“Therapeutic hypothermia” for post-cardiac arrest is officially over
New AHA guidelines end the targeted temperature management era
New guidelines from the American Heart Association have effectively ended the two decade era of cooling comatose patients post-cardiac arrest. The new advisement: prevent fevers by keeping core temperature ≤37.5° C; cooling below that is optional. Read on for details. PulmCCM is not affiliated with AHA or ILCOR.
Therapeutic hypothermia, later rebranded as targeted temperature management, became a standard post-cardiac arrest therapy for comatose patients after two 2002 NEJM trials (n=273 and n=77) suggested reducing core temperature to 32°C to 34°C markedly improved neurologic outcomes and survival. Although those studies only included out of hospital cardiac arrests due to ventricular fibrillation, induced hypothermia was rapidly and widely adopted and applied to most cardiac arrest patients, regardless of initial rhythm or location of arrest. This was in large part due to the recommendations from a task force of the International Liaison Committee on Resuscitation (ILCOR), and their enshrinement in the American Heart Association’s 2005, 2010 and 2015 post-cardiac arrest management guidelines.
Failure to Replicate
Induced hypothermia’s original purported benefits were not convincingly replicated. Most large subsequent trials showed no beneficial effects of TTM:
In 2013, a trial more than three times larger than the original trials (NEJM, n=950) showed no benefit of cooling to 33°C compared to maintenance at 36°C, in either survival or neurologic outcome. The trial was later dubbed “TTM1”.
A 2017 trial (JAMA, n=355) did not find benefit in neurologic outcome with prolonged cooling to 33°C for 48 hours, but may have been underpowered (the intervention group had a nonsignificant absolute 5% improvement).
In 2019, the HYPERION trial (NEJM, n=584) found neurologic benefits of cooling to 33°C during the first 24 hours (compared to maintenance at 37°C). There were numerous concerns regarding the validity of these findings, which we described here after its publication.
Also in 2019, the PRINCESS trial showed no improvement in neurologic outcome with nasal-evaporative cooling after cardiac arrest among 677 patients.
The largest-yet TTM2 trial (NEJM 2021, n=1,900) tested temperature reduction to 33°C in all out of hospital cardiac arrests, and found no benefit compared to controls who received active fever prevention (application of automated cooling if core temperature reached 37.8°C despite acetaminophen, etc).
The Ice Cracks
In 2015, in response to the mounting data casting doubt on any benefits of deep cooling, ILCOR and AHA modified their guidance to recommend targeted temperature management between 32°C and 36°C (increasing the upper temperature range considered to be therapeutic).
This dramatically reduced the complexity of caring for post-cardiac arrest patients, as patients kept at 36°C (96.8° F) don’t require the same degree of burdensome antishivering regimens, deep sedation, etc. required at lower temperatures.
Noting the TTM1 and TTM2 trial results, authors for the American College of Cardiology went further than their AHA colleagues, arguing in 2021 that cooling patients below normothermia may no longer be necessary in any out-of-hospital cardiac arrest patients.
In 2022, ILCOR made things even simpler, weakly recommending active fever prevention alone (actively maintaining temperature ≤37.5° C) as appropriate for post-cardiac arrest care (with low certainty of evidence).
AHA Effectively Ends the “Induced Hypothermia” Era
In December 2023, AHA issued a “focused update” to its standard-of-care-defining guidelines, endorsing fever prevention ≤37.5° C alone in post-cardiac arrest patients not following commands. (Temperatures as low as 32° C were also endorsed.)
They did not recommend cooler temperatures for any patient subgroup (e.g., deeply comatose, shockable rhythm, etc.), due to insufficient evidence.
How Long to Maintain Fever Prevention?
AHA only advised that it’s “reasonable” to maintain fever prevention “for at least” 24 hours after post-cardiac arrest care is begun. It’s also “reasonable” to maintain temperature control for longer periods in comatose patients, they said.
Active temperature management was suggested (such as with external automated cooling), as opposed to waiting and reacting to fevers.
The evidence here is very scant. A prior study showed no difference in outcome between cooling to 33°C for 24 vs 48 hours.
An ongoing randomized trial, dubbed ICECAP, will compare durations of hypothermia—with patients being randomized in multiple arms ranging from 6 hours up to 72 hours of cooling to 33°C. There will apparently be no normothermia control group, which is disappointing and hard to understand at this stage of the game. Both shockable and nonshockable rhythms will be included.
Never Speak of This Again
Remember how “therapeutic hypothermia” became “targeted temperature management” after randomized trials made the hypothermia part start to seem maybe not-so-therapeutic?
The new AHA guideline now nudges us to abandon the phrase “targeted temperature management” in favor of “temperature control.”
Temperature control, you see, “which encompasses hypothermic temperature control, normothermic temperature control, and temperature control with fever prevention.”
(Because we’ve got to keep that “hypothermic temperature control” in there, or what were we doing for the last 20 years?)
Cooling is Still Cool, Says AHA. But Why?
The original two randomized trials (now 20 years old) did not maintain normothermia in the control arms—so they arguably provide no information about cooling as a therapy itself (as distinct from fever prevention). The one major positive trial that showed benefits of cooling compared against normothermia, HYPERION’s results should be viewed with skepticism, as we detailed here.
So I ask, is there sufficient evidence for any benefit of cooling to 32° C to 34° C after cardiac arrest that justifies the risks of bleeding, arrhythmias, adverse effects from deep sedation and neuromuscular blockade, metabolic and electrolyte derangements, delayed accurate neurologic assessments and prognostication, prolonging family uncertainty and distress, not to mention inappropriately burdening treatment teams with unproven, probably excessive intensity of care?
Hypothermia should work: it’s been found to be neuroprotective, again and again, in recent reviews of animal models of cardiac arrest. (Although come to think of it, there was a risk of publication and other bias in those studies as well … )
But today, are there any conditions, a specific patient profile, etc., that would create a rationale strong enough to justify the downsides of cooling? I’m asking sincerely, not rhetorically; please comment below if you have something to share.
The guideline authors don’t even mention a ‘maybe’ use case. The authors of a widely used decision support tool continue to advocate for hypothermia (less strenuously than previously)—citing one study that suggested deeper encephalopathy (on EEG) correlated with favorable neurologic response to induced hypothermia, and also potential benefits of hypothermia seen in unrelated illnesses (traumatic brain injury, stroke, hepatic encephalopathy etc).
The main criticism of the multiple large negative randomized trials seems to be that cooling was too long delayed after arrest, and that it must be started sooner to be beneficial (and that it will someday yet be proven to be so).
Coda: Cochrane Confusion
Also in 2023, a Cochrane meta-analysis confusingly concluded “conventional cooling methods to induce therapeutic hypothermia” (between 32°C and 34°C) “may improve neurological outcomes after cardiac arrest.” But when they analyzed separately the three trials (n=1,044) in which controls were maintained at 36°C, no benefit was observed.
The Cochrane authors’ certainty was low for all their conclusions, due to a risk of bias in two studies.
While Cochrane reviews are considered by many to be the most comprehensive and definitive systematic reviews in medicine, their inclusiveness of small (n=30) and even unpublished, un-peer-reviewed data sets (both of which occurred in this review) can diminish the certainty in their results, as they themselves acknowledge.
A 2021 meta-analysis by ILCOR that excluded some of what Cochrane included, led them to conclude that “targeted temperature management at 32–34 °C, when compared to normothermia, did not result in improved outcomes.” (Their certainty was also low.)
With the publication of AHA’s new guidelines, cooling comatose patients below normothermia after cardiac arrest (of any type, occurring in any location) should no longer be considered standard care.
An active approach to fever prevention (maintaining normothermia) should be pursued as soon as possible post-arrest for patients not following commands. This could mean applying external automated cooling to maintain temperatures 36.5 - 37.5°C, for example.
Normothermia should be maintained for at least 24 hours. Longer periods of temperature control for persistently comatose patients have unknown benefits, if any, but seem to have little risk.
Patients who present with hypothermia post-cardiac arrest should not be warmed faster than 0.5°C per hour, AHA suggested.