Accepted Manuscript

Title: Therapeutic hypothermia after cardiac arrest: a

systematic review/meta-analysis exploring the impact of
expanded criteria and targeted temperature

Author: Aldo L. Schenone MD Aaron Cohen DO Gabriel

Patarroyo MD Logan Harper MD XiaoFeng Wang PhD Mehdi
H. Shishehbor DO MPH Venu Menon MD Abhijit Duggal
MD MPH

PII: S0300-9572(16)30394-X
DOI: http://dx.doi.org/doi:10.1016/j.resuscitation.2016.07.238
Reference: RESUS 6872

To appear in: Resuscitation

Received date: 26-5-2016

Revised date: 19-7-2016
Accepted date: 25-7-2016

Please cite this article as: Schenone Aldo L, Cohen Aaron, Patarroyo Gabriel,
Harper Logan, Wang XiaoFeng, Shishehbor Mehdi H, Menon Venu, Duggal
Abhijit.Therapeutic hypothermia after cardiac arrest: a systematic review/meta-analysis
exploring the impact of expanded criteria and targeted temperature.Resuscitation
http://dx.doi.org/10.1016/j.resuscitation.2016.07.238

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Therapeutic hypothermia after cardiac arrest: a systematic
review/meta-analysis exploring the impact of expanded criteria and
targeted temperature
Aldo L Schenone, MD1, Aaron Cohen, DO1, Gabriel Patarroyo, MD2, Logan Harper, MD1, XiaoFeng
Wang, PhD3, Mehdi H Shishehbor, DO MPH4, Venu Menon, MD4, Abhijit Duggal, MD MPH5

1
Internal Medicine, Cleveland Clinic, Ohio USA
2
Nephrology Department, University Hospital Case Western Reserve University, Ohio, USA
3
Department of Quantitative Health Sciences, Cleveland Clinic, Ohio, USA
4
Cardiology Department, Cleveland Clinic, Cleveland, Ohio, USA
5
Pulmonary and Critical Care Department, Cleveland Clinic, Cleveland, Ohio, USA

Corresponding Author:
Aldo Schenone, MD
Chief Medical Resident
Internal Medicine Department, Medicine Institute
Cleveland Clinic, Cleveland, Ohio
Phone: 216-339-7336
schenoa@ccf.org

Word Count
Abstract: 248 words
Text: 3672 words
Abstract
Aims of the study: We aimed to determine the benefit of an expanded use of TH. We also described the
impact of a targeted temperature management on outcomes at discharge.

Data sources: We identified studies by searching MEDLINE, EMBASE and Cochrane Library databases.
We included RCTs and observational studies restricted to those reporting achieved temperature during
TH after OHCA. No other patient, cardiac arrest or hypothermia protocol restrictions were applied.
Outcomes of interest were hospital mortality and neurological outcome at discharge. Appropriate risk of
bias assessment for meta-analyzed studies was conducted. Studies contrasting hypothermia and
normothermia outcomes were meta-analyzed using a random-effect model. Outcomes of cooling arms,
obtained from enrolled studies, were pooled and compared across achieved temperatures.

Results: Search strategy yielded 32,275 citations of which 24 articles met inclusion criteria. Eleven
studies were meta-analyzed. The use of TH after OHCA, even within an expanded use, decreased the
mortality (OR 0.51, 95%CI [0.41-0.64]) and improved the odds of good neurological outcome (OR 2.48,
95%CI [1.91-3.22]). No statistical heterogeneity was found for either mortality (I2=4.0%) or neurological
outcome (I2=0.0%). No differences in hospital mortality (p=0.86) or neurological outcomes at discharge
(p=0.32) were found when pooled outcomes of 34 hypothermia arms grouped by cooling temperature
were compared.

Conclusion: The use of TH after OHCA is associated with a survival and neuroprotective benefit, even
when including patients with non-shockable rhythms, more lenient downtimes, unwitnessed arrest and/or
persistent shock. We found no evidence to support one specific temperature over another during
hypothermia.

Keywords: Therapeutic hypothermia; Out-hospital Cardiac arrest; Targeted temperature

management; expanded hypothermia use
Background

The use of cooling as therapeutic agent was initially utilized for the treatment of head wounds, and was

first described in the Edwin Smith papyrus over 5000 years ago.1 First reports of the use of therapeutic

hypothermia (TH) during resuscitative efforts date back to 1803 when patients where covered in snow as

way to facilitate resuscitation.2 Subsequently, the use of TH after cardiopulmonary arrest was described

50 years ago in a case series limited to patients with in-hospital cardiac arrest.3,4 Due to a consistent

benefit of hypothermia on brain cell ischemia in experimental models5, these earlier studies targeted much

lower temperatures than those recommended in current guidelines6. Despite demonstration of benefit,

generalization of results was often hampered by the high incidence of adverse events associated with TH

at the time.7 Therapeutic hypothermia became a more accepted therapeutic intervention after two

landmark trials in the early 2000s demonstrated significant survival benefit and improved neurological

outcomes associated with TH after out-hospital cardiac arrest (OHCA). 8-10

TH is strongly recommended in unconscious survivors of OHCA stemming from ventricular fibrillation

(Class Ib).6 Outside of this cohort of patients, there is a great degree of uncertainty and variation in

practice. There is a significant paucity of data for the use of TH in all cases of cardiac arrest due to small

size of study subjects, substantial differences in cooling protocols and strict inclusion criteria.10 Due to

this lack of data, most of the recommendations on the specifics of TH are based on observational studies

and expert opinion. Current guidelines support cooling of comatose survivor of cardiac arrests stemming

from non-shockable rhythms and in-hospital settings (Class IIb).8,9,11,12 These recommendations have

significantly modified clinical practice, and have expanded the use of TH beyond the inclusion criteria of

initial trials.13 Nonetheless, the data showing the benefits of TH when applied to a broader spectrum of

patients remain scattered at best.

Since the publication of the first TH trials, efforts have been made to define an ideal cooling

temperature.14-17 But after almost two decades, there still exists significant uncertainty about the optimal

temperature for TH.18,19 While clinical guidelines have advised a target temperature ranging from 32oC to
34oC, a recent update from the International Liaison Committee on Resuscitation (ILCOR) suggests that

the temperature range should be liberalized to 32oC-36oC.19 Beyond the issue of temperature, there is also

growing skepticism about the benefits of cooling and some researchers have suggested that the improved

outcomes in these patients are really due to a practice of fever avoidance.20

We conducted an updated meta-analysis to assess the performance of TH after OHCA on hospital

mortality and good neurological outcome at hospital discharge when treating patients within and beyond

the initial criteria of the landmark trials. We also performed a systematic review and pooled analysis to

determine the impact of different temperature goals on outcomes at discharge.

Methods

The results of this meta-analysis were written using the PRISMA (Preferred Reporting Items for

Systematic Reviews and Meta-Analysis) reporting guidelines for randomized trials and MOOSE (Meta-

analysis Of Observational Studies in Epidemiology) reporting guidelines for observational studies. We

used the PICOS (Participants, Intervention, Comparison, Outcomes and Study Design) model to generate

the research questions, create a search strategy and guide study selection. 21,22

Search Strategy

We performed a systematic search of the MEDLINE, OVID, EMBASE, and Cochrane Library databases

for studies published through July 2014 using a predefined search strategy (Appendix 1). We also

reviewed the references of all retrieved studies and the pertinent review articles to identify any additional

studies. The full text of any citation considered potentially relevant was reviewed.

Study selection

We included studies that enrolled patients 15 years or older who remained comatose after OHCA with

any initial rhythm. No restrictions were applied to downtime, resuscitation by bystander, witnessed arrest

or persistent shock after resuscitation. The enrollment of studies with this more inclusive patient selection

was meant to capture the interest in expanding the population undergoing TH beyond the current

guidelines. This represents an expanded use of TH in current clinical practice.13 The intervention of
interest was TH but the studies had to report the achieved cooling temperature to be considered. We

systematically reviewed all the studies, but only studies that compared TH versus normothermia were

considered for our meta-analysis. Additionally, we performed a pooled analysis of grouped cohorts from

included studies based on the average temperature achieved during TH. Outcomes of interest were

hospital mortality and neurological status at discharge based on Cerebral Performance Category (CPC).

We included randomized controlled trials (RCT) and cohort studies with 10 or more patients. No

language restrictions were applied. We excluded abstracts, material presented at medical conferences,

unpublished data and animal studies. Four authors independently reviewed the full articles (AS, AC, GP,

HL) to determine eligibility. Any discrepancies were resolved after discussion with an independent

reviewer (AD). All raters were well versed in the investigation guidelines, purpose and criteria.

Data extraction

We developed a pre-defined data collection tool for data extraction. The tool was pilot-tested on four

randomly selected studies. Each author (AS, AC, GP, LH) collected data independently from the enrolled

studies, and any inconsistency was addressed by group-review of the paper with the senior author (AD).

The data collection tool included patient demographics, patient comorbidities, enrollment setting (field,

emergency department or Intensive Care Unit), cardiac arrest characteristics based on Utstein format and

TH protocol.23 The TH protocol extraction form included induction setting (field or hospital), method of

cooling, target temperature, time at target temperature and rewarming. The achieved temperatures during

cooling were extracted from the study reports or obtained from included figures. Neurological outcome

was reported using the CPC scale (CPC 1-2 described as good neurological outcome); an author

designation of good neurological outcome was considered to be comparable to CPC 1 or 2.

Assessment of Risk of Bias

We used the Cochrane Collaboration tool to assess the risk of bias of meta-analyzed randomized trials,

while cohort studies included in the meta-analysis were evaluated for risk of bias using the Newcastle-

Ottawa assessment tool. 23,24 Similarly, we assessed the risk of bias in the hypothermia arms in the pooled

analysis using a modified Newcastle-Ottawa assessment tool. (Appendix 2). We graded the risk of bias as

being low, high or unclear. Three reviewers (AS, AC, GP) worked independently using the pre-defined
assessment tools to assess bias risk. Any inconsistency across reviewers was resolved by reviewing the

paper with senior author (AD).

Statistics

Patient characteristics were described using means and standard deviations for all continuous variables

and counts with percentages for all categorical variables. All analyses were two-tailed and performed at a

significance level of 0.05. Meta-analysis was performed using a random effects model to obtain summary

outcome point estimates with 95% confidence intervals. We assessed heterogeneity between studies by

visual inspection of a Forest plot and using Q statistics as well as an I2 index. The I2 values of 50% or less

and a p-value greater than 0.1 indicate no statistical heterogeneity were observed. We assessed

publication bias and heterogeneity using funnel plots. In the absence of publication bias and

heterogeneity, one would then expect the funnel plot to show points forming a funnel shape, with the

majority of the points falling inside of the pseudo-confidence region with bounds 1.96SE, where is

the estimated effect and SE is the standard error value from the y-axis.25 All these statistical analysis

were conducted on RStudio software. The pooled analysis was conducted on SAS 9.3 software (SAS

Institute, Cary, NC). We performed direct outcomes comparisons across temperature groups using both

nonparametric Kruskal-Wallis and unequal variance t-test. We also performed a univariate logistic

regression between: Hospital mortality and temperature as well as good neurological outcomes (CPC 1-2)

and temperature. Analysis was fitted to test the heterogeneity of slopes across temperature groups. We

treated variable temperature as continuous and categorical. Finally, a stepwise multivariate regression was

performed by outcome of interest.

Results
Study Characteristics

Our search strategy yielded 32,275 citations after de-duplication. We retrieved 149 articles for a detailed

evaluation and based on our inclusion criteria included 24 articles for our final review and analysis

(Figure 1).
Eleven studies (three randomized controlled trials and eight cohort studies) compared TH versus

normothermia and were included in our meta-analysis. (Table 1)8,9,26-34 These 11 studies contained 1381

patients with a mean age of 61.3 years. Shockable rhythms accounted for 51.6% of all arrests with arrest

downtimes ranging from 20-34.6 minutes (mean 24.6 minutes). In addition, most patients did not receive

bystander resuscitation. The mean temperature of the control groups was 37C (range 36.1-37.5C). All

studies reported hospital mortality, but only 10 described neurological outcomes at discharge. Only 3

studies compared two different levels of targeted temperature during TH and are described below.14,16,17

(Appendix 3)

We identified 34 separate cohorts of specific temperatures during TH among the 24 included studies.

(Appendix 4) The target temperature most commonly achieved was 33 reported as the endpoint in 20

cooling cohorts, 10 hypothermia arms used 34C, two achieved 32C, while one cohort each achieved

35C and 36C respectively. In total there were 4373 patients included by these pooled cohorts. Eleven

hypothermia arms were from RCTs whereas the other 23 were from observational cohort studies.

Risk of Bias

Only two trials included in the meta-analysis performed proper randomization whereas one conducted a

quasi-randomized study. None of the three studies blinded participants due to notable procedural

differences between the intervention and control groups. Only two of the three trials blinded the assessor

to outcomes. Complete outcome data were reported in two trials whereas it was unclear whether there was

incomplete data in the third trial. There was no selective reporting. We were unable to detect other

potential bias. Among the meta-analyzed observational studies, the intended exposed cohort was

represented appropriately by all studies. Studies selected their control from populations that were similar

to the exposed cohort. Yet, the criteria used for control group allocation were unclear in two of the cohort

studies. All but a single study studies had comparable baseline patient characteristics across the

intervention and control groups. Meanwhile, all except three studies had similar cardiac arrest

characteristics between the arms. One had significant differences that could have impacted the results and

two had subtle differences of unclear significance. All studies obtained patient and outcome information
via direct observation or review of medical records. Only three studies did not explicitly stated that they

excluded patients with a terminal condition while seven did not mentioned baseline neurological status.

All studies were conducted with proper assessment of the outcomes and adequate follow-up. (Appendix

5). A Funnel plot including randomized trial and observational studies did not show a significant risk of

publication bias. (Appendix 6 and 7)

From the 34 cooling arms included in pooled analysis, we identified five hypothermia cohorts at high risk

for cohort representativeness of intended population. All the arms obtained their data by direct

observation or medical records review. Only 15 cohorts explicitly reported the exclusion of patients with

terminal conditions, while 12 stated they excluded patients with poor baseline neurological status. The

remaining cohorts did not specify details regarding patients with terminal status and poor baseline

neurological status. Follow-up time was sufficient for the outcome to occur in all cohorts. Two cohorts

excluded moribund patients after TH was applied while nine arms excluded data from the final analysis

with no explanation. (Appendix 8)

Impact of expanded inclusion criteria on the benefit of therapeutic hypothermia

The use of TH after OHCA, even in the setting of expanded inclusion criteria, decreased the mortality

(OR 0.51, 95%CI [0.41-0.64]), and improved the odds of survival with good neurological outcome (OR

2.48, 95%CI [1.91-3.22]). (Figure 2). Separate analysis of both randomized trials and cohort studies

revealed a reduction in mortality (RCTOR 0.56 95%CI [0.37-0.84]; Observational study OR 0.46 95%CI

[0.33-0.65]) and good neurological outcome (RCT OR 2.18 95%CI [1.40-3.39]) Observational study OR

2.72 95%CI [1.89-3.92]. The addition of cohort studies did not introduce statistical heterogeneity across

studies for either mortality (I2=4.0%) or neurological outcome (I2=0.0%).

Impact of temperature on therapeutic hypothermia

Studies comparing two temperatures

Among the 24 included studies, three evaluated the differences in mortality and neurological outcomes

across different temperatures. (Appendix 3). Kim et al (2011) described outcomes in comatose survivors
of OHCA who were still alive after 24 hours of ROSC. The study reported no difference in outcomes

across 32C, 33C and 34C. Two thirds of the arrests stemmed from non-shockable rhythms and half of

them were non-cardiac in origin. Overall mortality was 39.6% while good neurological outcomes were

achieved in 22.6%. The study reported no difference in outcomes across 32C, 33C and 34C.

Lopez-de-Sa et al (2012) compared 32C versus 34C in a pilot RCT. The study reported an improved 6-

months mortality (p=0.03) favoring 32C. When restricting analysis to those with shockable rhythms, the

data showed improved good outcomes (CPC 1-2) favoring 32C (p=0.029). However, the patients

assigned to the 34C arm received less CPR by bystanders, had longer downtimes and had a tendency to

have worse admission Glasgow Scores.

Finally, Nielsen et al performed a large randomized trial contrasting 33C with 36C on comatose

survivors of OHCA. There was no difference in the composite endpoints of mortality and poor

neurological outcome at 6 months. It is important to note that there were subtle differences across groups

with those cooled at 33C having slightly less shockable rhythms, less witnessed arrest and more

circulatory shock compared to patients at 36C.

Studies comparing hypothermia versus normothermia

We also summarized the impact of TH at different levels of cooling compared to normothermia, (Table

1). Only one study cooled patients to 32C and it reported a non-significant trend for improved hospital

mortality (OR 0.43, 95%CI [0.11-1.61]). No neurological outcomes at discharge were reported. This

study reported mortality but did not confer a neuroprotective benefit at 6 months. Aggregate evidence for

cooling at 33C demonstrated a decrease in hospital mortality (OR=0.44 95%CI [0.33-0.60]) and

improved neurological outcomes at discharge (OR=2.57 95%CI [1.73-3.81]). Meanwhile, a single study

cooling at 34C showed a trend for improved in-hospital mortality (OR=0.70 95%CI [0.46-1.05] and a

significant neuroprotective benefit at discharge (OR=2.02 95%CI [1.21-3.38]). However, when analyzing

shockable rhythms only, cooling had a benefit on both hospital mortality (OR 0.53 (95CI 0.29-0.97) and

neurological outcome at discharge (OR2.07 95CI 1.26-3.41). Similarly, a single study consisting of
cardiac arrest patients with shockable rhythms who were cooled at 35C reported a mortality (OR0.44

(95%CI 0.166-1.165) and neuro-protective benefit (OR 2.86 95%CI [1.04-7.85]) at hospital discharge.

There was no data comparing cooling at 36C and normothermia.

Degree of hypothermia and associated outcomes

Outcome data was pooled from the hypothermia arms and grouped by achieved cooling temperature

(Table 3). We found no difference in hospital mortality (p=0.86) or neurological outcomes at discharge

(p=0.32) across temperatures using univariate logistic regression (Figure 3). Stepwise multivariate

regression determined that only the type of rhythm (p=0.0001) was associated with in-hospital mortality.

Both type of rhythm (p=0.01) and downtime (p=0.014) were independently associated with neurological

outcomes.

Discussion

This meta-analysis examined the benefit of TH post-cardiac arrest, beyond just the unconscious patient

who had a witnessed OHCA stemming from a shockable rhythm, with less than 30 min of downtime and

no persistent shock upon ROSC. We expanded our cohort to comatose survivors of OHCA with any

initial rhythm, more flexible downtimes, without restriction in regards of witnessed arrest or shock after

ROSC. This approach is radically different from previous meta-analyses that have reported outcomes with

use of TH in cardiac arrest. Despite expanding our inclusion criteria, and introducing hither to unexplored

cohorts TH continues to show a mortality benefit with decreased odds of in hospital mortality (OR 0.51

95% CI 0.41-0.64) while increasing the chances of survival with a good neurological outcome (OR 2.48

95% CI 1.91-3.22). The inclusion of observational cohort studies did not introduce statistical

heterogeneity across studies for either mortality (I2=4.0%) or neurological outcome (I2=0.0%).

Furthermore, sensitivity analysis by study design mirrored our main results, and did not differ from prior

meta-analyses which restricted their analysis to only randomized trials.10 Additionally, based on a

quantitative assessment of available evidence our study showed that there was no difference in reported
outcomes based on the targeted temperature level during TH. Our analysis did not support the superiority

of one temperature level over another during TH.

The addition of cohort studies in this meta-analysis provides a better representation of the current clinical

practice around TH. In contrast to landmark randomized trials, arrest survivors included in these studies

had higher proportion of unwitnessed arrests, more arrests stemming from non-shockable rhythms,

slightly longer downtimes and a more heterogeneous application of hypothermia than those in landmark

trials. This study helps elucidate the use of TH in daily clinical practice. To date, no study has evaluated

the impact of TH in a broader sense where outcomes of patients with non-shockable rhythms, more

lenient downtimes, inclusion of unwitnessed arrest and patients with persistent shock were evaluated

along with the larger trials.6,9 Our analysis proves that there is a benefit of TH in a much broader spectrum

of patients, and perhaps clinical guidelines need to strongly favor TH in all patients with cardiac arrest.

While we propose the inclusion of more patients for the consideration of TH, we need to be cautious that

this intervention is not used in inappropriate populations. This is a resource consuming intervention,

which needs to be provided in the appropriate population. Strict inclusion criteria should still be

maintained, but strong consideration should be provided to the excluded populations. Perhaps, the

exclusion of these patients should be based on pre-existing functional status, severity of illness in the

acute setting, risk of coagulopathy or other adverse events, and not based on the initial rhythm, witness of

arrest, or persist shock after ROSC.

In the current era of uncertainty about optimal cooling temperature, it is important to note that the

majority of studies reporting on the use of TH used a target a temperature of 33C to 34C, which is

consistent with recommendations from current guidelines. We did not see a difference in outcomes when

we compared33C to 34C, but there is a significant paucity of literature around outcomes associated with

other target temperatures. At this time there is no evidence to support the use of one temperature level

over other across the entire range of included temperatures of 32C and 36C during TH after OHCA.

This finding aligns with results from recent TTM trial.14 The variability in cooling protocols, and the

reporting of achieved targeted temperatures might have a significant impact on the reported outcomes, but
due to the lack of patient level data, we were not able to assess these differences. Based upon these

results there may not be an optimal hypothermia protocol with a fixed target temperature for all arrest

patients. It still remains unknown whether there is a potential benefit of a targeted temperature approach

during TH for certain subgroups of arrest survivors, such as the finding in this study of a non-significant

potential benefit when applying lower temperature to non-shockable rhythms (Appendix 8). It is unclear

the reason behind this trend, but this warrants further investigation.

This meta-analysis has several important limitations. The addition of observational cohort studies leads

to potential for a higher risk of bias. The study enrollment was restricted to only those studies reporting

achieved temperatures during cooling, which might have led to selection bias. However, the quantitative

analysis of these studies was supported by the lack of visual and statistical heterogeneity. The exclusion

of abstracts or unpublished research could have introduced publication bias. In regards to pooled analysis,

the number of studies cooling at a temperature other than 33C or 34C was small thus there is unlikely

enough power to detect difference when using such temperature at the end of spectrum.

Conclusion

The use of therapeutic hypothermia is associated with a survival and neuroprotective benefit after OHCA,

even when including patients with non-shockable rhythms, more lenient downtimes, unwitnessed arrest

and/or persistent shock after ROSC. An expanded use of this intervention in daily practice might be

appropriate in patients with good baseline functional status, regardless of the initial rhythm, unwitnessed

arrest status or persistent shock. However, we found no evidence to support one specific temperature over

another as long as the patients underwent hypothermia. The available evidence is not robust enough to

form a definitive conclusion regarding the impact of targeting a specific temperature during TH. Whether

there is a potential benefit of using lower temperatures in some patient groups requires further

exploration.
Conflict of interest statement

None to declare.

Declaration of interests
None of the authors involved with this manuscript have any financial and personal relationships with
other people or organizations that could inappropriately influence or bias their work.

Sources of Funding
No funding to disclose

Authors and Contributors

Aldo Schenone, Aaron Cohen, Gabriel Patarroyo and Abhijit Duggal had substantial contributions to the
conception and design of the work. Aldo Schenone, Aaron Cohen, Gabriel Patarroyo and Logan Harper
were responsible for the acquisition of data. Aldo Schenone, Aaron Cohen, Mehdi Shishehbor, Venu
Menon and Abhijit Duggal were responsible for the analysis and interpretation of data for this manuscript.
Aldo Schenone, Aaron Cohen, Mehdi Shishehbor, Venu Menon and Abhijit Duggal were responsible for
drafting and key revisions of this manuscript. All authors have approved the final version of the
manuscript and are accountable for all aspects related to the accuracy or integrity of this manuscript.

References

1. Wang H, Olivero W, Wang D, Lanzino G. Cold as a therapeutic agent. Acta Neurochir (Wien).
2006;148(5):565-570; discussion 569-570.
2. Varon J, Acosta P. Therapeutic hypothermia: past, present, and future. Chest. 2008;133(5):1267-1274.
3. Williams G, Jr., Spencer FC. The Clinical use of hypothermia following cardiac arrest. Ann Surg.
1958;148(3):462-468.
4. Benson DW WG, JR., Smith CR, Conn AW, Barker GA. The use of hypothermia after cardiac arrest.
Anesth Analg. 1959;14(6):423-428.
5. Colbourne F, Corbett D. Delayed postischemic hypothermia: a six month survival study using behavioral
and histological assessments of neuroprotection. J Neurosci. 1995;15(11):7250-7260.
6. Scirica BM. Therapeutic hypothermia after cardiac arrest. Circulation. 2013;127(2):244-250.
7. Bohn DJ, Biggar WD, Smith CR, Conn AW, Barker GA. Influence of hypothermia, barbiturate therapy,
and intracranial pressure monitoring on morbidity and mortality after near-drowning. Crit Care Med.
1986;14(6):529-534.
8. Hypothermia after Cardiac Arrest Study G. Mild therapeutic hypothermia to improve the neurologic
outcome after cardiac arrest.[Erratum appears in N Engl J Med 2002 May 30;346(22):1756]. New England
Journal of Medicine. 2002;346(8):549-556.
9. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest
with induced hypothermia. New England Journal of Medicine. 2002;346(8):557-563.
10. Arrich J, Holzer M, Herkner H, Mullner M. Cochrane corner: hypothermia for neuroprotection in adults
after cardiopulmonary resuscitation. Anesth Analg. 2010;110(4):1239.
11. 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 Cardiovascular 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(3):350-379.
12. Arrich J, Holzer M, Havel C, Mullner M, Herkner H. Hypothermia for neuroprotection in adults after
cardiopulmonary resuscitation. Cochrane Database Syst Rev. 2012;9:CD004128.
13. Bernard S. Hypothermia after cardiac arrest: expanding the therapeutic scope. Crit Care Med. 2009;37(7
Suppl):S227-233.
14. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 33degreeC versus
36degreeC after cardiac arrest. New England Journal of Medicine. 2013;369(23):2197-2206.
15. Drabek T, Kochanek PM. Therapeutic Hypothermia on Its 10th Anniversary: Is it Time to Turn the
Thermostat Down? Circulation. 2012;126(24):2803-2805.
16. Kim JJ, Yang HJ, Lim YS, et al. Effectiveness of each target body temperature during therapeutic
hypothermia after cardiac arrest. American Journal of Emergency Medicine. 2011;29(2):148-154.
17. Lopez-de-Sa E, Rey JR, Armada E, et al. Hypothermia in comatose survivors from out-of-hospital cardiac
arrest: pilot trial comparing 2 levels of target temperature. Circulation. 2012;126(24):2826-2833.
18. Deakin CD, Morrison LJ, Morley PT, et al. Part 8: Advanced life support: 2010 International Consensus on
Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment
Recommendations. Resuscitation. 2010;81 Suppl 1:e93-e174.
19. Donnino MW, Andersen LW, Berg KM, et al. Temperature Management After Cardiac Arrest An
Advisory Statement by the Advanced Life Support Task Force of the International Liaison Committee on
Resuscitation and the American Heart Association Emergency Cardiovascular Care Committee and the
Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation. Resuscitation. 2015.
20. Rittenberger JC, Callaway CW. Temperature management and modern post-cardiac arrest care. N Engl J
Med. 2013;369(23):2262-2263.
21. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and
meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535.
22. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal
for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA.
2000;283(15):2008-2012.
23. Deeks JJ, Dinnes J, D'Amico R, et al. Evaluating non-randomised intervention studies. Health Technol
Assess. 2003;7(27):iii-x, 1-173.
24. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in
randomised trials. BMJ. 2011;343:d5928.
25. Sterne JA, Egger M. Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin
Epidemiol. 2001;54(10):1046-1055.
26. Laurent I, Adrie C, Vinsonneau C, et al. High-volume hemofiltration after out-of-hospital cardiac arrest: a
randomized study. Journal of the American College of Cardiology. 2005;46(3):432-437.
27. Bernard SA, Jones BM, Horne MK. Clinical trial of induced hypothermia in comatose survivors of out-of-
hospital cardiac arrest. Annals of Emergency Medicine. 1997;30(2):146-153.
28. Oddo M, Schaller MD, Feihl F, Ribordy V, Liaudet L. From evidence to clinical practice: effective
implementation of therapeutic hypothermia to improve patient outcome after cardiac arrest. Critical Care
Medicine. 2006;34(7):1865-1873.
29. Derwall M, Stoppe C, Brucken D, Rossaint R, Fries M. Changes in S-100 protein serum levels in survivors
of out-of-hospital cardiac arrest treated with mild therapeutic hypothermia: a prospective, observational
study. Critical Care (London, England). 2009;13(2):R58.
30. Ferreira I, Schutte M, Oosterloo E, et al. Therapeutic mild hypothermia improves outcome after out-of-
hospital cardiac arrest. Netherlands Heart Journal. 2009;17(10):378-384.
31. Petrovic M, Panic G, Jovelic A, et al. Therapeutic hypothermia and neurological outcome after cardiac
arrest. Vojnosanitetski Pregled. 2011;68(6):495-499.
32. Walters EL, Morawski K, Dorotta I, et al. Implementation of a post-cardiac arrest care bundle including
therapeutic hypothermia and hemodynamic optimization in comatose patients with return of spontaneous
circulation after out-of-hospital cardiac arrest: a feasibility study. Shock. 2011;35(4):360-366.
33. Don CW, Longstreth WT, Jr., Maynard C, et al. Active surface cooling protocol to induce mild therapeutic
hypothermia after out-of-hospital cardiac arrest: a retrospective before-and-after comparison in a single
hospital. Critical Care Medicine. 2009;37(12):3062-3069.
34. Belliard G, Catez E, Charron C, et al. Efficacy of therapeutic hypothermia after out-of-hospital cardiac
arrest due to ventricular fibrillation. Resuscitation. 2007;75(2):252-259.
33. Heard KJ, Peberdy MA, Sayre MR. A randomized controlled trial comparing the Arctic Sun to standard
cooling for induction of hypothermia after cardiac arrest. Resuscitation. 2010 Jan;81(1):9-14
34. Bernard SA, Smith K, Cameron P. Rapid Infusion of Cold Hartmanns (RICH) Investigators. Induction of
therapeutic hypothermia by paramedics after resuscitation from out-of-hospital ventricular fibrillation
cardiac arrest: a randomized controlled trial. Circulation. 2010 Aug 17;122(7):737-42.
35. Bernard SA, Smith K, Cameron P. Rapid Infusion of Cold Hartmanns Investigators. Induction of
prehospital therapeutic hypothermia after resuscitation from non-ventricular fibrillation cardiac arrest. Crit
Care Med. 2012 Mar;40(3):747.
36. Zeiner A, Holzer M, Sterz F, Mild resuscitative hypothermia to improve neurological outcome after cardiac
arrest. Stroke. 2000 Jan;31(1):86-94.
37. Bruel C, Parienti JJ, Marie W. Mild hypothermia during advanced life support: a preliminary study in out-
of-hospital cardiac arrest. Crit Care. 2008;12(1).
38. Nielsen N1, Hovdenes J, Nilsson F. Outcome, timing and adverse events in therapeutic hypothermia after
out-of-hospital cardiac arrest. Acta Anaesthesiol Scand. 2009 Aug;53(7):926-34.
39. Benz-Woerner J, Delodder F, Benz R. Body temperature regulation and outcome after cardiac arrest and
therapeutic hypothermia. Resuscitation. 2012 Mar;83(3):338-42.
40. Vaahersalo J, Hiltunen P, Tiainen M. Therapeutic hypothermia after out-of-hospital cardiac arrest in
Finnish intensive care units: the FINNRESUSCI study. Intensive Care Med. 2013 May;39(5):826-37.
41. Woo JH, Lim YS, Yang HJ. Factors associated with pneumonia in post-cardiac arrest patients receiving
therapeutic hypothermia. Am J Emerg Med. 2014 Feb;32(2):150-5.
42. Yokoyama H, Nagao K, Hase M, Tahara Y. Impact of therapeutic hypothermia in the treatment of patients
with out-of-hospital cardiac arrest from the J-PULSE-HYPO study registry. Circ J. 2011;75(5):1063-70.
43. Torgersen C, Meichtry J, Schmittinger CA. Haemodynamic variables and functional outcome in
hypothermic patients following out-of-hospital cardiac arrest. Resuscitation. 2013 Jun;84(6):798-804.
44. Gal R, Slezak M, Zimova I, Cundrle I. Therapeutic hypothermia after out-of-hospital cardiac arrest with the
target temperature 34-35 degrees C. Bratisl Lek Listy. 2009;110(4):222-5.
Legends of figures
Figure 1. PRISMA flow diagram of included studies
This is a graphical representation of the flow of citations reviewed in the course of a Systematic Review
Figure 2. Impact of expanded use of TH after OHCA on mortality and neurological outcomes at
hospital discharge.
A. Results of 11 studies (3 RCTs and 8 cohort studies) on the impact of expanded use of TH after OHCA
on hospital mortality. Summary OR= 0.51 [95% CI (0.40-0.64)]. Test for heterogeneity: I2(%)= 4.0604,
CI = (0.0, 52.5916). B. Results of 10 studies (2 RCT and 8 cohort studies) on the impact of expanded
application of of TH after OHCA on neurological outcomes at hospital discharge. Summary OR= 2.48
[95% CI (1.91-3.22)]. Test for heterogeneity: I^2(%) = 0.0, CI = (0.0-69.84). Meta-analysis was
performed using a random effects model to obtain summary outcome point estimates with 95%
confidence intervals.
Figure 3. Pooled analysis on the impact of cooling temperature during TH post-OHCA on hospital
mortality and neurological outcome at discharge
Pooled outcomes across temperature groups were contrasted using univariate logistic regression fitted to
test heterogeneity of slope. No differences were identified between A. Hospital mortality and temperature
(p=0.86) or B. Good neurological outcome (CPC1-2) and temperature (p=0.32).
 Table 1: Characteristics of the Studies Included in the Meta-Analysis
Study Rhythm Sample Down Temperature Follow-up Mortality CPC 1-2
TH/control time TH/control Assessment TH/control TH/control
(n) (min) (C)

RCT
55% / 74%
Both Discharge p=0.16 NA
Laurent24 Shockable 72.7% 22 / 19 20 31.7 / 37.4
2005 Non-shockable 27.2%
68% / 79% 45% / 26%
6-months p=0.018 p=0.21

51% / 68% 49% / 26%

Bernard7 Shockable 43 / 34 26.5 32.9 / 37.4 Discharge p=0.145 p=0.046
2002
37% / 50% 47% / 30%
Discharge p=0.02 p=0.004
HACA6 Shockable
2002 137 / 138 21 33 / 37.5
41% / 55% 55% / 39%
6 months p=0.02 p=0.009

Cohort studies
Study Rhythm Sample Down Temp Follow up Mortality CPC 1-2
TH/control time TH/control Assessment TH/control TH/control
(n) (min) (C) (%) (%)

Bernard11 Shockable 22 / 22 25.2 32.8 / 36.9 Discharge 45 / 77.2 50 / 13.6

1997 p=0.03 p=0.014
Oddo26 Shockable 43 / 43 23.1 33.1 / 37.4 Discharge 44.2 / 60.5 55.8 / 25.6
2006 p=0.28 p=0.004
Non-shockable 12 / 11 34.6 33 / 37.4 Discharge 83 / 90.9 16.6 / 0
p=0.595 p=0.29
Derwall27 Both 37 / 31 NA 33.4 / 37.5 Discharge 21.6 / 45.2 56.8 / 45.2
2009 Shockable 67.6 % p=0.067 p=0.341
Non-shockable 32.4%

Ferreira Both 49 / 26 NA 33 / 36.5 Discharge 32.7 / 57.7 51 / 19.2

2009 Shockable 93.7% p=0.04 p=0.008
Non-shockable 6.1%

Petrovic29 Both 45 / 37 26.9 33.3 / NA Discharge 44.4 / 81.1 46.7 / 18.9

2011 Shockable 60% p=0.006 p=0.006
Non-shockable 40%

Walters30 Both 29 / 26 NA 33 / 37 Discharge 55.2 /. 69.2 31 / 12

2011 Shockable 27.6% p=0.29 p=0.08
Non-shockable 72.4%

Don31 Shockable 81 / 93 NA 34.3 / 36.1 Discharge 45.7 / 61.3 34.6 / 15

2009 p=0.04 p=0.003
Non-shockable 122 / 191 NA 34.1 / 36.1 Discharge 78.7 / 80.6 11.5 / 8.9
p=0.67 p=0.45
Belliard32 Shockable 32 / 36 20 35.3 / 37.8 Discharge 44 / 64 72 / 46
2007 p=0.04 p=0.02

Studies were grouped by study designed and sorted in ascending order by cooling temperature.
Temp: temperature, RCT: randomized control trial, NA: not available
Table 2: Pooled descriptive statistics of hypothermia data across temperature groups

Achieved Temperature
32C 33C 34C 35C 36C Total
(N=2) (N=20) (N=10) (N=1) (N=1) (N=34)
N Patients 35 2610 1258 32 466 4373

Mean Age (y) 60 60 61 54 64 60

Male (%) 77 75 63 81 79 73

Shockable rhythm 57 72 63 100 81 70

(%)

Downtime (min) 26 26.2 26.9 20 25 26.1

Cardiac cause (%)

77 74 79 81 100 78
Mortality
N 2 20 10 1 1 34
Mean (SD) 39.0 (22.6) 45.9 (13.3)1 52.0 (25.4) 1 44.0 (.) 44.0 (.) 47.2 (17.5)
Median 39.0 44.8 46.2 44.0 44.0 44.8
Q1, Q3 23.0, 55.0 37.7, 51.7 35.7, 78.7 44.0, 44.0 44.0, 44.0 36.4, 52.5
Range (23.0-55.0) (21.6-83.3) (19.9-91.4) (44.0-44.0) (44.0-44.0) (19.9-91.4)

CPC 1-2
N 1 13 10 1 1 26
2 2
Mean (SD) 7.7 (.) 41.6 (13.2) 35.3 (18.5) 40.6 (.) 46.0 (.) 38.0 (16.0)
Median 7.7 46.7 41.1 40.6 46.0 45.0
Q1, Q3 7.7, 7.7 31.0, 50.0 12.2, 50.0 40.6, 40.6 46.0, 46.0 30.0, 50.0
Range (7.7-7.7) (16.6-56.8) (8.6-55.3) (40.6-40.6) (46.0-46.0) (7.7-56.8)

1
Unequal Variance T-Test contrasting in-hospital mortality between 33C and 34C (p=0.49). 2
Unequal Variance T-Test contrasting good neurological outcome at discharge between 33C and
34C (p=0.38)