A post last year discussed the top 10 reasons to stop cooling to 33C. It was based largely on the Nielsen trial, which showed similar outcomes between therapeutic hypothermia (TH33) and therapeutic temperature management (TTM36). However, this trial left some questions about how these protocols would perform outside the context of a RCT (external validity). Last year's post speculated that since TTM36 is easier to achieve, it would out-perform TH33 in real-world conditions (see #2).
Casamento, Young, Bellomo et al. 2016
This is a retrospective study comparing patients treated at two hospitals before and after transitioning from TH33 to TTM36. Both groups contained 69 patients and were well matched:
The primary outcome was the amount of time spent in the prescribed temperature ranges during the first 24 hours. There were no differences. However, there were differences in several secondary outcomes:
(1) Fever control and “dose” of hypothermia
Patients in the TTM36 group had a lower rate of fever during the first 96 hours after cardiac arrest (33% vs. 55%, p=0.01). This may relate to differences in the protocols:
- TH33 involved targeting 33C for 24 hours followed by gradual rewarming.
- TTM36 involved targeting 36C for 28 hours, rewarming gradually to 37C, and then maintaining patients below 37.5C until 72 hours after arrest.
The key to targeted temperature management might be avoidance of fever. If that were the case, then the TTM36 protocol would be superior because it provides a longer duration of temperature control. Thus, the best way to maximize the “dose” of temperature therapy may be to extend the duration of therapy (e.g., normothermia for 72 hours) rather than dropping the target temperature (e.g., hypothermia for 24 hours).
(2) Shivering, sedatives, and paralytics
Patients treated with TH33 had a markedly greater rate of shivering (52% vs. 19%). Patients in the TH33 group also were more likely to be paralyzed, and received higher doses of fentanyl and midazolam. It seems likely that paralysis and deeper sedation was required in the TH33 group to manage shivering (for example, there are few other reasons that these patients should require paralysis).
Deeper sedation associated with TH33 is potentially harmful for several reasons, including increased risk of delirium, clouding of neuroprognostication, and delaying extubation. Paralysis is especially undesirable in these patients because it could obscure diagnosis of a seizure.
(3) Hemodynamic stability
Patients in the TH33 group had trends towards increased hemodynamic instability and arrhythmia. This is consistent with data from the Nielsen trial showing that patients in the TH33 group had worse SOFA-Cardiovascular scores (more here, see #7). This is also consistent with the clinical observation that hypothermia causes hypotension and bradycardia.
Given that this is a retrospective cohort study, it is subject to numerous confounders. For example, it is possible that minor improvements in critical care over time could have led to better outcomes in the TTM36 group.
However, overall these results are consistent with the Nielsen study. The only potentially surprising result is a reduced rate of shivering in the TTM36 group, a finding which was not observed in the Nielsen study. However, this seems to be consistent with my experience (specifically, patients managed with TTM36 rarely have severe shivering that requires paralysis).
Neurointensive care: Time to stop being over-achievers?
The history of neurocritical care is replete with attempts to push physiology beyond the limits of normalcy:
- Sustained hypocapnia (in efforts to reduce ICP)
- “Triple-H therapy” (hypertension, hypervolemia, and hemodilution) for cerebral vasospasm
- Hypertonic saline infusions
- Therapeutic hypothermia
Unfortunately, these heroic efforts tend to fail. The injured brain is a delicate organ, which doesn’t seem to like being pushed beyond the limits of its usual homeostasis. Thus, it is possible that the best general approach to neurocritical care is aggressive defense of normal physiology: normocapnia, normonatremia, normothermia, and normal cerebral perfusion pressure.
- Casamento 2016 is a retrospective cohort study comparing patients at two hospitals before and after transitioning from therapeutic hypothermia to therapeutic normothermia.
- Normothermia was associated with less shivering, less paralysis, less sedation, less hemodynamic instability, and fewer fevers.
- Despite the possibility of confounding, these results support the adoption of therapeutic normothermia.
- Top 10 reasons to stop cooling to 33C (PulmCrit)
- The targeted temperature trial changes everything (EMCrit)
- Post-cardiac arrest care in 2013 with Stephen Bernard: Parts I and II (EMCrit)
- Hypertonic saline infusion rant (PulmCrit)
Opening image: From Frozen film (2010)
- IBCC chapter:Guide to APRV for COVID-19 - April 8, 2020
- PulmCrit Theoretical Post – The COVID Severity Index (CSI 1.0) - April 2, 2020
- PulmCrit wee – Why the SCCM/AARC/ASA/APSF/AACN/CHEST joint statement on split ventilators is wrong. - March 29, 2020