The Cold Truth: Rethinking Temperature Management after Cardiac Arrest

Published - Written by Rena Xu

Over a decade ago, two pivotal trials altered the standard of care for unconscious patients after cardiac arrest.  Many of these patients suffer severe neurologic injury and death; researchers sought a treatment that could reduce global brain ischemia and prevent neurologic damage even when initiated after cardiopulmonary resuscitation and the restoration of circulation.

The trial results, published in 2002, suggested that induced hypothermia, also known as “targeted temperature management,” could serve as such a treatment.  Patients treated with hypothermia (temperatures of 32°C to 34°C) were found to have improved survival and neurological function as compared to patients receiving routine care.  Since then, international resuscitation guidelines have incorporated therapeutic hypothermia into their recommendations, and the practice has become widely adopted.

More than a decade later, however, criticism of this guideline persists.  In the original trials, many patients in the routine treatment arm developed fever, which is known to be associated with poor outcomes.  Did the patients in the intervention arm do better because they were treated with hypothermia, or simply because they were kept from developing hyperthermia?

In this week’s NEJM, Nielsen et al report findings from a large randomized controlled trial that addresses this question.  The trial, conducted in intensive care units across Europe and Australia, randomly assigned 950 unconscious adults who had suffered out-of-hospital cardiac arrest to targeted temperature management at either 33°C (“induced hypothermia”) or 36°C (“near-normal temperature”) for 36 hours.  In contrast to the previous studies, which only included patients with ventricular fibrillation, this study enrolled patients with both shockable and nonshockable rhythms.  Importantly, fever prevention measures were used to keep all patients below 37.5°C until 72 hours after cardiac arrest.  The primary outcome was all-cause mortality, and the main secondary outcome was a composite of neurologic function and death at 180 days.

Unlike the previous studies, this trial established a clear protocol for when life-sustaining therapy should be withdrawn.  A physician blinded to the intervention assignments performed neurological evaluations 72 hours after the intervention period ended and recommended either continuing or withdrawing life-sustaining therapy, according to pre-specified criteria.

During the first seven days, life-sustaining therapy was withdrawn in 132 patients in the 33°C group and 115 in the 36°C group, for reasons including brain death and multi-organ failure.  By the end of the trial, overall mortality did not differ significantly between the two study arms: 50% of the patients in the 33°C group had died, as compared to 48% of those in the 36°C group (HR 1.06, p=0.51).  There was also no difference between the two groups in neurological function (measured either by the Cerebral Performance Category: HR 1.02, p=0.78; or by the modified Rankin scale: HR 1.01, p=0.87) or death at 180 days (HR 1.01, p=0.92).

Is therapeutic hypothermia actually therapeutic, then?  In this study, neither benefit nor harm was evident at 33°C as compared to 36°C.  Several factors may have contributed to the departure of results from those of older studies. This trial had fewer exclusion criteria, resulting in a broader, more heterogeneous patient population; 20% of participants, for example, had nonshockable rhythms.  In addition, improvements in pre-hospital and critical care management over the last decade may have made it harder to isolate the incremental benefits of a single intervention.  Finally, as some have argued all along, it is possible that the real value of targeted temperature management lies not in the induction of hypothermia but rather in the prevention of hyperthermia.

In an accompanying editorial, Drs. Jon Rittenberger and Clifton Callaway of the University of Pittsburgh write, “The exceptional rates of good outcomes in both the 33°C and 36°C groups in the present trial may reflect the active prevention of hyperthermia. Whatever the mechanisms, it seems clear that we should not regress to a pre-2002 style of care that does not manage temperature at all.”  They conclude, “Perhaps the most important message to take from this trial is that modern, aggressive care that includes attention to temperature works, making survival more likely than death when a patient is hospitalized after CPR.”

NEJM Deputy Editor Dr. John Jarcho states, “This trial illustrates the importance of carefully testing and evaluating each individual component of patient management in critical care, to be certain that our practice is not only evidence-based, but based on evidence that is derived from adequate testing of appropriate options.”

Further investigation is needed to clarify whether, and to what extent, induced hypothermia can be therapeutic.  If benefits can be demonstrated in future studies, additional research will also be needed to identify an optimal target temperature.  It is encouraging that outcomes for patients who suffer cardiac arrest have improved considerably over the last decade.  It is equally encouraging that even widely accepted treatment guidelines can be challenged by new evidence so that outcomes may continue to improve in the decades to come.

What is your current approach to temperature management for unconscious patients brought into the hospital after cardiac arrest? Is there a specific temperature you target?  How will the findings of this study affect your approach?