Prove it: Early use of therapeutic hypothermia improves cardiac arrest outcomes

Researchers conducted an observational prospective investigation to compare outcome between patients receiving early versus late therapeutic hypothermia


Engine and Medic 42 arrive at a local residence to find a male age 66 lying in the front yard with neighbors performing cardiopulmonary resuscitation (CPR). Two firefighters take over CPR while Medic DaVilla attaches a manual monitor/defibrillator to the patient's chest.
During a very short interruption in chest compressions, DaVilla recognizes ventricular fibrillation (VF), charges the defibrillator, and delivers a countershock. The two firefighters immediately resume chest compressions.

During the two-minute period of CPR, Medic Williamson establishes intravenous access. Before administering epinephrine, Williamson notices that during the ventilation portion of CPR, the rhythm displayed on the electrocardiogram (ECG) appears organized. He decides to withhold the drug until after the next pulse check.

At the end of the two-minute cycle of CPR, DaVilla confirms return of spontaneous circulation (ROSC) through the presence of a carotid pulse. The patient remains unconscious and apneic. Williamson successfully intubates the patient and confirms the tube location with quantitative waveform capnography.

The patient is hemodynamically stable and the 12-lead ECG confirms the presence of an ST-segment myocardial infarction (STEMI). DaVilla activates the cardiac catheterization lab from the scene and Medic 42 begins transport.
Upon arrival, the emergency department (ED) physician verifies endotracheal tube placement and the crew proceeds to the catheterization suite. There, the staff replaces your IV with a bag of chilled saline and begin surface cooling measures.

One member of the team comments, "I'll be glad when you guys start cooling in the field. It will be better for the patient and it will save us some time."


Italian researchers conducted an observational prospective investigation to compare outcome between patients receiving early versus late therapeutic hypothermia following successful resuscitation from cardiac arrest (The Italian Cooling Experience (ICE) Study Group, 2011).

To obtain the study sample, researchers included all adult patients who achieved ROSC long enough for admission to an intensive care unit (ICU). The researchers excluded patients

  • Under the age of 18 years
  • With hemodynamic instability, defined as systolic blood pressure below 80 mmHg despite the administration of a vasopressor or an inotrope
  •  With sustained cardiac arrhythmia
  • Who developed a severe blood clotting disorder with active bleeding
  •  Who suffered a traumatic cardiac arrest
  • Who were pregnant
  • Who suffered an acute intoxication.

Physicians in each ICU chose which patients received therapeutic hypothermia and managed the cases according to facility-specific guidelines and equipment, which included ice packs and chilled saline infusion, intravascular cooling devices, or surface cooling techniques.

The target temperature for each of patient was 32 C to 34 C, which is consistent with the current American Heart Association (AHA) guidelines (Peberdy et al., 2010). The researchers arbitrarily defined early hypothermia as implementation within two hours following the onset of cardiac arrest.

The outcome measures of most importance to the research team included survival and neurological status at two different time points. The research team defined short-term survival as survival to ICU discharge and long-term survival as survival for at least 6-months after the cardiac arrest occurred.

The researchers measured the patient's neurological status using the Cerebral Performance Categories (CPC) scale, which uses a range of values from one to five (Jennett & Bond, 1975). Cardiac arrest researchers often categorize scores of one or two as favorable outcomes.

A score of one actually means that the patient is conscious, alert and is able to work. There may be some minor psychological or neurological deficits but nothing so severe that it interferes with daily activities of normal life. A score of two indicates that the patient is conscious, is independently capable of daily activities such as dressing and self-feeding, but can only work part-time.

Patients with a CPC score of two may suffer from seizures, hemiplegia, or permanent memory or mental changes. On the other end of the scale, a score of four indicates a vegetative state while a score of five is consistent with brain death (Cummings et al., 1991).


During a nine-month period in 2009, 174 patients achieved ROSC long enough to be admitted into one of 17 ICUs in Italy. On average, each patient achieved ROSC 20 minutes after the onset of pulselessness, with the range of cases between 10 to 30 minutes. Three quarters of the cardiac arrests occurred in the out-of-hospital environment.

Physicians at each institution initiated therapeutic hypothermia in 122 patients, which represents about 70 percent of all the ROSC admissions. The primary reasons for not initiating therapeutic hypothermia were severe coexisting disease, unfavorable decision by the attending physician, and patients who regained consciousness following ROSC.

Table 1 displays the results of a simple outcome comparison between patients who received therapeutic hypothermia and those who did not. Although there is a clear difference in the absolute value of the numbers in the two comparison columns, one must consider the influence that random chance had on outcome. That information is conveyed in the last column.

The p-value represents the probability of the difference occurring by chance. By convention, researchers use 0.05 as the threshold for statistical significance. If the p-value is less than 0.05, there is a less than one in 20 chance that randomness caused the difference.

In this study, both p-values associated with mortality are greater than the threshold, indicating that it is possible that the observed differences were caused by chance and not the intervention (therapeutic hypothermia).

However, the p-values associated with the CPC score are below the threshold, suggesting that randomness played a very small role and the neurological improvement more likely resulted from the intervention.

Table 1:

Outcome for patients who received TH vs. those who did not receive TH


(n = 122)


(n = 52)


ICU Mortality




6-month Mortality




Median CPC score at ICU discharge




Median CPC score at 6 months




TH = Therapeutic Hypothermia; CPC = Cerebral Performance Categories

A CPC score of 1 indicates good cerebral performance while a score of 5 means brain death.

More importantly, the purpose of this investigation was to examine whether early versus later therapeutic hypothermia influenced outcome. For the early hypothermia group, the hospital staff began the cooling process about 60 minutes (range 34 - 90 minutes) and achieved the target temperature about 180 minutes (range 100 - 284 minutes) after the onset of cardiac arrest.

Cooling in the late hypothermia group began about 192 minutes (range 160 - 255) after the onset of cardiac arrest, with the staff achieving target temperature at the 410-minute mark (range 300 - 490 minutes).

There were no significant differences in age, sex, presenting rhythm, interval from arrest onset to arrival of a basic life support (BLS) rescue team, or the interval from arrest onset until ROSC between the two groups.

Table 2 displays the outcome data of interest. Observed mortality at both outcome time points was significantly higher in the patients who received therapeutic hypothermia within two hours of cardiac arrest onset compared to those who received therapeutic hypothermia later. For the survivors, neurological status at the end points was not different between the two groups.


Table 2: Outcome for patients who received early TH vs. later TH

Early TH

Later TH


Mortality at ICU Discharge




Mortality at 6 Months




CPC score at ICU discharge




CPC score at 6 Months




TH = Therapeutic Hypothermia; CPC = Cerebral Performance Categories

A CPC score of 1 indicates good cerebral performance while a score of 5 means brain death.

The study authors conclude that initiation of therapeutic hypothermia within the first two hours is associated with increased mortality. For survivors, there is no association between early hypothermia and worsening neurological outcome. The authors acknowledge the unexpectedness of the results and further conclude that other researchers will need to conduct additional studies to see if the results are consistent.

What this means for you

Although available evidence suggests that initiating mild therapeutic hypothermia in patients who achieve ROSC following cardiac arrest improves both short term survival and neurological outcome (Arrich, Holzer, Herkner, & Müllner, 2009; Belliard et al., 2007; Bernard et al., 2002; Bernard, Jones, & Horne, 1997; Busch, Soreide, Lossius, Lexow, & Dickstein, 2006; Castrejon et al, 2009; Don et al., 2009; Felberg et al., 2001; Hachimi-Idrissi, Corne, Ebinger, Michotte, & Huyghens, 2001; Hypothermia after Cardiac Arrest Study Group, 2002; Oddo, Schaller, Feihl, Ribordy, & Liaudet, 2006; Storm et al., 2008; Sunde et al., 2007), there are still many unanswered questions related to optimal patient selection, onset and optimal cooling method, and duration of therapy (Peberdy et al., 2010).

Animal cardiac arrest models suggest that initiating hypothermia within the first 15 to 20 minutes after the cardiac arrest produces neurological benefits not observed if the therapy is delayed until a later time (Abella et al., 2004; Kuboyama et al., 1993; Takata et al., 2005).

Those models kept the animal cool for less than one hour before measuring the effects. A different animal model found that implementing therapeutic hypothermia within the first hour following ROSC also improves outcome even with longer duration cooling (> 12 hours) (Hicks, DeFranco, & Callaway, 2000).

However, benefits observed in animal models do not always translate into human benefits. An analysis of data in an out-of-hospital cardiac arrest registry could not demonstrate differences in neurological outcome related to when cooling began or how long it took to achieve the target temperature following ROSC, although cooling began in under two hours and reached the target temperature in less than seven hours for every patient (Nielsen et al., 2009).

A small case series of both in- and out-of-hospital patients who achieved ROSC following cardiac arrest could not demonstrate a significant difference in neurological outcome related to either how quickly cooling began or when the patient reached the target temperature (Wolff, Machill, Schumacher, Schulzki, & Werner, 2009).

Paramedics infusing chilled saline following ROSC in adult patients presenting with VF lowered the patient's body temperature an average of 3.46 F (0.8 C) by arrival in the ED, but this intervention did not result in improved outcome at hospital discharge (Bernard et al., 2010).

Physicians staffing an EMS helicopter who administered chilled saline to adult patients regardless of the presenting rhythm saw similar results (Kämäräinen, Virkkunen, Tenhunen, Yli-Hankala, & Silfvast, 2009).

Paramedics using an intranasal cooling device demonstrated that starting the cooling process during the resuscitation attempt on scene reduced the interval necessary to reach target temperatures; however, there were no improvements in ROSC, survival to hospital admission, or survival to hospital discharge (Castrén et al., 2010).

These studies suggests that for most patients, initiating the cooling process on scene or en route to the hospital may not provide improved survival benefits over waiting until the patient arrives in the ED.


This investigation was an observational study and not a controlled experiment. Observational studies are useful for establishing correlations and associations, which can help drive future research projects.

However, observational studies cannot establish causation. Readers are cautioned not to assume that this study proves early hypothermia results in increased mortality. There is only an association and many other factors may explain why.

For example, the early therapeutic hypothermia group did have a higher Simplified Acute Physiology Score II (SAPS II), which is a measure of disease severity (Le Gall, Lemeshow, & Saulnier, 1993). Higher scores suggest more severe disease and it is reasonable to expect the early group to have higher mortality because of the increased severity of their disease.

Patients in the early hypothermia group also had significantly lower mean arterial blood pressures during their first two ICU days. The authors speculate that slowing metabolism too early may interfere with the body's ability to self-stabilize its hemodynamic status.

In fact, on day two and day four ICU residency, almost twice as many patients in the early hypothermia group needed vasopressor support to maintain adequate perfusion than the patients in the later hypothermia group.

It is reasonable to expect increased mortality in groups of patients who display hemodynamic instability at an earlier moment.

Although the authors do not report the raw data, they do point out that in a very small subgroup of patients with available blood lactate levels prior to cooling, patients who received early hypothermia had higher values than the later hypothermia group.

Elevated blood lactate levels indicate tissue ischemia or hypoxemia. Researchers have established a link between early effective lactate clearance and survival following cardiac arrest (Donnino et al., 2007; Kliegel et al., 2004). It is reasonable to expect higher mortality in the early cooling group because of the elevated blood lactate levels.

This investigation is also limited by several other factors related to the study design. The researchers only sampled a small number of patients from the larger population of people who suffer cardiac arrest. Trying to generalize results from a small sample to a larger population is always problematic. The researchers could also not control the cooling process.

As a result, each institution was free chose both the patients who would receive therapeutic hypothermia and the cooling method, thus creating variables with potential outcome effects.

Finally, the researchers did not provide data on other variables known to effect outcome, such as the bystander CPR, time interval until the first shock, and post-cardiac arrest blood glucose levels.

To date, therapeutic hypothermia is the only intervention with a demonstrated benefit at improving neurological recovery (Peberdy et al., 2010). However, there is still much that the resuscitation community does not know about the therapy, including the optimal time to begin the cooling process.

So far, there is no conclusive data that demonstrates improved survival or neurological status for patients who receive initial cooling measures in the field.

Recommended for you

Join the discussion

Copyright © 2020 EMS1. All rights reserved.