Resuscitation 89 (2015) 129136

Contents lists available at ScienceDirect

Resuscitation

j ourna l h o me pa g e : www.elsev ier .com/ locate / resusc i ta t ion

Clinical Paper

The effect of mild therapeutic hypothermia on gorecovery after out-of-hospital cardiac arrest accoreturn of spontaneous circulation: A nationwide

Kwang S JunYu Jin Lea Department ob Laboratory of nstitu

a r t i c l e i n f o

Article history:Received 26 October 2014Received in revised form 15 January 2015Accepted 20 Ja

Keywords:Cardiac arrestHypothermiaReturn of sponOutcomes

a b s t r a c t

Background: Mild therapeutic hypothermia (MTH) has been known to be associated with good neurolog-ical recovery after out-of-hospital cardiac arrest (OHCA). Prehospital return of spontaneous circulation(P-ROSC) is associated with better hospital outcomes than ROSC at emergency department (ED-ROSC).

1. Introdu

Out of hwide, serioumedical ser

A Spanish in the nal on

CorresponUniversity ColKorea.

E-mail addshinsangdo@mskciva@gmail.(Y.J. Lee), suhg

http://dx.doi.o0300-9572/ nuary 2015

taneous circulation

The study aims to examine the association between MTH by location of ROSC and good neurologicalrecovery after OHCA.Methods: Adult OHCA cases with presumed cardiac etiology who survived to hospital admission werecollected from a nationwide cardiac registry between 2008 and 2013. MTH was dened as a case receiv-ing hypothermia procedure regardless of procedure method. Primary outcome was good neurologicalrecovery with cerebral performance category score of 1 and 2. Multivariable logistic regression analysiswas performed adjusting for potential confounders with an interaction term between MTH and locationof ROSC to calculate adjusted odds ratios (AORs) and 95% condence intervals (CIs).Results: Among 11,158 patients survived to admission, good neurological recovery was 23.6% (399/1691)in MTH vs. 15.0% (1400/9316) in non-MTH (p < 0.001), and 58.2% (1074/1864) in P-ROSC vs. 7.9%(725/9161) in ED-ROSC (p < 0.001). There was a signicant association between MTH and good neu-rological recovery (AOR = 1.32, 95% CI = 1.111.57). In the interaction model, AOR of MTH and interactioneffect with P-ROSC and ED-ROSC was 0.78 (0.582.70) and 1.68 (1.341.98), respectively.Conclusion: MTH was signicantly associated with good neurological recovery among OHCA survivors. Inthe interaction model, MTH showed signicant benets in patient group with ROSC at ED, not in P-ROSCgroup.

2015 Elsevier Ireland Ltd. All rights reserved.

ction

ospital cardiac arrest (OHCA) is recognized as a world-s public health problem. The incidence of emergencyvices (EMS) treated OHCA was reported to be between

translated version of the abstract of this article appears as Appendixline version at http://dx.doi.org/10.1016/j.resuscitation.2015.01.024.ding author at: Department of Emergency Medicine, Seoul Nationallege of Medicine, 101 Daehak-Ro, Jongno-Gu, Seoul 110-744, South

resses: zorbathekorean@gmail.com (K.S. Bae),edimail.co.kr (S.D. Shin), Ro.youngsun@gmail.com (Y.S. Ro),

com (K.J. Song), ironlyj@gmail.com (E.J. Lee), eyeblack99@gmail.comil@snu.ac.kr (G.J. Suh), yhkwak@snuh.org (Y.H. Kwak).

50 and 130 cases per 100,000.1 In the US, only 9.8% of EMS-treatedadult OHCAs survived to discharge from hospital.2 In Korea, wehave also reported that the incidence of EMS-assessed OHCA wasapproximately 45 per 100,000, and 3.5% of those patients survivedto discharge from hospital.3

Previous studies have reported that factors such as wit-nessed arrest, initial shockable rhythm, bystander CPR, shorterEMS response time, use of public access debrillator (PAD) pro-gram play a key role in survival to hospital discharge afterOHCA.46 Hospital interventions (i.e., post cardiac arrest care)comprised of targeted temperature management, optimal venti-lation/oxygenation, optimizing hemodynamic parameter, seizurecontrol, and electrolyte/glucose control have been developed toimprove neurological outcome and survival of patients after cardiacarrest.7

rg/10.1016/j.resuscitation.2015.01.0242015 Elsevier Ireland Ltd. All rights reserved.oo Baea, Sang Do Shina,, Young Sun Rob, Kyoung ea, Gil Joon Suha, Young Ho Kwaka

f Emergency Medicine, Seoul National University College of Medicine, South Korea Emergency Medical Services, Seoul National University Hospital Biomedical Research Iod neurologicalrding to location ofobservational study

Songa, Eui Jung Leea,

te, South Korea

130 K.S. Bae et al. / Resuscitation 89 (2015) 129136

Over the past decades, researchers have found that hypothermiacan decrease cerebral oxygen demand8 and block chemical cascadeleading to further cerebral injury.9,10 Mild therapeutic hypothermia(MTH) is the core treatment that has demonstrated neurologicalbenets in cIn 2002, twapy in patiefavorable npatients whas initial rharrest and favorable nrevealed thas initial rhat dischargethe effectivneurologicashockable rof standardsome reseaagement (3good neuro

On the tal ROSC (Poutcomes tHowever, tbetween Mcome. We aresuscitatedat prehospiof MTH ondepending to investigain comatosaccording t

2. Method

This studSeoul Natio

2.1. Data so

The Caris a nationdatabase inthe electroEmergencyinclude chiarrest, bystables of reshospital infplaint was CPR duringKorea Centsecondary dcustomized(QMC) teammedical recepidemiolo

2.2. Study s

Korea hatem operat

department. Ambulance crews can give pre-hospital cares com-parable to that of intermediate emergency medical technician(EMT-I) level in the US. However, advanced life support is onlyavailable in hospitals. The EMTs cannot declare death or stop CPR

eldientsaccet debulanbriluch t valital

reso ED

el of h ancare anc

ittee.n acuols fent

udy s

incluor oldissio

All caied prdiacobviots wiord

udy v

coll witnnedtime(ECGy, anf EDency, locacilityntio).

ain e

H waing ddinavasrd, the min

regahospn orucedlly aomatose survivors with shockable cardiac arrest.7,1113

o randomized trials have shown that hypothermia ther-nts with initial shockable rhythm increased the rates ofeurological outcome One study enrolled adults OHCAo had ventricular brillation or ventricular tachycardiaythm and achieved ROSC within 60 min after cardiacreported MTH was signicantly associated with moreeurological outcome at 6 month.12 The other one alsoat MTH gave OHCA patients with ventricular brillationythm more chances to have good neurological outcome.11 Moreover, even with the controversial debate overeness of hypothermia therapy, MTH still has possiblel benets in comatose adult OHCA patients with non-hythm.1417 As a result, MTH is incorporated as a part

post-cardiac arrest care.7 Furthermore, in recent years,rchers have veried that targeted temperature man-6 C) without signicant variation had similar effect onlogical outcome compared with MTH.18

other hand, previous studies reported that prehospi--ROSC) is associated with more favorable neurologicalhan ROSC at emergency department (ED-ROSC).1921

hese studies did not account for the interaction effectTH and location of ROSC on favorable neurological out-ssumed that ischemic brain damage of OHCA patients

at ED is much greater than that of patients who ROSCtal stage. Therefore, we hypothesized that the effect

neurological outcome could be different in patientson the location of ROSC. The purpose of this study waste the effect of MTH on favorable neurological outcomee adult patients after an out-of-hospital cardiac arresto the location of ROSC.

s

y was approved by the Institutional Review Board of thenal University Hospital in 2013 (No. 1206-007-412).

urce

diovascular Disease Surveillance (CAVAS) databasewide, population-based, EMS-assessed OHCA cohort

Korea.22 The primary data were accumulated fromnic ambulance run sheet database of the National

Management Agency (NEMA). Ambulance run sheetsef complaints, sociological data, location of cardiacander cardiopulmonary resuscitation (CPR), time vari-uscitation efforts, prehospital cares and the destinationormation. Cases were coded as OHCA if the chief com-cardiac or respiratory arrest or if the patient received

transport. Designated medical record reviewers of theers for Disease Control and Prevention (CDC) collectedata for hospital outcomes and related information using

Utstein style. The Data Quality Management Control composed of study co-investigators, statistics experts,ord review experts, re department managers and angist conducted quality control activities.3

etting

s a single-tier, basic life support (BLS) ambulance sys-ed by 16 provincial headquarters of the national re

in theall patPublic so mosby amnal decases sact, buof hospity andLevel 1est levfor 24gency performcommlines oprotocimplem

2.3. St

Weof age tal admtotal). who dnon-caother patienical rec

2.4. St

Wearrest,tan (descene gram therapLevel oEmergcenterical fainterve(CABG

2.5. M

MTmia usice paor intrstandaand thbut weif the duratioing indgeneraKorea. or during transport unless ROSC occurs. Therefore, with OHCA should be transported to the nearest ED.ss debrillator program was not available until 2008,brillation procedures in prehospital stage were givence crews. Every ambulance has an automatic exter-

lator. A withdrawal guideline for use in unsalvageableas evidence of rigor mortis was included in the EMSidity of such decision-making was not studied. LevelsEDs are designated by the government based on capac-urce availability such as facility, stafng, and devices.s (n = 20) and level 2 EDs (n = 110) provide the high-emergency services by emergency medicine physiciansd 7 days. Level 3 EDs (n = 310) provide basic emer-by general physicians. All EDs are subject to mandatorye evaluation every year in keeping with public audit3 Hospital EDs generally accept international guide-te cardiac care and resuscitation. However, no standardor post-resuscitation care at hospitals care have beened.

ubjects

ded all EMS-assessed OHCA patients who are 15 yearser with presumed cardiac cause and survived to hospi-n from January 1, 2008 to December 31, 2013 (6 yearsses were conrmed by medical record review. Patientsrior to hospital admission as well as patients who had

etiology (i.e., trauma, drowning, asphyxia, hanging orus non-cardiac causes) were excluded. Furthermore,th unknown neurological status due to incomplete med-were also excluded.

ariables

ected all potential confounders: sex, age, location ofessed status, bystander CPR, residence in metropoli-

as population over 1 million), response time interval, interval, transport time interval, initial electrocardio-), pre-hospital debrillation, level of ED, reperfusiond location of any ROSC (i.e., P-ROSC or ED-ROSC).

was divided into 3 groups according to the Korean Medical Service Act as regional emergency medicall emergency medical center, and local emergency med-. Reperfusion therapy included emergency coronaryn, thrombolysis, and coronary artery bypass grafting

xposure

s dened as a case receiving mild therapeutic hypother-methods such as external cooling (water, fanning, org), internal cooling (gastric lavage, bladder cooling,cular cooling using a catheter) or mixed cooling. Ine target temperature should be between 32 and 34 Cimum duration of hypothermia should be at least 12 h,rded the case as part of the hypothermia-treated groupital attempted the procedure regardless of the actual

early withdrawal of the procedure due to death dur- hypothermia. During the study period, hospitals hadccepted 3234 C as the targeted temperature range in

K.S. Bae et al. / Resuscitation 89 (2015) 129136 131

2.6. Study outcome

The primary outcome was survival with favorable neurologicaloutcome at discharge after cardiac arrest, dened as having a Cere-bral Performneurologicave point sity, dependscore was don the disccal records.drafted by ireviewed acal record rnational hethere was nparticipatincoded by thsupervised ambiguous

2.7. Analysi

Study pacation of MROSC (P-ROline characcategorical variables. Moutcome winteraction formed to were perforCary, NC). Asignicant.

3. Results

From 13were younghad non-caprior to ad(N = 78), andThe remainincluded ininduced MT(16.6%) wertively (Tabl

GenerallsignicantlyROSC. Patiemale, have and survivaCompared were morewere less liNo signicaED-ROSC gr

In the san associatratio = 1.32,details morinteraction come differfrom arrest

tudy prrest.

ogica 0.78,wereantltio =

cussi

ardleitatiose ation,rhyththe aHCAssocer, Mcom

in E posted

cheme mchemterval from arrest to ROSC was beyond 8 min.24,25 Addition-stori et al.26 reported that when the duration of completetory standstill increased, the positive effects of MTH inpatients increased. These ndings implicate early ROSC asking factor for patients who display diminished effect of

case series studies revealed the shorter time interval frome to ROSC as an important prognostic factor for good neuro-

outcome after cardiac arrest.27,28 In our data, time from callrture from scene was 13 min (median) in P-ROSC and 12 minn) in ED-ROSC group. When considering that many patientsSC group were resuscitated before leaving the scene in ouractice, it is possible to predict that time from arrest to ROSCSC group was much shorter than that of ED-ROSC group.

sumption is supported by other studies.21 The study settingS system follows scoop and run model. The model encour-S providers to deliver shorter CPR at the eld by providingan 34 cycles of CPR with rhythm analysis and to trans-e patient to ED while providing ambulance CPR unless theance Category(CPC) score of 1 (good performance, nol disability) or 2 (moderate disability, can work) on acale; the remaining categories were 3 (severe disabil-ent), 4 (a vegetative state), and 5 (dead).23 The CPCetermined by the medical recorder reviewers based

harge summary abstracts or documentations in medi- Discharge summary abstracts or documentations werenpatient care doctors, and the summaries were usuallynd revised for greater validity by the hospital medi-eview team because the summaries were utilized foralth insurance claim data after discharge. However,o standard recording frame for the CPC score among theg hospitals. Therefore, the scores were determined ande medical record reviewers of the Korea CDC who wereby the project QMC via on- and off-line consultation forcases.

s

tients were divided into 2 groups based on the appli-TH (MTH group and non-MTH group) and location ofSC group and ED-ROSC group). These two types of base-teristics were compared using the chi-square test forvariables and the MannWhitney U test for continuousultivariable logistic regression analyses for the primaryith (interaction model) or without (simple model) theterm between MTH and location of ROSC were per-adjust for potential confounders. Statistical analysesmed with SAS software version 9.3 (SAS Institute Inc.,

2-sided value of p < 0.05 was regarded as statistically

8,410 eligible OHCA patients, we excluded patients whoer than 15 years old at the time of arrest (N = 2814),rdiac causes of arrest (N = 38,675), had been deadmission (N = 85,763), had incomplete medical record

had unknown neurological status at discharge (N = 73).ing 11,007 patients who survived to admission were

the analysis (Fig. 1). Among those, 1691 (15.4%) wereH. In MTH(+) and MTH() group, 298 (17.6%) and 1548e resuscitated from arrest at pre-hospital stage, respec-e 1).y, the two groups divided by the application of MTH had

different baseline characteristics except the location ofnts treated with MTH were more likely to be youngera witness, receive bystander CPR, reperfusion therapyl to discharge compared with non-MTH group (Table 1).with P-ROSC group, Patients resuscitated in ED stage

likely to be older female, less likely to be witnessed,kely to receive bystander CPR and reperfusion therapy.nt difference was there on MTH between P-ROSC andoup (Table 2).imple model without the interaction term, MTH hadion with good neurological outcome (adjusted odds

95% condence interval = 1.111.57) (Table 3) (seee information on full model in Appendix 1). In themodel, the effect of MTH on favorable neurological out-ed by the location of ROSC. If patients were resuscitated

at prehospital stage, the benecial effect of MTH on

Fig. 1. Scardiac a

neurolratio =ever, signicodds ra

4. Dis

Regresusccomatobrillainitial mine after Oicant ahowevcal outbenet

Onebe relaless istal stagthat istime inally, TecirculaOHCA the lacMTH.

Twocollapslogicalto depa(mediain P-ROEMS prin P-ROThis ason EMage EMless thport thopulation. EMS: emergency medical services; OHCA: out-of-hospital

l outcome could not be identied (adjusted odds 95% condence interval = 0.582.70). If patients, how-

resuscitated at emergency department, MTH wasy associated with good neurological outcome (adjusted

1.68, 95%, condence interval = 1.341.98) (Table 4).

on

ss of location of ROSC, the current guideline for post-n care strongly recommends application of MTH todult patients resuscitated from OHCA of ventricular

and weakly recommends to those with non-shockablem.7,17 The primary purpose of this study was to deter-ssociation between MTH and neurological outcome

according to the location of ROSC. We found a signif-iation between MTH and good neurological outcome;TH did not have positive effects on good neurologi-

e in prehospital ROSC group as opposed to signicantD-ROSC group.sible explanation for this result is that P-ROSC canto early ROSC. Early ROSC may be associated withic brain injury, so patients resuscitated in prehospi-ay benet less from MTH. Animal studies have shownic brain damage after cardiac arrest increased when

132 K.S. Bae et al. / Resuscitation 89 (2015) 129136

Table 1Demographic ndings of study population between mild therapeutic hypothermia and non-mild therapeutic hypothermia group.

Variables All Non-MTH MTH p-Value

N % N % N %

All, Total 11,007 100.0 9316 100.0 1691 100.0

Gender

K.S. Bae et al. / Resuscitation 89 (2015) 129136 133

Table 1 (Continued)

Variables All Non-MTH MTH p-Value

N % N % N %

Prehospital ROSC 0.308No 9161 83.2 7768 83.4 1393 82.4Yes 1846 16.8 1548 16.6 298 17.6

Survival to discharge

134 K.S. Bae et al. / Resuscitation 89 (2015) 129136

Table 2 (Continued)

Variables All P-ROSC ED-ROSC p-Value

N % N % N %

Primary ECG

K.S. Bae et al. / Resuscitation 89 (2015) 129136 135

Table 4Logistic regression analysis on outcomes by mild therapeutic hypothermia across the location of return of spontaneous circulation.

Outcomes Location of ROSC MTH vs. non-MTH

Adjusted ORa 95% CI

Survival to d

CPC 1 or 2 at

MTH: mild the partmcondence int

a Adjusted f ardioemergency de fusioncirculation (RO OSC.

to be generand was con

Current as possibleies of cardcooling migcome. A rechypothermas possible with standalogical recoto dene wto ROSC or time to ROSresuscitatio

To date,researcher get temperacontrolled be consideradd one evpractice.

4.1. Limitat

There arall potentiaof MTH prcol for postcould not mpotentially rological oudischarge acian who pspecied. Wvariables suor anticonvtus on discunnecessarand cause provided stlection, an we did notThe most itics. The stis far differshould be cinterpreted

clus

H waOHCMTH

P-RO

t of

re ar

wled

s stunt Al and0093300

dix A

plem in itatio

nces

owskital caies. ReAS, M13 u;127:

KO, Shof-hosischargePrehospital 0.84 ED 2.04

dischargePrehospital 0.78 ED 1.68

rapeutic hypothermia; ROSC: return of spontaneous circulation; ED: emergency deerval.or gender, age group, ECG group, metropolitan, place of event, witness, bystander cpartment, response time interval, scene time interval, transport time interval, reperSC), and interaction term between mild therapeutic hypothermia and location of R

alized as the study included a small number of patientsducted in a single hospital.

guideline also recommends MTH to be induced as soon after ROSC based on previous animal model stud-iac arrest.2931 Walters et al.17 reported that earlierht not have important impact on good neurological out-ent randomized clinical trial on effect of prehospitalia by infusing up to 2 L of 4 C normal saline as soonfollowing return of spontaneous at the eld comparedrd prehospital care showed similar survival and neuro-very.32 However, no clinical studies have been reportedho benets more or less from MTH in terms of timelocation of ROSC. Based on our results, the shorter theC, the greater benet MTH may disappear on cerebraln.

optimal hypothermia protocol remains unclear. Manyand practitioner considered 3224 C as optimal tar-ture17 Recently, Nielsen et al.18 performed randomizedtrial and revealed that hypothermia protocol coulded targeted temperature management. Our study canidence on who will be the target of MTH to current

ions

e several limitations to this study. We did not adjustl confounders such as quality of CPR or specic typeocedure. Hospital EDs do not have the same proto-

resuscitation care and hypothermia interventions. Weeasure the exact time for ROSC, which could have

biased our analysis. We measured the clinical neu-tcomes obtained by medical record reviewers frombstract or medical record, not by specialties of physi-erformed clinical neurological examination were not

5. Con

MTery in ROSC. not for

Conic

The

Ackno

ThiagemeControCDC (22011-E

Appen

Supfound,resusc

Refere

1. Berdhospstud

2. Go 202013

3. Ahnout-e also did not collect data concerning the importantch as seizure status, administrating sedative drugsulsant drugs which might inuence neurological sta-harge. Another limitation is on selection bias. Casesy to be cooled might be included in the MTH groupbias for the effect of MTH. In addition, though, werict data quality assurance program during data col-inter-rater agreement issue could be raised because

test the agreement among medical record reviewers.mportant limitation is the EMS system characteris-udy setting accepts scoop and run model, whichent from the Western EMS programs. The differenceonsidered when the study results are generalized to be.

citation 204. Spaite DW

witnessedresponse t

5. Richardsobased pu2005;12:6

6. Iwami T, cardiac-onculation 2

7. Peberdy M2010 Ameand emerg

8. Hegnauer hypotherm

9. Busto R, Gmild hypofatty acids0.59 1.191.79 2.32

0.58 2.701.34 1.98

ent; CPC: cerebral performance category; OR: odds ratio; 95% CI: 95%

pulmonary resuscitation, debrillation at prehopsital period, level of therapy, cardiopulmonary bypass, prehospital return of spontaneous

ion

s signicantly associated with good neurological recov-A survivors, but the effect was different by location of

was signicantly benecial for ED-ROSC patients, butSC patients in this EMS system.

interest statement

e no conicts of interest for all authors in this study.

gements

dy was supported by the National Emergency Man-gency of Korea and the Korea Centers for Disease

Prevention (CDC). The study was funded by the Korea2013) (Grant No.; 2009-E00543-00; 2010-E33022-00;4-00; 2012-E33010-00; 2013-E33015-00).

. Supplementary data

entary data associated with this article can bethe online version, at http://dx.doi.org/10.1016/j.n.2015.01.024.

i J, Berg RA, Tijssen JGP, Koster RW. Global incidences of out-of-rdiac arrest and survival rates: systematic review of 67 prospectivesuscitation 2010;81:147987.ozaffarian D, Roger VL, et al. Heart disease and stroke statisticspdate: a report from the American Heart Association. Circulatione6245.in SD, Suh GJ, et al. Epidemiology and outcomes from non-traumaticpital cardiac arrest in Korea: a nationwide observational study. Resus-

10;81:97481., Hanlon T, Criss EA, et al. Prehospital cardiac arrest: the impact of

collapse and bystander CPR in a metropolitan EMS system with shortimes. Ann Emerg Med 1990;19:12649.n LD, Gunnels MD, Groh WJ, et al. Implementation of community-blic access debrillation in the PAD trial. Acad Emerg Med8897.Kawamura T, Hiraide A, et al. Effectiveness of bystander-initiatedly resuscitation for patients with out-of-hospital cardiac arrest. Cir-007;116:29007.A, Callaway CW, Neumar RW, et al. Part 9: post-cardiac arrest care:rican Heart Association guidelines for cardiopulmonary resuscitationency cardiovascular care. Circulation 2010;122:S76886.AH, DAmato HE. Oxygen consumption and cardiac output in theic dog. Am J Physiol 1954;178:13842.lobus MY, Dietrich WD, Martinez E, Valdes I, Ginsberg MD. Effect ofthermia on ischemia-induced release of neurotransmitters and free

in rat brain. Stroke 1989;20:90410.

136 K.S. Bae et al. / Resuscitation 89 (2015) 129136

10. Morimoto Y, Kemmotsu O, Kitami K, Matsubara I, Tedo I. Acute brain swellingafter out-of-hospital cardiac arrest: pathogenesis and outcome. Crit Care Med1993;21:10410.

11. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivorsof out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med2002;346:55763.

12. Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypother-mia to improve the neurologic outcome after cardiac arrest. N Engl J Med2002;346:54956.

13. Arrich J, Holzer M, Herkner H, Mullner M. Hypothermia for neuroprotectionin adults after cardiopulmonary resuscitation. Cochrane Database Syst Rev2009:CD004128.

14. Bernard SA, Jones BM, Horne MK. Clinical trial of induced hypothermiain comatose survivors of out-of-hospital cardiac arrest. Ann Emerg Med1997;30:14653.

15. Sunde K, Pytte M, Jacobsen D, et al. Implementation of a standardised treat-ment protocol for post resuscitation care after out-of-hospital cardiac arrest.Resuscitation 2007;73:2939.

16. Testori C, Sterz F, Behringer W, et al. Mild therapeutic hypothermia is associatedwith favourable outcome in patients after cardiac arrest with non-shockablerhythms. Resuscitation 2011;82:11627.

17. Walters JH, Morley PT, Nolan JP. The role of hypothermia in post-cardiac arrestpatients with return of spontaneous circulation: a systematic review. Resusci-tation 2011;82:50816.

18. Nielsen N, Wettersley J, Cronberg T, et al. Targeted temperature management at33 C versus 36 C after cardiac arrest. N Engl J Med 2013;369:2197206.

19. Hayakawa K, Tasaki O, Hamasaki T, et al. Prognostic indicators and outcomeprediction model for patients with return of spontaneous circulation from car-diopulmonary arrest: the Utstein Osaka Project. Resuscitation 2011;82:87480.

20. Yanagawa Y, Sakamoto T. Analysis of prehospital care for cardiac arrest in anurban setting in Japan. J Emerg Med 2010;38:3405.

21. Shinada T, Hata N, Kobayashi N, et al. Efcacy of therapeutic hypothermia forneurological salvage in patients with cardiogenic sudden cardiac arrest: theimportance of prehospital return of spontaneous circulation. J Nippon Med Sch2013;80:28795.

22. Shin SD, Suh GJ, Ahn KO, Song KJ. Cardiopulmonary resuscitation outcome ofout-of-hospital cardiac arrest in low-volume versus high-volume emergencydepartments: an observational study and propensity score matching analysis.Resuscitation 2011;82:329.

23. Group BRCTIS. A randomized clinical study of a calcium-entry blocker (lid-oazine) in the treatment of comatose survivors of cardiac arrest. BrainResuscitation Clinical Trial II Study Group. N Engl J Med 1991;324:122531.

24. Radovsky A, Safar P, Sterz F, Leonov Y, Reich H, Kuboyama K. Regional prevalenceand distribution of ischemic neurons in dog brains 96 hours after cardiac arrestof 0 to 20 minutes. Stroke 1995;26:212733 [discussion 334].

25. Sanders AB, Kern KB, Bragg S, Ewy GA. Neurologic benets from the use of earlycardiopulmonary resuscitation. Ann Emerg Med 1987;16:1426.

26. Testori C, Sterz F, Holzer M, et al. The benecial effect of mild therapeutichypothermia depends on the time of complete circulatory standstill in patientswith cardiac arrest. Resuscitation 2012;83:596601.

27. Nielsen N, Hovdenes J, Nilsson F, et al. Outcome, timing and adverse events intherapeutic hypothermia after out-of-hospital cardiac arrest. Acta AnaesthesiolScand 2009;53:92634.

28. Soga T, Nagao K, Sawano H, et al. Neurological benet of therapeutichypothermia following return of spontaneous circulation for out-of-hospitalnon-shockable cardiac arrest. Circ J 2012;76:257985.

29. Kuboyama K, Safar P, Radovsky A, Tisherman SA, Stezoski SW, Alexander H. Delayin cooling negates the benecial effect of mild resuscitative cerebral hypother-mia after cardiac arrest in dogs: a prospective, randomized study. Crit Care Med1993;21:134858.

30. Abella BS, Zhao D, Alvarado J, Hamann K, Vanden Hoek TL, Becker LB. Intra-arrest cooling improves outcomes in a murine cardiac arrest model. Circulation2004;109:278691.

31. Hicks SD, DeFranco DB, Callaway CW. Hypothermia during reperfusion afterasphyxial cardiac arrest improves functional recovery and selectively altersstress-induced protein expression. J Cereb Blood Flow Metab 2000;20:52030.

32. Kim F, Nichol G, Maynard C, et al. Effect of prehospital induction of mildhypothermia on survival and neurological status among adults with cardiacarrest: a randomized clinical trial. JAMA 2014;311:4552.

The effect of mild therapeutic hypothermia on good neurological recovery after out-of-hospital cardiac arrest according to...1 Introduction2 Methods2.1 Data source2.2 Study setting2.3 Study subjects2.4 Study variables2.5 Main exposure2.6 Study outcome2.7 Analysis

3 Results4 Discussion4.1 Limitations

5 ConclusionConflict of interest statementAcknowledgementsAppendix A Supplementary dataReferences