Emergency vs Delayed Coronary Angiogram in Survivors of Out-of-Hospital Cardiac Arrest: Results of the Randomized, Multicentric EMERGE Trial

Key Points

Question  Does performing an emergency coronary angiogram (CAG) vs a delayed CAG after an out-of-hospital cardiac arrest (OHCA) without ST-segment elevation on the postresuscitation electrocardiogram (ECG) lead to a better outcome?

Findings  In this randomized clinical trial including 279 patients, there was no difference in the 180-day survival rate with Cerebral Performance Category 1 or 2, which included 34.1% of patients in the emergency CAG group and 30.7% of patients in the delayed CAG group. However, the study was underpowered; a combined analysis on pooled published data of 4 randomized trials on the same topic revealed the same conclusion.

Meaning  Results of this study do not support emergency CAG after OHCA without ST-segment elevation.

Importance  Although an emergency coronary angiogram (CAG) is recommended for patients who experience an out-of-hospital cardiac arrest (OHCA) with ST-segment elevation on the postresuscitation electrocardiogram (ECG), this strategy is still debated in patients without ST-segment elevation.

Objective  To assess the 180-day survival rate with Cerebral Performance Category (CPC) 1 or 2 of patients who experience an OHCA without ST-segment elevation on ECG and undergo emergency CAG vs delayed CAG.

Design, Setting, and Participants  The Emergency vs Delayed Coronary Angiogram in Survivors of Out-of-Hospital Cardiac Arrest (EMERGE) trial randomly assigned survivors of an OHCA without ST-segment elevation on ECG to either emergency or delayed (48 to 96 hours) CAG in 22 French centers. The trial took place from January 19, 2017, to November 23, 2020. Data were analyzed from November 24, 2020, to July 30, 2021.

Main Outcomes and Measures  The primary outcome was the 180-day survival rate with CPC of 2 or less. The secondary end points were occurrence of shock, ventricular tachycardia, and/or fibrillation within 48 hours, change in left ventricular ejection fraction between baseline and 180 days, CPC scale at intensive care unit discharge and day 90, survival rate, and hospital length of stay.

Results  A total of 279 patients (mean [SD] age, 64.7 [14.6] years; 195 men [69.9%]) were enrolled, with 141 (50.5%) in the emergency CAG group and 138 (49.5%) in the delayed CAG group. The study was underpowered. The mean (SD) time delay between randomization and CAG was 0.6 (3.7) hours in the emergency CAG group and 55.1 (37.2) hours in the delayed CAG group. The 180-day survival rates among patients with a CPC of 2 or less were 34.1% (47 of 141) in the emergency CAG group and 30.7% (42 of 138) in the delayed CAG group (hazard ratio [HR], 0.87; 95% CI, 0.65-1.15; P = .32). There was no difference in the overall survival rate at 180 days (emergency CAG, 36.2% [51 of 141] vs delayed CAG, 33.3% [46 of 138]; HR, 0.86; 95% CI, 0.64-1.15; P = .31) and in secondary outcomes between the 2 groups.

Conclusions and Relevance  In this randomized clinical trial, for patients who experience an OHCA without ST-segment elevation on ECG, a strategy of emergency CAG was not better than a strategy of delayed CAG with respect to 180-day survival rate and minimal neurologic sequelae.

Trial Registration  ClinicalTrials.gov Identifier: NCT02876458

Sudden cardiac death remains a major public health issue. The rate of patients discharged from the hospital with no or minimal neurologic sequelae is low, varying between 2% and 18%.^1-5 The most common cause of sudden cardiac death is ischemic cardiovascular disease, mostly acute coronary artery occlusion.^6-9 However, selection of patients for an invasive strategy is still debated. Several retrospective studies show that the probability of finding an acute coronary artery lesion during an early coronary angiogram (CAG) is high (70%-80%) if ST-segment elevation on the postresuscitation electrocardiogram (ECG) is present. Therefore, guidelines recommend performing an emergent CAG in survivors of sudden cardiac death with no obvious noncardiac cause of arrest and ST-segment elevation. In patients with cardiac arrest without ST-segment elevation on postresuscitation ECG, the benefit of an emergency CAG is still a matter of debate. In these patients, the rate of acute coronary artery lesion is much lower (15%-20%).⁸ Three recent randomized studies^10-12 performed in patients without ST-segment elevation after an out-of-hospital cardiac arrest (OHCA) showed no difference in survival whether CAG was performed early or at a deferred time. However, 1 study excluded patients with nonshockable rhythm¹⁰ and the second trial was underpowered.¹¹ There is, therefore, still a need for data from randomized trials on this issue. The Emergency vs Delayed Coronary Angiogram in Survivors of Out-of-Hospital Cardiac Arrest (EMERGE) trial tested the hypothesis that the 180-day survival rate of patients with no or minimal neurologic sequelae who experience OHCA without ST-segment elevation on the postresuscitation ECG and no obvious noncardiac cause of arrest would be improved when they underwent emergency CAG compared with those who received delayed CAG.

The EMERGE trial was a national, multicenter, randomized open-label parallel-group trial, in which survivors of OHCA were randomly assigned (1:1) to either emergency CAG or delayed CAG (within 48 to 96 hours). The trial was conducted from January 19, 2017, to November 23, 2020, and the trial protocol has been published previously.¹³ Survivors of OHCA are usually comatose after resuscitation, and they cannot provide an informed consent for participation in a trial. From January 2017 to 2019, if proxies were present on the site of cardiac arrest, they were asked to provide an informed consent before inclusion and randomization. Owing to difficulties in obtaining such a consent during prehospital care, an amendment was added to the protocol on September 2019 stating that a signed consent was no longer required for inclusion. The patient and/or family members (or next of kin) were informed as soon as possible, and their consent was sought for research to be continued (Supplements 1, 2, and 3).

The protocol was approved by the Agence Nationale de Sécurité du Médicament et des Produits de Santé and the French Ethics Committee (Comité de Protection des Personnes Ile de France IV). This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.

Eligible for inclusion were patients older than 18 years who experienced an OHCA with return of spontaneous circulation (ROSC), without an obvious noncardiac cause of arrest¹⁴ and who were admitted to a center with an intensive care unit and a 24/7 interventional cardiology department. Exclusion criteria were as follows: patients younger than 18 years, those with in-hospital cardiac arrest, patients without ROSC, the presence of ST-segment elevation on the postresuscitation ECG according to European Society of Cardiology guidelines,¹⁵ suspected noncardiac etiology, presence of comorbidities with life expectancy of less than 1 year, pregnancy, adults subject to a legal protection measure (guardianship or curatorship), and participation in another interventional trial.

In France, management of OHCA is performed by the mobile intensive care unit system, with 1 physician specialized in emergency medicine, 1 nurse, and 1 ambulance driver. Calls for OHCA are received by a physician at the medical dispatch center. The included patients were from 22 French centers and were monitored for up to 180 days. Randomization was performed electronically, with a centralized randomization system (Cleanweb software; Telemedicine Technologies SAS) by the dispatch center. Randomization could also be done at hospital arrival if it had not been performed before. In this case, randomization had to be performed within the hour after the patient’s admission. Patients allocated to the emergency CAG group were transferred directly to the catheterization laboratory. Patients randomly assigned to the delayed CAG group were admitted to the intensive care unit, and a CAG was planned 48 to 96 hours after admission.

The primary study outcome was the 180-day survival rate with no or minimal neurologic sequelae. All survivors were evaluated by an independent physician who was blinded to the randomization to assess the Cerebral Performance Category (CPC) level at 180 days. A patient was assigned CPC categories 1 or 2 if they had no or minimal neurologic sequelae 180 days after the event.

Secondary end points were shock, tachycardia, and fibrillation episodes during the first 48 hours after hospital admission, changes in left ventricular ejection fraction between baseline and 180 days as assessed by echocardiogram, major neurologic sequelae defined by CPC score 3 or 4 (assessed at intensive care unit discharge, 90 days, and 180 days), all-cause death, and hospital length of stay. Definitions of the trial outcomes are presented in the eAppendix in Supplement 4.

The treatments implemented in the intensive care unit were in accordance with European guidelines, adjusted to each unit’s practices, particularly regarding targeted temperature management (between 32 °C and 36 °C), and hemodynamic management.¹⁵ Investigators were allowed to perform a CAG sooner than 48 to 96 hours in the delayed group if the following events appeared: ST-segment elevation or new left bundle branch block on the ECG, shock unresponsive to inotropes, electrical storm (repeated ventricular fibrillation or tachycardia), or new segmental hypokinesia or akinesia on an echocardiogram.

The CAG and percutaneous coronary intervention (PCI) were performed according to the strategies described in the consensus statement from the European Association for Percutaneous Cardiovascular Interventions/Stent for Life groups.¹⁶ In summary, use of the radial approach was encouraged. PCI was performed when a culprit lesion was clearly identified using angiographic criteria (recent occlusion, angiographic features suggesting an unstable lesions). Revascularization of nonculprit lesions was discouraged. The use of drug-eluting stents and preprocedure unfractionated heparin was recommended. The postprocedure regimen included aspirin and an antiplatelet P2Y12 inhibitor.

The first patient was randomly assigned on January 19, 2017, and the last 6-month follow-up was performed on November 23, 2020. Follow-up was conducted during outpatient clinic visits scheduled at 90 (range, 80-100) days and 180 (range, 170-190) days after cardiac arrest, during a medical consultation in the center which had included the patient. Patients for whom no outpatient visit was possible were contacted by telephone or followed up through contact with the patient’s relatives to determine neurologic status with the CPC score.

Based on results of the Parisian Region Out-of-Hospital Cardiac Arrest (PROCAT) study,⁸ the 180-day survival rate with no or minimal sequelae in the delayed CAG group was estimated at 36%. Assuming an absolute difference of 10% between the 2 groups, ie, a 180-day survival rate with CPC 1 or 2 of 46% in the emergency CAG group, with a 2-sided 5% type I error, 80% power, and a 180-day maximum length of follow-up, a total number of 678 patients was required. A total of 970 patients included were planned to account for a 10% lost-to-follow-up rate and a crossover rate of 30%.

All analyses were performed on an intention-to-treat basis. Clinical event rates and other categorical data are summarized as percentages. Continuous data are presented as mean (SD) or as median (IQR). Kaplan-Meier plots were constructed for time-to-event outcomes, with treatment effects estimated using Cox models and results presented as hazard ratios (HRs) with their 95% CIs. A Schoenfeld test was used to check the proportional-hazards assumption. Secondary outcomes are presented with HRs, risk ratios, and 95% CIs. The widths of the CIs have not been adjusted for multiplicity, and any inferences drawn from these intervals may not be reproducible. Analyses were performed using SAS software, version 9.2 (SAS Institute) and R statistical software, version 4.0.2 (R Core Team).

We performed a search for randomized clinical trials comparing early vs delayed CAG in patients who experience cardiac arrest without ST-segment elevation. Databases including PubMed, Embase, and Google Scholar were used to obtain all relevant studies up to October 2021, using key words including “cardiac arrest” and “coronary angiography” or “coronary angiogram” and “randomized controlled trial,” which worked as Medical Subject Headings and free-style words in searching retrieval. All publications containing the relevant key words were included in this study. Risk ratios and corresponding 95% CIs as well of the summary risk ratio (random-effects model) were obtained using the R-package metaphor (metafor 2021), version 3.0-2 (R Foundation). Statistical heterogeneity was tested using the Q statistic generated from the χ² test and was considered statistically significant for a 2-sided P value < .05. Data were analyzed from November 24, 2020, to July 30, 2021.

A total of 279 patients (mean [SD] age, 64.7 [14.6] years; 195 men [69.9%]; 84 women [30.1%]) were enrolled during the study period from January 2017 to June 2019. At that time, funding was terminated according to prespecified rules regarding the rate of inclusions, requiring at least 50% of inclusions in 2 years. The enrollment objective was therefore not achieved. A total of 141 patients (50.5%) were randomly assigned to the emergency CAG group, and 138 patients (49.5%) were assigned to the delayed CAG group (Figure 1).

There were no differences between the 2 groups regarding cardiopulmonary resuscitation characteristics (Table 1). The initial rhythm was nonshockable in 65.2% of patients (90 of 138) in the early CAG group and 69.9% of patients (93 of 133) in the delayed CAG group.

CAG was performed in 126 of 141 patients (89.4%) in the emergency CAG group vs 74 of 138 patients (53.6%) in the delayed CAG group. The mean (SD) time delay between randomization and CAG was 0.6 (3.7) hours in the emergency CAG group and 55.1 (37.2) hours in the delayed CAG group. CAG was performed within a median of 65.5 (IQR, 40.8-74.8) hours after the OHCA in the delayed CAG group compared with 2 (IQR, 2-3) hours in the emergency CAG group. The main reason for not performing CAG in the delayed CAG group was early death from a neurologic cause in 14 patients, a cardiorespiratory etiology in 25 patients, and an unknown cause in 1 patient. In the delayed CAG group, CAG was performed urgently in 23 patients among the 138 patients (16.7%) of this group within 14.7 hours, and 7 underwent a percutaneous intervention. No significant coronary artery disease was found in 45.2% of patients (57 of 126) in the emergency CAG group and in 55.4% of patients (41 of 74) in the delayed CAG group. The rate of PCI among patients who underwent CAG was higher in the emergency CAG group (38 [30.2%]) than in the delayed CAG group (17 [23%]). The main reason for not performing PCI was the absence of significant lesion (35 of 88 patients [39.8%] in the early CAG group and 20 of 57 patients [35.1%] in the delayed CAG group) or the absence of clear culprit lesion (42.1% in both groups).

There was no difference in the 180-day survival rates with CPC 1 or 2: 34.1% (47 of 141) in the emergency CAG group and 30.7% (42 of 138) in the delayed CAG group (HR, 0.87; 95% CI, 0.65-1.15; P = .32) (Table 2, Figure 2). There was no difference in the overall survival rate at 180 days (emergency CAG, 36.2% [51 of 141] vs delayed CAG, 33.3% [46 of 138]; HR, 0.86; 95% CI, 0.64-1.15; P = .31) and in secondary outcomes between the 2 groups (Table 2).

The search for randomized clinical trials comparing early vs delayed CAG in patients who experienced cardiac arrest without ST-segment elevation found 3 studies^10-12 published in the literature, performed with similar methods. We summarized their results with those of the EMERGE study. As shown in Figure 3, the results of this combined analysis, comprising a total of 1446 patients, are in accordance with those of the EMERGE study alone, showing no benefit of an early CAG in this population.

In the EMERGE randomized multicenter clinical trial, 2 strategies were compared: emergency CAG vs delayed CAG, in survivors of an OHCA with no obvious noncardiac cause of arrest and no ST-segment elevation on ECG. No significant difference was found between the 2 groups in the 180-day survival rate with CPC 1 or 2. However, the study patient enrollment target goal was not achieved, and thus, the EMERGE study was underpowered to adequately assess the primary end point.

Nonrandomized studies suggested a benefit of an emergency invasive strategy in survivors of OHCA regardless of ECG findings. Spaulding et al⁷ showed that coronary artery disease was the cause of more than 70% of OHCA and that performing an emergency CAG seemed to improve patient outcome. The Larsen and Ravkilde meta-analysis¹⁷ showed that the prevalence of an acute coronary artery lesion ranged from 59% to 71% in patients with OHCA and without an obvious noncardiac cause of arrest. In 2015, the European Resuscitation Council and European Society of Intensive Care Medicine defined in their guidelines for postresuscitation care that all survivors of OHCA should be sent to a specialized center that can perform an emergent CAG in the event of an ST-segment elevation and can consider performance of a CAG as soon as possible in the event of a non–ST-segment elevation.¹⁵ However, several recent studies have challenged those recommendations in survivors of an OHCA without ST-segment elevation on the post-ROSC ECG, who represent the majority of OHCA survivors.¹⁸

Our results are in line with 3 recent randomized clinical trials, even though the Coronary Angiography After Cardiac Arrest (COACT) trial used survival at 90 days as a primary end point,¹⁰ the Randomized Pilot Clinical Trial of Early Coronary Angiography vs No Early Coronary Angiography for Postcardiac Arrest Patients Without ST-Segment Elevation (PEARL) study used a composite primary outcome,¹¹ and the Immediate Unselected Coronary Angiography vs Delayed Triage in Survivors of Out-of-Hospital Cardiac Arrest Without ST-Segment Elevation (TOMAHAWK) study used all-cause death at 30 days as the primary end point.¹² Because these studies were performed with similar methods, we summarized their results with those of the EMERGE study. The combined analysis on 1446 patients shows no benefit of an early strategy.

Our study adds new insights as EMERGE included mostly patients with a nonshockable rhythm at presentation. These patients were excluded from randomization in the COACT trial and represented only 23.5% of patients in the PEARL study. The EMERGE trial randomized a nonselected population that did not exclude patients with shock after ROSC in contrast to COACT, which could have lowered the expected benefit of an emergency CAG. Indeed, patients with cardiogenic shock and myocardial infarction are known to benefit the most from early revascularization.^19,20 Of note, in France, the first medical evaluation is done very early in the course of the OHCA, as the first aid unit is led by a physician. Data from EMERGE suggest that in this specific population, even when medical evaluation and randomization between emergency vs delayed CAG is done early, a direct admission to a catheterization laboratory is debatable and may challenge strategy in terms of organization and resources. The absence of superiority of emergency CAG has important consequences on the organization of the chain of survival.

Future research should focus on identifying the subgroup of patients who experience an OHCA with non–ST-segment elevation who would benefit from an early CAG. The PROCAT 2 cohort study²¹ found that younger age, a shorter resuscitation length (<20 minutes), an initial shockable rhythm, and low-dose epinephrine were significantly associated with a better neurologic outcome after a PCI in an OHCA population with non–ST segment elevation on ECG after ROSC but that the subgroup of older men (50 years or older) resuscitated from an initially shockable rhythm mostly benefitted from an emergent CAG strategy.

In randomized trials, patients are selected based on predefined variables but not on clinical history, warning signs, and OHCA circumstances. A meta-analysis of the 3 randomized trials and EMERGE, based on such individual factors, could help to define more precisely the subpopulation that could benefit from an early CAG after an OHCA without an extracardiac cause and without ST-segment elevation on ECG. These patients could be triaged and admitted directly to the catheterization laboratory. However, if CAG and PCI can benefit a specific population, the best timing of this procedure remains unknown.

In EMERGE, the main reason for not performing PCI was the absence of a significant or clear culprit lesion. Fifty percent of patients presented with potential signs of ischemia on the ECG. Ischemic ECG changes alone seem to be a poor marker of the presence of a culprit lesion.²²The absence of benefit of an emergent CAG strategy may, therefore, be attributable to the low rate of acute lesions. Furthermore, the main cause of death in OHCA is neurologic damage. Of interest, the higher rate of PCI in the emergency CAG group suggests that death occurred in patients with significant coronary artery disease despite emergency CAG and PCI when indicated. Future research could focus on diagnostic methods and algorithms to select survivors of OHCA with no ST-segment elevation who present with unstable lesions and who may benefit from an early invasive strategy. A strategy using ECG findings has been proposed but has not been evaluated on a large number of patients.²³

This study had several limitations. Preplanned sample size was not achieved, and thus, the EMERGE study was underpowered to adequately assess the primary and secondary end points. Overestimation of the number of cardiac arrests meeting the inclusion and exclusion criteria and difficulties in obtaining informed consent were the main reasons for not achieving the preplanned sample size. Similar studies are ongoing and are having similar recruitment difficulties.^24,25 The physicians were not blinded to randomized treatment allocation but were not involved in the research process. Determination of culprit lesions is subjective, and the angiograms were not analyzed by a core laboratory. Patients with refractory cardiac arrest were excluded, and therefore, no conclusion on the value of CAG in this group of patients can be drawn. Finally, the echocardiograms and the follow-up visits, including neurologic assessment, were not evaluated by a core laboratory.

In this randomized clinical trial of patients successfully resuscitated after an OHCA without ST-segment elevation, a strategy of emergency CAG was not found to be better than a strategy of delayed CAG with respect to the 180-day survival rate with no or minimal neurologic sequelae. The study patient enrollment goal was not achieved, and therefore, the study was underpowered to adequately assess the primary and secondary end points. However, our results are consistent with previously published studies and do not support emergency CAG in survivors of OHCA without ST elevation.

Accepted for Publication: April 12, 2022.

Published Online: June 8, 2022. doi:10.1001/jamacardio.2022.1416

Corresponding Author: Christian Spaulding, MD, PhD, European Hospital Georges Pompidou, Assistance Publique–Hôpitaux de Paris, Paris Cité University, Sudden Cardiac Death Expert Center, INSERM U 971, PARCC, 20 Rue Leblanc, 75015 Paris, France (christian.spaulding@aphp.fr).

Author Contributions: Dr Spaulding had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hauw-Berlemont, Lamhaut, Vilfaillot, Chatellier, Cariou, Spaulding.

Acquisition, analysis, or interpretation of data: Hauw-Berlemont, Diehl, Andreotti, Varenne, Leroux, Lascarrou, Guerin, Loeb, Daubin, Beygui, Boissier, Marjanovic, Christiaens, Vilfaillot, Glippa, Djadi Prat, Chatellier, Cariou, Spaulding.

Drafting of the manuscript: Hauw-Berlemont, Lascarrou, Vilfaillot, Glippa, Chatellier, Cariou, Spaulding.

Critical revision of the manuscript for important intellectual content: Hauw-Berlemont, Lamhaut, Diehl, Andreotti, Varenne, Leroux, Guerin, Loeb, Daubin, Beygui, Boissier, Marjanovic, Christiaens, Djadi Prat, Chatellier, Cariou, Spaulding.

Statistical analysis: Hauw-Berlemont, Vilfaillot, Chatellier, Spaulding.

Obtained funding: Hauw-Berlemont, Diehl, Djadi Prat, Spaulding.

Administrative, technical, or material support: Hauw-Berlemont, Lascarrou, Glippa, Djadi Prat, Chatellier.

Supervision: Hauw-Berlemont, Lamhaut, Glippa, Djadi Prat, Cariou, Spaulding.

Conflict of Interest Disclosures: Dr Hauw-Berlemont reported receiving grants from the French Ministry of Health during the conduct of the study. Dr Lamhaut reported receiving grants from the French Ministry of Health during the conduct of the study. Dr Varenne reported receiving grants from Abbott Medical and Boston Scientific during the conduct of the study; personal fees from Servier; and nonfinancial support from Astra Zeneca and Biosensors outside the submitted work. Dr Lascarrou reported receiving personal fees from Bard Pharmaceuticals and Zoll during the conduct of the study. Dr Guerin reported receiving personal fees from Abbott Medical and grants from Edwards during the conduct of the study. Ms Glippa and Dr Prat reported receiving grants from the French Ministry of Health during the conduct of the study. Dr Cariou reported receiving lecture fees from Bard during the conduct of the study. No other disclosures were reported.

Funding/Support: This work was supported in part by Assistance Publique– Hôpitaux de Paris and the French Ministry of Health, through the national Programme Hospitalier de Recherche Clinique.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

The EMERGE Investigators: The EMERGE Investigators are listed in Supplement 5.

Data Sharing Statement: See Supplement 6.

Additional Information: Coauthor Eric Roupie, MD, PhD, died February 2, 2022.