Chapter 3: Targeted Temperature Management After Cardiac Arrest

The implementation of therapeutic hypothermia (TH) and targeted temperature management postcardiac arrest has arguably been one of the most significant contributions to resuscitation care. It should be viewed as the most important intervention in the postresuscitation period.

The modern application of TH had its origins in the 1950s. In 1956, Bigelow pioneered the use of hypothermia for neurologic protection during cardiac surgeries, a practice that has since become standard.¹ In 1964 Peter Safar's historic paper “the first ABC's of resuscitation” recommended the use of hypothermia after cardiac arrest in patients who do not regain consciousness after return of spontaneous circulation.² Physicians hypothesized that hypothermia's ability to suppress metabolic activity would translate into a tissue preservative capacity. Low target temperatures were used to accentuate the reduction in metabolism. However, a core body temperature below 30°C which we now refer to as “deep hypothermia,” exposes patients to the dangers of cardiac instability. Presumably whatever beneficial effects existed at these temperatures were outweighed by harm, and the therapy was abandoned. Then the 1980s and 1990s saw the emergence of animal research on the beneficial effects of “mild hypothermia.” Mild TH, that is a temperature between 32°C and 34°C, has profound effects on stemming reperfusion injury, without the deleterious consequences of cardiac instability.

The modern framework of TH postcardiac arrest stems from two landmark trials in the New England Journal of Medicine in 2002. The larger study known as the “Hypothermia after Cardiac Arrest (HACA) trial” was a multicenter randomized controlled trial (RCT) in Europe involving 273 postventricular fibrillation (VF)/ventricular tachycardia (VT) cardiac arrest patients who were randomized to either cooling to 32°C to 34°C for 24 hours, or “normothermia.” The normothermia control group did not have any temperature management. The neurologic outcome at 6 months was favorable in 55% of the hypothermia patients versus 39% of the control group.³ The second major trial enrolled 77 patients in Australia with out-of-hospital arrest from VF. At hospital discharge 49% of patients who were cooled went home or to rehabilitation, versus only 26% in the control group.⁴ The odds ratio for a favorable neurologic outcome with TH was 5.25 (95% confidence interval [CI], 1.47-18.76; P = 0.01), after adjustment for age and duration of the arrest.

It is important to note that two studies investigated patients with out-of-hospital electrical arrests who remained unconscious. The International Liaison Committee on Resuscitation (ILCOR) published a recommendation to use TH after electrical arrests in 2003. The most recent recommendations from the American Heart Association (AHA) in 2010 gave hypothermia after VF/VT a class I recommendation and a class IIb for pulseless electrical activity (PEA)/asystole.⁵

Since the two initial RCTs targeted a temperature range between 32°C and 34°C, the majority of centers began targeting a temperature of 33°C. In late 2013, the large multicenter trial known as the Targeted Temperature Management (TTM) trial randomized 950 postarrest patients to a temperature target of 33°C versus 36°C for 28 hours, and demonstrated equivalent mortality and neurologic outcome.⁶ It is emphasized that temperature was tightly controlled in both arms, as opposed to the earlier two trials in which the control group was allowed to become hyperthermic. The study included all rhythms, but a prespecified subgroup analysis of VF/VT also failed to show a difference between the two temperatures. Given how recent this important article was published, it remains to be seen how this will impact provision of TH. Most of the knowledge surrounding the practice of temperature management after cardiac arrest derives from studies and experience targeting a temperature of 33°C. One option many will take is to change to a target of 36°C in all patients. There could also be more variability in targeted temperatures going forward. While it is reasonable to continue to cool to 33°C in many patients with VF or VT as the presenting rhythm, for those with higher cooling risk (ie, more hemodynamically unstable, more bleeding concerns) and/or less clear clinical benefit (PEA, asystole), a temperature target of 36°C may be a better alternative (Table 3–1).

To clarify terminology, TTM will refer to the global practice of controlling temperature after cardiac arrest and the “TTM trial” will refer to this above study. TH will refer to a target temperature of 33°C, and minimal hypothermia (MH) will refer to a target temperature of 36°C.

The therapeutic value of TH is believed to lie mainly in its ability to decrease reperfusion injury. Cells subject to ischemia become injured and stunned, but do not necessarily undergo cell death immediately. A large proportion of the cytotoxicity from broad ischemic injury occurs in the period of time following tissue reperfusion. Ischemic cells have lost normal oxidative and metabolic function and an immediate burst of oxygenated free radicals and mitochondrial damage occurs on reperfusion. A second phase of injury involves inflammation, apoptosis and necrosis, which can play out over hours to days. Numerous complicated pathways of reperfusion injury include but are not limited to intracellular calcium and glutamate dysregulation, proinflammatory cytokines, and complement activation.¹ Hypothermia impacts this cascade at multiple steps, stemming calcium and glutamate imbalance, reducing oxidative damage, and ultimately decreasing apoptosis.⁷ The brain is particularly susceptible to ischemia-reperfusion injury, and particularly sensitive to temperature. However, in cardiac arrest, the ischemia is global, and damage is widespread. There is reason to believe that systemic cooling as opposed to local brain cooling accounts for some of the impact of TH in reducing mortality. It is not clear how low of a temperature is required to have clinical benefit, and the TTM trial suggests minimal hypothermia may be enough. Hyperthermia (fever) exacerbates the reperfusion injury, and may explain why its prevention alone has therapeutic potential. The TTM trial also raises the question of whether simply preventing hyperthermia has enough therapeutic power as to negate any beneficial effects of lower temperature.

In summary the positive effects of TTM postarrest are believed to be related to its effects in suppressing the whole body reperfusion injury, and preventing exacerbation of that injury by hyperthermia.

Neurologic Status After Arrest

TTM is not recommended for patients who have clear return to baseline neurologic status after their arrest. Criteria for enrollment in the HACA trial was the inability to follow verbal commands; the Australian study used the more vague description of “unconscious,” and the recent TTM trial specified a Glasgow Coma Scale (GCS) less than 8. It is not known whether there is any benefit to TTM in patients who have less severe neurological examinations postreturn of spontaneous circulation (ROSC). The 2010 AHA guidelines recommend TH for patients who have a lack of meaningful response to verbal commands,⁵ and it is reasonable to apply this criteria to TTM in general. As for patients who may have received sedatives thereby confounding their presenting examination, it may be prudent to wait a reasonable period to see if they regain the ability to follow commands before deciding to cool. It is suggested not to provide TTM in patients who are severely neurologically impaired at baseline, terminally ill, those for whom intensive care does not seem appropriate, or in whom they or their surrogates have indicated that they would not want intensive care.⁷

Location and Other Rhythms

In terms of location, the core studies were limited to out-of-hospital arrests. Accumulating evidence however supports TH for in-hospital arrests, and aside from the higher proportion of nonarrhythmic arrests there is no reason to believe the difference in physical location would have any detrimental impact on the therapy. Although TH is best established in VF and VT arrests, there has always been interest in its application to nonshockable rhythms. Irrespective of the dominant rhythm during cardiac arrest, ischemia and reperfusion injury occurs and may therefore be responsive to TH. The 2015 AHA guidelines gave a class I recommendation that all comatose patients post ROSC irrespective of rhythm receive TTM in the range between 32°C and 36°C.⁶ A large retrospective study in 2011 showed that TH was associated with improved outcomes in out-of-hospital nonshockable rhythms.⁸ However, a recent prospective review of a French registry from 2000 to 2009, did not find improved outcomes of TH in PEA/asystole patients.⁹ The authors raised the possibility that the anoxic/hypoxic insult in nonarrhythmic arrests is likely much greater and therefore may require deeper or longer periods of hypothermia to address. On the other hand, it is also possible that the risk-benefit ratio may simply be altered in this high-risk population and that PEA and asystole are better served by targeting a (potentially safer) temperature of 36°C. In conclusion, data supporting TH (33°C) in PEA/asystole are lacking and the benefit of TTM in general in this population is not clear. However, given the physiologic rationale for benefit of TH, it is arguable to target a potentially safer goal of 36°C.

Physiologic Effects of Cooling

The following sections on the physiologic effects of cooling refer mainly to a temperature of 33°C. The effects of minimal hypothermia (36°C) are less well characterized, but presumably milder.

Cardiovascular Effects

Hypothermia raises systemic vascular resistance (SVR) but its effects on blood pressure are variable as are its effects on cardiac output. The hypothermic heart develops diastolic dysfunction and there is an expected and physiologically adaptive bradycardia. It is not uncommon to see heart rates (HRs) in the low 40s (occasionally high 30s) with core temperatures of 33°C. Artificially elevating the HR could result in worsening contractility and cardiac output.¹⁰ Therefore, it is not recommended to treat bradycardia (with catecholamine infusions or pacing) unless evidence suggests it is directly responsible for poor perfusion. Mild TH has an antiarrhythmic effect and suppresses the development of ventricular arrhythmias. But overcooling should be avoided because a core temperature under 30°C frequently results in cardiac arrhythmias including spontaneous VF.¹¹

Although initial studies excluded hemodynamically unstable patients, increasing experience shows the safety in TH even in very unstable patients. A case series of patients in cardiogenic shock postcardiac arrest from acute coronary syndrome (ACS) on multiple vasopressors/inotropes including a large percentage requiring intra-aortic balloon pumps reported the safety of TH.¹² Another study of cardiac arrest patients from ACS suggests that TH may have a favorable impact on the hemodynamics of cardiogenic shock. Hypothermic patients had a higher SVR and HR. But more surprisingly, they also had improved cardiac output and index¹³ (as measured by a minimally invasive cardiac output monitoring device and echocardiography).

Neurologic

The brain is particularly sensitive to ischemic injury and temperature. For every decrease in 1°C, brain metabolism decreases by 6% to 7%.¹ There is abundant evidence that hyperthermia is associated with poor outcomes after many different brain injured states (stroke, traumatic brain injury). Postcardiac arrest cerebral edema may be more common than readily recognized and hypothermia is very effective in lowering intracranial pressure (ICP). Lastly, TH may provide an anti-antiepileptic effect.¹⁴

Hematologic

It has been long recognized that hypothermia potentiates coagulopathy during trauma and surgery. Accidental hypothermia is widely considered a marker of mortality in trauma victims, as well as a marker of morbidity during surgery. The oft-repeated bloody vicious triad of trauma “hypothermia, coagulopathy, and acidosis,” is thought critical to avoid. Studies of TH effects on coagulation demonstrate a prolongation of prothrombin time (PT) and activated partial thromboplastin time (aPTT), as well as reversible thrombocytopenia and platelet dysfunction.¹⁵ A mild prolongation of time to clot formation, but not clot propagation or clot firmness using more sophisticated clotting testing (ROTEM) has also been found.¹⁶ However, while unintentional hypothermia may be important to avoid during massive resuscitation of a bleeding patient, the clinical relevance of its effects on coagulopathy in controlled hypothermic conditions however is less clear. Importantly, major clinical studies have not shown a significant increase in bleeding. There is accumulated experience of TH safely being applied concurrent with heparinization for pulmonary embolism, as well as in extremely anticoagulated patients with ACS (who may receive aspirin, clopidogrel, or heparin). Routine intensive care unit (ICU) procedures such as arterial and venous catheterization should not be affected by TH. In summary, the coagulopathy of TH is well tolerated, and safe with concurrent anticoagulation, but should be considered contraindicated in patients who arrest secondary to bleeding, especially those with ongoing uncontrolled bleeding. TH at 36°C is expected to have trivial effects on coagulation.

Renal/Electrolytes

A major concern is the effect of TH on potassium. Induction of hypothermia drives intracellular flux of potassium and the reverse occurs upon rewarming. Hypokalemia on induction is usually not a huge problem, but caution and possibly intervention should be taken in the hyperkalemic patient before rewarming. Magnesium levels mimic potassium, falling on cooling and rising again on rewarming. Hypophosphatemia is also common but rarely a problem. Hypothermia also inhibits insulin sensitivity and release leading to hyperglycemia.¹⁵ TH may lead to an increase in urine output, dubbed the “cold diuresis.” Whether this is related to a more centrally distributed circulation or a direct tubular effect is not clear.

Immune Suppression

TH is thought to be immunosuppressive but the degree is not well characterized, and may do not have clinical relevance. Pneumonia is common after cardiac arrest in general, and it is unclear if TH increases that risk. It is possible that TH can indirectly lead to an increase in infections by requiring sedation, paralysis, and mechanical ventilation.

Acid Base

As temperature falls the solubility of gases in the blood decreases. Temperature uncorrected arterial blood gases (ABGs) therefore overestimate both Pco₂ and Po₂. The patient therefore may have a slight respiratory alkalosis and be slightly more hypoxemic. The significance of this effect is unknown, and it is controversial whether to aim for a normal pH (pH stat) or normal CO₂ (alpha stat) level during TH. It is likely that the impact is more important at deeper levels of hypothermia used during cardiopulmonary bypass. This author recommends addressing the pH/Pco₂ on the uncorrected ABG as you would any other critically ill patient, but aim for a slightly higher Po₂ than normal (ie, 70 mm Hg) to give buffer room.

In summary, TH routinely leads to well-tolerated bradycardia and electrolyte shifts. Other effects such as coagulopathy, immune suppression, and acid-base imbalances, are uncommon severe problems. TH effects on any of these conditions are expected to be minimal.

1. Measuring temperature

2. Device therapy

3. Controlling shivering

Measuring Temperature

Surface or tympanic temperature measurements are not accurate enough for implementation of TTM. Core temperature monitoring via bladder or esophagus is recommended. Esophageal temperature rapidly reflects changes to core temperature, whereas bladder temperature may lag by 15 to 30 minutes as it equilibrates. Concern is raised in using bladder temperature probes with low or absent urine output, and therefore in a dialysis patient for example, esophageal monitoring is recommended.

Device Therapy

Good outcomes are possible with simple surface methods (ice packs, blankets); however the practice is very labor-intensive, messy, and less precise. Simple surface methods frequently lead to both over- and undershooting in temperature.¹⁷ Device therapy is recommended, as it provides strict control of temperature during the maintenance phase, while also allowing for a very steadily controlled rewarming rate. It is not clear that TH is easier (less deep temperature) to achieve without device therapy, and therefore device therapy is still recommended.

Device therapy can easily be divided into surface cooling and invasive technology. Surface cooling devices typically circulate water through a cooling pad or blanket which relies on heat dissipation through the skin. Modern sophisticated devices work off feedback from a continuous core temperature monitor. Popular options include the Arctic Sun by Medivance and Medi-Therm by Gaymar. There is also a cooling pad by EMCOOLS that is studded with multiple square-shaped conductors containing a proprietary water graphite mixture called hydrocarbon.

Invasive devices require the insertion of a cooling catheter into a central vein (internal jugular femoral or subclavian). The device circulates cold fluid through the catheter cooling the blood by contact. Popular models include the CoolGuard and Thermogard systems made by Zoll, and the InnerCool systems by Philips.

Studies directly comparing simple surface cooling methods to intravascular cooling showed equivalent outcomes but better temperature control with device therapy.^18,19 But when compared against sophisticated surface devices there is no evidence that invasive cooling results in either faster cooling or decreased shivering. More important than the choice of device is being familiar with operating the device that you choose.

Extracorporeal cooling is very effective and convenient if the patient is already being placed on extracorporeal membrane oxygenation (ECMO) as part of their resuscitation, but this is not routine.

Controlling Shivering

If a patient is not achieving goal temperature target, the first consideration should be uncontrolled shivering. Shivering is not only counterproductive to cooling efforts but also detrimental to the critically ill patient who is unable to tolerate the high metabolic energy expenditure required to generate heat. This author has seen shivering lead to a drop in saturation of the central venous oxygenation (S_CVO₂) to the 50s only to see it normalize after the shivering was addressed with paralysis. Shivering can occur in very obvious gross myoclonic form, or may be so subtle as to not be visible to the eye. Sometimes shivering can only be recognized by looking for baseline vibration artifact on the cardiac monitor tracing, or with electroencephalogram (EEG) monitoring (if present). Nonpharmacologic methods of shivering control such as skin counter warming have been explored, but virtually every patient undergoing TH will need pharmacologic control of shivering. Many medications have been successfully used to control shivering including opiates, benzodiazepines, propofol, and dexmedetomidine. Although there are subtle differences in response to one medication over another, there is no proof of a superior drug. Neuromuscular blockade (NMB) will definitively stop shivering and was used routinely in both of the 2002 trials and widely in the 2013 TTM trial. Although many have argued against using NMB during TTM, there is no proof of a deleterious effect, and this author believes that the downside of uncontrolled shivering far outweighs the potential of critical illness neuropathy. Many have reported an ability to stop paralytics or decrease sedatives once at 33°C, presumably due to suppression of the shivering reflex at a lower core temperature. Notably, in the TTM trial there was no difference in shivering between the 33°C and 36°C groups. It remains to be seen whether real-world practice will see less shivering at 36°C (because it is closer to normothermia) or more (because there will no longer be suppression of the shivering reflex). As always, the use of NMB should be accompanied by continuous sedation. For TH, this author favors starting continuous sedation and NMB on induction and continuing up through the rewarming phase.

Induction

Induction Techniques

Hypothermia devices can certainly be used for the induction phase of cooling. If cooling to 33°C, the only potential disadvantage of device therapy is a slow cooling rate (1°C-1.5°C per hour on average). A faster induction can be achieved with chilled IV saline. Kim et al showed that infusion of intravenous saline by emergency medical service (EMS) shortly after ROSC was a safe and effective way to lower core body temperature by on average 1.5°C.²⁰ There was no evidence that cold saline led to recurrent arrests or significant pulmonary edema. Another small series of the use of intra-arrest cold saline showed that this could safely be done with the effect of lowering temperature by 2°C,²¹ with a median time to reach 33°C of only 16 minutes. There is no evidence that mild TH inhibits ROSC and in fact one study showed that TH improves defibrillation success.²² Years of experience since the initial studies also supports the safety of IV chilled saline.

Chilled saline can either be kept ready in a refrigerator or can be made by placing a saline bag in ice water for 15 minutes. Induction can also be enhanced by placing ice packs over superficial vascular points including the neck, axilla, and groin.

Safety and Efficacy of Early Cooling

How early to start cooling, how fast to bring down the temperature, and how long to cool are areas of ongoing academic study. There is a multitude of animal data suggesting that earlier cooling is beneficial. One mouse study showed that intra-arrest cooling had better neurologic effects when compared with cooling initiated postarrest.²³ Certainly there is physiologic rationale to slow and prevent the reperfusion injury earlier on in the process, and like most interventions in resuscitation, the earlier the better.

In one recent trial, out-of-hospital arrest patients were randomized to prehospital cooling with IV saline versus cooling initiated in the emergency department (ED). There was no difference found in outcome, but due to the fast EMS response time and aggressive cooling in the ED, within 1 hour of arrival to the ED both groups were at the same temperature. The study, however, bolsters the safety of early IV cooling, and draws attention to the need for more trials addressing the topic.²⁴

The normal physiologic response to maintain thermostasis appears to be disturbed in sicker patients. Patients who are passively cooler post-ROSC, and who are less resistant to the induction of cooling tend to do worse overall. This phenomenon has complicated efforts to study time to goal temperature and outcome. Studies that have emerged seemingly linking earlier cooling to worse outcomes may reflect a selection bias toward sicker patients.²⁵

Most patients will have a significant drop in body temperature upon ROSC presumably due to heat loss, low metabolism, and low-flow state during the arrest. Rather than allowing the body to rewarm, the clinician can take advantage of the lower temperature and build upon it by either rapidly inducing TH, or maintaining the temperature at 36°C. The Fire Department of New York began intra-arrest cold saline administration in 2010.

How Late After Cardiac Arrest is too Late to Start Treating?

In the HACA trial it is important to note that target temperature was not reached for an average of 6 to 8 hours postarrest, a very long induction time compared to modern practice. There is animal research suggesting loss of benefit of hypothermia when started beyond 4 to 6 hours of ROSC.²⁵ However reperfusion injury postarrest occurs over hours to days, so if there is a delay in therapy for whatever reason, an impact may still be possible. Although it is reasonable to consider not treating if the patient is 8 hours post-ROSC, it is also important to emphasize that the upper limit of delay to implementation is not known.

Maintenance

Once at goal temperature, patients tend to stabilize. Temperature is continually monitored and controlled. Similarly, shivering is monitored and addressed. Aggressive critical care including ventilator management, replacing nonsterile lines, fluid resuscitation, and vasopressor agent titration is crucial. In the HACA trial patients were maintained at goal for 24 hours after arrival to the ED, whereas in the Australian trial induction of TH was earlier but maintenance was only for 12 hours. In the TTM trial patients were rewarmed 28 hours after randomization (which roughly corresponded to 24 hours at target temperature). The luxury of having RCTs with different intervals is that it provides a basis for modifying the treatment interval. There has been experience of using longer cooling intervals (up to 72 hours) in conjunction with extracorporeal life support in Japan.²⁷ A case report also details a good outcome in a patient cooled to 33°C for 48 hours postarrest.²⁸ It is possible that like many critical care therapies, the maintenance interval of TTM should be adjusted on a sliding scale tailored to the particular injury of the patient. However, until more data are available, the consensus opinion is to maintain 24 hours at target temperature.

Rewarming

Although there is no consensus on the optimal rate of rewarming it is recognized that uncontrolled or rapid rewarming can lead to vasodilation, hemodynamic instability, and dangerous electrolyte shifts. There is also limited evidence that rapid rewarming may trigger deleterious cellular injury, obviating the potential benefits of the initial cooling.¹⁵ Herein lies a major advantage in using a TH device, as rewarming can be extremely difficult to control otherwise. Studies have shown that you can safely rewarm between 0.1°C and 0.5°C per hour although the latter may be too rapid for less stable patients. Seizures which are common in the postarrest period tend to occur during rewarming. This author will frequently pause or decrease the rate of rewarming if the patient shows hemodynamic instability or develops seizures until these problems can be addressed. It is likely that the principles of practice listed above are more important when rewarming from a target of 33°C as opposed to 36°C.

Normothermia After Rewarming

It is a common practice to maintain 24 to 48 hours of normothermia (37°C) after rewarming. Rebound hyperthermia is common and the injured brain may be susceptible to further damage. However, despite the logic, there is currently little clinical evidence to support its benefit.²⁷ The only potential harm of normothermia maintenance is that some patients require ongoing sedation to prevent shivering which could delay extubation and neurologic prognostication.

Neurologic prognostication is arguably the most complicated and least well-understood aspect of care revolving around the application of TTM. The medical world is still in its infancies in understanding how temperature management affects healing of the injured brain. Much of what is known about anoxic encephalopathy originated from studies published prior to the era of TTM. It is crucial to avoid an errant negative prognosis that might lead to withdrawal of care in a patient who could go on to have a good outcome. On the other hand, many individuals would not opt for ongoing care if they were in a persistent vegetative state. Additionally, ongoing care for neurologically devastated patients leads to undue burden on the medical system. Tests should ideally have a zero false-positive rate for determining a poor prognosis, and most of the research has thus been focused on prediction of poor outcomes. Unfortunately, many studies of neurologic tests after TH suffer from the “self-fulfilling prophecy bias.” It is hard to draw conclusions regarding the predictive value of tests that were in turn used as criteria to withdraw care (Table 3–2).

Neurologic Examination: The basic neurologic examination arguably remains the centerpiece for prognostication. Rosetti et al showed that motor response (ability to withdraw from pain and follow motor commands) is often delayed by TH. Absence of motor response on day 3 had a high false-positive rate (24%), and should not be used alone as criteria to withdraw care.³⁰ However, several studies support that the absence of either pupillary or corneal reflexes at 72 hours postcardiac arrest consistently portends a very poor neurologic prognosis irrespective of the use of TH.^31,32 A recent meta-analysis suggested that absence of pupillary reflex was a better negative predictor than absence of corneal reflex, and reinforced the lack of reliability of the motor examination.³³

Advanced Testing and Monitoring

EEG

Although there is no RCT showing mortality or neurologic benefit to routine (continuous or spot) EEG monitoring during TTM, EEG helps in the identification and treatment of seizures and also in neurologic prognostication. Seizures are common postcardiac arrest, occurring in 30% to 50% of patients, and may contribute to further brain damage and prolong coma.¹⁵ They are often difficult to detect, due to the high percentage of nonconvulsive status epilepticus as well as the use of sedatives and NMB during TH. However, the abnormal EEG patterns seen after severe brain injury may be difficult to interpret and can be very refractory to treatment. Some believe that treatment refractory nonconvulsive status on rewarming may in fact be “electrical chaos” from a severely injured dying brain and not seizures in the conventional sense. In terms of prognosis, a flat EEG, nonpharmacologic burst suppression, and electrographic status, are all patterns associated with a poor outcome. MSE, bilaterally synchronous twitches of limb, trunk or facial muscles jerks associated with EEG correlate has a well-established association with poor outcome in many forms of anoxic brain injury including cardiac arrest.³⁴ Although this association still largely holds, some reports have suggested that MSE may not be as dismal as a marker in a patient treated with TH.³⁵ The presence of EEG reactivity (an electrographic response to visual, auditory, or painful stimuli) during TH, portends a good outcome while the lack of reactivity portends a bad outcome.

Somatosensory-evoked Potentials

SSEPs are particularly helpful as negative prognostic markers. The median nerve is stimulated electrically at the wrist and a cranial electrode records a response wave called the N20 which reflects stimulation of the somatosensory cortex. The bilateral absence of N20 response on SSEP is possibly the most accurate predictor of a poor neurologic prognosis, with many early studies showing nearly a 0% false-positive response.³⁶ The 2006 American Academy of Neurology guidelines recommended withdrawing care if there is an absent bilateral N20 response.³⁷ The accuracy of SSEP was again supported by a meta-analysis of over 1000 patients.³² However, there does exist a very small false-positive rate, and there is still a need for caution in interpreting SSEPs in isolation.³⁸

Brain Imaging

There is not enough information to support neurologic prognostication based on head computed tomography (CT) or magnetic resonance imaging (MRI) alone. There are situations in which the imaging is dramatically abnormal and shows for example diffuse cerebral edema, loss of gray-white differentiation, or herniation. However in these cases usually the examination also reflects the extent of injury, so imaging ends up playing a supportive role.

Biomarkers

Neuron-specific enolase is a biomarker of neuronal injury, and there is accumulating evidence of its predictive value during TH. Levels above 33 generally correlate with poor outcome but there have been exceptions.^39,40

When to Prognosticate

Caution needs to be exercised against holding to rigid time frames postcardiac arrest. There is wide variability in time to awakening post-TH. In one study the median time to awakening post-TH was 3.2 days but the variability was wide and patients often awaken after 3 days. In this study, 15% of the patients woke more than 72 hours after their arrest.⁴¹ Sedative choices, accumulative doses, and renal function all potentially can impact time to awakening.

The TTM trial used a combination of different criteria for prognostication, and this is a well-advised approach. The following were used as criteria to recommend withdrawal of care: MSE plus negative SSEP at 24 hours, GCS motor score 1 to 2 at 72 hours plus negative SSEP, or GCS motor score 1 to 2 at 72 hours plus treatment refractory status epilepticus.⁶ An important general premise is that it is easier to predict who will do very well (eg, the patient who is waking and following commands soon after TH), and easy to predict who will do very poorly (those with lack of brainstem reflexes at 3 days, absent SSEPs). Nevertheless, it is hard to predict long-term outcomes for those who fall in between. It is strongly recommended to take precautions that sedatives are not negatively influencing the neurologic examination post-TH. The lack of pupillary reflexes 3 days postcardiac arrest, the presence of MSE, and the absence of SSEPs all very likely portend an extremely poor prognosis. However, caution should be exercised in using any one test in isolation. Information from imaging, EEG, and neuron-specific enolase levels may be helpful adjuncts, especially in those patients who fall into the gray zone.