Targeted temperature management at 33C versus 36C after cardiac arrest
Nielsen et alNew England Journal of Medicine, 12 2013, vol./is. 369/23(2197-206), 0028-4793;1533-4406
“Unconscious survivors of out-of-hospital cardiac arrest have a high risk of death or poor neurologic function. Therapeutic hypothermia is recommended by international guidelines, but the supporting evidence is limited, and the target temperature associated with the best outcome is unknown. Our objective was to compare two target temperatures, both intended to prevent fever.”
“…randomly assigned 950 unconscious adults after out-of-hospital cardiac arrest of presumed cardiac cause to targeted temperature management at either 33C or 36C. The primary outcome was all-cause mortality through the end of the trial. Secondary outcomes included a composite of poor neurologic function or death at 180 days.”
“In unconscious survivors of out-of-hospital cardiac arrest of presumed cardiac cause, hypothermia at a targeted temperature of 33C did not confer a benefit as compared with a targeted temperature of 36C.”
Big news in the emergency and critical care worlds. Having read the article I have since listened to a couple of podcasts which have discussed the subject, and for my own learning, and hopefully for anyone else who wants to read my blog I have tried to summarise some of the things that were said.
The first is Matt MacPartlins’ interview with Niklas Nielsen and Anders Aneman, authors of the above trial. As temperature was still controlled even with the patients at 36 degrees Dr Nielsen did not feel that TTM was no longer applicable. Fever was still therefore actively treated. Cooling is still important and temperature should still be managed.
Broad inclusion criteria in the TTM trial was felt to help with its validity, although it is still specific to out of hospital cardiac arrest with presumed cardiac cause. Excluded drowning hanging and trauma. Ended up with 80% shockable rhythms in the trial. Included 87% of the patients that were screened.
Power calculations were based on looking for a mortality relative risk reduction of 20%. In relation to previous trials this was quite modest. Powered for mortality and not for neurological outcome.
More work to be done on neurological outcomes rather than just mortality. There are some other outcome measures which are currently not validated in the cardiac arrest population, which they plan to look into for more neurological outcome. Plan for more trials.
Have proven with this trial that it is possible to do large cardiac arrest trials. Collaboration is needed across borders.
It might be time to look at the patients with the higher temperatures to see the effects of this i.e. temps higher than 36 degrees as in previous studies.
It now might be a dose range rather than a specific number is more appropriate. There might be other temperatures which need to be investigated. There may be some patient types who benefit from 33 degrees and some from 36 degrees.
Is timing of the intervention important? TTM trial cannot answer this question. Other trials did not demonstrate any benefit of cooling earlier. Indication is that earlier cooling would not have mattered.
Protocolisation of withdrawal of therapy in the trial has been praised.
Both would expect their policies to change so that they aim for 36 degrees and avoid pyrexia. Would not seem to be any justification for aiming for the lower temperature.
The second is Scott Weingarts interview with Stephen Bernard author of “Treatment of Comatose Survivors of Out-of-Hospital Cardiac Arrest with Induced Hypothermia” NEJM (2002). This is a discussion which covers not just therapeutic hypothermia but also many other aspects of post cardiac arrest care.
Dr Bernard opens by saying that the Nielsen paper is an excellent, well-conceived and well run trial, and compliments them on their approach to prognostication during their trial. As a consequence he has gone back to his own unit and adopted a target temperature of 36 degrees and the method of prognostication discussed. He asks the question as to why not just stick to 33 degrees if they are equivalent, but answers this by commenting on how difficult it can be to achieve that degree of cooling, difficulty in both cost and time.
Dr Weingart asks if there is any group of patients which should be cooled to 33, and Dr Bernard talks about a trial currently running in Australia where cooling takes place during CPR although the steering committee is meeting soon to discuss their options as it was originally based on the fact that survivors would be cooled. It may be that cooling during CPR is where we go? From previous studies it is known that cooling on the way to hospital is not worthwhile. The other group which may still benefit from cooling is the refractory cardiac arrest patient going onto ECMO. Difficult to test in an RCT but animal data is supportive.
Dr Weingart asks if there is anything to support prognostication as the patient rolls through the door or time zero as he calls it. Using data from the Victoria Cardiac Arrest Data, Dr Bernard would be cautious with the un witnessed arrest with an initial asystole (often admitting this type of patient to his own unit but with a view towards palliative or bridging care for organ donation), or the older patient, 80 years plus.
The discussion then moves onto which patients one takes to the cath lab. At Dr Bernard’s hospital a post VF arrest patient will go to the lab unless there is a compelling reason not to go. After hours becomes a little more difficult as the following days’ work has to be rescheduled. Patients then tend to be kept in ITU overnight. They have the data which demonstrates that 24/7 cath lab facilities are beneficial.
Unnecessary oxygen is harmful, shown in animal studies, in the reperfusion period. However it would be a change of practice for paramedics to give low, or no oxygen, whilst avoiding hypoxia. Dr Bernard will actively reduce oxygen being given to these patients, not infrequently ending up on air! Not yet proven but may appropriate to test in a RCT. Best evidence shows that lower oxygen is better. Some pilot work started in New Zealand.
There is also some animal data and observation studies which may show some improved outcomes with higher CO2. Dr Bernard not persuaded at present though.
Now aiming for systolic BP of 120. Recommendations from ILCOR have varying results, some with a MAP as high as 110 in 2000. In 2010 it changed to a systolic BP of 90 being associated with improved outcomes. They have some evidence that a systolic BP of 120 is associated with the best outcomes. Reached statistical significance at 90. This work has been submitted for publication. Would the post arrest patient benefit from a pressor? May need a study. Best evidence now is that 120 is the target.
Aiming for a temperature of 36 will require some intervention as they will be inclined to go hotter so surface cooling technology will need to be applied which will induce shivering so some sedation will be needed for a full 24 hours and then switch sedation off. Dr Bernard tends to use Morphine and Midazolam as Propofol tends to need a vasopressor with it. Starting to use more remifentanil as a pose to fentanyl, as it wears off very rapidly. But the idea is still that prognostication does not take place for 96 hours.
During swine flu in Australia critical care physicians were trained how to get patients onto ECMO. Once they became proficient they felt they could do it during chest compressions, and now have done 16 patients with 7 making full recoveries. Technique of placing the lines should be something most emergency physicians should be able to do.
An alternative approach for refractory cardiac arrest would be the use of automated compressions. Some units in Sydney are looking to take this forward. May need to have ECMO support as well.
With the advent of ECMO and automated compression devices should paramedics scoop and run? This is something that is being trialled by Dr Bernard with a donation from Zoll. However still asking paramedics to through their normal routine of placing lines, securing airway etc, which should take about 10 minutes and then leave the scene at approx. 20 mins, but only with someone who is an ECMO candidate.