Skip to main content

Foundational Trials

EMERGE

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Emergency vs Delayed Coronary Angiogram in Survivors of Out-of-Hospital Cardiac Arrest: Results of the Randomized, Multicentric EMERGE Trial
  • Acronym: EMERGE
  • Year: 2022
  • Journal published in: JAMA Cardiology
  • Citation: Hauw-Berlemont C, Lamhaut L, Diehl JL, et al; EMERGE Trial Investigators. Emergency vs delayed coronary angiogram in survivors of out-of-hospital cardiac arrest: results of the randomized, multicentric EMERGE trial. JAMA Cardiol. 2022;7(7):700-707.

Context & Rationale

  • Background
    • Out-of-hospital cardiac arrest (OHCA) survivors without ST-segment elevation can still have an acute coronary culprit lesion, but post-return of spontaneous circulation (ROSC) ECG has imperfect sensitivity for acute coronary occlusion.
    • Observational registries and early invasive “cardiac arrest centre” pathways supported early coronary angiography (CAG) and percutaneous coronary intervention (PCI) as potentially time-critical and modifiable contributors to outcome.
    • Competing priorities in early post-arrest care include haemodynamic stabilisation, ventilation, temperature management, and neuroprognostication; routine emergency CAG could delay ICU-based care, add iatrogenic risk, and consume system resources.
    • Before EMERGE, equipoise persisted regarding whether a routine “go-to-cath-lab” strategy improves patient-centred outcomes in OHCA without ST-elevation, as opposed to a selective or delayed approach.
  • Research Question/Hypothesis
    • Whether an emergency CAG strategy is superior to a delayed CAG strategy (48–96 hours) for improving 180-day survival free of major neurological sequelae in OHCA survivors without ST-segment elevation and without an obvious non-cardiac cause.
  • Why This Matters
    • Routine emergency CAG is a high-resource, high-consequence pathway that drives prehospital triage and hospital system design.
    • Demonstrating benefit would justify system-wide cath-lab activation and early invasive treatment for a broad OHCA cohort; demonstrating no benefit would support an ICU-first, selective invasive strategy.
    • Neurologically intact survival is the outcome that matters to patients and families, and it is plausibly affected by both early coronary reperfusion (if indicated) and uninterrupted high-quality post-resuscitation ICU care.

Design & Methods

  • Research Question: In adult survivors of OHCA without ST-segment elevation and without an obvious non-cardiac cause, does an emergency coronary angiography strategy (vs delayed coronary angiography at 48–96 hours) improve 180-day survival free of major neurological sequelae?
  • Study Type: Investigator-initiated, multicentre, open-label, randomised, parallel-group, superiority trial; national (France); randomisation in the prehospital/early hospital phase; intervention in the cardiac catheterisation laboratory; post-resuscitation care in ICU.
  • Population:
    • Adults (≥18 years) with out-of-hospital sudden cardiac arrest and sustained ROSC.
    • No obvious non-cardiac cause identified at presentation (eg, trauma/neurological/respiratory cause).
    • Absence of ischaemic ST-segment elevation on the post-ROSC ECG (left bundle branch block was not treated as a STEMI equivalent for inclusion).
    • Admitted to a centre with ICU capability and 24/7 interventional cardiology service.
    • Key exclusions: in-hospital cardiac arrest; no ROSC; ischaemic ST-segment elevation; suspected non-cardiac aetiology; major comorbidity with life expectancy <1 year; pregnancy; legal protection status; participation in another interventional trial.
    • Deferred consent framework: patients could be excluded post-randomisation if consent could not be obtained (per trial flow diagram).
    • Randomisation stratified by pre-specified prognostic strata: “Group 1” required all of age <75 years, initial shockable rhythm, no-flow <5 minutes, low-flow <20 minutes, and epinephrine dose <1 mg; “Group 2” otherwise.
  • Intervention:
    • Emergency coronary angiography strategy with immediate transfer to the cardiac catheterisation laboratory as soon as feasible after admission.
    • PCI performed during the index procedure when an angiographic culprit lesion was identified, consistent with contemporary invasive coronary treatment strategies for OHCA survivors.
    • Post-resuscitation ICU care otherwise delivered according to local practice (eg, temperature management, ventilatory/haemodynamic support).
  • Comparison:
    • Delayed coronary angiography strategy planned for 48–96 hours after ICU admission.
    • Urgent coronary angiography before the planned window permitted when clinically indicated (eg, deterioration or suspected ongoing coronary ischaemia), at treating team discretion.
    • Post-resuscitation ICU care according to local practice, as per the intervention group.
  • Blinding: Open-label for clinicians and sites; primary endpoint assessed by an independent physician masked to randomisation allocation.
  • Statistics: Planned total sample size 970 to detect a 10% absolute improvement in the primary endpoint (from 36% to 46%) with 80% power at a two-sided 5% significance level (allowing for 30% crossover and 10% loss to follow-up); primary analysis by intention-to-treat with time-to-event methods (Kaplan–Meier/log-rank; Cox proportional hazards models adjusted for stratification group and accounting for centre effects).
  • Follow-Up Period: 180 days (with additional neurological outcome ascertainment at ICU discharge and 90 days).

Key Results

This trial was stopped early. Recruitment ceased after funding termination when a prespecified recruitment-rate threshold was not reached; 279 patients were included in analyses (planned 970).

Outcome Emergency CAG strategy Delayed CAG strategy Effect p value / 95% CI Notes
Primary: 180-day survival free of major neurological sequelae (CPC 1–2) 47/141 (34.1%) 42/138 (30.7%) HR 0.87 95% CI 0.65 to 1.15; P=0.32 Time-to-event analysis; major neurological sequelae defined as CPC 3–5
180-day survival 51/141 (36.2%) 46/138 (33.3%) HR 0.86 95% CI 0.64 to 1.15; P=0.31 No between-group difference detected
Coronary angiography performed 126/141 (89.4%) 74/138 (53.6%) Not reported Not reported In the delayed group, 23/74 (31.1%) underwent “urgent” angiography before the planned 48–96 h window
Time from OHCA to coronary angiography, median (IQR), hours 2 (2 to 3) 65.5 (40.8 to 74.8) Not reported Not reported Demonstrates protocol separation in timing of angiography
Shock during first 48 hours 50/129 (38.8%) 53/133 (39.8%) RR 1.03 95% CI 0.76 to 1.39; P=0.86 Risk ratio reported in trial table; definition as per protocol
Ventricular tachycardia/ventricular fibrillation during first 48 hours 10/141 (7.1%) 5/138 (3.6%) RR 0.51 95% CI 0.18 to 1.46; P=0.21 No statistically significant difference
Withdrawal of life-sustaining care 56/141 (39.7%) 65/138 (47.1%) RR 1.19 95% CI 0.91 to 1.55; P=0.22 Potentially important competing pathway for outcome; no between-group difference detected
  • Emergency CAG did not improve 180-day survival with favourable neurological outcome: 34.1% vs 30.7% (HR 0.87; 95% CI 0.65 to 1.15; P=0.32).
  • There was large separation in angiography timing (median 2 hours vs 65.5 hours), yet no signal of improved patient-centred outcomes in the analysed sample.
  • Early stopping meant substantially reduced statistical power versus the planned sample size, limiting precision around clinically important effect sizes.

Internal Validity

  • Randomisation and Allocation:
    • Centralised, web-based 1:1 randomisation with variable block sizes; allocation sequence generated in advance and held by the sponsor.
    • Randomisation performed by an emergency physician at the dispatch centre (independent of recruiting clinicians), or within 1 hour of admission if not randomised prehospital.
    • Stratification by a prognostic stratum (Group 1 vs Group 2) based on age, initial rhythm, no-flow/low-flow times, and epinephrine dose.
  • Drop out or exclusions:
    • 338 patients were randomised; 59 (17.5%) were excluded after randomisation (37 lacked consent; 16 did not meet inclusion criteria; 6 other reasons), leaving 279 in the analysed cohort.
    • Deferred consent with post-randomisation exclusion introduces the possibility of selection after allocation, and represents a departure from a strict intention-to-treat principle.
    • Follow-up attrition at 180 days was low but present (patients lost to follow-up or declining follow-up were reported in the participant flow diagram); primary analysis used time-to-event methods with censoring.
  • Performance/Detection Bias:
    • Open-label allocation could influence co-interventions (eg, haemodynamic support decisions, timing of neuroprognostication) and clinician behaviour.
    • Primary endpoint was assessed by an independent physician masked to allocation, mitigating detection bias for neurological outcome classification.
  • Protocol Adherence:
    • Emergency group: coronary angiography performed in 126/141 (89.4%); 15/141 (10.6%) did not undergo angiography.
    • Delayed group: coronary angiography performed in 74/138 (53.6%); 23/74 (31.1%) underwent angiography earlier than the planned window; 64/138 (46.4%) had no angiography.
    • PCI among those undergoing angiography: 38/126 (30.2%) vs 17/74 (23.0%).
  • Baseline Characteristics:
    • Groups were broadly similar: mean age 65.4 (13.1) vs 63.9 (12.7) years; male sex 103/141 (73.1%) vs 92/138 (66.7%).
    • Initial rhythm was predominantly non-shockable: 92/141 (65.2%) vs 96/138 (69.9%).
    • Post-resuscitation coma severity was similar: Glasgow Coma Scale median 3 (IQR 3–3) in both groups.
    • Targeted temperature management use was similar: 73/141 (52.0%) vs 74/138 (53.7%).
  • Heterogeneity:
    • Multicentre design increases clinical heterogeneity in ICU and cath-lab practice; analyses accounted for centre effects in modelling.
    • Case-mix included shockable and non-shockable rhythms and a substantial burden of non-coronary precipitants that were not “obvious” at enrolment, reflecting pragmatic enrolment but also diluting a potential coronary-specific treatment effect.
  • Timing:
    • Separation in timing was achieved: time from OHCA to angiography median 2 (2 to 3) hours vs 65.5 (40.8 to 74.8) hours.
    • Mean delay from randomisation to angiography was 0.6 (3.7) hours vs 55.1 (37.2) hours.
  • Dose:
    • The “dose” of invasive management, in terms of angiography completion and revascularisation, differed less than timing alone might suggest because many delayed-strategy patients died before angiography and a subset underwent urgent early angiography.
    • No significant coronary disease was common among those catheterised: 57/126 (45.2%) vs 41/74 (55.4%).
  • Separation of the Variable of Interest:
    • Angiography exposure: 126/141 (89.4%) vs 74/138 (53.6%).
    • Timing exposure: 2 (2 to 3) hours vs 65.5 (40.8 to 74.8) hours from arrest to angiography.
    • Strategy contamination: “urgent” angiography occurred in 23/138 delayed-strategy patients (performed within 14.7 hours), with 7 undergoing intervention.
  • Key Delivery Aspects:
    • Enrolment occurred early (prehospital or within 1 hour of admission), which is appropriate for testing an early cath-lab pathway.
    • Control strategy was an explicit delayed pathway (48–96 hours) rather than an uncontrolled “usual care” approach, improving interpretability.
  • Outcome Assessment:
    • Primary endpoint was clinically meaningful (neurologically intact survival) and assessed at 180 days.
    • Cerebral Performance Category is a standard post-arrest measure but remains partly subjective; masking of assessors mitigates (but does not eliminate) classification variability.
  • Statistical Rigor:
    • Time-to-event analysis with pre-specified covariate adjustment was appropriate for the endpoint definition.
    • Early stopping substantially reduced the ability to exclude clinically important effects, and the wide confidence intervals reflect imprecision.

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomisation and timing separation were strong, and outcome assessment was masked, but early stopping, substantial post-randomisation exclusions related to consent/eligibility, and meaningful protocol non-adherence in the delayed group reduce precision and raise residual bias concerns.

External Validity

  • Population Representativeness:
    • Includes a pragmatic OHCA cohort without ST-elevation, with a majority presenting with non-shockable rhythms, which mirrors many real-world OHCA case-mixes in contemporary systems.
    • Enrolment required admission to centres with 24/7 interventional cardiology and ICU capability, which may select for higher-resource systems and structured regional pathways.
  • Applicability:
    • Findings apply to OHCA survivors without ST-elevation and without an obvious non-cardiac cause; they do not address patients with ST-elevation (where emergency angiography is standard).
    • In systems with limited cath-lab access, results support that a delayed/selective approach may be acceptable for many stable post-arrest patients without ST-elevation, provided urgent angiography remains available when clinically indicated.
    • Generalisation to pathways with different neuroprognostication practices, withdrawal-of-care norms, or post-arrest ICU bundles should be made cautiously.

Conclusion on External Validity: Generalisability is moderate: the OHCA population is clinically representative, but applicability is strongest in systems able to deliver both rapid ICU care and selective/urgent cath-lab access within a structured regional network.

Strengths & Limitations

  • Strengths:
    • Randomised design addressing a high-impact systems-of-care question in OHCA without ST-elevation.
    • Clinically meaningful primary endpoint prioritising neurological outcome at 180 days.
    • Early randomisation (prehospital/early hospital) testing the real-world cath-lab pathway decision point.
    • Masked endpoint assessment by an independent physician.
    • Substantial achieved separation in angiography timing (median 2 hours vs 65.5 hours).
  • Limitations:
    • Early stopping with marked under-recruitment versus the planned sample size, resulting in reduced power and imprecision.
    • Post-randomisation exclusions for deferred consent/eligibility, deviating from strict intention-to-treat principles.
    • Open-label intervention with potential co-intervention and care-pathway effects.
    • Protocol non-adherence and contamination in the delayed group: only 53.6% underwent angiography, and 23 patients underwent “urgent” early angiography.
    • High competing risk of death from neurological injury limits the proportion of patients for whom coronary revascularisation can plausibly be outcome-modifying.

Interpretation & Why It Matters

  • Clinical practice
    • A routine emergency angiography strategy did not improve neurologically intact survival compared with a delayed (48–96 h) strategy in this OHCA cohort without ST-elevation.
    • These data support an ICU-first, selective invasive strategy in many patients, while preserving urgent angiography for evolving clinical indications.
  • Trialist/methodologist lens
    • The pragmatic case-mix increases relevance but likely dilutes a coronary-specific treatment effect in a population where neurological injury dominates outcome.
    • Deferred consent and early stopping are pivotal design/operational constraints in emergency care trials that meaningfully shape interpretability and precision.
  • What’s next
    • Future work should sharpen selection for potentially reversible coronary occlusion (eg, physiology, imaging, biomarkers, refined ECG timing) rather than broad routine cath-lab activation.
    • Systems research should integrate patient-centred outcomes and competing risks (withdrawal-of-care practices, neuroprognostication) into pathway evaluation.

Controversies & Subsequent Evidence

  • Precision and interpretability after early stopping:
    • Stopping far short of the planned sample size limits the ability to rule in or rule out clinically important benefit or harm, despite point estimates close to neutrality.
  • Deferred consent and post-randomisation exclusions:
    • Excluding 59 randomised participants after allocation (notably for lack of consent) creates a “modified ITT” cohort and can introduce selection bias if exclusions differ by prognosis or pathway effects.
  • Contamination of the control strategy:
    • The delayed strategy included both “no angiography” (often due to early death) and “urgent early angiography” (23 patients), complicating causal attribution to timing alone.
  • How EMERGE aligns with other randomised evidence:
    • Other randomised trials in OHCA without ST-elevation also reported no improvement in key outcomes with routine immediate angiography strategies (COACT, PEARL, TOMAHAWK).123
  • Tension between earlier observational support and trial results:
    • Consensus and registry-driven pathways historically supported early invasive management after OHCA, based on perceived prevalence of treatable culprit lesions and observational survival associations.45
    • EMERGE’s angiographic findings reinforce the challenge of case selection: nearly half of catheterised patients had no significant coronary disease, and PCI occurred in 30.2% vs 23.0% among those undergoing angiography.

Summary

  • In OHCA survivors without ST-elevation and without an obvious non-cardiac cause, an emergency angiography strategy did not improve 180-day survival free of major neurological sequelae compared with a delayed (48–96 h) strategy.
  • The trial stopped early for funding/recruitment reasons, enrolling 279 of a planned 970 participants, limiting statistical power and precision.
  • Separation in angiography timing was substantial (median 2 hours vs 65.5 hours), yet patient-centred outcomes were similar.
  • Protocol adherence differed: angiography was performed in 89.4% vs 53.6%, and urgent early angiography occurred in a subset of delayed-strategy patients.
  • Findings support an ICU-first approach with selective/urgent angiography when clinically indicated, rather than routine immediate angiography for all such patients.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Where risk ratios are presented, interpret direction in line with the original trial tables; no additional calculations were performed for this summary.

Overall Takeaway

EMERGE tested a systems-defining question in post-cardiac arrest care: whether routine emergency coronary angiography improves neurologically intact survival in OHCA survivors without ST-elevation. Despite substantial separation in angiography timing, the trial found no improvement in 180-day neurological outcome or survival, and early stopping with non-adherence limits precision rather than reversing the overall message. In the context of concordant randomised evidence, EMERGE supports an ICU-first, selective invasive strategy for many such patients, reserving urgent coronary angiography for evolving clinical indications.

Overall Summary

  • Routine emergency angiography after OHCA without ST-elevation did not improve 180-day neurologically intact survival versus a delayed (48–96 h) strategy; trial stopped early and was underpowered.

Bibliography