SGEM#391: Is it Time for a Cool Change (Hypothermia After In-Hospital Cardiac Arrest)?

Guest Skeptic: Dr. Justin Morgenstern is an emergency physician and the creator of the #FOAMed project called www.First10EM.com

Case: You are working an overnight shift at a small rural hospital. You are tidying your things in anticipation of the arrival of the dayshift when a code blue is called. A 50-year-old man who was admitted to the hospital with a non-ST elevated myocardial infarction (NSTEMI) overnight was found unconscious and without a pulse. The nurses started CPR immediately and place pads before you even arrived. The patient is in ventricular fibrillation, and you achieve return of spontaneous circulation (ROSC) on the second shock. The patient is still unconscious. A post-arrest ECG doesn’t show any signs of STEMI. At this point, the dayshift doc walks into the room and asks, “I can’t keep up with all the evidence. Are we supposed to be starting hypothermia?”

Background: “Therapeutic” hypothermia took the critical care world by storm in 2002, with the simultaneous publication of two randomized control trials (RCTs) in the same issue of the New England Journal of Medicine – the Hypothermia after Cardiac Arrest (HACA) study and the Bernard study.

As a very brief recap, the HACA study randomized 275 comatose adult patients with ROSC after a witnessed cardiac arrest with a shockable rhythm, a presumed cardiac origin of arrest, and a short downtime. The hypothermia group was cooled using an external device to a target temperature between 32 and 34 degrees Celsius and maintained there for 24 hours.

The primary outcome was a good neurologic outcome within six months and occurred in 55% of the hypothermia group and 39% of the normothermia group (p=0.009, RR 1.40, 95% CI 1.08-1.81). This translated into an impressive NNT of 6. The six-month mortality was also improved in the hypothermia group (41% vs 55%, p=0.02) NNT 7. Key issues with this study were possible selection bias, early stopping without a clear endpoint, and a subjective outcome in a trial that was only partially blinded.

The Bernard study included 77 patients with an initial cardiac rhythm of ventricular fibrillation who had achieved ROSC but were persistently comatose. It was not randomized to individual patients, but rather based on the day of the week. It was also not blinded.

The primary outcome, patients with neurologic function good enough to be sent home or to a rehabilitation facility, occurred in 49% of the hypothermia group and 26% of the normothermia group (p=0.046, although when you plug the numbers into a fragility index calculator, you get a fragility index of 0 and a p value of 0.06). This give a very impressive NNT of 4. There was an NNT of 6 for mortality (51% vs 68%, p=0.16) but it was not statistically significant. For more information on the fragility index (FI) click on this LINK.

Therefore, therapeutic hypothermia was introduced into clinical practice based on two small trials with multiple sources of bias. Since 2002, we have seen several larger trials that have raised questions about the value of hypothermia.

We have covered the issue of cooling patients post OHCA sever times on the SGEM including the original Targeted Temperature Management (TTM) trial (SGEM#82). TTM was a multicentre RCT from 36 intensive care units (ICUs) in Europe and Australia, which enrolled 950 comatose adult patients on arrival to hospital after out of hospital cardiac arrest, regardless of the presenting rhythm. Patients all had their temperatures controlled, but they were randomized to a target of either 33 or 36 degrees Celsius. There were no statistical differences between the groups in mortality, Cerebral Performance Category (CPC), modified Rankin Score (mRS) or mortality at 180 days.

The TTM2 trial was covered on SGEM#336. It was another multicenter RCT, and that time hypothermia was compared to normothermia (a goal of keeping temperatures less than 37.5). Once again, there was no statistical difference in outcomes between the two groups in all-cause mortality or neurologic function at six months.

On the other hand, the HYPERION trial, which was covered in SGEM#275, was an RCT which include both OHCAs (73%) and IHCAs (27%). The trial comparing hypothermia (33 degrees) to normothermia. In that trial, there was a statistically significant improvement in their primary outcome of neurologically intact survival (10.2% TTM vs. 5.7% usual Care (absolute difference 4.5%), p=0.047 which gives an NNT of 22). The fragility index was 1 and the trial was unblinded, leaving us with significant uncertainty.

Therapeutic hypothermia has also been trialed in the prehospital environment and not been found to be superior to usual care (SGEM#21 and SGEM#54). Therefore, there remains significant uncertainty about the value of therapeutic hypothermia after cardiac arrest, especially in the inpatient environment, where patients generally have better outcomes than the OHCAs we usually see in the emergency department.

Clinical Question: Does hypothermia improve all-cause mortality in adult patients who remain comatose after inpatient cardiac arrest?

Reference: Wolfrum et al. Temperature Control After In-Hospital Cardiac Arrest: A Randomized Clinical Trial. Circulation. September 2022

• Population: Adult patients who remained unconscious (GCS <9) more than 45 minutes after inpatient cardiac arrest. Patients were eligible irrespective of cardiac rhythm or etiology of arrest.
  • Exclusions: These can only be found in the supplemental appendix and include the following: active bleeding, suspected intracranial bleeding, immunodeficiency, severe heart rhythm disorders, known severe cognitive deficit, pregnancy, any condition that makes 6-month survival unlikely, major hemodynamic instability.
• Intervention: Temperature control with a target between 32 and 34 degrees Celsius for 24 hours, followed by slow rewarming.
• Comparison: Temperature control with a target of normothermia. No specific protocol was followed; it was just strongly recommended to avoid temperatures greater than 37.5 Celsius.
• Outcome:
  • Primary Outcome: All-cause mortality at 180 days
  • Secondary Outcomes:In-hospital mortality and favorable functional outcome after 180 days using the Cerebral Performance Categories (CPC) score <3
• Type of Study: Multicentred, open-label, blinded-outcome-assessor, randomized controlled trial

Authors’ Conclusions: “Hypothermic temperature control as compared with normothermia did not improve survival nor functional outcome at day 180 in patients presenting with coma after IHCA. The HACA-IHCA (Hypothermia After In-Hospital Cardiac Arrest) trial was underpowered and may have failed to detect clinically important differences between hypothermic temperature control and normothermia.”

Quality Checklist for Randomized Clinical Trials:

1. The study population included or focused on those in the ED. No
2. The patients were adequately randomized. Yes
3. The randomization process was concealed. Yes
4. The patients were analyzed in the groups to which they were randomized. Yes
5. The study patients were recruited consecutively (i.e. no selection bias). Unsure
6. The patients in both groups were similar with respect to prognostic factors. No
7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
8. All groups were treated equally except for the intervention. Unsure
9. Follow-up was complete (i.e. at least 80% for both groups). Yes
10. All patient-important outcomes were considered. Yes
11. The treatment effect was large enough and precise enough to be clinically significant. No
12. Financial Conflicts of Interest. Several authors declared fCOIs

Results: Of 1,055 patients assessed for eligibility, 249 were randomized, and 238 are included in the final analysis. The mean age was 73 years, 64% were male, 54% were on the medical ward, and 73% were witnessed arrests.

Key Result: No statistical difference in all-cause mortality at six months.

• Primary Outcome: All-cause mortality at 180 days. 72.5% with hypothermia and 71.2% with normothermia, RR 1.03, 95% CI 0.79-1.40, p=0.89.
• Secondary Outcomes:No statistical difference for in-hospital mortality and favorable functional outcome after 180 days. in hospital mortality, ICU length of stay, and hospital length of stay were all also statistically insignificant.

1. Lack of Blinding: Although blinding a trial of hypothermia would be incredibly difficult, the lack of blinding could have a significant impact on the results. In a modern ICU, decisions about life and death are contingent on the choices of physicians, and the way we present prognosis to patients. Clinicians beliefs about the efficacy of hypothermia might have shaped their clinical decisions or how they counselled patients, and that could have biased even a seemingly objective outcome like all cause mortality.

2. Small Study Stopped Early: This trial was supposed to include 440 patients based upon their sample size, but they only enrolled 249. The trial was stopped early for futility, but there were no predetermined criteria for this decision. The result is a study less than half the size that they calculated was required to identify a 16% absolute decrease in all-cause mortality. Even without stopping early, this study was probably under-powered. A 16% absolute decrease in all cause mortality is a completely unheard-of benefit in modern critical care, so designing a trial with that goal seems overly optimistic. (To be fair to the researchers, that was the benefit supposedly seen in the original HACA trial.)

3. Imbalanced Groups: Perhaps because it was stopped early, the groups are not balanced at baseline. The hypothermia group had fewer patients with shockable rhythms, which is probably one of the most important prognostic factors in cardiac arrest.

4. Selection Bias and Generalizability: Like many efficacy trials this trial excluded far more patients than it included. Out of 1,055 patients screened, they enrolled 249, and included 238 in their final analysis. That means that the results may not generalize to the average patient you are treating. Also, with somewhat subjective exclusion criteria, the inclusion of less than 25% of patients screened raises concerns for selection bias.

5. Slow to Target Temperature: If you believe hypothermia is helping these patients, it makes sense to get them to their target temperature as soon as possible. In this trial, they failed to hit their target temperature in the hypothermia group in 28% of patients. The mean temperature at 12 hours was 34.0 degrees, which could bias against the treatment group. On the other hand, we have multiple trials of prehospital hypothermia showing no benefit, so maybe time to cooling is not an important factor. The burden of proof in on those claiming TTM has a patient-oriented outcome of benefit and not on others to prove that TTM does not work.

6. Time to Publication: This trial took a surprisingly long time to be published. The last patient was enrolled in May of 2014. That was nine years ago! Even with the 180 days of follow-up data that was required, it seems crazy that it took this long to get published. Perhaps this is an example of the publication bias that often occurs with negative results? Either way, it is important to recognize that these patients were enrolled between 2007 and 2014. Many aspects of treatment, such as neuro prognostication, have changed over the last 15 years, which again might limit the applicability of these results to your patients.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree that hypothermia did not improve survival nor functional outcome at day 180 in patients presenting with coma after IHCA.  Despite the many short comings of this trial, it is important data, as the bulk of the prior research on hypothermia has focused on out of hospital cardiac arrest.

SGEM Bottom Line: There are many priorities when managing a post-arrest patient. For now, the best available evidence seems to suggest that hypothermia is not one of them, and that we should aim to keep our patients in a normothermic range.

Case Resolution: You do not start a hypothermic protocol on the patient and are thrilled when you return a day later to find the patient sitting up in bed and talking to his family. You make a note of thanking his nurse for rapidly initiating CPR and defibrillation, which probably saved both his life and brain.

Clinically Application: Although there is remaining uncertainty, this trial adds to the body of research that suggests there is nothing therapeutic about therapeutic hypothermia. We don’t know if fever avoidance is important, but for now the best advice is to target normothermia is post-arrest patients.

What Do I Tell the Patient? Unfortunately, if this question arises, the patient is critically ill and comatose. When talking to their family, I might say something like, “You might have heard in the past the cooling patients down, or inducing hypothermia, could save lives. The best available research seems to suggest that it doesn’t work, so we are instead focusing our efforts on the many other aspects of critical care that are essential to your loved one’s survival.”

Keener Kontest: There was no winner last time. Dexamethasone was first synthesized by Philip Showalter Hench in 1957.

Listen to the podcast to hear this weeks’ trivia question. If you know the answer, send an email to TheSGEM@gmail.com with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.