The Prehospital Therapeutic Hypothermia Health And Social Care Essay

Background: Therapeutic hypothermia has been recommended for the treatment of cardiac arrest patients who remain comatose after the return of spontaneous circulation. However, the optimal time to initiate therapeutic hypothermia remains unclear. The objective of the present study is to assess the effectiveness and safety of prehospital therapeutic hypothermia after cardiac arrest.

Methods: Databases such as MEDLINE, Embase, and Cochrane Library were searched from their establishment date to May of 2012 to retrieve randomized control trials on prehospital therapeutic hypothermia after cardiac arrest. Thereafter, the studies retrieved were screened based on predefined inclusion and exclusion criteria. Data were extracted and the quality of the included studies was evaluated. A meta-analysis was performed by using the Cochrane Collaboration Review Manager 5.1.6 software.

Results: Five studies involving 633 cases were included, among which 314 cases were assigned to the treatment group and the other 319 cases to the control group. The meta-analysis indicated that prehospital therapeutic hypothermia after cardiac arrest produced significant differences in temperature on hospital admission compared with in-hospital therapeutic hypothermia or normothermia (patient data; mean difference=-0.95; 95% confidence interval -1.15 to -0.75; I2 = 0%). However, no significant differences were observed in the survival to the hospital discharge, favorable neurological outcome at hospital discharge, and rearrest. The risk of bias was low; however, the quality of the evidence was very low.

Conclusion: This review demonstrates that prehospital therapeutic hypothermia after cardiac arrest can decrease temperature on hospital admission. On the other hand, regarding the survival to hospital discharge, favorable neurological outcome at hospital discharge, and rearrest, our meta-analysis and review produces non-significant results. Using the Grading of Recommendations, Assessment, Development and Evaluation methodology, we conclude that the quality of evidence is very low.

Keywords: Meta-analysis Prehospital Hypothermia Cardiac arrest

1 Introduction

Therapeutic hypothermia (33 °C–35 °C) is typically induced after the return of spontaneous circulation (ROSC). In two previous randomized clinical trials in 2002, maintained application of therapeutic hypothermia for 12 to 24 hours significantly improves survival and neurological outcomes in patients resuscitated after cardiac arrest.[1, 2] Consequently, the treatment is now recommended for cardiac arrest patients who remain comatose after ROSC and for those with ventricular fibrillation (VF) or ventricular tachycardia as initial cardiac rhythms.[3]

The protective effects of therapeutic hypothermia include the reduction of cerebral metabolism and oxygen free-radical production, inhibition of excitatory amino acid release, attenuation of the immune response during reperfusion and brain edema, and apoptosis inhibition.[4] However, the optimal time to initiate therapeutic hypothermia remains unclear. Based on the data obtained from animal models, prehospital therapeutic hypothermia after ROSC produces both better brain function and survival rate than in-hospital therapeutic hypothermia or normothermia,[5-9]whereas delayed therapeutic hypothermia significantly negates these beneficial effects.[10]Therapeutic hypothermia, especially in a prehospital setting,[11, 12] is feasible, safe, and effective.

In the present study, a systematic review and meta-analysis was performed to assess the effectiveness and safety of prehospital therapeutic hypothermia after cardiac arrest compared with in-hospital therapeutic hypothermia or normothermia. The temperature on hospital admission, survival to hospital discharge, favorable neurological outcome at hospital discharge, and rearrest were the main outcome parameters.

2 Methods

The structured guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses[13] for randomized controlled trials (RCTs) were followed in conducting this review.

2.1 Search methods 

An extensive literature search was conducted using electronic databases, manual searching, and correspondence with authors of included studies. The MEDLINE (through PubMed), Cochrane Central Register of Controlled Trials (CENTRAL), and EMBASE databases were searched from their establishment dates up to May of 2012. No language restrictions were applied. The search strategy used the following terms: "arrest" OR "cardiac arrest" OR "heart arrest" OR "asystole" OR "cardiopulmonary arrest" AND "hypothermia" OR "therapeutic hypothermia" OR "induced hypothermia" OR "cooling" AND "randomized controlled trial" OR "controlled clinical trial." References from the studies identified and other relevant review articles were also searched to identify other potentially eligible citations.

2.2 Data collection and analysis

2.2.1 Selection of studies

Two authors independently reviewed the citations, abstracts, and full-text articles and determined the eligibility of all the studies identified in the initial search.

The inclusion criteria were as follows:

1) RCTs evaluating prehospital therapeutic hypothermia (prehospital or out-of-hospital induction of hypothermia) versus in-hospital therapeutic hypothermia (induction of hypothermia after hospital admission) or versus no hypothermia after cardiac arrest;

2) Studies on adult patients who suffered cardiac arrest (only those who suffered out-of-hospital cardiac arrest). Cardiac arrest is defined as the absence of cardiac mechanical activity and/or pulse, and therapeutic hypothermia as any target body temperature between 32 and 35 °C;

3) Outcome data, including body temperature on hospital admission, survival to the hospital discharge, favorable neurological outcome at hospital discharge, and rearrest.

The exclusion criteria are as follows: 1) review articles; 2) case reports; 3) experimental studies (for example, cell culture and isolated organs); 4) animal studies; 5) studies on children and adolescents (aged <14 years); and 6) quasi-RCTs and non-randomized studies, such as cohort or cross-over design studies.

In cases of disagreements, a third author was consulted. The agreement regarding the study inclusion was assessed using the Cohen kappa statistic.[14]

2.2.2 Data extraction

The required data were extracted independently using an extraction form. For each eligible study, two authors independently abstracted data regarding a) study design, b) study population characteristics, c) sample size, d) outcome measurements, and e) study quality.

2.2.3 Assessment of the risk of bias and quality of evidence in the included studies

Assessment of the internal validity of the included studies was performed using the methodology recommended by the Cochrane Collaboration.[15] This method involves the assessment of the risk of bias across the following domains: 1) random sequence generation; 2) allocation concealment; 3) blinding of participants and personnel; 4) incomplete outcome data; 5) selective reporting; and 6) other bias on the potential threats to validity, such as the baseline, source of funding, and academic biases.

The evidence was summarized by applying Grading of Recommendations Assessment, Development and Evaluation (GRADE) levels[16] of high, moderate, low, and very low based on the assessment of the design, limitation, inconsistency, indirectness, imprecision, and possible publication bias of the included studies using the GRADE Pro version 3.6 software.

2.2.4 Outcomes

The primary outcome was the temperature on hospital admission. The secondary outcomes were 1) survival up to the hospital discharge, 2) favorable neurological outcome at hospital discharge, and 3) rearrest.

Favorable neurological outcome is defined as a Cerebral Performance Category score of either 1 (good recovery) or 2 (moderate disabilities).

2.2.5 Data synthesis methods 

The meta-analysis of the included studies was performed using the Review Manager 5.1.6[17] to provide the mean difference (MD) with 95% confidence intervals (95% CI) using a fixed-effects model for the primary outcome of temperature on hospital admission, as well as the relative risk ratio (RR) with 95% confidence intervals (95% CI) using a fixed-effects model for the secondary outcomes of survival to the hospital discharge, favorable neurological outcome at hospital discharge, and rearrest. The heterogeneity among the trials was quantified by the visual inspection of the forest plots using a chi-square test with n-1 degrees of freedom (df), which is also expressed as I2. The statistical heterogeneity was considered relevant if I2<50%.

3 Results

3.1 Description of studies

3.1.1 Literature search results

In the present study, the initial search yielded 321 citations, of which 291 were eliminated for various reasons based on the title and abstract. The full texts of the remaining six articles were further evaluated in detail. One study was excluded, given it was a trial protocol.[18] Therefore, only five studies[19-23] fulfilled our eligibility criteria (Fig. 1).

3.1.2 Characteristics of Included studies

The characteristics of the five eligible studies are summarized in Table 1.The studies were published between 2007 and 2012, and all were RCTs. One study was a multicenter investigation,[21] whereas all the others were single-center. Of the five studies, two were conducted in Europe,[21, 22]one in North America,[23]and two in Australia.[19, 20]All studies were published in English. The initial cardiac rhythms in three studies were ventricular fibrillation (VF) and non-VF,[21-23]one was VF,[20] and the last one was non-VF.[19] In one study, therapeutic hypothermia was initiated before ROSC,[21]and, in all the other studies, after ROSC. Two studies induced therapeutic hypothermia by infusing 4 °C normal saline solution[23] or Ringer’s acetate[22] solution, two studies by ice-cold Ringer’s solution,[19, 20] and one by transnasal evaporative cooling.[21]

3.2 Risk of bias and quality in included studies

The risks of bias in the included studies are outlined in Table 2. The risk of randomization in the sequence generation and allocation concealment of all five studies was low. Considering that blinding in the intervention of therapeutic hypothermia is difficult and unfeasible, the risk of blinding was considered low if the outcome assessors were blinded to the allocation groups. Only one study was funded by a company. Overall, a low risk of bias was assessed across the included studies.

All trials had substantial risks of bias. In one study,[19] the mean and 95% CI obtained were converted to mean ± SD by formula method. The core temperature was measured from different body parts. Two studies included a subset (VF[20] or non-VF[19]) of the target cardiac arrest population, such that an indirectness existed (when patients, intervention tested, outcomes may differ from those of interest, evidence can be indirect). The total sample size was limited, and the event rates were low. Each of the four trials conducted[19-22] had a wide confidence interval spanning both the potential for benefit and harm, suggesting a serious imprecision. One study[21] was funded by a company, hence, the study has a questionable reporting bias. Based on the summary of the GRADE methodology, the accumulated qualities were very low (Table 3).

3.3 Effects of interventions

3.3.1 Temperature on hospital admission

All five eligible studies reported the primary outcome of temperature on hospital admission (involving 308 cases and 294 controls). The pooled results showed a significant difference in temperature on hospital admission under prehospital therapeutic hypothermia compared with that under normothermia in a prehospital or out-of-hospital setting (patient data; MD=-0.95; 95% CI -1.15 to -0.75; I2=0%; Fig. 2).

3.3.2 Survival to hospital discharge

All five included studies reported a secondary outcome of survival to hospital discharge (involving 378 cases and 381 controls). The pooled results showed no significant difference in the survival to the hospital discharge between prehospital therapeutic hypothermia and in-hospital therapeutic hypothermia or normothermia (patient data; RR=1.01; 95% CI 0.82 to 1.25; I2=0%; Fig. 3). Furthermore, three studies included a VF subset [20,21,23] and the pooled results of these three studies also showed no significant difference in the survival to the hospital discharge between prehospital therapeutic hypothermia and in-hospital therapeutic hypothermia or normothermia (patient data; RR=1.00; 95% CI 0.80 to 1.24; I2=31%; Fig. 4). Three studies included a non-VF subset [19,21,23] and the pooled results of these three studies indicated no significant difference in the survival to the hospital discharge between prehospital therapeutic hypothermia and in-hospital therapeutic hypothermia or normothermia (patient data; RR=1.00; 95% CI 0.53 to 1.90; I2=45%; Fig. 4).

3.3.3 Favorable neurological outcome at hospital discharge

Two of the eligible studies[21, 22] reported a secondary outcome of favorable neurological outcome at hospital discharge (involving 51 cases and 60 controls). The pooled results showed no significant difference in the favorable neurological outcome at hospital discharge between prehospital therapeutic hypothermia and in-hospital therapeutic hypothermia or normothermia (patient data; RR=1.14; 95% CI 0.65 to 2.01; I2=0%; Fig. 5).

3.3.4 Rearrest

Three of the eligible studies[21-23] reported a secondary outcome of rearrest (involving 175 cases and 181controls). The pooled results showed no significant difference in the rearrest between prehospital therapeutic hypothermia and in-hospital therapeutic hypothermia or normothermia (patient data; RR = 1.10; 95% CI of 0.62 to 1.96; I2=0%; Fig. 6).

3.3.5 Adverse outcomes

Only one study[21] reported 17 medical device-related adverse events such as periorbital emphysema (1 case), epistaxis (3 cases), and nasal whitening (13 cases).

4 Discussion

4.1 Summary of main results

We found five studies[19-23] had investigated prehospital therapeutic hypothermia after cardiac arrest involving 759 patients (378 cases and 381 controls).

Bernard et al. randomized out-of-hospital VF cardiac arrest survivors in 2010[20] and non-VF survivors in 2012[19] respectively. Both studies demonstrated prehospital cooling with a rapid infusion of large volume, ice-cold intravenous fluid could significantly decrease core temperature on hospital admission compared with control group (34.6 °C vs 35.4°C, p=0.01 [20]; 34.4 °C vs 35.7°C, p<0.001 [19]) . The risk ratio of 0.89 with a [95% CI: 0.69, 1.15] [20] and the difference risk ratio of 1.55 with a [95% CI: 0.63, 3.80] [19] showed no statistical difference in rate of survival to hospital discharge. No adverse effects of rearrest, hemodynamic instability, or pulmonary edema were reported in both studies and the method was shown to be safe and effective.

Castren et al. [21] assessed the safety, feasibility, and cooling efficacy of pre-hospital cooling applied by intra-arrest transnasal evaporative cooling. The study was the only included study which initiated cooling before ROSC. On arrival at the hospital, the mean tympanic temperature was significantly lower in the treatment group (34.2°C versus 35.5°C, P<0.001). However, on regards of rearrest, survival to hospital discharge, and favorable neurological outcome at hospital discharge, the study was not adequately powered to detect changes in the outcomes. Therefore, there were statistically insignificant. No adverse effects were reported, but they reported 17 medical device-related adverse events. Research proved pre-hospital intra-arrest transnasal cooling to be safe, feasible and effective.

On the purpose of evaluating the efficacy and safety of intravenous infusion of ice-cold fluid to induce therapeutic hypothermia in the pre-hospital setting, Kamarainen et al. [22] randomized out-of-hospital cardiac arrest survivors in comparison with conventional therapy. Hypothermia group showed lower core temperature at hospital admission compared with control (p < 0.001). No significant differences regarding safety or secondary outcome measures such as neurological outcome and mortality were observed between the groups. There were also not statistically significant with respect to survival to hospital discharge and favorable outcome at hospital discharge. Increased rate of rearrest was not observed, meanwhile adverse effects of hemodynamic instability or pulmonary edema were not reported and this method was considered to be safe and effective. This study was not designed or powered to assess neurological status or survival at hospital discharge.

Kim et al. [23] examined the feasibility, safety, and efficacy of pre-hospital hypothermia. Patients who survived from out-of-hospital cardiac arrest randomized to pre-hospital cooling rapidly infused 4°C normal saline. The temperature at hospital admission differed significantly (p < 0.0001). This study presented the outcome of survival to discharge based on VF and non-VF groups. VF patients randomized to pre-hospital cooling showed improved survival to discharge. Non-VF associated with pre-hospital cooling was associated with poor outcomes. Otherwise, when combined the two groups, the result showed no improvement. This cooling method was not associated with adverse effects in terms of hemodynamic instability, pulmonary edema assessed by chest x-ray or arrest. Although these results suggest that this method of cooling is safe, confirmation is needed in a larger, more adequately powered study.

Based on our meta-analysis, a significant difference in the temperature on hospital admission existed, but no differences in survival to hospital discharge, favorable neurological outcome at hospital discharge, and rearrest were observed. These results may be attributable to the following reasons: Firstly, the intravenous infusion was not completed on arrival at the hospital; secondly, the small number of RCTs; thirdly, some studies in this review did not have standardization to continue the cooling process after hospital arrival to maintain induced hypothermia, which contributed to the inability to determine if pre-hospital cooling was beneficial to increased survival and improved neurological outcomes[24]. Recent target temperature management study indicates that temperature control rather than cooling per se may be the beneficial intervention for mortality and neurologic function[25].

In addition, our findings are also different from the data from animal model. These may be explained due to the complexity and severity of the animal model, the different durations of VF, the methods used to induce hypothermia (for example, systemic vs. selective brain), and limited numbers of experimental animals used. Speed of achieving target temperature in experimental studies is dependent on animal size and the duration of exposure to hypothermia during circulatory arrest. Importantly, many animal experiments apply hypothermia initiated during cardiopulmonary resuscitation, that is, intra-arrest, which has been demonstrated to be superior to cooling initiated after ROSC, both in terms of increased rates of successful CPR , improved survival and neurological outcome [6,7].

4.2 Quality of the evidence

All studies were academia-initiated. The quality of the studies was generally good. All studies reported almost all essential quality criteria, and the number of patients lost to follow-up was within an acceptable range. After the literature search, the data in all eligible studies retrieved were included. In one study,[19]obtaining individual patient data on the temperature on hospital admission was impossible, although the study obtained a mean and 95% CI that were converted to mean ± SD by formula method. In two other studies,[22, 23] data on whether cooling was continuous or not was according to physician preferences during hospital stay were clinically too heterogeneous to be combined.

In all studies, patients were enrolled on a consecutive basis. In one study,[23] 34.2% of all eligible patients were not randomized because of problems such as patient hemodynamic instability, equipment problems, temporary suspension of the study, and overlooked patients during the enrollment process. In the studies conducted by Bernard et al., 41.2%[20] and 47.2%[19] of all eligible patients were not randomized. The reasons were not reported.

In one study,[22] when the allocation to the hypothermia group intervention was discontinued in three patients because of one mortality, one rearrest prior to intervention, and one protocol violation, one patient in the hypothermia group and two patients in the control group were lost to follow-up. In another study,[21] three patients in each group were lost to follow-up during the prehospital cooling, whereas one patient in the intra-arrest group was lost to follow-up during the hospital stay. These patients were included in the intention-to-treat analysis. All other studies showed complete patient follow-up records.

All studies with individual patient data reported the same outcome of temperature on hospital admission and survival to hospital discharge. Two studies reported favorable neurological outcome at hospital discharge, and three studies reported rearrest. In three studies,[19-21] the outcome assessors were blinded to the treatment.

4.3 Limitations

The major limitation of this review is the small number of randomized controlled trials, hence, the small number of included patients. Therefore, the precision of the outcome parameters obtained is generally low. All included studies were RCTs, but one study was a multicenter one, whereas the others were single-center studies. Thus, methodological heterogeneity exists.

Clinical heterogeneity in the volume of the intravenous fluid infusion by hospital admission time was also present by the time of hospital admission. Four studies initiated therapeutic hypothermia by the intravenous infusion of normal saline or Ringer’s acetate solution, but, with most of these patients, the intravenous infusion was not completed on arrival at the hospital due to lack of time, recurrent arrest, and mortality while in the field. Another area of clinical heterogeneity was the measurement of core temperature from different body parts, including the esophageal, nasopharyngeal, and tympanic cavity areas. In many clinical studies, the temperature of each of the above sections is used in place of core temperature, but these temperature readings cannot accurately reflect the core temperature.[26, 27]

Our systematic search for relevant studies was not limited by language, but all of the included studies were published in English. Thus, publication bias was possible.

Considering individual patient data were not reported in each study, the data used in our meta-analysis was extracted directly from the published papers, which may subsequently affect the validity of the pooled results. Gaining access to individual patient data in each study and conducting individual patient meta-analysis are some ways of reducing the chances for bias.

Bernard et al.[1] firstly reported benefit of therapeutic hypothermia applied in cardiac arrest survivors. But two clinical studies of his groups recently demonstrated no statistical difference in survival to hospital discharge. Because both of the studies contributed most of the patients, we could not neglect the effects of the two papers. We considered that the most reason was short transport times, recurrent arrest, and mortality which lead to be lack of the full volume to be infused prior to arrival[24]. In the future research, we should perfect this method or find new internal or external means of cooling.

5 Conclusions

This review demonstrates that prehospital therapeutic hypothermia after cardiac arrest can decrease temperature on hospital admission. Nevertheless, with regards survival to hospital discharge, favorable neurological outcome at hospital discharge, and rearrest, our meta-analysis and review produces non-significant results. The risk of bias is low, but assessment using GRADE-methodology reveals that the quality of evidence is very low. Thus, further study should be performed with larger RCTs which should include pre-hospital cooling associated with continued in-hospital cooling and be powered to determine discharge outcomes.

Fig. 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart of the search for relevant references.

Fig. 2. Summary of data on temperature on hospital admission for prehospital therapeutic hypothermia versus normothermia in prehospital or out-of-hospital setting.

Fig. 3. Summary of data on survival to hospital discharge for prehospital therapeutic hypothermia versus in-hospital therapeutic hypothermia or normothermia.

Fig. 4. Summary of data on VF survival and non-VF survival to hospital discharge for prehospital therapeutic hypothermia versus in-hospital therapeutic hypothermia or normothermia.

Fig. 5. Summary of data on favorable neurological outcome at hospital discharge in prehospital therapeutic hypothermia versus in-hospital therapeutic hypothermia or normothermia.

Fig. 6 Summary of data on rearrest in prehospital therapeutic hypothermia versus in-hospital therapeutic hypothermia or normothermia.

Table 1 Characteristics of the eligible studies.

Table 2 Risk of bias assessment across the eligible studies.

Table 3 GRADE profile for assessing the quality of evidence for the studies including in the meta-analysis.

1 All trials had substantial risks of bias.

2 In one study, the mean and 95% CI were converted to mean ± SD by formula method. The core temperature was measured from different body parts. The two studies included subsets (VF and non-VF) of the target cardiac arrest population, respectively.

3 The total sample size is limited.

4 One study was funded by a company.

5 The two studies included a subset (VF or non-VF) of the target cardiac arrest population.

6 The total sample size was limited and event rates were low. One trial had a wide confidence interval spanning both the potential for benefit and harm.

7 The total sample size was limited and event rates were low. One trial had a wide confidence interval spanning both the potential for benefit and harm.

8 The total sample size was limited and event rates were low. Two trials had wide confidence intervals spanning both the potential for benefit and harm.