American Journal of Emergency Medicine (2012) 30, 1838-1844

Original Contribution

Duration of well-controlled core temperature correlates with neurological outcome in patients with Post-cardiac arrest syndrome

Koichiro Shinozaki PhD, MD ^a,?, Shigeto Oda PhD, MD ^a, Tomohito Sadahiro PhD, MD ^a, Masataka Nakamura MD ^a, Yo Hirayama PhD, MD ^a, Eizo Watanabe PhD, MD ^a, Yoshihisa Tateishi PhD, MD ^a, Kasuya Nakanishi PhD, MD ^b,

Nobuya Kitamura PhD, MD ^c, Hiroyuki Hirasawa PhD, MD ^a

^aDepartment of Emergency and Critical Care Medicine, Graduate School of Medicine, Chiba University, 1-8-1 Inohana,

Chuo-ku, Chiba City, Chiba, 260-8677, Japan

^bDepartment of Emergency and Critical Care Medicine, Narita Red Cross Hospital, Japan

^cDepartment of Emergency and Critical Care Medicine, Kimitsu Chuo Hospital, Japan

Received 7 March 2012; revised 22 March 2012; accepted 22 March 2012

Abstract

Purpose: Detailed procedures for optimal Therapeutic hypothermia have yet to be established. We examined how duration of well-controlled core temperature within the first 24 hours after cardiac arrests (CA) correlated with neurological outcomes of successfully resuscitated out-of-hospital CA (OHCA) patients. Methods: OHCA patients who survived over 24 hours and treated with TH were included. Core temperature was measured every hour. Physicians intended to maintain temperature at 33?C +- 1?C for 24 hours. Cerebral Performance Categories (CPC) of patients at 6 months were recorded and patients were retrospectively divided into favorable (CPC1,2) and poor (CPC3-5) neurological outcome groups. Total time while the core temperature reached to 33?C +- 1?C within the first 24 hours after CA was measured and this duration was defined that of well-controlled temperature. receiver-operating characteristic analysis was performed on duration of well-controlled temperature to select the optimal cutoff value. Neurological outcome predictors were investigated by logistic regression analysis.

Results: Fifty-six patients were included. Optimal cutoff value of duration of well-controlled temperature was

18 hours. Ratio of male sex, witnessed by emergency medical service (EMS) personnel, first electrocardiogram as shockable, and duration of well-controlled core temperature >=18 h of favorable neurological outcome group (n = 21) were significantly larger than that of poor neurological outcome group (n = 35). Logistic regression analysis identified “witnessed by EMS”, “performed bystander CPR,” and “the duration >=18 h” as independent predictors of favorable neurological outcome. Conclusion: TH maintained at target temperature of 33?C +- 1?C over 18 hours independently correlated with favorable neurological outcome. Therefore, stable core Temperature control may improve neurological outcome of successfully resuscitated OHCA.

(C) 2012

* Corresponding author. Tel.: +81 43 226 2341; fax: +81 43 226 2371.

E-mail address: [email protected] (K. Shinozaki).

0735-6757/$ - see front matter (C) 2012 http://dx.doi.org/10.1016/j.ajem.2012.03.024

Introduction

Approximately 160 000 cases of emergency medical services (EMS)-treated sudden cardiac arrests (CA) are estimated to occur in the United States annually [1]. Regional variation is reported in survival outcome of EMS-treated CA: patients with “any return of spontaneous circulation (ROSC)” account for approximately 30% on average of those in whom resuscitation is attempted [2], with the percentage of patients “discharged alive” ranging from 3.0% to 16.3% [2].

An advisory statement by the advanced life support task force of the International Liaison Committee on Resuscita- tion [3] in 2003 suggested that cooling to 32?C to 34?C should be initiated as soon as possible after ROSC but appears to be successful even if delayed (eg, 4-6 hours). However, in 2005 American Heart Association (AHA) Guidelines, it was simply suggested that Therapeutic hypothermia of 32?C to 34?C should be initiated in unconscious adult patients with ROSC when the initial rhythm was ventricular fibrillation (class IIa) [4]. In 2010 AHA Guidelines [5], the classification of recommendation for TH was up to Class I. However the detailed procedures were not defined.

From the results of an observational study, Haugk et al [6] demonstrated that in CA patients treated with TH after ROSC, a faster decline in body temperature to the 34?C target appears to predict an unfavorable neurological outcome. In consideration of this controversial subject, although clinical efficacy of TH has been demonstrated in a number of studies [7,8], detailed procedures for effective TH (eg, optimal induction technique, target temperature, dura- tion and Rewarming rate) have yet to be established [3,9]. Few extensive investigations of detailed procedures for TH and the resulting outcome of treated patients have been documented [9]. In particular, although there are some reports [6,10,11] which investigated on the early induction of TH, there are few investigations which indicate optimal duration of TH within the first 24 hours after CA.

We hypothesized that differences in duration of well-

controlled body temperature within targeted temperature induced by TH determine neurological outcomes in successfully resuscitated CA patients. Therefore, we retro- spectively investigated how duration of well-controlled core temperature within the first 24 hours after CA correlated with neurological outcomes of successfully resuscitated out-of- hospital CA (OHCA) patients.

Methods

Study design and patients

We conducted a retrospective Multicenter observational study between May 2006 and March 2009 at 3 medical

institutions located in Chiba Prefecture, Japan. One of them (Chiba University Hospital) is located in a midsized urban region with a population of 940 000 and has 6 EMS systems. The other 2 institutions are located in two different medical districts: one with a population of 330,000 and 4 EMS systems consisting of four suburban and rural regions (Kimitsu Chuo Hospital), and the other with a population of 860 000 and 8 EMS systems consisting of 17 suburban and rural regions (Narita Red Cross Hospital).

All patients aged 18 years or older who presented with OHCA were eligible for inclusion. Patients were excluded if the cause of CA was trauma, if they died within 24 hours after CA, or if TH was not indicated.

In the study period, TH was performed in all patients except those meeting any of the following exclusion criteria: response to verbal command on arrival at hospital, absence of independent daily life prior to the event, CA caused by stroke or hanging, or age over 80 years with an initial rhythm of asystole or pulseless electrical activity.

Interventions

Basic life support and advanced cardiovascular life support were provided according to the 2005 AHA Guidelines for Cardiopulmonary Resuscitation and Emer- gency Cardiovascular Care [12]. All patients included in the present study were admitted to the intensive care unit and received standard intensive care including invasive monitor- ing, hemodynamic support, mechanical ventilation, analge- sia and sedation. Five to 10 mg/h of midazolam and 0.02 to

0.03 mg/h of buprenorphine were used as analgesia and sedation. After the attending physician decided to perform TH, every possible effort was made to adjust the core temperature of the patient as soon as possible and maintain it at 33?C +- 1?C for 24 hours using a cooling device. The device was chosen by the attending physician: Surface cooling (using a blanket, Meditherm; IMI Co, Ltd, Tokyo, Japan), core cooling (using extracorporeal circulation; KTEK, Kawasumi Laboratories, Inc, Tokyo, Japan), or both. All patients underwent intravenous infusion of ice-cold ftuids (1500-2000 mL of 0.9% saline) on admission to the ED. To prevent shivering, vecuronium bromide was administered to all patients while TH was performed.

Data collection and processing

Patients’ characteristics were extracted from EMS reports and from medical records of each hospital. CA was defined as the cessation of cardiac mechanical activity, and was confirmed by the absence of a detectable pulse, unrespon- siveness, and apnea (or agonal, gasping respiration) [13]. A witnessed CA was defined as a CA in which the patient’s collapse was seen or heard by a bystander or emergency personnel. Basic life support for CA witnessed by emergency personnel was not defined as bystander cardiopulmonary

resuscitation (CPR) [13]. Shockable rhythms were defined as electrocardiograms showing ventricular fibrillation or pulse- less ventricular tachycardia, while Non-shockable rhythms were defined as those showing asystole or pulseless electrical activity. The time of onset of CA was defined as witnessed collapse time in cases of witnessed CA or as emergency call receipt time in cases of non-witnessed CA.

Core temperature was measured in all patients every hour, with the probe inserted in the rectum or bladder. Total time while the core temperature was reached to 33?C +- 1?C within the first 24 hours after CA was measured.

Outcome measures

The cerebral performances of individual patients were evaluated using Glasgow-Pittsburgh cerebral performance categories (CPC) as recommended by the Utstein templates [13]. Individual CPC categories are defined as follows: CPC1, “conscious and alert with normal neurological function or only slight cerebral disability”; CPC2, “con- scious with moderate cerebral disability but capable of part- time work in a sheltered environment or independent existence”; CPC3, “conscious with severe cerebral disability precluding independent existence”; CPC4, “comatose or in a Persistent vegetative state“; and CPC5, “brain dead”. The CPCs for those discharged were determined by follow-up telephone interview of the patients themselves, their family members, or personnel of the patients’ rehabilitation facilities. In the present study, all patients were followed up to 6 months after CA. We recorded the CPC at the end of the follow-up period. If the patient had died before 6 months follow-up, their CPC was defined as CPC5. The patients were then classified into two groups on the basis of ethical aspects“>neurological outcome: a “neurologically favorable outcome group” (CPC1 and CPC2) and a “neurologically poor outcome group” (CPC3 to CPC5).

In this study, the subjects were divided into 2 groups by

neurological outcomes. The duration of well-controlled core temperature was compared between these two groups and factors significantly correlated with neurological outcome were investigated by comparing baseline characteristics and the duration of well-controlled core temperature.

Blinding

Patients’ characteristics and outcomes were recorded from EMS persons and attending physicians blinded from the study hypothesis, and the investigators of data collection were also blinded to study hypothesis.

Statistical analysis

We reported continuous variables as medians and 25% to 75% interquartile ranges because they were not generally normally distributed. The Mann-Whitney U test was used to

test the null hypothesis of no difference. The sensitivity and specificity of prediction of neurological outcomes were calculated for different cutoff values of the duration of well- controlled temperature, and the results are depicted as a receiver-operating characteristics (ROC) curve with an area under the curve (AUC). The position of the “left upper corner” of the curve indicates the optimal cutoff value. Baseline variables were compared using ?2 test or Fisher’s exact test. Where the expected value in a 2×2 contingency table was less than five, the Fisher’s exact test was used. ?2 test was performed for age, the sole continuous variable, after categorization using 65 years as the cutoff value.

After the classification of the subjects with their neurological outcomes, the detailed characteristics were compared between the two groups. Furthermore, for the purpose of investigation about factors which were thought to significantly correlate with neurological outcome, patient baseline characteristics and the status of core temperature control were compared between the two different outcome groups using a multivariate analysis.

To select the best combination of predictors, the data set for the patients finally included in the study was analyzed using a stepwise selection procedure in a multivariate logistic regression model. The stepwise procedure used a threshold of 0.05 for inclusion in and 0.15 for exclusion from the final model considering all possible combinations of main effects and interactions, and a model maximizing the area under the ROC (AUC) [14] was chosen as the final model. The goodness of fit of the final model was examined using the Hosmer-Lemeshow test [15].

Two-tailed P values of less than .05 were considered significant. All statistical analyses were carried out using SPSS 13.0 package for Windows (SPSS, Tokyo, Japan, Inc).

Ethical aspects

The study was approved by the ethical review boards of all participating hospitals. The procedures described above were either performed as a part of routine management of patients or involved minimal risk to the patients. Each institutional review board therefore waived the need for informed consent from the patient.

Results

All 1026 nontraumatic OHCA patients received basic life support in pre-hospital settings and arrived at one of the three participating hospitals. Of these 1026 patients, 326 (31.8%) gained any ROSC. However, 222 of 326 resuscitated patients died within the first 24 hours. Of the remaining 104 (10.1%) patients that survived more than 24 hours, 48 did not receive TH based on the exclusion criteria. Therefore, 56 patients (41 men and 15 women; median and inter-quartile range of age, 62 [54-66] years) were the subjects of this study (Fig. 1).

Received TH

Included in analysis

n = 56

n = 1,026

Excluded: n = 700 Died without ROSC

n = 326

Excluded: n = 222

Died within 24 hours after CA

Excluded: n = 48

Met exclusion criteria for performing TH

response to verbal command on arrival at hospital, absence of independent daily life prior to the event, CA caused by stroke or hanging, or age over 80 years with an initial rhythm of asystole or PEA

n = 104

Survive more than 24 hours after CA

Individuals gained any ROSC

Individuals aged >= 18 years with non-traumatic OHCA

Fig. 1 Study population profile.

Main results

The box plots of duration (total time counts) of well- controlled core temperature are shown in Fig. 2. With regard to the duration of well-controlled core temperature, there is significant difference between favorable (n = 21) and poor (n = 35) neurological outcome groups (P = .03). The result of ROC analysis to investigate the optimal cutoff value for prediction of favorable neurological outcome is shown in Fig. 3. From this ROC analysis, 18 hours that was pointed out as left upper the corner of the curve was considered the best cutoff values.

Fifty-six patients were divided by this cutoff point. Twenty eight patients were categorized as “longer than 18 hours,” and the other 28 patients were categorized as “shorter than 17 hours.” The number of patients categorized as Favorable neurological outcomes in “longer than 18 hours” group is 16 (57%), and that in “shorter than 17 hours” group is 5 (18%), respectively. The difference was statistically significant (P = .002), and the odds ratio for poor neurological outcome was 0.16 (CI, 0.05-0.55).

Changes of core temperatures within the first 24 hours

after CA in the patients categorized as “longer than 18 hours”

(hrs) 24

18

Total ti e counts

12

6

0

P = .03

Favorable Neurological Poor Neurological Outcome Group (n = 21) Outcome Group (n = 35)

90% Percentile

75% Percentile Median

25% Percentile

10% Percentile

are shown in Fig. 4A. Those in the patients categorized as

“shorter than 17 hours” are shown in Fig. 4B. The number of

Fig. 2 Duration of well-controlled core temperature compared between favorable and poor neurological outcome groups.

1.00

18 hrs

.75

.50

sensitivity

.25

0

.25

.50

1-specificity

.75

1.00

cooled, was 8 in 28 patients who were categorized as “shorter than 17 hours.”

Twenty-eight patients categorized “longer than 18 hours” were distributed to 17, 5, and 6 of each institution, and 28 patients categorized “shorter than 17 hours” were distributed to 14, 11, and 3 of each institution, respectively. The control of TH was not significantly different between institutions.

The comparisons of baseline characteristics and the duration of well-controlled core temperature between “favorable” (n = 21) and “poor” (n = 35) neurological outcome groups are shown in Table 1. Statistically significant differences were found for the following vari- ables: gender, witnessed by EMS personnel, first ECG as shockable, duration of well-controlled core temperature

>=18 hours. There is no significant difference for the methods of cooling.

All variables included the categorized variable were inserted into the logistic regression analysis. Logistic regression analysis identified “not witnessed by EMS personnel”, “not performed bystander CPR”, and “duration

AUC 0.644 [95% CI 0.491-0.798]

Fig. 3 ROC curve of duration of well-controlled core tempera- ture for prediction of favorable neurological outcome.

the patients who had no time while core temperature was reached to 33?C +- 1?C, even though who were tried to be

of well-controlled core temperature b18 hours” as indepen- dent predictors of poor neurological outcome (Table 2). The remaining variables tested (age, gender, Cardiac origin, witnessed by bystander, first ECG as shockable, using surface cooling device and using core cooling device) either had no effect or lost independent Predictive capacity in the multivariate model.

(oC) A

38

37

36

35

34

33

32

0 6 12 18 24 (hours)

3

(oC) B

8

7

6

5

4

3

2

3

3

3

3

3

3

0 6 12

18 24 (hours)

Fig. 4 Changes of core temperatures within the first 24 hours after CA compared between the two groups categorized by the duration of well-controlled core temperature. The patients categorized as “longer than 18 hours” are shown in Fig. 4A. Those categorized as “shorter than 17 hours” are shown in Fig. 4B.

Total number, 56	Favorable (n = 21)	Poor (n = 35)		P	OR (95%CI)
Age, y	64 (48-74)	60	(54-65)	.610	0.75 (0.24-2.31)
Male gender, n (%)	19 (91)	22	(63)	.024	0.18 (0.04-0.89)
Cardiac origin, n (%)	16 (76)	25	(71)	.697	0.78 (0.23-2.71)
Witnessed by EMS personnel, n (%)	6 (29)	3	(9)	.066	0.23 (0.05-1.07)
Witnessed by bystander, n (%)	12 (57)	22	(63)	.672	1.27 (0.42-3.83)
Bystander CPR, n (%)	8 (38)	8	(23)	.240	0.48 (0.15-1.57)
First ECG as shockable, n (%)	16 (76)	15	(43)	.015	0.23 (0.07-0.78)
Duration of well-controlled core temperature >=18 h, n (%)	16 (76)	12	(34)	.002	0.16 (0.05-0.55)
Using surface cooling device, n (%)	17 (81)	26	(74)	.747	0.68 (0.18-2.56)
Using core cooling device, n (%)	9 (43)	12	(34)	.521	0.70 (0.23-2.11)
Values are numbers, percentages, median and inter-quartile range. Odds ratios as predictors of “poor neurological outcome” are shown. Baseline variables were compared using ?2 test or Fisher exact test. ECG indicates electrocardiogram; OR, odds ratio; CI, confidence interval.

Discussion

Table 1 Baseline characteristics-favorable vs. poor neurological outcome group

This observational study examined whether the duration of well-controlled core temperature inftuenced the neuro- logical outcome of successfully resuscitated OHCA patients treated with TH. Logistic regression analysis indicated that TH maintained at target temperature of 33?C +- 1?C over 18 hours independently correlated with favorable neurological outcome. Interestingly, no correlation was observed between the neurological outcomes and the particular cooling methods used for TH.

Few extensive investigations on detailed procedures for TH and the resulting outcome of treated patients have been documented [9]. Infusion of ice-cold 0.9% saline or Ringer’s lactate is recommended to maximally increase the rate of fall in core temperature to the target value [9]. However, two randomized controlled trials investigating the clinical efficacy of infusion of ice-cold water in pre-hospital settings failed to demonstrate benefits of this intervention [10,11].

An advisory statement by the advanced life support task force of the International Liaison Committee on Resuscita- tion [3] in 2003 suggested that cooling should probably be

Table 2 Results of logistic regression analysis for prediction of poor neurological outcome

Variables Adjusted OR 95%CI P AUC

well-controlled core temperature

>=18 h

Results of the Hosmer-Lemeshow test indicated that the p value of this Model was 0.950.

Witnessed by	0.090	0.013-0.637	.016	0.82
EMS
personnel
Bystander CPR	0.165	0.034-0.811	.027
Duration of	0.101	0.023-0.448	.003

initiated as soon as possible after ROSC but appears to be successful even if delayed (eg, 4-6 hours). The present study did not show the salutary effect of early attainment but suggested that the stable control of core temperature over a longer period during the first 24 hours after CA is important for improvement of neurological outcome.

We believe that cooling should be initiated as soon as possible after ROSC at least within 6 hours after CA, because early induction might regulate the early inftammatory response caused by systematic ischemia/reperfusion. How- ever, not only earlier cooling but also stable control after achievement of target temperature at least for the first 24 hours after CA may be important. It is considered that the inftammatory response and the systemic and brain insult caused by this response might be suppressed by the stable and continuous control of core temperature. We encourage that the core temperature of successfully resuscitated CA patients should be aggressively reduced and carefully maintained at the target temperature.

Previous papers raised the question whether core or surface cooling techniques differ in effectiveness, in terms of time spent at target temperature, induction of cooling, and maintenance of core temperature? And also it has not yet been clearly proved whether the choice of cooling technique affect incidence of complications or outcome. Gillies et al

[16] investigated the difference of endovascular versus surface cooling following CA in the retrospective study. After adjustment for confounders, they could not demon- strate a difference in outcome between the groups. However, they advocated that endovascular cooling appears more efficacious than surface cooling in terms of temperature control for TH. Tomte et al [17] also indicated that surface and core cooling of OHCA patients following the same established post-resuscitation treatment protocol resulted in similar survival to hospital discharge and comparable neurological function. From our results, there was no difference in the choice of cooling technique compared between favorable and poor neurological outcome groups.

Therefore, we also can not demonstrate what method is better for TH and outcome of the patient.

Limitations

Since this study was observational, it has the limitation that potential confounders not actually examined (ie, more detailed Utstein templates parameters describing resuscita- tion and post-cardiac arrest care such as total dose of epinephrine, duration of CA, and values of blood pressure in the intensive care unit) could have inftuenced its results.

This study included 8 patients who had no time while core temperature was reached to 33?C +- 1?C, even though who were tried to be cooled. The reasons were various; bradycardia, spending long time on coronary angiography and Percutaneous coronary interventions and mainly techni- cal problems. It is difficult to clarify all the reasons that explain for the failure to maintain hypothermia. Therefore, these factors might be potential confounders.

The other limitation is the small sample size. The remaining variables that lost independent predictive capacity in the multivariate model of this study may have had the power to determine the outcome, if the sample size would be large enough.

The present findings emphasize the importance of the strict control of core temperature during TH in improvement of neurological outcome. We will perform the other prospective study to verify generalization of the results of this study. Future investigations that will be clearly randomized and have a large sample size are needed to confirm our findings.

Conclusions

TH maintained at target temperature of 33?C +- 1?C over 18 hours independently correlated with favorable neurolo- gical outcome. Therefore, present study suggests that if the core temperature was kept within target range longer than 18 hours within the first 24 hours, neurological outcome of successfully resuscitated CA patients might be improved.

References

1. Rea TD, Eisenberg MS, Sinibaldi G, et al. Incidence of EMS-treated out-of-hospital cardiac arrest in the United States. Resuscitation 2004;63:17-24.
2. Nichol G, Thomas E, Callaway CW, et al. Regional variation in out-of- hospital cardiac arrest incidence and outcome. JAMA 2008;300: 1423-31.
3. Nolan JP, Morley PT, Vanden Hoek TL, et al. Therapeutic hypothermia after cardiac arrest: an advisory statement by the advanced life support task force of the International Liaison Committee on Resuscitation. Circulation 2003;108:118-21.
4. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 7.5: Post- resuscitation Support. Circulation 2005;112:IV84-8.
5. Peberdy MA, Callaway CW, Neumar RW, et al. Part 9: post-cardiac arrest care: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S768-86.
6. Haugk M, Testori C, Sterz F, et al. Relationship between time to target temperature and outcome in patients treated with therapeutic hypothermia after cardiac arrest. Crit Care 2011;15:R101.
7. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557-63.
8. The Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549-56.
9. Nolan JP, Neumar RW, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication: a Scientific Statement from the International Liaison Committee on Resuscitation; the American Heart Association Emergency Cardio- vascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; the Council on Stroke. Resuscitation 2008;79:350-79.
10. Kim F, Olsufka M, Longstreth Jr WT, et al. Pilot randomized clinical trial of prehospital induction of Mild hypothermia in out-of-hospital cardiac arrest patients with a rapid infusion of 4 degrees C normal saline. Circulation 2007;115:3064-70.
11. Bernard SA, Smith K, Cameron P, et al. Induction of therapeutic hypothermia by paramedics after resuscitation from out-of-hospital ventricular fibrillation cardiac arrest: a randomized controlled trial. Circulation 2010;122:737-42.
12. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2005;112:IV1-203.
13. Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein style. A statement for health professionals from a task force of the American Heart Association, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, and the Australian Resuscitation Council. Circulation 1991;84: 960-75.
14. Schwarz G. Estimating the dimension of a model. Ann Stat 1978;6: 461-4.
15. Hosmer DW, Hosmer T, Le Cessie S, et al. A comparison of goodness- of-Fit tests for the logistic regression model. Stat Med 1997;16:965-80.
16. Gillies MA, Pratt R, Whiteley C, et al. Therapeutic hypothermia after cardiac arrest: a retrospective comparison of surface and endovascular cooling techniques. Resuscitation 2010;81:1117-22.
17. Tomte O, Draegni T, Mangschau A, et al. A comparison of intravascular and surface cooling techniques in comatose cardiac arrest survivors. Crit Care Med 2011;39:443-9.