Cardiac arrest in extreme hypothermia

Colloquially known as "freezing to death", cardiac arrest in extreme hypothermia is still a part of the critical care syllabus even in this severely overcooked country. This has appeared in the exam only a couple of times, as part of Question 7 from the second paper of 2019 and in  Question 3 from the first paper of 2020

In summary:

Prognostic implications of cardiac arrest with severe hypothermia:

  • Prognosis may be better than expected given the usually prolonged duration of CPR
  • A large percentage of survivors (~ 40%) have a good neurological outcome

Changes to basic life support:

  • It may take longer than normal to detect signs of life (up to 1 minute)
  • CPR should ideally be performed mechanically (prolonged CPR is to be expected)
  • Intermittent CPR (stopping for 5 minutes every 5 minutes) is reasonable for prehospital and retrieval staff, particularly when interruption facilitates retrieval
  • Manual or mechanical ventilation may encounter poor lung compliance

Changes to advanced life support

  • Do not defibrillate until core temperature is over 30°C.
    • If you decide to defibrillate and after three shocks the rhythm remains VF, withold further attempts until core temperature is over 30°C.
  • Do not give adrenaline until core temperature is over 30°C.
  • After 30°C is achieved, double the interval between adrenaline doses until 35°C
  • Use a low-reading thermometer to record core body temperature

Rewarming techniques

  • Extracorporeal circuit rewarming is the ideal
  • Warmed fluids and peritoneal lavage is the next best option
  • External warming is least effective
  • Remember that intubation will produce a increase in the rate of cooling by interruption of shivering though paralysis and anaesthesia.

The best resources for this are the large-scale society guidelines statements, the partly digested remains of which one can find in the discussion below. One should definitely become familiar with these as the primary resources:

Expectations of survival

"You are not dead until you are warm and dead", say ALS instructors worldwide as they teach the Special Circumstances station in their respective courses (and all respectable courses should have one of those). This is supported by the evidence. Saczkowski et al (2018) demonstrated that good neurological outcome survival in severe accidental hypothermia is around 40%, with a low percentage of survivors with a poor outcome (5.6%). 

Given the above, CPR may need to continue for a very long time. In fact, getting a mechanical device (eg. a LUCAS) is definitely warranted, as no team of rescuers is ever large enough for this sort of challenge. Everybody always seems to quote the case report by Kristian Lexow (1991),  published in a supplement to Arctic Medical Research, where the victim was CPRed for 6 hours and 30 minutes (while being rewarmed). The article is not available anywhere online and that journal has been out of print for decades, making this case report somewhat difficult to discuss; however other (newer) data is available, and is even more startling. For instance, Meyer et al (2014) reported a case of a 65-year old woman who had spent 8 hours and 40 minutes in cardiac arrest. "At 7:00 AM, hunters unexpectedly found the woman gasping on a snow-covered riverbank" after she (presumably) overdosed on SSRIs and jumped into the river. The lowest recorded core temperature in this patient was a minimum of 20.8ºC. They ended up rewarming her on bypass, but prior to this circulation was supported with manual CPR (for almost five hours). In summary, get a LUCAS, or line up about fifty burly orderlies. 

Basic life support in severe hypothermia

The recommendations for BLS in hypothermia can be summarised as "all things considered, do what you would normally do in a normothermic person". 

Are they really dead, or just nearly dead?

This is a serious question. Checking for signs of life is frustrated by virtually all the clinical features of hypothermia. A profoundly hypothermic patient appears dead for all intents and purposes. They are cold, they have fixed dilated pupils, and no peripheral pulse. Reflecting on this matter, the European Resuscitation Council issued guidelines recommending to spend one whole minute checking for signs of life. "Palpate a central artery and assess the cardiac rhythm" they say. Do a bedside echo. In response, one may argue that any situation where it takes an entire minute to find some evidence of cardiac activity is functionally indistinguishable from cardiac arrest, in terms of its need for urgent CPR. One would surely not be satisfied that everything is perfectly peachy if, after feeling around for some time, one finally finds a feeble carotid pulse with a heart rate of 10. 

Airway management

Many would be surprised to learn that there was at one stage a belief among critical care professionals that intubation of the severely hypothermic patient might give rise to fatal arrhythmias. From the practical perspective this defies logic, considering that the patient is already experiencing cardiac arrest, and some might say that things could not get much worse. However, somehow this thinking had propagated to the point that a review of historical cases (Danzi et al, 1987) had to make a specific entry to debunk this myth. It appears to have its origin in the 1960s, when one group of authors (Fell et al, 1968) misquoted another (Lee & Ames, 1965). "The factors which are believed to aggravate the tendency to ventricular fibrillation in hearts which are cold include ... stimulation of the throat", wrote Fell et al, supporting this with the citation that "endotracheal intubation was followed by cardiac arrest in a large proportion of the cases of barbiturate poisoning with hypothermia described by Lee and Ames". However, Lee and Ames did not describe any such events. Theirs was a case series describing the use of haemodialysis in hypothermic barbiturate poisoning, and they did not report any association between intubation and cardiac arrest; it only came up in their discussion where they cautioned against intubating people purely for the purposes of gastric lavage. In summary: do not delay tracheal intubation. Manage the airway as you normally would.

Respiratory support

As one's metabolic rate drops with decreasing temperature, so does the utilisation of metabolic substrates. As such, one might conclude that fewer breaths per minute would be enough for the hypothermic patient. At this stage, nobody has spoken up in support of this hypothesis.  It will, of course, do no harm to continue doing the normal thing (i.e. giving a random number of breaths while pretending to count compressions). Generally, experts agree that "during CPR, ventilation should be provided as in normothermic CA patients" (Paal et al, 2016). Given that lung compliance deteriorates with hypothermia, one might find oneself applying lung-protective principles, which is fine because not a lot of CO2 is being produced by the patient and so a low minute volume should be reasonably well tolerated. 

External cardiac compressions

Following on from the "what are they going to do with it anyway" argument, one might conclude that fewer cardiac compressions per minute would be enough. Also the chest is harder to compress because of muscle rigidity, so surely one may be able to spare the rescuers and slow things down a little? Interestingly, there is little support for brady-CPR in the literature. Even weirder is the suggestion that CPR continues at a normal rate, but only intermittently, something which flies in the face of normal minimise-time-off-the-chest dogma. However, this is also a serious recommendation of the ERC. Particularly where the temperature is below 28°C, they recommend that should receive 5 min of CPR, alternating with periods ≤5 min without CPR. This is supported by work done by Gordon et al (2015), who came to the conclusion that "inefficient continuous CPR during evacuation is potentially worse than intermittent good-quality CPR" and recommended that rescuers focus instead on bringing the patient to some place of warmth and ECMO.

Case reports abound regarding the success of delayed, intermittent or totally absent CPR in profound hypothermia. For one example, Oberhammer et al (2008) describe a skier extracted from an avalanche who, with a core body temperature of 22.0 °C, was flown in a helicopter across the Eastern Alps in VF, with nobody doing any CPR. When they arrived at Bruneck Hospital (a 15 minute flight), he was rewarmed on bypass and went on to have a complete recovery. In summary, though local recommendations emphasise the need for uninterrupted CPR, in severe hypothermia the destination is all-important and one should not feel too bad about being off the chest for a few minutes if it makes the retrieval process faster.

Circulatory access

Much is made of the arrhythmogenicity of J-tipped guidewires in hypothermia. Most guideline writers seem to acknowledge this at least on some superficial level. For instance, the AHA document on resuscitation in special circumstances recommends that we "do not delay urgent interventions such as airway management and insertion of vascular catheters regardless of evidence of cardiac irritability"; but a supplement cautions one to "perform them gently and monitor cardiac rhythm closely", as if the routine practice were to violently cram them in while ignoring the monitors. In general, most people will agree that if a central line is needed it should be placed. Some recommendations from non-guidelines publications recommend you focus on femoral vessels, but this is probably going to be impractical because the groins should be spared for ECMO cannulation.

Advanced life support in severe hypothermia


In general, recommendations regarding defibrillation in the profoundly hypothermic patient can be summarised as "it's pointless". The 20 °C patient in VF will be resistant to shocks. The VF will remain refractory because the hypothermia itself is driving the propensity toward VF, by lowering the fibrillation threshold (Mortensen et al, 1993). If you do by some chance manage to defibrillate successfully, the rhythm will again degenerate into VF shortly thereafter. So, some might argue that it is ok not to defibrillate until the temperature has risen to ≥30 °C. If you do decide to defibrillate for whatever reason, you may still be successful for some sustained ROSC, but if you don't succeed feel free to give up after three shocks: the ERC recommends that "if VF persists after three shocks, delay further attempts until core temperature is ≥30 °C". The energy you use should remain the same (i.e. maximum), as suggested by some pig data (Ujhelyi et al, 2001)

ACLS drugs

These days, this really just means adrenaline and amiodarone. These are drugs with nontrivial toxicities, and people have a legitimate concern that in hypothermic cardiac arrest they will accumulate and cause harm. Consider: if every two cycles one gives the hypothermic patient some adrenaline and it takes 60 minutes to rewarm them, that's potentially a total dose of 15mg of adrenaline. The AHA and ERC both recommend waiting for a temperature ≥30 °C before one starts giving these drugs. That's not to say that we have very strong evidence for this: the data comes from an animal study by Kornberger et al (2001), who gave repeated adrenaline doses to hypothermic pigs and found them becoming more acidaemic over the course of external cardiac massage. 

Amiodarone, a drug whose slippery grip on its place in the resuscitation guidelines is maintained only by the suspicion that it might convert refractory VF into being slightly less refractory, is also probably pointless in severe hypothermia. Again, there is little direct human evidence, but a canine model by Stoner et al (2003) clearly demonstrated that amiodarone had no effect on the success rate of defibrillation when the myocardium is cooled. 


Lexow, Kristian. "Severe accidental hypothermia: survival after 6 hours 30 minutes of cardiopulmonary resuscitation." Arctic medical research 50 (1991): 112-114.

Meyer, Marie, et al. "Sequela-free long-term survival of a 65-year-old woman after 8 hours and 40 minutes of cardiac arrest from deep accidental hypothermia." The Journal of thoracic and cardiovascular surgery 147.1 (2014): e1-e2.

Saczkowski, Richard S., et al. "Prediction and risk stratification of survival in accidental hypothermia requiring extracorporeal life support: An individual patient data meta-analysis." Resuscitation 127 (2018): 51-57.

Danzl, Daniel F., et al. "Multicenter hypothermia survey." Annals of emergency medicine 16.9 (1987): 1042-1055.

Fell, R. H., et al. "Severe hypothermia as a result of barbiturate overdose complicated by cardiac arrest." The Lancet 291.7539 (1968): 392-394.

Lee, H. A., and A. C. Ames. "Haemodialysis in severe barbiturate poisoning." British medical journal 1.5444 (1965): 1217.

Paal, Peter, et al. "Accidental hypothermia–an update." Scandinavian journal of trauma, resuscitation and emergency medicine 24.1 (2016): 111.

Gordon, Les, et al. "Delayed and intermittent CPR for severe accidental hypothermia." Resuscitation 90 (2015): 46-49.

Oberhammer, Rosmarie, et al. "Full recovery of an avalanche victim with profound hypothermia and prolonged cardiac arrest treated by extracorporeal re-warming." Resuscitation 76.3 (2008): 474-480.

Lee, Christopher H., et al. "Advanced cardiac life support and defibrillation in severe hypothermic cardiac arrest." Prehospital Emergency Care 13.1 (2009): 85-89.

Mortensen, Elin, et al. "Changes in ventricular fibrillation threshold during acute hypothermia. A model for future studies." Journal of basic and clinical physiology and pharmacology 4.4 (1993): 313-320.

Kornberger, Elisabeth, et al. "Effects of epinephrine in a pig model of hypothermic cardiac arrest and closed-chest cardiopulmonary resuscitation combined with active rewarming." Resuscitation 50.3 (2001): 301-308.

Ujhelyi, Michael R., et al. "Defibrillation energy requirements and electrical heterogeneity during total body hypothermia." Critical care medicine 29.5 (2001): 1006-1011.

Stoner, Jason, et al. "Amiodarone and bretylium in the treatment of hypothermic ventricular fibrillation in a canine model." Academic emergency medicine 10.3 (2003): 187-191.