DOI: https://doi.org/10.21203/rs.3.rs-1697239/v1
Aim: To characterize and summarize the studies regarding the influence of post-rewarming fever on neurological outcome and mortality in cardiac arrest patients.
Methods: Two investigators separately screened relevant articles in EMBASE, PubMed, and Cochrane Central databases. Randomized clinical trials (RCTs) and cohort studies that evaluated the influence of post-rewarming fever (PRF) and normothermia in cardiac arrest patients were included. The meta-analysis was performed using a random effects model or a fixed effects model to calculate the pooled odds ratios (ORs) and corresponding 95% confidence intervals (CIs). The primary outcome was the unfavorable neurological outcome and the secondary outcome was the mortality.
Results: The meta-analysis included 12 studies involving 2,991 patients. Results of quantitative analysis suggested that PRF (body temperature >37.5°C) did not affect the unfavorable neurological outcome of patients with cardiac arrest (OR, 0.82; 95% CI, 0.54-1.25; I2, 81%). Also, PRF (body temperature >37.8°C) was not related to higher mortality of patients with cardiac arrest (OR, 0.86; 95% CI, 0.55-1.34; I2, 74%). However, PRF with higher body temperature (>38.5℃) was associated with higher mortality (OR, 2.22; 95%CI, 1.40- 3.35; I2, 0%).
Conclusions: This study suggests that PRF (body temperature >37.5°C) is not related to neurological outcome. And no significant association is found between PRF (body temperature >37.8°C) and mortality. However, PRF is associated with higher mortality when PRF was defined as >38.5℃.
Cardiac arrest (CA) is an important global public health issue and associated with high mortality. [1] In-hospital CA occurs in more than 290,000 adults each year in the United States.[2] Cardiopulmonary resuscitation (CPR) is an effective emergency intervention for cardiac arrest.[3] Researches on CPR and post-resuscitation care are still developing in recent years.[4–7] After the treatment of CA and the successful return of spontaneous circulation (ROSC), the neurological outcome of patients tends to be undesirable.[8, 9] Approximately 50% of survivors suffer different degree of neurologic disability.[10] The targeted temperature management (TTM), which maintains the patient’s temperature from 32℃ to 36℃ for at least 24 h, is recommended by international guidelines to attenuate brain injury and improve the neurological outcome.[3, 11] With the implementation of TTM, post-arrest fever in the immediate period can be prevented in recent years. Development of fever before the TTM occurs frequently and has previously been found to be associated with unfavorable outcomes.[12, 13] In this study, we mainly focus on the fever in the patients who completed TTM.
Post-rewarming fever (PRF), or rebound hyperthermia (RH), was observed in many patients who received TTM.[14–26] The PRF was defined as a phenomenon of increased body temperature (༞38℃ or greater) after the rewarming period following TTM in most studies.[17–26] The recent guidelines reckon the prevention of fever in comatose patients after the TTM as a Class IIb recommendation.[3] However, studies about the influence of PRF on neurological outcome showed conflicting results. Some studies suggested that the PRF was relative to unfavorable neurological outcome.[14, 15, 24, 25, 27] But recent studies indicated that PRF may be a symbol of good neurological outcome.[17, 21] Therefore, the clinical significance of PRF should be better understood .
The aim of this systematic review and meta-analysis was to investigate whether the PRF could have an impact on clinically relevant outcomes, including neurological outcome and mortality, in patients suffering from CA.
We conducted the systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis — Protocols (PRISMA-P) guidelines.[28] The protocol of this study was registered in www.inplasy.com (INPLASY202240052).
We performed the literature retrieval up to 13th March, 2022 in the PubMed, EMBASE and CENTRAL databases using the following Medical Subject Headings (MeSH) terms: “heart arrest” AND “targeted temperature management” AND (“fever” OR “hyperthermia”). (Specific search strategies see Supplementary data 1)
The studies that were selected for the meta-analysis according to the following requirements: (1) the participants of the study were adult patients with CA; (2) the research should have at least one control group treated with TTM; (3) the study observed and defined the phenomenon of PRF; (4) the neurological outcome and mortality of patients was available; (5) the studies were cohort studies or randomized controlled trials (RCTs) ; (6) the study articles were written in English. The studies were excluded if the following conditions were met: (1) the full text of the article could not be accessed; (2) only the abstract of articles was accessible; (3) only a citation or report on the study could be found in another publication; (4) the outcome and mortality data was not available.
We reviewed the articles that conformed to the above-mentioned criteria and extracted the relevant data about neurological outcome and/or mortality. In this meta-analysis, the definition of unfavorable neurological outcome was Glasgow–Pittsburgh cerebral performance category (CPC) > 2 or modified Rankin Scale (mRS) > 2. [29, 30] The primary outcome is unfavorable neurological outcome at discharge or at the end of the follow-up period. The secondary outcome is mortality rate at discharge or at the end of the follow-up period.
Two separate researchers (GGQ and MYN) screened all the accessible articles based on the above-mentioned inclusion and exclusion criteria and extracted the relevant data of the articles. During the progress of the data extraction, the arising discrepancies about articles from two researchers were discussed with XS. The first author's first name, the study's year, the study's host country, the study design type, the study period, the study participants, the number of PRF patients and control patients, the definition of PRF, the type of neurological outcome assessment, and mortality of the original studies were all extracted.
No RCTs were selected into in this meta-analysis. For the 12 non-random study, the Newcastle-Ottawa Scale (NOS) was used to assess risk of bias (ROB). [31] The NOS had a total score of 9 points and included cohort selection, cohort comparability, and outcome evaluation. Studies that scored at least 6 points were included in the meta-analysis (See Table 1)
Cohort Studies |
Selection |
Comparability |
Outcome |
Total |
|||||
---|---|---|---|---|---|---|---|---|---|
Representative of exposed cohort |
Selection of the non-exposed cohort |
Ascertainment of exposure |
Endpoint not present at start |
Comparability (Confounding) |
Assessment of outcome |
Follow-up duration |
Adequacy follow-up |
||
Lee 2020 |
* |
* |
* |
* |
** |
* |
* |
* |
9 |
Makker 2017 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Grossestreuer 2017 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Lee 2015 |
* |
* |
* |
* |
** |
* |
* |
* |
9 |
Nobile 2015 |
* |
* |
* |
* |
** |
* |
* |
* |
9 |
Cocchi 2014 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Pellis 2014 |
* |
* |
* |
* |
** |
* |
* |
* |
9 |
Winters 2013 |
* |
* |
* |
* |
*- |
* |
* |
* |
8 |
Leary 2013 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Bro-Jeppesen2013 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Aldhoon 2012 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
Benz-Woerner 2012 |
* |
* |
* |
* |
-- |
* |
* |
* |
7 |
The degree of variability and the heterogeneity of the meta-analysis were assessed using I2 and P values. When I2 was 50%, 51–75%, or > 76%, it was classified as low, moderate, or high heterogeneity, respectively. For the high heterogeneity, the random effects model was used to calculate the merged odds ratios (ORs) and 95% confidence intervals (CIs), while the fixed effects model was used for low or moderate heterogeneity. The funnel plot approach was used to assess possible publication bias. [32] The study model's robustness was tested by sensitivity analysis. In this study, Review Manager 5.4 software was used to conduct statistical analysis.
After electronic database searches and bibliographical inspections, 330 records were screened according to the retrieval strategies. 42 records of them were duplicated. 276 of the remaining records were excluded on account of title, abstract, study design, unavailable data and other factors. As a result, 12 cohort studies were selected into this meta-analysis. [14, 16–26] (Fig. 1).
A total of 2,991 cardiac arrest patients were selected in this meta-analysis, of which 1,195 patients developed PRF and 1,796 kept the normothermia. Twelve studies were included in this meta-analysis.[14, 16–26] These studies had different definitions of PRF. Eligible studies included seven articles defined the PRF as body temperature > 38.0℃ after the TTM.[14, 17–19, 21, 23, 26] Three articles defined the PRF as > 38.5℃ after the TTM.[20, 24, 25] One of the eligible study defined the PRF as > 37.8℃[22] and the remaining one study defined PRF as > 37.5℃[16]. The neurological outcome of patients was assessed in 11 studies[14, 16–25], while mortality was calculated in 9 studies[14, 18, 19, 21–26]. More details of study characteristics of included articles are presented in Table 2.
Author year |
Country |
Study design |
Definition of PRF |
Type of CA |
Outcome assessment timing |
Mortality |
Patients (n) |
PRF(n)/ normothermia(n) |
|
---|---|---|---|---|---|---|---|---|---|
Lee 2020 |
Korea |
cohort study |
38℃ |
OHCA |
CPC at 6 months |
- |
1,031 |
389/642 |
|
Makker 2017 |
America |
Cohort study |
38℃ |
OHCA |
CPC in-hospital |
discharge |
97 |
54/43 |
|
Grossestreuer 2017 |
America |
cohort study |
38℃ |
Both** |
CPC at discharge |
discharge |
465 |
179/286 |
|
Lee 2015 |
Korea |
cohort study |
38℃ |
NA* |
CPC at discharge |
in hospital |
277 |
81/196 |
|
Nobile 2015 |
Belgium |
cohort study |
38.5℃ |
NA |
CPC at 3 months |
- |
229 |
41/188 |
|
Cocchi 2014 |
America |
cohort study |
38℃ |
OHCA |
CPC in-hospital |
in hospital |
54 |
28/26 |
|
Pellis 2014 |
America |
cohort study |
38℃ |
Both |
CPC at discharge |
discharge |
42 |
19/23 |
|
Winters 2013 |
America |
cohort study |
38.5℃ |
OHCA |
mRS at discharge |
discharge |
141 |
42/99 |
|
Leary 2013 |
America |
cohort study |
37.8℃ |
Both |
CPC at discharge |
discharge |
167 |
69/98 |
|
Bro-Jeppesen 2013 |
Denmark |
cohort study |
38.5℃ |
OHCA |
CPC at discharge/ 1 year |
30 days |
270 |
136/134 |
|
Aldhoon 2012 |
Czech Republic |
cohort study |
37.5℃ |
OHCA |
CPC at 1 month |
- |
41 |
22/19 |
|
Benz-Woerner 2012 |
Switzerland |
cohort study |
38℃ |
Both |
- |
in hospital |
177 |
135/42 |
|
*The article didn’t specify the type of CA. | |||||||||
**Both the OHCA and IHCA were included in the article. | |||||||||
OHCA out-of-hospital cardiac arrest | |||||||||
IHCA in-hospital cardiac arrest |
Twelve studies included were observational cohort studies and no RCTs were selected into this meta-analysis. Scores of all the eligible cohort studies were greater than or equal to 6 points on the basis of NOS (Table 1).
Unfavorable neurological outcome
Eleven of the including studies reported the neurological outcome of patients. The unfavorable neurological outcome was defined as CPC > 2 or mRS > 2. Six studies defined PRF as > 38°C[14, 17–19, 21, 23], three as > 38.5°C[20, 24, 25], and the other two as > 37.8°C[22] and > 37.5°C[16], respectively. Eleven studies (all the definition of PRF) including neurological outcomes were analyzed with random-effects models. The results showed that, compared to normothermia, PRF did not affect the unfavorable neurological outcome of patients with CA (OR, 0.82; 95% CI, 0.54–1.25; I2, 81%). (Fig. 2).
Articles defining PRF as > 38°C and > 38.5℃ were selected for different meta-analysis about PRF and neurological outcome. Nine articles defined PRF as > 38°C (three articles defining PRF as > 38.5℃ was also included) [17–25] were analyzed with random effects model. Our analysis of these nine articles showed that PRF had no significant effect on unfavorable neurological outcome of CA patients (OR, 0.95; 95% CI, 0.57–1.58; I2, 86%). (Fig. 3)
In this meta-analysis, three articles defined PRF as > 38.5°C and involved neurological outcome.[20, 24, 25] One study Makker et al[18] defined the PRF as > 38°C but contained a subgroup that researched the relationship between PRF > 38.5℃ and neurological outcome. We used random effects model to analyze data from the four studies. The result showed that PRF with higher body temperature (PRF > 38.5℃) had no significant effect on unfavorable neurological outcome of CA patients. (OR, 1.45; 95% CI, 0.81–2.60; I2, 59%) (Fig. 4)
Significant heterogeneity was observed among studies for PRF and neurological outcome. Therefore, sensitivity analysis was used to test the robustness of the results of the meta- analysis. (Supplementary Table 1–3)
Mortality
Mortality data were available in nine of the studies included in this meta-analysis. Six studies defined PRF as > 38°C[18, 19, 21–23, 26], two as > 38.5°C[24, 25], and the other one as > 37.8°C[14]. We examined these studies (all the definition of PRF) using a random-effect model. The results showed no significant association between PRF and mortality (OR, 0.86; 95% CI, 0.55–1.34; I2, 74%). (Fig. 5)
In this meta-analysis, two studies involved PRF > 38.5℃ and mortality.[24, 25] One study Makker et al[18] had a subgroup that researched the relationship between PRF > 38.5℃ and mortality. A fixed-effect model was used to analyze the three studies. Results showed that PRF with higher body temperature was associated with higher mortality (OR, 2.22; 95%CI, 1.40–3.35; I2, 0%). (Fig. 6)
Meanwhile, sensitivity analysis was used to test the robustness of the results of the meta- analysis. (Supplementary Table 4, Supplementary Table 5)
None of the 12 studies were evaluated as high risk of bias. Therefore, sensitivity analysis was performed by removing the included studies one by one to see how they affected the pooled OR, 95% CI and heterogeneity. The results of sensitivity analysis were showed in table form. (Supplementary Table 1–5)
Eleven studies were included into the meta-analysis to analyze the influence of PRF on unfavorable neurological outcome compared to normothermia. (Fig. 2) Funnel plots were conducted to evaluate the possible publication bias for the included studies. Evaluating the unfavorable neurological outcome of PRF and normothermia suggested that there was no significant publication bias. (Supplementary Fig. 1)
In this meta-analysis, 12 trials involving 2,991 patients analyzed the relationship between PRF and the neurological outcomes and/or mortality of patients. Results of this meta-analysis suggest that the patients who experienced PRF after TTM had similar neurological outcomes compared with the normothermia patients regardless of the definition body temperature of PRF. Meanwhile, results suggest that PRF with a broader definition (body temperature > 37.8°C or greater) after TTM have no significant association with mortality. However, the patients who experienced PRF defined as a higher body temperature (> 38.5℃) possessed a significantly higher mortality than the normothermia.
A concise meta-analysis conducted by Makker et al. [33] had found that PRF was associated with a significantly worse neurological outcome, which findings were not consistent with the present authors’ study. Possible reasons for the above discrepancy results may include the following: (1) all the studies included in the previous meta-analysis were outdated. Some studies have emerged in recent years suggesting that PRF may not be associated with unfavorable neurological outcomes. Several studies even suggested that PRF was associated with favorable neurological outcomes.[17, 21] The inclusion of new studies influenced the results of previous meta-analysis; (2) according to the meta-analysis operation requirements, the previous meta-analysis was a less rigorous meta-analysis, which only extracted the OR value of the studies but did not concern the specific details and data of studies.[31]
The present meta-analysis reveals that PRF with a broader definition (body temperature > 37.5°C, > 37.8℃ or greater) is not significantly associated with neurological outcome. (Fig. 2, Fig. 3) It is notable that in the analysis, when PRF was defined as a higher body temperature (> 38.5℃), four included studies showed the same result. (Fig. 4) However, sensitivity analysis suggest that the study conducted by Nobile et al. [20] was worthy of attention. The study obviously influenced the result and heterogeneity of the analysis. (Supplementary Table 3). The possible reason is that study conducted by Nobile et al. [20] involved CA patients who stayed in ICU, while the other three studies [18, 24, 25] focused on OHCA patients.
In this meta-analysis, when PRF was defined as body temperature > 37.8℃ or > 38℃, PRF was not significantly associated with mortality. (Fig. 5) Notably, the PRF was associated with higher mortality when PRF was defined as > 38.5℃. (Fig. 6) Therefore, after the TTM, the results suggest that active temperature control is necessary when the body temperature was > 38.5℃. However, one study indicated that the implementation of controlled normothermia to prevent PRF was not associated with favorable neurological outcome.[34] And in an in vitro study, the prevention of PRF obviously aggravated apoptosis of cells and release of inflammatory factors. [35] This study suggested that PRF was related to activation of inflammatory response and programmed cell death following the ischemia-reperfusion injury caused by CA. Since most works are focused on the clinical application of TTM, including optimal cooling temperature[5, 36, 37], practical methods of cooling for temperature control[6, 38, 39] and rate of rewarming following TTM[40, 41], it is necessary to investigate the PRF with different body temperature and body temperature control in the future.
The following are the advantages of this study: (1) this meta-analysis evaluated PRF after TTM in CA patients and assisted physicians to recognize the clinical effects of body temperature management and patient prognosis following ROSC. (2) the PRF group was divided into the > 38°C and > 38.5°C subgroups in order to provide a better understanding for physicians to make respective clinical strategies about different body temperature.
Nonetheless, the following are the limitations of this meta-analysis: (1) meta-analysis was a secondary analysis of original studies. High heterogeneity existed between different studies and was caused by many reasons, such as sample size, type of CA, TTM treatment strategies, etc. The difference between the original studies could have influenced statistical analysis. (2) all of the studies selected into this meta-analysis were observational studies. The data of observational studies were analyzed for certain purpose, which was considered a methodological flaw. Therefore, more RCTs were needed in the future to investigate the affect of PRF.
In the authors’ view, this study provides a specific cognition of PRF phenomenon and suggests that PRF with a broader definition (body temperature > 37.5°C, > 37.8℃ or greater) is not associated with neurological outcome and mortality. However, PRF is associated with higher mortality when PRF was defined as > 38.5℃. In the future, more studies are required to further understand the influence of PRF with different body temperature.
Ethics Approval and Consent to participate
Not applicable.
Availability of data and materials
All data generated or analyzed during this study are included in this published article.
Authors’ contributions
GGQ and SP gave the study concept and design. GGQ, MYN and XS carried out the statistical analysis. GGQ, MYN, XS, ZHR and SP drafted the manuscript. ZHR and SP revised the manuscript. GGQ, MYN, XS, ZHR and SP supervised the study. All authors read and approved the final manuscript.
Conflict of interest statement
The authors declare that there are no conflicting interests.
Funding
The systematic review was funded by National Natural Science Foundation of China (82072137).
Consent for publication
Consent for publication was provided by all authors.
Acknowledgements
None.