In 2002, two seminal papers were published simultaneously in the New England Journal of Medicine, demonstrating that therapeutic hypothermia for comatose patients could result in improved survival and improved neurological outcome after out-of-hospital cardiac arrest, which altered the course of ICU history.
One after another, ICUs across the United States and around the world adopted protocols for rapidly achieving and maintaining body temperatures of 33°C. By 2010, the practice was considered mainstream and well-evidenced enough that the American Heart Association included therapeutic hypothermia in its revision of the Advanced Cardiac Life Support guidelines for post-arrest care.
Although the original studies enrolled subjects who had suffered out-of-hospital ventricular fibrillation (v-fib) or ventricular tachycardia (v-tach) arrests, many centers extended the technique to patients who had been resuscitated from pulseless electrical activity or, perhaps more reasonably, to patients who experienced cardiac arrest in the hospital. At times one would have to stay the hand of exuberant house staff who wanted to institute hypothermia in patients who were not only not comatose but who were well-oriented and conversant after a particularly successful resuscitation. It was a miraculous extravaganza of modern medicine and a tribute to our ever-advancing knowledge of human pathophysiology — and our ability to manipulate it to our patients’ advantage.
But wait. Hold the phone!
For a decade, the hypothermia train roared into the night, never stopping, except to pick up more passengers and more freight. Even PBS’ NOVA extolled the virtues of therapeutic hypothermia. That is until last month, when the New England Journal of Medicine published a study by Scandinavian investigators that included more than twice the number of patients of the two original papers combined and may have set the air brakes on that speeding train. In a prospective comparison of therapeutic temperature management, there was no difference in outcome in groups targeted at either 33°C or 36°C, with favorable outcomes of approximately 50% in both groups. The study is well-controlled, multicentered, and large enough to be nearly indisputable. So, what happened, what does it mean, and how should we proceed?
How do the outcomes of this study compare with historical outcomes for out-of-hospital arrest? The most recent numbers from the American Heart Association (AHA) suggest that overall survival from out-of-hospital arrest was 9.5% for all comers and 28.4% for v-fib arrest in 2010; however, that number does not provide information regarding how many patients made it to the hospital and might have been eligible for therapeutic temperature management. The CDC gives similar statistics for the period of 2005-2010; survival was 9.6% overall, but 26.3% of patients survived long enough to be hospitalized. A little math, then, tells us that the survival to discharge those who may have been eligible for therapeutic temperature management was 36.5%, lower than either arm of the newest temperature management study.
Neither the AHA nor the CDC provides data on satisfactory neurological outcome, which is undoubtedly even lower. Based on these numbers, it seems that therapeutic temperature management could be useful. But not so fast: a 2003 population-based study from Olmsted County, Minnesota, demonstrated that early defibrillation in out-of-hospital arrest was associated with 53.5% survival among those who were admitted to the hospital, though 38% of patients with v-fib arrest did not survive to hospitalization. Neurological outcome was generally good among long-term survivors.
Which bandwagon should you hop on?
When we take all of this together, it looks as if there are two possible ways to achieve good outcome for patients who suffer out-of-hospital cardiac arrest. The first, the “Olmsted way,” would be to achieve rapid and definitive treatment for the condition. The second would be to receive specialized care involving, or at least including, therapeutic temperature management. Either path appears to lead to better long-term outcomes than an undefined “standard of care.”
We are left with the question of why there is no difference in outcomes between targets of 33°C and 36°C. One theory expressed even before this trial is that the actual benefit may relate to avoidance of fever, rather than in setting a low core temperature. This seems plausible, especially since one trial of therapeutic, externally applied normothermia was associated with more rapid shock resolution in patients with septic shock and persistent fever. It also seems clear that the degree of hypothermia needed to achieve a therapeutic effect may have been overestimated; 36°C is still a touch below normal core temperature.
Based on the evidence we have, a reasonable practice would seem to be to use therapeutic temperature management targeted at 36°C. This should help us to avoid hypocoagulation associated with the lower temperature, as well as some issues associated with rewarming. A wise attending physician taught me in medical school that one should never be the first on the bandwagon, but also one should never be the last. In keeping with that philosophy, perhaps one could say that when the bandwagon loses a wheel, neither be the first to jump off, nor wait until you’re alone in the wreckage.
Steven Q. Simpson is vice chair, CHEST’s Council of NetWorks and former chair, Critical Care NetWork. He can be reached on Twitter @sqsimp.