Article Text

Download PDFPDF

Therapeutic hypothermia for out-of-hospital cardiac arrest: implementation in a district general hospital emergency department
  1. Shashank Patil,
  2. Sadiq Bhayani,
  3. John M Denton,
  4. Jerry Nolan
  1. Royal United Hospital, Bath, UK
  1. Correspondence to Shashank Patil, 24 Pavilion Terrace, Wood Lane, London W12 0HT, UK; shashank.patil{at}


Background The use of therapeutic hypothermia is recommended for unconscious adult patients with return of spontaneous circulation (ROSC) after out-of-hospital ventricular fibrillation cardiac arrest. There is evidence that the time taken to achieve target temperature impacts survival.

Objectives To audit the performance of an emergency department (ED) in implementing therapeutic hypothermia and achieving target temperature in survivors of out-of-hospital cardiac arrest admitted to the intensive care unit (ICU).

Methods Data were extracted from the medical records of patients admitted to the ICU from the ED in the Royal United Hospital following out-of-hospital cardiac arrest (OHCA) between June 2002 and October 2008. The intervals between ROSC and initiation of cooling and between initiation of cooling and achieving the core temperature of 34°C were recorded.

Results During this period, 83 patients were admitted to the ICU following OHCA. Of these, 67 (81%) were actively cooled. All 16 patients who were not cooled had recognised exclusion criteria. The median time (IQR) from ROSC to initiation of cooling was 60 (40–165) minutes and the median time (IQR) to reach 34°C was 175 (40–420) minutes. Of the 67 who were cooled, 44 (66%) achieved the temperature of 34°C within 4 h, the audit standard published by the Royal College of Anaesthetists. In 29 (43%) patients, the temperature increased after leaving the ED.

Conclusions Among OHCA patients who met recognised inclusion criteria, therapeutic hypothermia was implemented successfully by the ED staff. The temperature should be measured continuously from the same site in both the ED and the ICU. This will provide consistent and continuous temperature monitoring between the ED and the ICU and will enable prompt intervention to prevent temperature increases.

  • Therapeutic hypothermia
  • cardiac arrest
  • ventricular fibrillation
  • return of spontaneous circulation (ROSC)
  • environmental medicine
  • hypothermia
View Full Text

Statistics from


In 2003, the International Liaison Committee on Resuscitation (ILCOR) recommended that unconscious adult patients with spontaneous circulation after out-of-hospital cardiac arrest (OHCA) should be cooled to 32–34°C for 12–24 h when the initial rhythm was ventricular fibrillation.1 The ILCOR scientific statement also stated that cooling may also be beneficial for other rhythms or in-hospital cardiac arrest and these indications are now supported by clinical studies with historical controls.2 3 Hypothermia reduces cerebral oxygen consumption and suppresses the effects of damaging chemical cascades initiated by reperfusion—so called ‘reperfusion injury’.4 5

In recent years, the use of therapeutic hypothermia (TH) in the management of OHCA has become common practice throughout the UK.6 Two recent observational clinical trials suggest that more favourable neurological outcomes are associated with early initiation of TH and shorter times to reach the target temperature range of 32–34°C.7 8 The Royal College of Anaesthetists audits compendium cites achieving a target temperature of 34°C within 240 min as an audit standard.9

We have examined retrospectively the implementation of TH in patients who presented to a district general hospital emergency department (ED) following OHCA between 1 January 2002 and 31 December 2008. The aim of this study was to analyse how well therapeutic hypothermia has been implemented locally and to determine whether target temperature is being achieved within an acceptable time from return of spontaneous circulation (ROSC).



Staff on the intensive care unit (ICU) at the Royal United Hospital (RUH) recorded the name, sex, hospital number, and the date and time of arrival onto the ICU of all patients admitted to the ICU for therapeutic hypothermia following OHCA between 1 January 2002 and 31 December 2008. The medical records for these patients were obtained and the following information was retrieved for each case:

  1. Time of 999 call and time at which ambulance arrived on the scene.

  2. Initial cardiac rhythm and treatment measures.

  3. Time of ROSC (estimated if not recorded).

  4. Time of arrival in the ED.

  5. Time and place of the decision to initiate (or not to initiate) TH (data collected from ED, ICU notes or observation charts).

  6. Method of cooling (external or intravascular).

  7. Temperature on arrival on the ICU.

  8. Time at which a temperature of 34°C was achieved.

  9. The maximum temperature recorded once cooling was initiated.

  10. Outcome—whether or not patient survived hospital admission.

From these data the following time intervals were calculated:

  1. The times between ROSC, arrival in ED and initiation of cooling.

  2. The time between the initiation of cooling and reaching 34°C.

In some cases the precise time of ROSC was not documented by ambulance staff. For these cases, ROSC time was estimated based on the time of arrival at the ED, the time at which the ambulance staff arrived on the scene and the geographical location of the scene of the arrest.

Inclusion criteria

Out-of-hospital cardiac arrest with ROSC, Glasgow Coma Scale score <9 in the ED.

Exclusion criteria

Cardiogenic shock, multiple organ failure, severe co-morbidities, pregnancy, sepsis, pre-existing medical coagulopathy, a valid do-not-attempt resuscitation order.


The appendix outlines the therapeutic hypothermia protocol used at the RUH. All mechanically ventilated patients who were transferred to the ICU following OHCA were sedated with propofol and alfentanil. Cooling was initiated either in the ED or on the ICU. The mode of cooling in the ED was primarily with ice packs in the groin, axillae and neck and 2 l of intravenous ice cold Hartmann's solution (4°C) over 30 min. Cooling in the ICU was achieved with an endovascular device, which consisted of a closed system of a femoral central venous catheter (Icy, Alsius Corp., Irvine, California, USA) placed in the inferior vena cava and an external heat exchange system (CoolGard 3000). The system contains circulating saline within the catheter (0–42°C) and was adjusted to maintain a target temperature of 33°C for 24 h. The temperature was measured with a tympanic probe in the ED and with a bladder probe on the ICU. Rewarming was at a rate of 0.25°C per hour and then maintained at 36.5°C using the endovascular device. If necessary, shivering was controlled by giving bolus doses of atracurium. Plasma potassium concentration was checked 1–2 hourly and maintained at >4 mmol/l; magnesium was checked 6 hourly and maintained at 1.0–1.5 mmol/l. Blood glucose values were maintained strictly between 4 and 7 mmol/l using an intensive insulin therapy protocol, which has been published previously.10


The neurological status of hospital survivors was recorded using the Cerebral Performance Category (CPC) scale (table 1) based on the patient's best neurological status documented in the medical records before hospital discharge.11

Table 1

Cerebral Performance Category (CPC) scale


Eighty-three OHCA patients were admitted to the ICU for TH over the 7-year period of study (1 January 2002 to 31 December 2008).

Sixty-seven (81%) patients were cooled according to ILCOR recommendations and departmental protocol. The 16 (19%) patients who were not cooled met the recognised exclusion criteria (table 2). The time to ROSC was estimated in 25% patients (±5–10 min).

Table 2

Number of patients not cooled and reasons for exclusion

Of 67 patients who were cooled, 32 (48%) were discharged from the ICU and had good neurological recovery (CPC 1). Of the 16 patients who were not cooled, 13 (81%) died and 3 (19%) survived with CPC 1.

Of the 67 patients who were cooled, 44 were cooled within 240 min of ROSC; 23 (52%) patients survived in this group. Twenty-three patients were cooled at least 240 min after ROSC; 9 (39%) patients survived in this group (p=0.7).

In 29 (43%) patients the temperature increased after arrival on the ICU before decreasing again. Thirteen (44%) patients survived in this group, all with good neurological recovery (CPC 1).

Table 3 shows the survival rate by age. Table 4 shows the survival rates by presenting rhythm.

Table 3

Survival in different age groups

Table 4

Presenting cardiac rhythm and survival in each group

The median time (IQR) from ROSC to initiation of cooling was 60 (40–165) minutes (figure 1) and the median time to reach 34°C was 175 (40–420) minutes (figure 2). The time taken to reach the target temperature reduced year by year during the study period (figure 3). During the study period, the main location for initiating cooled shifted from the ICU to the ED (figure 4).

Figure 1

Time to initiate cooling from return of spontaneous circulation (ROSC).

Figure 2

Time to reach target temperature.

Figure 3

Median time from return of spontaneous circulation (ROSC) to target temperature by year. Ranges represent IQRs.

Figure 4

Place where cooling was initiated year on year. ED, emergency department; ITU, intensive care unit.


Therapeutic hypothermia has been successfully implemented at the RUH: the therapy was initiated in all patients who met the ILCOR criteria. TH training was included in the educational programmes of ED doctors and nurses. The median time taken to achieve a temperature of 34°C decreased from 420 min in 2002 to 50 min in 2008. Nagao et al and Wolff et al have shown that earlier cooling is associated with improved neurological outcome after OHCA.7 8 12 13 The dramatic reduction in the median time to achieve 34°C has been achieved by shifting the initiation of TH from the ICU to the ED. This change in strategy was implemented following collaboration between ICU and ED consultants and nurses and through an education programme for all ED staff. A recent qualitative study has identified many potential barriers to the implementation of TH in the ED; these include lack of a TH protocol, lack of interdisciplinary collaboration between the ICU and ED, and variable nursing awareness.14 Our study has shown that effective collaboration between the ICU and ED staff can achieve full implementation of a TH protocol. Although we have shown that the time to achieve target temperature has been reducing, there are still some cases in which there are significant delays. These delays can be reduced by initiating cooling earlier and by using more efficient cooling methods.

Ultimately, the most effective way of shortening the time to target temperature is to start cooling prehospital as soon as ROSC is achieved15–18 or even during CPR before ROSC is achieved.8 19 20

The use of intravenous cold saline to initiate cooling is well established21–24 and is a very inexpensive and efficient method for reducing the core temperature rapidly by about 1.5°C. In our study, there was no evidence of harm from the infusion of 2 l of ice cold fluid, and the safety of this technique is consistent with the observations of several investigators who have studied this cooling technique.21–24 Ice cold fluids alone cannot be used to maintain hypothermia,25 but even the addition of simple ice packs may control the temperature adequately.26 Simple external cooling systems can easily be applied in the ED, and on the ICU cooling can be maintained by either external or internal (intravascular) systems. At present, there are no data to show whether the type of cooling system used influences survival, but tighter temperature control is achieved with the internal systems.27 28

A weakness of our study is that different sites and techniques were used to monitor temperature in the ED and in the ICU. In the ED, intermittent tympanic temperature monitoring was used, while in the ICU bladder temperature was monitored continuously from a thermister located on the urethral catheter. The tympanic site is known to be unreliable. Bladder temperature can be monitored continuously although there is about a 20 min lag behind true core temperature.5 We have since implemented bladder temperature monitoring in the ED. In 29 (43%) of our patients the temperature initially increased before decreasing again as cooling continued. This may have reflected a genuine increase caused by the delay in inserting the intravascular cooling catheter in the ICU and/or it may simply reflect the different site used for temperature monitoring. A genuine increase in temperature at this stage may be harmful to injured neurological tissue.29 In this study there was no evidence that the outcome in these patients was any worse than those who did not have an early temperature increase, but our sample size is far too small for us to draw firm conclusions. We have subsequently implemented an educational programme to reinforce the importance of continuing cooling with simple external techniques while waiting for the intravascular system to be started.


Following close collaboration and teamwork between ICU and ED personnel, we have demonstrated very successful implementation of TH in the ED of a general hospital. The temperature should be measured continuously from the same site in both the ED and the ICU; this should enable prompt intervention to prevent temperature increases.


Initiation of therapeutic hypothermia (checklist for emergency department)

  1. Patient should satisfy the inclusion criteria: (cardiac arrest (in and out of hospital), loss of consciousness, subsequent return of spontaneous circulation). Exclusion criteria: cardiogenic shock, multiple organ failure, severe co morbidities, pregnancy, sepsis, pre-existing medical coagulopathy, a valid do-not-attempt resuscitation order.

  2. Keep defibrillator pads on patient, as ventricular arrhythmia can occur during induction of hypothermia.

  3. Thrombolyse if indicated. If there is doubt about the cause of the collapse, the patient should have a CT head on the way to the intensive care unit (ICU) and be thrombolysed immediately after the scan if it is normal.

  4. Sedate with propofol infusion with or without alfentanil.

  5. Take baseline bloods including full blood count, urea and electrolytes, liver function tests, magnesium and glucose (aim: K >4–5 mmol/l, Mg 1–1.5 mmol/l).

  6. Take a baseline temperature.

  7. Method of cooling: infusion of 2 l (or 30 ml/kg) of ice cold (4°C) Hartmann's solution over 30 min. Give through blood giving set under pressure if necessary. It is quick and avoids fluids getting warmed up. Ice packs can be placed in the groin, axillae and around the neck; be careful for frostbite (target temperature 32–34°C).

  8. Shivering can be reduced by giving a muscle relaxant (atracurium); give stat doses only and avoid infusions as it may mask seizure activity. The role of magnesium infusion is still under review.

  9. Insert a urinary catheter with a bladder temperature probe as it is more accurate than ear probe temperature. It is also a single procedure, and the chances of dislodging, unlike with an oesophageal temperature probe, are much less.

  10. Start insulin using the insulin protocol for any patient with blood glucose >7 mmol/l.

  11. Transfer to the ICU.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Primary survey
    Jonathan Wyatt