Post‐cardiac arrest intensive care in Sweden: A survey of current clinical practice

European guidelines recommend targeted temperature management (TTM) in post‐cardiac arrest care. A large multicentre clinical trial, however, showed no difference in mortality and neurological outcome when comparing hypothermia to normothermia with early treatment of fever. The study results were valid given a strict protocol for the assessment of prognosis using defined neurological examinations. With the current range of recommended TTM temperatures, and applicable neurological examinations, procedures may differ between hospitals and the variation of clinical practice in Sweden is not known.


| INTRODUCTION
Targeted temperature management (TTM) is recommended in comatose survivors after resuscitation from cardiac arrest. 1 The two randomised TTM trials ('TTM trial' and 'TTM2 trial') studied hypothermia at 33 versus 36 C, and 33 C versus normothermia with early treatment of fever in out-of-hospital cardiac arrest (OHCA). 2,3Both trials found similar survival and neurological outcomes in the respective temperature groups, also when pooled in a meta-analysis defining normothermia as 36.0-37.7 C. 4 In contrast, the HYPERION trial, including both OHCA and in-hospital cardiac arrest found a better neurological outcome in patients treated with 33 as compared to 37 C. 5 The updated European Resuscitation Council and European Society of Intensive Care Medicine (ESICM) guidelines recommend to actively prevent fever (core temperature >37.7 C) targeting 37.5 C but state that evidence for targets between 32 and 36 C are insufficient for recommendation for or against. 6However, given the equipoise between hypothermia and normothermia in OHCA and the wide range of target temperatures having been recommended before the last guideline update, current practice may vary.Also, a lack of positive effect from hypothermia in the first TTM trial may have been falsely interpreted as a lack of need for any TTM in everyday practice. 7,8In Sweden, a nationwide registry study found decreasing use of TTM during 2010-2015, suggesting that the TTM approach was partly abandoned with a lower adherence to guidelines as consequence. 9The current application of target temperature management practice in post-cardiac arrest care in Sweden has not been investigated.
For the assessment of neurological prognosis in survivors after cardiac arrest remaining comatose, a multimodal approach is recommended. 1For a robust assessment of the prognosis, the neurological function of the patient should be examined using several different methods.However, resources and availability of such methods may vary between hospitals.The guidelines do not specify a minimum of methods 1 to be applied and the current clinical practice in Sweden is unknown.
The aim of the present survey was to describe current practice in post-resuscitation care after cardiac arrest in Swedish intensive care units (ICUs) as to TTM target temperature and assessment of neurological prognosis.To spare responders from repeated contacts and because standard operating procedures (SOPs) for post-resuscitation care normally contain both TTM targets and guidance for the assessment of neurological prognosis after cardiac arrest, both research questions were chosen for one survey and one report.

| METHODS
The study did not include any patient related or personal sensitive data and was not reviewed by the Swedish Ethical Review Authority.
Respondents were invited to participate in the survey in their professional capacity for their respective site.Normothermia was defined as 36.0-37.7 C in line with a recent meta-analysis. 4 Data on number of admitted patients in 2022 were collected from the open data portal of the Swedish Intensive Care Registry. 10structured telephone survey was conducted in all (= 53) Swedish Levels 2 and 3 ICUs during April and May 2022.A Level 2 ICU corresponds to a general ICU able to handle multi-organ failure patients.A Level 3 unit is a specialised ICU, such as thoracic-or neuro-ICU.Some respondents were not available for a telephone interview and responded to an e-mail questionnaire instead.The questions used were identical.Unit telephone numbers were found using the hospitals respective public websites, or the regional hospital telephone switchboard.Unclear answers regarding technical method for application of TTM from 10 units were clarified via a follow-up telephone call during November 2022.The primary survey consisted of 23 questions; 4 on background data; 5 on OHCA; 10 on TTM in OHCA patients; and 4 on assessment of neurological prognosis (see Appendix).The questions were essentially based on previous questionnaire studies to ease comparison.[11][12][13] To clarify uncertainties that appeared in the first analysis, four additional questions were addressed to all sites in April 2023 regarding target temperature, temperature trigger for treatment, date of routine update and thoughts on national guidelines.All answers were registered in an online form by the interviewer (Google Forms, Google LLC, Mountain View, CA, USA) and data saved locally.Analysis was made using Excel (Version 16.61, Microsoft, Redmont, Washington, DC, USA).

| RESULTS
From the 53 ICUs contacted, 5 were excluded as they did not manage post-cardiac arrest patients.Of the eligible 48 units, 43 responded to the survey (90%).The hospitals that did not to participate consisted of one university hospital, and four local hospitals.Survey results are found in Table 1.Most ICUs were 5-10 bed units.The post-cardiac arrest admission rate varied between units and ranged from 0 to 70 admissions, with a total of 873 admissions in 2022 according to the Swedish Intensive Care Registry. 10

| Targeted temperature management
All 43 responding units used TTM in post-cardiac arrest care, and all used a predefined post-resuscitation care protocol (100%).A surface cooling device was used in 38 (88%) units while an endovascular device was used in 3 (7%).A feedback device was available in 36 (84%) units.
In April of 2022, 30/43 (70%) sites had changed temperature target as a result of the TTM2 trial publication. 2The sites that did not either already targeted 36-37.7 C (four sites), were on their way to change to 36-37.7 C target temperature (five site), or had a wellestablished SOP with 36 C target temperature.When surveyed again in April 2023, all sites used variations of normothermia (36-37.7 C and <37.8 C).Median time from the publication of the TTM2 trial to SOP change was 6 months, ranging from 0 (immediate) to 18 months.

| Assessment of neurological prognosis
Of the 43 responding ICUs, 38 (88%) reported that they used a detailed SOP for the assessment of neurological prognosis.These sites admitted a total of 835 OHCA cases in 2022.Results are found in Table 2.The time for the assessment ranged from 24 to >96 h after return of spontaneous circulation (ROSC) with 72-96 h being most common, managed by a neurologist and an intensivist together.

| Attitudes toward national guidelines
Almost all senior consultants/chief physicians who commented (no.= 38) were positive to nation-wide treatment recommendations.
Three were uncertain of its value, and two reserved against such T A B L E 1 Survey results from (n = 43) Levels 2 and 3 intensive care units in Sweden.
There were previous surveys investigating TTM routine and postcardiac arrest care conducted in Swedish, Nordic, and European ICUs.
Regarding temperature management, a survey of Nordic TTM use found most centres using a predefined protocol for TTM (59% Nordic, 49% Sweden, 2015). 11Another survey among ESICM members found 89% of respondents having predefined SOPs for OCHA patients in 2012. 12In the present study, 100% had predefined TTM SOPs, applied to most patients regardless of initial rhythm (92%).In a second questionnaire, after TTM1 was published, 33% of Nordic sites had changed to 36 C target temperature, 11 also similar to our findings.In contrast to our results, a Swedish nationwide registry study found decreasing number of patients registered as receiving 'active hypothermia', defined as 33 or 36 C, after the publication of the TTM1 trial 3 ; while the rate of VT/VF patients receiving TTM 33 or 36 C were about 70% from 2010 to 2013, they decreased to around 55% in 2014-2015.However, survival was constant around 47%. 9 With regard to our findings on temperature targets, this might be due to the use of a normothermic approach instead.
Similar investigations were conducted in Australia, New Zealand, and the United Kingdom.The use of TTM and patient outcomes were studied before and after the publication of the TTM trial. 3The main findings included a higher average lowest temperature, along with an increased frequency of fever (>38 C) and a decline in rate of change in in-hospital mortality in the post-TTM trial cohort. 8,14,15In a US study, the authors found a trend of lower TTM usage, along with a trend toward lower overall risk-adjusted survival.However, the change in TTM use did not consistently explain the lower survival rate in the mediation analysis. 7It is possible that a local change in clinical routine may cause indirect unintentional effects such as increased fever rates and a lower compliance to higher temperature targets (36 C). 16,17 It is, however, unlikely that the variations found in the studies mentioned above were due to local, random events or changes in the translation of knowledge to clinical practice.Rather we believe the ICU community as a whole has made similar changes when adopting higher temperature targets.
When considering the results as to SOP protocols for the assessment of neurological prognosis, one must bear in mind that guidelines have changed during the last decade.Clinicians reported a high frequency of unreliable prognostic evaluations (30%) in the European survey, 12 while a recent evaluation of the guideline recommended algorithm found high specificity (100%) but a lower sensitivity (38%). 18In a previous survey of European post-cardiac arrest care, only one in two sites had a predefined routine for the assessment of neurological prognosis, 13 as opposed to 89% in our study.The use of EEG and somatosensory evoked potentials (SSEP) was relatively uncommon in the Nordic survey in 2012 (66% and 34%, respectively), 11 while our results reveal a much higher availability in T A B L E 2 Survey results regarding assessment of neurological prognosis in units with a detailed routine for assessment of neurological prognosis (n = 38).

Time of assessment of neurological prognosis (h)
• 24-48 4 (11) • 48-72 0 • 72-96 32 (84) Neurological examination performed by • Neurologist and intensive care physician together 24 (63) • Neurologist 4 (11) • Intensive care physician 9 (24) Sweden today (98% and 53%, respectively).Our results indicate a possible shift of the timing of the assessment of prognosis.In a European survey, 60% of assessments were made >72 h after ROSC and 30% of assessments earlier (24-72 h) in 2015. 13The corresponding numbers in our survey were 84% (>72 h) and 11% (24-72 h).We also found a high rate of co-assessment (63%) between intensivists and neurologists in our material, to our knowledge, not previously described.These findings can be interpreted as reassuring for post cardiac arrest patient care to be systematic, avoiding early pitfalls in neurological assessment.

| Limitations
First, the response rate was not 100% and one of the eligible university units declined to participate.Our use of a mixed methods for the interviews increased the response rate but may decrease the overall precision of data.When surveyed via telephone, the responder was asked to recite directly from the SOP if possible.Even if the questions were identical, potential misunderstandings could be better clarified via telephone than via e-mail.We saw several examples of errors and uncertain answers from e-mail responders during both primary and secondary surveys.Some units had listed all possible available technical methods and not the main method used in the unit.Another example is that all units had the possibility to take ice from a freezer to cool the patient but only one unit listed it as the primary method.In the primary survey, the question on temperature target was commonly misinterpreted as the trigger for treatment.It might have yield better data resolution if the question had primarily been divided into (1) temperature trigger of treatment (such as 37. 8), and (2) device temperature setting (such as 37.5), as during the second survey.These examples can act as examples of the inherent weakness of a survey study.

14 .
Did Your unit change target temperature after the results from the TTM1 trial (2013) was published?a. Yes, the temperature was changed to: _______________________ b.No 15.Did Your unit change target temperature after the results from the TTM2 trial (2021) was published?a. Yes, the temperature was changed to: _______________________ b.No 16.If the target temperature is 36 C or less, how long is the cooling phase in Your unit?there a detailed routine for assessment of neurological prognosis after cardiac arrest?a. Yes b.No 21.If the answer was 'yes' in the previous question, how many hours after return of spontaneous circulation is the assessment methods are available in Your unit for assessment of neurological prognosis?a. Electroencephalogram (EEG) b.Somatosensory evoked potential (SSEP) c.Neuron-specific enolase (NSE) d.Neurofilament light-chain e. Computed tomography/magnetic resonance imaging (CT/MRI) 23.Who performs the assessment of neurological prognosis, according to the detailed routine?a. Neurologist b.Intensive care physician c.Neurologist and intensive care physician together d.Other: ____________________ Secondary survey 24.When was your SOP updated?Answer: _____________________________________ 25.What temperature trigger does your site use for initiating TTM treatment, with device or otherwise?Answer: _____________________________________ 26.If applicable, why did your centre not change target temperature after the publication of the TTM2-trial?Answer: _____________________________________ 27.At the moment, no national guidelines exist for post-cardiac arrest care and assessment of neurological prognosis Sweden.What is your attitude towards such guidelines/recommendations?Answer: _____________________________________ . Nolan JP, Orzechowska I, Harrison DA, Soar J, Perkins GD, Shankar-Hari M. Changes in temperature management and outcome after out-of-hospital cardiac arrest in United Kingdom intensive care units following publication of the targeted temperature management trial.Resuscitation.2021;162:304-311.16.Bray JE, Stub D, Bloom JE, et al.Changing target temperature from 33 C to 36 C in the ICU management of out-of-hospital cardiac arrest: a before and after study.Resuscitation.2017;113:39-43. 15