a b s t r a c t

Objective: Targeted temperature management (TTM) with Therapeutic hypothermia has been used to improve neurological outcomes in patients after cardiac arrest; however, several trials have reported conflicting results regarding its effectiveness. This systematic review and meta-analysis assessed whether TH was associated with better Survival and neurological outcomes after cardiac arrest.

Method: We searched online databases for relevant studies published before May 2023. Randomized controlled trials comparing TH and normothermia in post-cardiac-arrest patients were selected. Neurological out- comes and all-cause mortality were assessed as the primary and secondary outcomes, respectively. A subgroup analysis according to initial Electrocardiography rhythm was performed.

Result: Nine RCTs (4058 patients) were included. The Neurological prognosis was significantly better in patients with an Initial shockable rhythm after cardiac arrest (RR = 0.87, 95% confidence interval [CI] = 0.76-0.99, P = 0.04), especially in those with earlier TH initiation (<120 min) and prolonged TH duration (>=24 h). However, the mortality rate after TH was not lower than that after normothermia (RR = 0.91, 95% CI = 0.79-1.05). In patients with an initial nonshockable rhythm, TH did not provide significantly more neurological or Survival benefits (RR = 0.98, 95% CI = 0.93-1.03 and RR = 1.00, 95% CI = 0.95-1.05, respectively).

Conclusion: Current evidence with a moderate level of certainty suggests that TH has potential neurological bene- fits for patients with an initial shockable rhythm after cardiac arrest, especially in those with faster TH initiation and longer TH maintenance.

(C) 2023

1. Introduction

Sudden cardiac arrest is a medical emergency, which, despite advances in medical treatment, continues to have a high mortality rate. According to the “Heart Disease and Stroke Statistics-2022 Update” by the American Heart Association, the incidence rate of out-of-

* Corresponding author at: Department of Neurology, Taipei Medical University, Shuang Ho Hospital, 291 Zhongzheng Road, Zhonghe District, New Taipei City 23561, Taiwan.

E-mail address: [email protected] (Y.-C. Kuan).

¹ These two authors contributed equally to this work.

hospital cardiac arrest was 140.7 individuals per 100,000 populations, and only 10% of these patients survived to discharge [1]. Among those who survived to hospital discharge, the majority of survivors remained neurologically impaired, costing considerable medical and social resources [2].

The concept of therapeutic hypothermia (TH) was introduced in the 1990s [3]. TH involves several mechanisms that may provide cerebroprotective effects after cardiac arrest, including the following:

(1) reducing toxicity from excessive glutamate, (2) reducing the cere- bral metabolic rate, (3) preventing intracellular calcium overload,

(4) maintaining Protein Synthesis, and (5) reducing oxidative stress from free radicals [4-6]. Two randomized control trials (RCTs) published

https://doi.org/10.1016/j.ajem.2023.06.040

0735-6757/(C) 2023

in 2002 revealed significant neurological or survival benefits of TH, and its adoption for patients after cardiac arrest was widely discussed [7,8]. However, the populations and interventions in these studies were heterogeneous, and the results of several other trials have been inconsistent [9-12]. The 2021 Targeted Temperature Management-2 (TTM-2) trial revealed no significant neurological or survival benefits of TH in patients after cardiac arrest [11]. Given the controversy over the Clinical effectiveness of TH, we performed an updated systematic re- view and meta-analysis of RCTs to investigate whether targeted tem- perature management (TTM) with TH benefits patient neurological and survival outcomes after cardiac arrest.

1. Materials and methods

Our systematic review was conducted in accordance with the 2020 PRISMA statement. Our review protocol was registered on PROSPERO (CRD42021268939). Ethics approval is not required for this meta- analysis because informed consent of included trials has already been obtained by the trial investigators.

• 1. Selection criteria

We reviewed studies evaluating the efficacy of TH for cardiac arrest and included RCTs comparing patients who received TH and normo- thermia after cardiac arrest. TH was defined as TTM with a targeted temperature of 32-34 ?C, and normothermia was defined as a targeted temperature of 36-37.8 ?C. We included studies that (1) reported neurological or Survival outcomes of TH and normothermia groups,

(2) enrolled adult patients (aged >=18 years) who remained unrespon- sive to stimulation after the return of spontaneous circulation (ROSC), and (3) performed TH in hospital units and not in a prehospital setting. We excluded studies that (1) conducted secondary analyses in patients from other trials, (2) involved a nonrandomized control cohort, and

(3) applied TH for <4 h.

• 1. Search strategy and study selection

We searched PubMed, Embase, Cochrane Library, and ClinicalTrials. gov databases without language restriction for studies published from their inception to May 2023 by using the following search terms: ((hypothermia) OR (targeted temperature management)) AND (cardiac arrest). Two authors (PYC and YCK) independently screened ti- tles and abstracts and then reviewed the full texts of potentially eligible studies. A third author (CCC) was consulted if any discrepancies occurred.

• 1. Data extraction

Two authors (PYC and YCK) independently extracted the informa- tion from the included studies by using a predefined data extraction form. Data including age and number of participants; eligibility criteria; time spans between cardiac arrest, ROSC, and TH initiation; intervention method; and outcomes of interest were collected. Any disagreements were resolved by consulting with the third author (CCC).

• 1. Bias assessment“>Risk of bias assessment

Two authors (PYC and YCK) assessed the included studies indepen- dently by using the Cochrane Risk of Bias 2 tool [12]. Five domains of risk were assessed: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of outcome, and (5) selection of reported result. Each domain was graded as having low risk, some concerns, or high risk of bias. To detect potential publication bias, funnel plot and egger’s test would be

performed if >10 studies were enrolled [13]. Any disagreements were resolved through mutual discussion and consensus of all authors.

• 1. Outcome assessment

The primary outcome was poor neurological outcome, which was defined as either (1) a Cerebral performance category score of 3 (severe disability), 4 (coma or vegetative state), or 5 (brain death) or

(2) a Modified Rankin scale (mRS) score of 4 (moderately severe disabil- ity), 5 (severe disability), or 6 (death). Secondary outcomes were all- cause mortality and incidence of adverse events during hospitalization.

• 1. Subgroup analysis

Because prognosis after cardiac arrest was better in patients with an initial shockable rhythm than those with a nonshockable rhythm [14-16], we performed subgroup analyses according to initial rhythm, including shockable rhythm (ventricular tachycardia and ventricular fibrillation) and nonshockable rhythm (pulseless electrical activity and asystole).

Additional subgroup analyses were conducted based on the time from cardiac arrest to TH initiation (>=120 min; <120 min), duration of TH maintenance (>=24 h; < 24 h), cardiac arrest location (in-hospital; out-of-hospital), and the last postintervention follow-up (at discharge; at 6 months). Some studies suggested that earlier initiation of TH was associated with better outcome [17,18]. Though the optimal timing was elusive, we set 120 min as cut point because the median time from cardiac arrest to TH initiation of included studies was about 120 min. Duration of TH maintenance for at least 24 h was suggested by current guideline [19]. Earlier study also indicated that it required 24 h of TH maintenance to provide benefit of cerebralprotection [6]. We also implemented subgroup analyses for cardiac arrest location and the time of postintervention follow-up to assess the effects of these factors on the outcomes. Additional subgroup analyses were con- ducted for the time from cardiac arrest to TH initiation, the duration of TH maintenance (>=24 h or <24 h), the cardiac arrest location (in hospi- tal or out of hospital), and the time of last postintervention follow-up.

• 1. Statistical analysis

All statistical analyses were conducted using Review Manager, version 5.4.1 (Cochrane Collaboration, Oxford, UK). Dichotomous outcomes were analyzed using risk ratios (RRs). All effect sizes were reported with 95% confidence intervals (CIs). The pooled risk ratio (RR) was calculated using a random-effects model with the Mantel-Haenszel method. Statistical significance was set at P < 0.05. The I square statistic (I²) was calculated to assess between-study heterogeneity.

A trial sequential analysis (TSA) was performed with the calculation of the required information size (RIS) to detect a 20% relative risk reduc- tion in poor neurological outcome and mortality in the hypothermia group. All monitoring boundaries were defined as two-sided with type I error of 5%, and power of 80%. The DerSimonian and Laird random- effects model was applied using TSA software, version 0.9.5.10 beta (Copenhagen Trial Unit, Copenhagen, Denmark) [20,21].

• 1. Quality of evidence

Two authors (PYC and YCK) assessed the quality of each outcome independently, using the Grading of Recommendations, Assessment, Development and Evaluation approach. The aspects of evidence certainty assessment included (1) risk of bias, (2) inconsistency, (3) indirectness,

(4) imprecision, and (5) other considerations, and the overall certainty was rated as “high,” “moderate,” “low,” or “very low.”

1. Results
   1. Search strategy

The initial search strategy yielded 23,697 results. After excluding duplicate studies, 17,546 studies were retained. After screening the titles and abstracts, 17,529 studies were excluded. Of the remaining 17 studies, eight were excluded after full-text review for the following reasons: nonrandomized studies (n = 4), review article (n = 1), and substudy of other trials (n = 3). The remaining nine studies were en- rolled for the analysis of outcomes of interest (Fig. 1) [7-12,22-24].

• 1. Characteristics of the included trials

The characteristics of the nine included RCT studies are summarized in Table S1. A total of 2034 patients were enrolled in the TH group and 2024 patients in the normothermia group. Seven trials predominantly en- rolled patients of out-of-hospital cardiac arrest (OHCA) [7-9,11,22-24], the HYPERION study [10] included patients of both in-hospital cardiac ar- rest (IHCA) and OHCA, and Hypothermia after Cardiac Arrest in-hospital (HACA in-hospital) study enrolled patients of IHCA exclusively. Two of the trials included patients with an initial shockable rhythm [7,23], two included patients with an initial nonshockable rhythm [10,22], and the remaining five included both types of patient [8,9,11,12,24]. The median

time from cardiac arrest to TH initiation was <120 min in three studies [8,22,23], >120 min in three [10-12], and unreported in the remaining three [7,9,24]. The average time from cardiac arrest to TH initiation of in the included studies was about approximately 120 min. Targeted temper- ature in the TH groups was 32-34 ?C in all the included studies. The con- trol group used normothermia witha temperature of 36-37.8 ?C in eight studies [7,8,10-12,22-24], whereas one study used a target temperature of 36 ?C in the control group [9]. One study reported mortality only [24], and the remaining eight reported both neurological outcomes and mortality [7-12,22,23].

• 1. Risk of bias assessment

Some risk of bias was observed in all nine trials (Table 1). One study used an odd or even number of dates for allocation (not randomly) and did not report the underlying comorbidities at the baseline for both groups (confounding), resulting in some risk of bias in the randomiza- tion process [7]. Because blinding to medical personnel is not feasible for therapeutic temperature therapy, there may have been unbalanced care and management beyond the protocol between active and control groups in the studies. No missing data were observed in any study, and outcome assessors were unaware of the assignment in eight studies (one study did not provide information on this) [22]. In studies before 2013, published protocols or registrations were lacking [7,8,22-24].

Fig. 1. PRISMA flowchart of study selection.

Table 1

Risk of bias assessment of the included studies.

Study	Randomization process	Deviations from intended intervention	Missing outcome data	Measurement of the outcome	Selection of reporting	Overall
Hachimi, 2001	Low	Some concernb	Low	Lowc	Some concernd	Some concernc
HACA, 2002	Low	Some concernb	Low	Low	Some concernd	Some concern
Bernard, 2002	Some concerna	Some concernb	Low	Low	Some concernd	High
Laurent, 2005	Low	Some concernb	Low	Low	Some concernd	Some concern
Hachimi, 2005	Low	Some concernb	Low	Low	Some concernd	Some concern
TTM, 2013	Low	Some concernb	Low	Low	Low	Some concern
HYPERION 2019	Low	Some concernb	Low	Low	Low	Some concern
Dankiewicz, 2021	Low	Some concernb	Low	Low	Low	Some concern
HACA in-hospital 2022	Low	Some concernb	Low	Low	Low	Some concern

^a Allocation based upon even or odd number of date of admission and unknown underlying comorbidities of both groups.

^b Medical personnel not blinded.

^c No information of blinding outcome assessor, so assessment of neurological outcome may be influenced by knowledge of intervention.

^d Neither published protocol nor registration available.

Two unpublished trials were registered on ClinicalTrials.gov. One trial (NCT01617291) was terminated because similar study was published with definitive result, and the other trial (NCT02578823) was not pub- lished with unknown status. Both unpublished trials might result in po- tential publication bias, while the actual effect could not be estimated accurately.

• 1. Primary outcome

Neurological outcomes were assessed at discharge in four studies [7,8,12,22], 90 days in one study [10], and 6 months in five studies [8,9,11,12,23]. Five studies reported neurological outcomes for patients

with an initial shockable rhythm [7-9,11,23], which demonstrated a sig- nificant neurological outcome benefit (RR = 0.87, 95% CI = 0.76-0.99), whereas no significant neurological benefit was observed in patients with an initial nonshockable rhythm in four studies (RR = 0.98, 95% CI = 0.93-1.03; Fig. 2). Furthermore, in patients with an initial shock- able rhythm, better neurological outcomes after TH were observed in the studies with a shorter median time from arrest to TH initiation (<120 min; RR = 0.74, 95% CI = 0.60-0.91), longer TH duration (>=24 h; RR = 0.74, 95% CI = 0.60-0.91), and the time of evaluation at discharge (RR = 0.75, 95% CI = 0.63-0.89; Fig. S1). However, in patients with an initial nonshockable rhythm, there was still no significant ben- efit in neurological outcome between the two groups under subgroup

Fig. 2. Forest plot for poor neurological outcome after therapeutic hypothermia compared with normothermia. Subgroup analysis according to the initial cardiac rhythm: shockable and nonshockable.

analysis (Fig. S2). The subgroup analysis for cardiac arrest location re- vealed no significant difference in neurological outcomes between IHCA and OHCA (Fig. S5-A).

• 1. Secondary outcome

Mortality was assessed at discharge in four studies [7,8,12,22], at 90 days in one [10], and at 6 months in six [8,9,11,12,23,24]. Compared with normothermia, TH was not associated with significantly more im- provement in mortality in patients with an initial shockable (RR = 0.91, 95% CI = 0.79-1.05) or nonshockable (RR = 1.00, 95% CI = 0.95-1.05)

rhythm (Fig. 3). In patients with an initial shockable rhythm, studies reporting a shorter median time from arrest to TH initiation (<120 min; RR = 0.74, 95% CI = 0.59-0.94), longer TH duration (>= 24 h; RR = 0.74, 95% CI = 0.59-0.94), and the time of evaluation at discharge (RR = 0.74, 95% CI = 0.6-0.93; Fig. S3) revealed a lower mor- tality rate in the TH group than in the normothermia group. For patients with an initial nonshockable rhythm, no significant between-group difference in survival was observed in the subgroup analysis (Fig. S4). The subgroup analysis for cardiac arrest location indicated no significant difference in survival between IHCA and OHCA (Fig. S5-B).

Regarding adverse events, five studies reported the incidence of bleeding, arrhythmia, and pneumonia [8-12], with no significance differences observed between groups (Fig. 4).

• 1. TSA

The cumulative Z-curve of the poor neurological outcome for pa- tients with an initial shockable rhythm crossed the conventional bound- ary and reached the RIS (Fig. S6-A), supporting a 20% relative risk reduction by using TH. However, the cumulative Z-curve of poor

neurological outcome for patients with an initial nonshockable rhythm and mortality for patients with an initial shockable or nonshockable rhythm did not cross the conventional boundary or the TSA boundary, but it did cross the futility boundary and reached the RIS at a 20% relative risk reduction (Fig. S6-B-D). This supported the findings of the conventional meta-analysis and also indicated that future trials are unlikely to provide a 20% relative risk reduction in mortality for patients with an initial shockable rhythm and in both poor neurological outcome and mortality for patients with an initial nonshockable rhythm.

• 1. Quality of evidence

The certainty of evidence of the neurological outcome and mortality in patients with an initial shockable or nonshockable rhythm was moderate (Table S2).

1. Discussion

Our findings revealed that TH led to significantly better neurological outcomes than normothermia in patients with an initial shockable rhythm. This benefit was more evident when the time from arrest to TH initiation was <120 min, TH duration was >24 h, or the assessment was at discharge but not at 6 months after TH. However, no significant benefit of TH on neurological or survival outcomes was observed in patients with an initial nonshockable rhythm.

Only two studies from 2002 (Bernard et al. [7] and the Hypothermia after Cardiac Arrest Study [HACA] trial [8]) have primarily demon- strated the benefits from TH in patients with an initial shockable rhythm. Four larger RCTs after 2013 concluded that TH did not significantly improve neurological or survival outcomes in patients with cardiac arrest, regardless of the initial cardiac rhythm [9-12].

Fig. 3. Forest plot for mortality after therapeutic hypothermia compared with normothermia. Subgroup analysis according to the initial cardiac rhythm: shockable and nonshockable.

Fig. 4. Forest plot for adverse events after therapeutic hypothermia compared with normothermia. Adverse events of bleeding, arrhythmia, and pneumonia analyzed.

Some differences exist between these two groups of RCTs. First, the baseline characteristics of the participants differed across the studies, which may have led to confounding effects. The study by Bernard et al. was not completely randomized and did not provide information on patients’ comorbidities [7]. In a large observational study, patients with diabetes and chronic pulmonary obstruction had significantly poorer outcomes after suffering from cardiac arrest [16]. Second, the earlier studies did not clearly describe whether body temperature was actively managed in the normothermia group, and the median temper- ature of the normothermia group was above 37 ?C [7,8], which was higher than that in other studies. For example, in the HYPERION and TTM2 studies, the normothermia group received active fever avoidance intervention [10,11]. Third, external cooling methods were used in the two 2002 studies [7,8], whereas internal cooling methods were used in more than half of the patients in more recent studies [9-12,24]. Some clinical trials have indicated that internal cooling provides no ad- ditional benefits compared with external cooling [25-27]. However, a meta-analysis including RCTs and observational studies indicated that compared with external cooling, internal cooling improved neurological outcomes among cardiac arrest survivors [28]. However, non-RCTs may have confounding or selection bias. Therefore, whether cooling methods affect the prognosis after cardiac arrest remains unclear. Finally, the mortality rates in normothermia groups with an initial shockable rhythm were 68% and 55% in the 2002 studies and were lower at 40% and 38% in studies published in 2013 and 2021, respec- tively [7-9,11], consistent with data released by the American Heart As- sociation indicating that age-adjusted Death rates for all-cause cardiac arrest declined by 30% from 1999 to 2018 [1]. The lower mortality rate might be associated with improvements in the ambulance system, the

promotion of CPCR, and advances in intensive care in both groups [29,30]. The aforementioned conditions may have influenced the effects of TH and normothermia.

In the early studies, TH was defined as maintaining a core body temperature of 32-34 ?C [7,8]. After Bernard et al. and the HACA study indicated that hypothermia was beneficial for neurological and mortal- ity outcomes in cardiac arrest patients, the 2005 European and American guidelines recommended this therapy as the standard of care [31]. The TTM trial revealed no significant prognostic difference with a target temperature of 32 ?C or 36 ?C; accordingly, the updated guidelines recommended a target temperature of 32-36 ?C [9,32]. The TTM2 study further indicated that compared with maintaining physio- logical normothermia, maintaining hypothermia at 32-34 ?C did not provide a significantly better prognosis [11]. Thus, the current guidelines for cardiac arrest recommend only active fever prevention in patients with ROSC after cardiac arrest while also mentioning that the benefits of TH over fever prevention in specific subpopulations re- main unexplored [19].

Our subgroup analysis, however, indicated a benefit for neurological outcomes in patients with an initial shockable rhythm, especially when the time from arrest to TH initiation was <120 min and the TH duration was >24 h. The average time from ROSC to achieve the target tempera- ture in the included trials was 4-8h [7-12,22-24]. However, evidence of the optimal time of initiating, achieving, and maintaining hypothermia remains inadequate, although a multicenter clinical trial comparing the duration of TH from a minimum of 6 h to a maximum of 72 h in co- matose cardiac arrest survivors to identify the optimal period of induced hypothermia for neuroprotection is ongoing [ClinicalTrials.Gov: NCT04217551].

The time from cardiac arrest to ROSC and patient age are also critical prognostic factors in patients with cardiac arrest. A longer time to ROSC and older age were significantly associated with poorer prognosis [29,30]. Nevertheless, stratified analyses of time to ROSC and age did not indicate different outcomes in TTM2 and TTM studies [9,11]. Although no single variable altered the efficacy of hypothermia in the subgroup analyses of clinical trials, the prognosis of cardiac arrest may be influenced by multiple factors. Studies have proposed different pre- dictive systems and included factors such as prehospital ROSC, age, ini- tial cardiac rhythm, and comorbidities to assess the prognosis of patients with cardiac arrest or receiving TH; these factors have accept- able predictive value [33-36]. Following the development of a clinically validated prediction system, multivariate subgroup analysis may help identify certain patients who benefit from TH.

Several studies had reviewed this issue in 2021 and 2022 [37-40]. Our study provides an updated systematic review and meta-analysis with the most recent study included [12], an additional TSA of the efficacy of TH after cardiac arrest, and focus on subgroup who benefits from TH. However, this study also has some limitations. First, blinding medical personnel is impossible in clinical studies of TH, and earlier studies’ lack of registration may have resulted in some risk of bias. Second, considerable between-study heterogeneity was noted in study populations, cooling methods, and intensive care protocols. To over- come this limitation, we performed a subgroup analysis of different ini- tial cardiac rhythms, the timing of hypothermia induction, and the duration of maintaining the targeted temperature. Third, the unavail- ability of Individual patient data makes it challenging to characterize patients who are likely to benefit from TH. Further studies based on var- ious subpopulations are required to provide more information.

1. Conclusion

Our meta-analysis indicated with a moderate certainty of evidence that TH has potential neurological benefits for patients with an initial shockable rhythm after cardiac arrest, especially in those with faster TH initiation and longer TH maintenance. However, no significant ben- efit of TH was found in patients with an initial nonshockable rhythm. The answer to whether TH should be used in certain subpopulations of patients with an initial shockable rhythm remains elusive and re- quires further research.

CRediT authorship contribution statement

Po-Yun Chiu: Writing - original draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation. Chen-Chih Chung: Supervision, Software, Formal analysis. Yu-Kang Tu: Visualiza- tion, Supervision, Formal analysis. Chien-Hua Tseng: Writing - review & editing, Visualization, Supervision, Conceptualization. Yi-Chun Kuan: Writing - review & editing, Visualization, Supervision, Software, Methodology, Investigation, Formal analysis, Conceptualization.

Declaration of Competing Interest

The study did not receive financial support from any institution or organization. All the authors have no potential conflicts of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2023.06.040.

References

1. Tsao CW, Aday AW, Almarzooq ZI, et al. Heart disease and stroke statistics-2022 up- date: a report from the American Heart Association. Circulation. 2022;145(8): e153-639.
2. Damluji AA, Al-Damluji MS, Pomenti S, et al. Health care costs after cardiac arrest in the United States. Circ Arrhythm Electrophysiol. 2018;11(4):e005689.
3. Sterz F, Safar P, Tisherman S, et al. Mild hypothermic cardiopulmonary resuscitation improves outcome after prolonged cardiac arrest in dogs. Crit Care Med. 1991;19 (3):379-89.
4. Ginsberg MD, Sternau LL, Globus MY, et al. Therapeutic modulation of brain temper- ature: relevance to ischemic brain injury. Cerebrovasc Brain Metab Rev. 1992;4(3): 189-225.
5. Coimbra C, Wieloch T. Moderate hypothermia mitigates neuronal damage in the rat brain when initiated several hours following transient cerebral ischemia. Acta Neuropathol. 1994;87(4):325-31.
6. Colbourne F, Sutherland G, Corbett D, Postischemic hypothermia.. A critical appraisal with implications for clinical treatment. Mol Neurobiol. 1997;14(3):171-201.
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(8): 557-63.
8. Hypothermia after Cardiac Arrest Study G. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med. 2002;346(8):549-56.
9. Nielsen N, Wetterslev J, Friberg H, et al. Targeted temperature management after cardiac arrest. N Engl J Med. 2014;370(14):1360.
10. Lascarrou JB, Merdji H, Le Gouge A, et al. Targeted temperature Management for Cardiac Arrest with nonshockable rhythm. N Engl J Med. 2019;381(24):2327-37.
11. Dankiewicz J, Cronberg T, Lilja G, et al. Hypothermia versus Normothermia after out- of-hospital cardiac arrest. N Engl J Med. 2021;384(24):2283-94.
12. Wolfrum S, Roedl K, Hanebutte A, et al. Temperature control after in-hospital cardiac arrest: a randomized clinical trial. Circulation. 2022;146(18):1357-66.
13. Sterne JA, Sutton AJ, Ioannidis JP, et al. Recommendations for examining and inter- preting funnel plot asymmetry in meta-analyses of Randomised controlled trials. BMJ. 2011;343:d4002.
14. Wibrandt I, Norsted K, Schmidt H, et al. Predictors for outcome among cardiac arrest patients: the importance of initial cardiac arrest rhythm versUS time to return of spontaneous circulation, a retrospective cohort study. BMC Emerg Med. 2015;15:3.
15. Choi SW, Shin SD, Ro YS, et al. Effect of therapeutic hypothermia on the outcomes after out-of-hospital cardiac arrest according to initial ECG rhythm and witnessed status: a nationwide observational interaction analysis. Resuscitation. 2016;100: 51-9.
16. Johnsson J, Wahlstrom J, Dankiewicz J, et al. functional outcomes associated with varying levels of targeted temperature management after out-of-hospital cardiac ar- rest - an INTCAR2 registry analysis. Resuscitation. 2020;146:229-36.
17. Abella BS, Zhao D, Alvarado J, et al. Intra-arrest cooling improves outcomes in a murine cardiac arrest model. Circulation. 2004;109(22):2786-91.
18. Nozari A, Safar P, Stezoski SW, et al. Critical time window for intra-arrest cooling with cold saline flush in a dog model of cardiopulmonary resuscitation. Circulation. 2006;113(23):2690-6.
19. Sandroni C, Nolan JP, Andersen LW, et al. ERC-ESICM guidelines on temperature con- trol after cardiac arrest in adults. Intensive Care Med. 2022;48(3):261-9.
20. Gartlehner G, Nussbaumer-Streit B, Wagner G, et al. Increased risks for random errors are common in outcomes graded as high certainty of evidence. J Clin Epidemiol. 2019;106:50-9.
21. Claire R, Gluud C, Berlin I, et al. Using trial sequential analysis for estimating the sam- ple sizes of further trials: example using smoking cessation intervention. BMC Med Res Methodol. 2020;20(1):284.
22. Hachimi-Idrissi S, Corne L, Ebinger G, et al. Mild hypothermia induced by a helmet device: a clinical feasibility study. Resuscitation. 2001;51(3):275-81.
23. Hachimi-Idrissi S, Zizi M, Nguyen DN, et al. The evolution of serum astroglial S-100 beta protein in patients with cardiac arrest treated with mild hypothermia. Resusci- tation. 2005;64(2):187-92.
24. Laurent I, Adrie C, Vinsonneau C, et al. High-volume hemofiltration after out-of- hospital cardiac arrest: a randomized study. J Am Coll Cardiol. 2005;46(3):432-7.
25. Pittl U, Schratter A, Desch S, et al. Invasive versus non-invasive cooling after in- and out-of-hospital cardiac arrest: a randomized trial. Clin Res Cardiol. 2013;102(8): 607-14.
26. Deye N, Cariou A, Girardie P, et al. Endovascular versus external targeted tempera- ture management for patients with out-of-hospital cardiac arrest: a randomized, controlled study. Circulation. 2015;132(3):182-93.
27. Look X, Li H, Ng M, et al. Randomized controlled trial of internal and external targeted temperature management methods in post- cardiac arrest patients. Am J Emerg Med. 2018;36(1):66-72.
28. Bartlett ES, Valenzuela T, Idris A, et al. Systematic review and meta-analysis of intra- vascular temperature management vs. Surface cooling in comatose patients resusci- tated from cardiac arrest. Resuscitation. 2020;146:82-95.
29. Wissenberg M, Lippert FK, Folke F, et al. Association of national initiatives to improve Cardiac arrest management with rates of bystander intervention and patient survival after out-of-hospital cardiac arrest. JAMA. 2013;310(13):1377-84.
30. Stromsoe A, Svensson L, Axelsson AB, et al. Improved outcome in Sweden after out- of-hospital cardiac arrest and possible association with improvements in every link in the Chain of survival. Eur Heart J. 2015;36(14):863-71.
31. Nolan JP, Deakin CD, Soar J, et al. European resuscitation council guidelines for resus- citation 2005. Section 4. Adult advanced life support. Resuscitation. 2005;67(Suppl. 1):S39-86.
32. 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 interna- tional liaison committee on resuscitation and the American Heart Association emer- gency cardiovascular care committee and the council on cardiopulmonary, critical care, perioperative and resuscitation. Circulation. 2015;132(25):2448-56.
33. Chiu WT, Chung CC, Huang CH, et al. Predicting the survivals and favorable neurol- ogic outcomes after targeted temperature management by artificial neural net- works. J Formos Med Assoc. 2021;121(2):490-9.
34. Chung CC, Chiu WT, Huang YH, et al. Identifying prognostic factors and developing accurate outcome predictions for in-hospital cardiac arrest by using artificial neural networks. J Neurol Sci. 2021;425:117445.
35. Wong XY, Ang YK, Li K, et al. Development and validation of the SARICA score to pre- dict survival after return of spontaneous circulation in out of hospital cardiac arrest using an interpretable machine learning framework. Resuscitation. 2022;170: 126-33.
36. Chou SY, Bamodu OA, Chiu WT, et al. artificial neural network-boosted cardiac arrest survival post-resuscitation in-hospital (CASPRI) score accurately predicts outcome in cardiac arrest patients treated with targeted temperature management. Sci Rep. 2022;12(1):7254.
37. Sandroni C, Natalini D, Nolan JP. Temperature control after cardiac arrest. Crit Care. 2022;26(1):361.
38. Granfeldt A, Holmberg MJ, Nolan JP, et al. Targeted temperature management in adult cardiac arrest: systematic review and meta-analysis. Resuscitation. 2021; 167:160-72.
39. Fernando SM, Di Santo P, Sadeghirad B, et al. Targeted temperature management fol- lowing out-of-hospital cardiac arrest: a systematic review and network meta- analysis of temperature targets. Intensive Care Med. 2021;47(10):1078-88.
40. Elbadawi A, Sedhom R, Baig B, et al. Targeted hypothermia vs targeted normother- mia in survivors of cardiac arrest: a systematic review and meta-analysis of random- ized trials. Am J Med. 2022;135(5):626-633.e4.