Do Out-of-Hospital Cardiac Arrest Patients Have Increased Chances of Survival When Transported to a Cardiac Resuscitation Center?

Background Patients suffering from an out-of-hospital cardiac arrest are often transported to the closest hospital. Although it has been suggested that these patients be transported to cardiac resuscitation centers, few jurisdictions have acted on this recommendation. To better evaluate the evidence on this subject, a systematic review and meta-analysis of the currently available literature evaluating the association between the destination hospital's capability (cardiac resuscitation center or not) and resuscitation outcomes for adult patients suffering from an out-of-hospital cardiac arrest was performed. Methods and Results PubMed, EMBASE , and the Cochrane Library databases were first searched using a specifically designed search strategy. Both original randomized controlled trials and observational studies were considered for inclusion. Cardiac resuscitation centers were defined as having on-site percutaneous coronary intervention and targeted temperature management capability at all times. The primary outcome measure was survival. Twelve nonrandomized observational studies were retained in this review. A total of 61 240 patients were included in the 10 studies that could be included in the meta-analysis regarding the survival outcome. Being transported to a cardiac resuscitation center was associated with an increase in survival (odds ratio=1.95 [95% confidence interval 1.47-2.59], P<0.001). Conclusions Adult patients suffering from an out-of-hospital cardiac arrest transported to cardiac resuscitation centers have better outcomes than their counterparts. When possible, it is reasonable to transport these patients directly to cardiac resuscitation centers (class II a, level of evidence B, nonrandomized). Clinical Trial Registration URL : www.crd.york.ac.uk/PROSPERO/ . Unique identifier: CRD 42018086608.

Clinical Perspective

What Is New?

This is the first systematic review to evaluate the association between destination hospital characteristics and resuscitation outcomes following an out‐of‐hospital cardiac arrest.

Direct transport to a cardiac resuscitation center is associated with improved survival and survival with a good neurologic outcomes for these patients.

This association was stronger among patients not having experienced a prehospital return of spontaneous circulation.

What Are the Clinical Implications?

When possible, it is reasonable to transport patients suffering from an out‐of‐hospital cardiac arrest directly to a cardiac resuscitation center.

A bypass delay of up to 15 minutes for patients not having experienced prehospital return of spontaneous circulation and of 30 minutes for patients having experienced prehospital return of spontaneous circulation is probably safe. This should be further tested in a prospective study.

Introduction

Out‐of‐hospital cardiac arrest (OHCA) is one of the leading causes of death in the United States, and it is a serious public health burden.1 Despite an improvement in prehospital resuscitation practices, including an increased access to early cardiopulmonary resuscitation and defibrillation, mortality rates remain high, with only 10% of patients surviving to hospital discharge.1, 2, 3, 4, 5 To further decrease the mortality from OHCA, the establishment of a regionalized approach for the treatment of OHCA, including direct transport to specialized cardiac resuscitation centers, such as in the case of an ST‐segment–elevation myocardial infarction, has been proposed (class IIb, level of evidence C‐limited data).5, 6, 7 For a hospital to be considered a cardiac resuscitation center, it must be able to provide diagnostic angiography and percutaneous coronary intervention (PCI) on site at all times as well as targeted temperature management (TTM).5 Indeed, most nontraumatic OHCA results from an acute coronary syndrome, and PCI is the preferred therapeutic procedure for that pathology.8, 9, 10, 11, 12 Patients remaining comatose following an OHCA also strongly benefit from some form of TTM.10, 13, 14

Despite these recommendations made by the American Heart Association, few jurisdictions have implemented a regionalized approach for OHCA patients with designated receiving centers. As a result, there remains significant variation with regard to treatment standards for these patients.15, 16 However, since these guidelines were published, multiple new studies have emerged, and their results could influence the decisions made for patients suffering from an OHCA.

To better evaluate the evidence on this subject, a systematic review and meta‐analysis of the currently available literature evaluating the association between the destination hospital capability (cardiac resuscitation center or not) and resuscitation outcomes (survival and survival with a good neurologic outcome) for patients suffering from an OHCA were performed.

Methods

This review was registered (Prospero CRD42018086608) before its initiation. Its results are presented as per the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines.17 Its main objective was to evaluate the association between the transport to a cardiac resuscitation center (defined as having on‐site PCI and TTM capability at all times) and resuscitation outcomes (survival and survival with a good neurologic outcome) for adult patients suffering from an OHCA. The data that support the findings of this study are available from the corresponding author on reasonable request. Because of its nature, this study did not need to be reviewed by an institutional review board.

Search Strategy

The search strategy aimed to find both published and unpublished studies. PubMed, EMBASE, and the Cochrane Library databases were first queried using a specifically designed search strategy. This search strategy included terms such as heart arrest, cardiac arrest, out‐of‐hospital cardiac arrest, cardiopulmonary arrest, ventricular fibrillation, pulseless electrical activity, hospital characteristics, critical care center, high‐volume hospital, regionalization of care, and high‐volume centers (Data S1). The search was limited to humans and English‐language publications. Gray literature was searched using Web of Science and Google Scholar. The references of all identified articles and main review articles were also searched for additional relevant studies. The search was performed initially on February 4, 2018 and repeated on July 24, 2018 to ensure that no new literature had been published in the interim.

Article Selection

Following the automatic removal of duplicates, remaining citations were screened by 2 independent reviewers (D.L., A.G.) for potentially pertinent publications using the Covidence online software (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia). Potentially eligible citations were then fully evaluated. Discrepancies regarding the selection of articles were resolved by consensus with a third reviewer (N.C.).

Original randomized controlled trials and observational studies were both considered for inclusion. Case series describing only 1 population were excluded. Studies published before 2008 were excluded because the evolution in treatment standard might make these results no longer applicable by today's standards.18 To be included, studies had to include adults suffering from a nontraumatic OHCA who were transported to the hospital. Studies reporting on traumatic OHCA or in‐hospital cardiac arrest were excluded. Included studies also had to report outcome data on patients being transported to a cardiac resuscitation center and those who were transported to a hospital that was not a cardiac resuscitation center. To be considered a cardiac center, a hospital was required to have both PCI capability and TTM capability as defined by the American Heart Association.5 If that information was not available, it was decided to exclude these studies from the review to limit the risk of bias.

Quality Assessment

The quality assessment of all retained articles was performed by 2 independent reviewers (D.L., A.C.). The risk of bias was evaluated using the Newcastle‐Ottawa scale (Table S1).19 Disagreements were resolved by consensus.

Data Abstraction

Data for the outcomes of interest were independently extracted from the included articles by 3 reviewers (N.C., N.G., and J.A.). In addition, the study design, population characteristics, sample sizes, and outcomes were also extracted. A standard template was created for the purpose of data extraction (Table S2).

Outcome Measures

The primary outcome measure was survival. The preferred timing of measurement was at hospital discharge. If that information was not available, survival at 30 or 90 days was used. The secondary outcome measure was survival with a good neurologic outcome (defined as a Cerebral Performance Category of 1 or 2).20 The preferred timings of measurement were the same as for the primary outcome.

Analyses

Adjusted odds ratio (OR) was the effect measure used whenever available. If these were not provided, unadjusted ORs were used or calculated from the available data instead. For outcomes reported in multiple studies, results were pooled in a meta‐analysis using Revman (Version 5.3. The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark, 2014) if appropriate.

Heterogeneity was assessed statistically using I2. Random effect models were preferred to better account for the differences in selection criteria and design among the included studies, but fixed‐effect models were also presented as supplementary analyses. All results are presented with their 95% confidence interval (CI).

For each meta‐analysis of more than 10 articles, a funnel plot was constructed to assess for a publication bias.21 When fewer than 10 articles were available, the reporting bias was assessed qualitatively.

Three sets of sensitivity analyses were performed to explore the heterogeneity, 1 excluding articles with some risk of bias (Newcastle‐Ottawa Scale ≤8), 1 including only patients having experienced prehospital return of spontaneous circulation (ROSC), and 1 including only those who did not. The same outcome measures (survival and survival with a good neurologic outcomes) were used for each set of sensitivity analyses.

Results

Search and Article Selection

The initial electronic search yielded 2727 references (Figure 1). A title and abstract screening left 31 potentially eligible citations. The search of gray literature, the second electronic search, and communications with authors of potentially eligible articles yielded an additional 5 citations for a total of 36 included for the full‐text review. Among these articles, a total of 24 were excluded for the following reasons: absence of comparison between cardiac centers and noncardiac centers (17), unknown availability of TTM (2), only abstract published (2), review article (1), availability of more recently published data from the same cohort (1), and included in‐hospital cardiac arrest (1). A total of 12 studies were included in the narrative review and 11 in the meta‐analysis.

Figure 1

Flow diagram of the systematic search.

Included Studies

All included articles were nonrandomized observational studies (Table).22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 It was possible to extract data regarding survival for 10 studies and regarding neurologic outcomes for 5 studies. One study only provided hazard ratios, which prevented its inclusion in the meta‐analysis.29 Despite some adjusted results being available for all included studies, unadjusted results needed to be used in some analyses or sensitivity analyses for 2 studies because they were not provided for all the outcomes or populations pertaining to the present study.22, 33

Table 1

Characteristics of Included Studies

Study	Level of Risk of Bias/NOS	Specific Inclusion/Exclusion Criteria	Number of Eligible Patients	Average Age (y)	Percentage of Initial Shockable Rhythm	Percentage of Prehospital ROSC	Outcomes of Interest Presented
Kajino 201022	Low/9	None	10 383	73	17%	8%	Survival at 30 d and survival with a good neurologic outcome at 30 d
Stub 201123	Low/8	Included only patients with prehospital ROSC	2706	67	57%	100%	Survival to hospital discharge
Wnent 201224	Low/9	None	889	69	26%	N/A	Survival to hospital discharge
Soholm 201325	Low/9	None	1020	65	45%	N/A	Survival at 30 d
Hunter 201626	Low/8	Included only patients with prehospital ROSC	1024	61	27%	100%	Survival to hospital discharge
Kragholm 201727	Low/8	Included only patients with prehospital ROSC	1507	65	39%	100%	Survival to hospital discharge
Matsuyama 201728	Low/8	Excluded patients with paramedic‐witnessed arrest	39 965	75	8%	6%	Survival with a good neurologic outcome at 30 d
Tranberg 201729, a	Low/9	None	41 186	70	21%	N/A	···
Tsai 201730	Low/8	Included only patients with an initial shockable rhythm without a prehospital ROSC	546	62	100%	0%	Survival to hospital discharge and survival with a good neurologic outcome at hospital discharge
Casey 201831	Low/8	Included only patients who survived to hospital admission	38 163	67	29%	N/A	Survival to hospital discharge and survival with a good neurologic outcome at hospital discharge
Cournoyer 201832	Low/9	None	4922	67	35%	34%	Survival to hospital discharge
McKenzie 201833	Low/8	Included only patients who survived to hospital admission	535	62	62%	86%	Survival to hospital discharge

N/A indicates not applicable; NOS, Newcastle‐Ottawa Scale; ROSC, return of spontaneous circulation.

Only included in the qualitative synthesis.

All included studies were considered at low risk of bias (Table). Seven studies lost a point regarding the representativeness of their cohort because it was composed of a selected subpopulation of nontraumatic OHCA transported to the hospital.23, 26, 27, 28, 30, 31, 33

Main Results

Survival

A total of 61 240 patients were included in the 10 studies that were part of this meta‐analysis.22, 23, 24, 25, 26, 27, 30, 31, 32, 33 Eight studies presented results regarding survival to hospital discharge, and the other 2 studies reported on survival at 30 days. This resulted in 2 independent subgroups.

Being transported to cardiac resuscitation centers was associated with an increase in survival (OR=1.93, 95% CI 1.48‐2.50, P<0.001) (Figures 2 and 3). There was no significant difference between the 2 subgroups (hospital discharge OR=1.81, 95% CI 1.33‐2.45, P<0.001; 30 days OR=2.35, 95% CI 2.06‐2.68, P<0.001; test for subgroup differences P=0.12).

Figure 2

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, performed using a random‐effect model.22, 23, 24, 25, 26, 27, 30, 31, 32, 33 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 3

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, performed using a fixed‐effect model.22, 23, 24, 25, 26, 27, 30, 31, 32, 33 CI indicates confidence interval; IV, inverse variance; SE, standard error.

The heterogeneity was high across this analysis and was hence explored using sensitivity analyses.

In addition, for the study that presented hazard ratios, which could not be mathematically included in this meta‐analysis, being transported to a cardiac resuscitation centers was also independently associated with better survival (adjusted hazard ratio 1.10, 95% CI 1.08‐1.12, P<0.001).29

Survival With a Good Neurologic Outcome

Five studies, including a total of 89 491 patients, reported rates of survival with good neurologic outcomes.22, 24, 28, 30, 31 Three of these studies presented results regarding survival with a good neurologic outcome at discharge, and the other 2 presented results regarding survival with a good neurologic outcome at 30 days, resulting in 2 independent subgroups.

Being transported to a cardiac resuscitation centers was associated with an increase in survival with a good neurologic outcome (OR=1.84, 95% CI 1.52‐2.21, P<0.001) (Figures 4 and 5). There was no significant difference between the 2 subgroups (hospital discharge OR=1.95, 95% CI 1.09‐3.49, P=0.02; 30 days OR=2.00, 95% CI 1.37‐2.92, P<0.001; test for subgroup differences P=0.95).

Figure 4

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, performed using a random‐effect model.22, 24, 28, 30, 31 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 5

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, performed using a fixed‐effect model.22, 24, 28, 30, 31 CI indicates confidence interval; IV, inverse variance; SE, standard error.

The heterogeneity was also high across this analysis. This was again explored using sensitivity analyses.

Publication Bias

There was no clear asymmetry in the funnel plot used to evaluate publication bias in the 10 studies addressing the survival outcome (Figure 6). It remains possible that some smaller studies with negative results might not have been published. For survival with a good neurologic outcome, after inspection of the results and nature of the studies, no evidence of a publication bias was observed.

Figure 6

Funnel plot for the evaluation of publication bias for the survival outcome.

Sensitivity Analyses

In the first set of sensitivity analyses, articles with some risk of bias were excluded. The results of these analyses did not differ from the ones presented in the main results (survival OR=2.13, 95% CI 1.73‐2.63, P<0.001; survival with a good neurologic outcome OR=2.50, 95% CI 2.06‐3.03, P<0.001) (Figures 7, 8, 9 through 10). However, the exclusion of these articles lowered the heterogeneity (survival I2 from 91% to 60%; survival with a good neurologic outcome I2 from 88% to 0%).

Figure 7

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, excluding studies with some risk of bias, performed using a random‐effect model.22, 23, 24, 25, 26, 27, 30, 31, 32, 33 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 8

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, excluding studies with some risk of bias, performed using a fixed‐effect model.22, 23, 24, 25, 26, 27, 30, 31, 32, 33 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 9

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, excluding studies with some risk of bias, performed using a random‐effect model.22, 24, 28, 30, 31 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 10

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, excluding studies with some risk of bias, performed using a fixed‐effect model.22, 24, 28, 30, 31 CI indicates confidence interval; IV, inverse variance; SE, standard error.

In the other 2 sets of sensitivity analyses, the association between the transport to cardiac resuscitation centers and both resuscitation outcomes seemed stronger among patients not having experienced prehospital ROSC than in those who did (survival OR 2.54, 95% CI 2.05‐3.15, P<0.001 versus OR 1.56, 95% CI 1.03‐2.36, P=0.04; survival with a good neurologic outcome OR=2.74, 95% CI 1.71‐4.38, P<0.001 versus OR=1.32, 95% CI 0.94‐1.86, P=0.11) (Figures 11, 12, 13, 14, 15, 16, 17 through 18). The number of articles that could be included in these sensitivity analyses was, however, limited.

Figure 11

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, including only patients who experienced prehospital return of spontaneous circulation, performed using a random‐effect model.22, 23, 26, 27, 32 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 12

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, including only patients who experienced prehospital return of spontaneous circulation, performed using a fixed‐effect model.22, 23, 26, 27, 32 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 13

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, including only patients who experienced prehospital return of spontaneous circulation, performed using a random‐effect model.22 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 14

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, including only patients who experienced prehospital return of spontaneous circulation, performed using a fixed‐effect model.22 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 15

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, including only patients who did not experience prehospital return of spontaneous circulation, performed using a random‐effect model.22, 25, 30, 32 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 16

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival, including only patients who did not experience prehospital return of spontaneous circulation, performed using a fixed‐effect model.22, 25, 30, 32 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 17

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, including only patients who did not experience prehospital return of spontaneous circulation, performed using a random‐effect model.22 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Figure 18

Meta‐analysis evaluating the association between transport to a cardiac resuscitation center and survival with a good neurologic outcome, including only patients who did not experience prehospital return of spontaneous circulation, performed using a fixed‐effect model.22 CI indicates confidence interval; IV, inverse variance; SE, standard error.

Discussion

The present systematic review and meta‐analysis sought to evaluate the association between the direct transport to cardiac resuscitation centers and resuscitation outcomes for patients suffering from an OHCA. Direct transport to a cardiac resuscitation center is associated with improved resuscitation outcomes for these patients. Interestingly, this association was stronger among patients without prehospital ROSC than among those who had experienced prehospital ROSC. Given the broad review performed, the presented results can now serve as the benchmark on this topic.

Despite the observational nature of the included articles, the quality of the evidence from which the present review's conclusions can be drawn is moderate.34 Indeed, all included articles were considered at low risk of bias and provided consistent results for 2 patient‐oriented outcomes. Despite some studies including only a selected subpopulation of OHCA, the global population assessed likely represents the population of interest. Given the large cohorts included, the obtained results were relatively precise for the main analyses. Although the observed association was significant, it did not reach the threshold required for it to be considered large.35 However, a dose‐response effect was observed, which upgrades the quality of the evidence.34 Because the vast majority of the included studies provided results adjusted for the Utstein criteria, which have been shown to predict most of the survival variability following OHCA, it is unlikely that any residual confounding would have significantly affected the main analyses.36 Finally, it is also unlikely that a publication bias would have altered significantly the presented results, given the consistency of the observed results and the absence of evidence of such a bias.

The potential benefits of being transported to a cardiac resuscitation center probably derive from the additional capabilities and experience that these centers have at treating patients suffering from an OHCA. Of note, acute coronary syndrome is the most common cause of OHCA, and its treatment of choice is PCI.1, 9, 37 Further, the timing of PCI also seems to be important for these patients because earlier treatment has been associated with better outcomes, even in the absence of an ST‐segment–elevation myocardial infarction on the initial ECG.38, 39, 40 Having on‐site access to this treatment all of the time increases the odds of providing this emergent intervention to patients.31 A similar argument can be made about TTM, which has been shown to increase survival among OHCA patients.10, 13, 14 In addition, many cardiac resuscitation centers are large, academic, tertiary or quaternary medical centers with increased exposure to and experience with patients suffering from severe disease. These characteristics have been associated less consistently with better outcomes for patients suffering from an OHCA, but it remains plausible that having more experienced professionals could be beneficial to these patients.26, 31, 41

The observation that the association between improved outcomes and direct transport to a cardiac resuscitation center is stronger among patients not having experienced prehospital ROSC had previously been made in 2 of the included studies.22, 32 Indeed, it is plausible that patients having the poorest prognosis are the ones who can benefit the most from the treatments available at cardiac resuscitation centers. However, this is based on a relative measure of effect. Given the observed difference in survival between patients having experienced prehospital ROSC and those who did not (≈50% versus ≈2%), it remains possible that patients having experienced prehospital ROSC could benefit the most in absolute terms from a direct transfer to cardiac resuscitation centers.22, 32

In light of the presented results, the remaining challenge is the operationalization of such a change in paradigm (transport to a cardiac resuscitation centers versus transport to the closest hospital) for emergency medical services. Multiple studies have concluded that there is no harm in prolonging the transport time of patients suffering from an OHCA, especially for patients having experienced prehospital ROSC.27, 29, 42, 43 The maximum tolerable bypass time for these patients remains uncertain, but delays of more than 30 minutes were still associated with improvements in survival for patients having experienced prehospital ROSC in 1 study.27 For patients not having experienced prehospital ROSC, this remains uncertain. In 1 study, a maximum bypass time of 14 minutes was proposed for a population consisting mostly of patients not having prehospital ROSC.32 The harm caused by prolonged transport for these patients was thought to be due to poor quality of the resuscitation during transport. However, 1 study observed that the quality of cardiopulmonary resuscitation did not decrease during transport.44 Because the observed benefit of direct transport to a cardiac resuscitation center in the study that proposed the maximum bypass time of 15 minutes was lower than what was observed in the meta‐analysis results, it is probably safe to tolerate a bypass time of 15 minutes for patients with ongoing resuscitation. This strategy should be tested in future prospective trials.

Limitations

The main limitation of this review is the observational nature of the articles it retained. In addition, although most of the literature is published in English, it is possible that a pertinent article was missed by the initial search. However, given the consistency of the observed results, this may still be unlikely to affect the overall conclusion. A minority of articles provided data for patients having and not having experienced prehospital ROSC. Albeit to a lesser extent, this is also true regarding survival with a good neurologic outcome. All of these analyses provided significant results, but the generalization of their results should still be made with caution. Some articles provided results while adjusting for other hospital characteristics in addition to being a cardiac resuscitation centers. Given the generally positive relationships between these other characteristics and resuscitation outcomes, this could have lowered the differences observed between the 2 groups in the present analysis.

Conclusions

Adult patients suffering from an OHCA transported to cardiac resuscitation centers seem to have better outcomes than their counterparts. It is reasonable to transport these patients directly to cardiac resuscitation centers (class IIa, level of evidence B‐nonrandomized). Future studies should further clarify how long a bypass time is tolerable for these patients, especially for the subpopulation of patients not having experienced prehospital ROSC.

Sources of Funding

This review received funding from the “Département de médecine familiale et de médecine d'urgence de l'Université de Montréal” and the “Fonds des Urgentistes de l'Hôpital du Sacré‐Cœur de Montréal.”

Disclosures

None.

Data S1. Search Strategies

Table S1. Newcastle‐Ottawa Quality Assessment Scale for Cohort Studies

Table S2. Template for Data Extraction

Click here for additional data file.