely on observational data and their own judgment in non-VF situations. If a patient appears to have a reasonable chance of survival after ROSC from a non-VF arrest, then TH should probably be undertaken.

Acknowledging these areas of uncertainty, the 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation strongly recommend that "all comatose adult patients with ROSC after out-of-hospital VF arrest be cooled to 32°-34° C for 12-24 hours."7 Induced hypothermia may also be "considered" for comatose patients with ROSC after in-hospital arrest of any initial rhythm, or out-of-hospital arrest with an initial rhythm of PEA or asystole. In addition, the guidelines state that active rewarming should be avoided in comatose patients who spontaneously develop mild hypothermia (> 32° C) during the initial 48 hours after ROSC.


The optimal technique for delivering TH is unclear. In general, there are two methods — surface cooling and endovascular cooling. However, there are many nuances to these approaches: cooling pads, cooling blankets, cooling mattresses, cooling helmets, ice packs, water immersion, intravascular catheters, gastric lavage, or cold intravenous fluids. In addition, cooling can be combined with hemofiltration or extracorporeal cardiopulmonary support. RCTs have described the safe use of cooling with IV cold saline in the prehospital setting. Regardless of the cooling method chosen, clinicians must continuously monitor the patient's core temperature using an esophageal thermometer, bladder catheter in nonanuric patients, or pulmonary artery catheter (if available). Oral, axillary, and tympanic temperatures are unreliable.

The International Liaison Committee on Resuscitation identified "optimal cooling technique — internal versus external" as a knowledge gap.8 In its 2008 consensus statement, the committee recommended that temperature control is best achieved with devices incorporating continuous temperature feedback. The committee highlighted several potential drawbacks of surface cooling (e.g., ice packs) including increased labor, greater temperature fluctuations, and inability to control rewarming. However, despite this recommendation, readers should recall that both of the landmark RCTs used external ice packs for cooling.2,3

Only a few small studies have compared cooling methods. Flint retrospectively evaluated endovascular cooling when used as an adjunct to a surface method in 42 patients.9 The combined endovascular and surface cooling method provided better temperature control with less overcooling and a lower incidence of bradycardia compared to surface cooling alone. Another study compared five cooling methods in 50 patients and also found endovascular techniques more efficacious.10 The largest study of endovascular vs surface cooling after cardiac arrest (n = 83) showed that endovascular cooling achieved more time in the target temperature range, less fluctuations, better control during rewarming, and fewer complications.11 No studies showed differences in ventilator days, mortality, or neurologic outcomes.

A recent observational study described the safety and low cost of a combined core-surface cooling approach for achieving mild hypothermia in 65 patients with cardiac arrest.12 Patients were cooled using a combination of rapid, cold saline infusion, evaporative surface cooling, and ice water gastric lavage. The key emphasis was on prompt hypothermia induction. Overall, the median time from induction to target temperature was 60 minutes. The authors achieved a median cooling rate of 2.6° C/hour, and 31% of patients recovered to a CPC score of 1-2.


Various adverse events have been reported with TH. These include coagulopathy, need for transfusions, pneumonia, sepsis, pancreatitis, renal failure, hemodialysis, pulmonary edema, seizures, arrhythmias, hyperglycemia, hypocalcemia, hypokalemia, and hypophosphatemia.13 Although the Cochrane analysis did not find a significant difference in adverse events between groups,5 it is well known that infections are common in post-arrest patients, and prolonged hypothermia decreases immune function. In addition, hypothermia impairs coagulation and any active bleeding must be controlled before undertaking TH. In general, clinicians need to closely monitor patients for potential complications.


No RCTs of TH have looked at long-term survival, dependency, quality of life, or cost-effectiveness. In addition, numerous questions remain unanswered about TH. For example, is there a "golden window" of time during which the patient must be cooled to capture the beneficial effects? Does it matter if cooling starts in the pre-hospital setting? Does cooling work equally well in every subgroup (e.g., patients with non-VF as primary rhythm, or patients with in-hospital arrest)? What is the optimal cooling protocol? What is the optimal rewarming protocol? What is the cost-benefit of TH?

In summary, there is strong evidence for the use of TH in patients who remain comatose after ROSC from VF arrest. For patients with out-of-hospital VF arrest, it is now standard of care. There is moderately robust observational evidence for the use of TH in patients who remain comatose after ROSC from non-VF arrest, albeit survival rates in such patients are usually lower and the implementation of TH should be considered in the context of the patient's overall chances for recovery. Numerous devices and protocols exist for cooling patients, but none has proven superior. Each ICU must carefully assess their available resources and expertise when choosing devices and implementing a TH program.


  1. Hachimi-Idrissi S, et al. Resuscitation 2001;51:275-281.
  2. Bernard SA, et al. N Engl J Med 2002;346:557-563.
  3. Hypothermia After Cardiac Arrest Study Group. N Engl J Med 2002;346:549-556.
  4. Holzer M, et al. Crit Care Med 2005;33:414-418.
  5. Arrich J, et al. Cochrane Database Syst Rev 2009:CD004128.
  6. van der Wal G, et al. Crit Care Med 2011;39:84-8.
  7. Peberdy MA, et al. Circulation 2010;122(18 Suppl 3):S768-786.
  8. Nolan JP, et al. Resuscitation 2008;79:350-379.
  9. Flint AC, et al. Neurocrit Care 2007;7:109-118.
  10. Hoedemaekers CW, et al. Crit Care 2007;11:R91.
  11. Gillies MA, et al. Resuscitation 2010;81:1117-1122.
  12. Kory P, et al. Resuscitation 2011;82:15-20.
  13. Nielsen N, et al. Acta Anaesthesiol Scand 2009;53:926-934.