Skip to main content

Foundational Trials

ARREST

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Expedited transfer to a cardiac arrest centre for non-ST-elevation out-of-hospital cardiac arrest (ARREST): a UK prospective, multicentre, parallel, randomised clinical trial
  • Acronym: ARREST
  • Year: 2023
  • Journal published in: The Lancet
  • Citation: Patterson T, Perkins GD, Perkins A, Clayton T, Evans R, Dodd M, et al.; ARREST trial collaborators. Expedited transfer to a cardiac arrest centre for non-ST-elevation out-of-hospital cardiac arrest (ARREST): a UK prospective, multicentre, parallel, randomised clinical trial. Lancet. 2023;402:1329-1337.

Context & Rationale

  • Background
    • “Cardiac arrest centre” (CAC) models aim to concentrate specialist post–cardiac arrest care (24/7 coronary angiography/PCI capability, ICU expertise, protocolised neuroprognostication and organ support).
    • Pre-existing evidence for CAC destination policies was largely observational, with substantial risk of confounding by case-mix (eg, witnessed arrest, shockable rhythm, early ROSC) and selection effects.
    • International guideline groups called for a pragmatic randomised evaluation of destination to CAC for resuscitated OHCA without ST-elevation, because the benefit (if any) is system-level rather than a single procedure.
  • Research Question/Hypothesis
    • In adults with ROSC after non-ST-elevation OHCA, does expedited transport to a CAC reduce 30-day all-cause mortality compared with standard transport to the nearest acute hospital?
    • Hypothesis: earlier access to coronary investigation/reperfusion and specialist post-arrest care at a CAC would improve survival and neurological outcomes.
  • Why This Matters
    • Destination policies affect transport times, cath-lab/ICU capacity, and equity of access across regions.
    • A null effect would question universal bypass strategies and support selective triage based on likely treatable coronary pathology or need for advanced circulatory support.
    • A positive effect would justify regionalisation investments and prehospital triage pathways.

Design & Methods

  • Research Question: Among adults with ROSC after non-ST-elevation OHCA, does expedited transfer to a CAC (vs standard destination) reduce 30-day all-cause mortality?
  • Study Type: Prospective, pragmatic, multicentre, parallel-group, open-label, randomised superiority trial; prehospital (ambulance) randomisation within a UK metropolitan EMS system; multiple receiving hospitals with a designated CAC.
  • Population:
    • Setting: London Ambulance Service with enrolment after ROSC in the prehospital phase; receiving hospitals included a designated CAC and multiple “standard care” acute hospitals.
    • Inclusion: Adults (≥18 years); OHCA with ROSC; no ST-elevation on the initial ECG performed by paramedics after ROSC.
    • Key exclusions: ST-elevation meeting STEMI criteria; DNACPR/refusal; suspected pregnancy; presumed non-cardiac aetiology (eg, trauma, drowning, suicide attempt, drug overdose); presumed significant trauma/injury.
  • Intervention:
    • Expedited transport to a designated CAC following ROSC without ST-elevation, using a prehospital destination pathway.
    • CAC care included access to immediate coronary angiography/PCI when clinically indicated, plus specialist post–cardiac arrest ICU management (organ support, temperature management, and structured prognostication practices according to local standards).
  • Comparison:
    • Standard destination to the nearest receiving acute hospital (ED-based pathway) following ROSC without ST-elevation, with subsequent coronary angiography/transfer at clinician discretion.
    • Inter-hospital transfer to the CAC remained permissible and occurred in a minority of participants.
  • Blinding: Open-label (destination and subsequent in-hospital care could not be blinded); primary outcome (mortality) was objective, while functional outcomes introduced potential ascertainment bias.
  • Statistics: Sample size 860 (430 per group) planned to detect a 10% absolute reduction in 30-day mortality (from 60% to 50%) with 80% power at the 5% significance level, allowing ~10% loss; primary analysis was intention-to-treat using a generalised linear model to estimate risk ratios (participants with unknown 30-day status excluded from the primary analysis).
  • Follow-Up Period: Mortality to 30 days (primary) and 3 months; functional outcome (modified Rankin scale) at discharge and 3 months; health-related quality of life (EQ-5D-5L) at 3 months.

Key Results

This trial was not stopped early. Recruitment occurred from 15 Jan 2018 to 1 Dec 2022, with the planned sample size achieved.

Outcome CAC strategy Standard destination Effect p value / 95% CI Notes
30-day all-cause mortality (primary) 258/411 (63%) 258/412 (63%) RR (survival) 1.00 95% CI 0.90 to 1.11; P=0.96 Adjusted OR 1.09; 95% CI 0.73 to 1.63
3-month all-cause mortality 267/411 (65%) 263/411 (64%) RR (survival) 1.02 95% CI 0.92 to 1.12; P=Not reported Risk difference 1.0%; 95% CI −5.6 to 7.5
Modified Rankin scale at discharge (ordinal) n=413 n=402 OR 1.00 95% CI 0.76 to 1.32; P=0.99 Includes death as mRS 6
Modified Rankin scale at 3 months (ordinal) n=399 n=390 OR 0.98 95% CI 0.73 to 1.31; P=0.87 Includes death as mRS 6
Favourable mRS (0–3) at discharge 130/413 (32%) 130/402 (32%) RR 1.01 95% CI 0.92 to 1.11; P=0.79 Risk difference 0.9%; 95% CI −5.5 to 7.3
Favourable mRS (0–3) at 3 months 119/399 (30%) 119/390 (31%) RR 1.01 95% CI 0.92 to 1.11; P=0.83 Risk difference 0.7%; 95% CI −5.7 to 7.1
EQ-5D-5L index score at 3 months (mean [SD]) 0.68 (0.32); n=97 0.72 (0.25); n=92 Mean difference −0.04 95% CI −0.12 to 0.05; P=Not reported Measured in a subset of survivors
Serious adverse events not related to OHCA management 8/414 (2%) 3/413 (1%) Not reported Not reported Serious adverse events related to OHCA management: 0 in both groups
  • Primary outcome was neutral: 30-day mortality was 62.8% vs 62.6% (RR for survival 1.00; 95% CI 0.90 to 1.11; P=0.96).
  • The CAC strategy increased coronary angiography (231/412 [56%] vs 153/410 [37%]) and shortened time to angiography (median 2.3 h [IQR 1.6–3.4] vs 5.7 h [3.0–9.8]), without improvement in mortality or functional outcomes.
  • Control-group contamination occurred: 71/413 (17.2%) allocated to standard care were transferred to the CAC after initial hospital arrival.

Internal Validity

  • Randomisation and allocation:
    • Randomisation occurred prehospital after ROSC and ECG confirmation of no ST-elevation, supporting allocation concealment up to the point of randomisation.
    • Permuted blocks of varying size (4 and 6) were used.
  • Dropout / exclusions:
    • 862 patients were randomised; 20 (CAC) and 19 (standard) withdrew or had unknown 30-day status, leaving 411 and 412 in the primary analysis.
    • Functional outcome denominators were smaller than randomised totals (mRS at discharge: 413 vs 402; mRS at 3 months: 399 vs 390; EQ-5D-5L: 97 vs 92), increasing risk of attrition/selection bias for patient-centred secondary outcomes.
  • Performance / detection bias:
    • Open-label design could influence co-interventions (eg, thresholds for angiography, ICU practices, and withdrawal-of-life-sustaining therapy).
    • Primary endpoint (mortality) is objective; mRS and EQ-5D-5L are more vulnerable to bias and missingness.
  • Protocol adherence and crossover:
    • Protocol deviations were recorded (n=19), including three crossovers between destination strategies.
    • Contamination in the standard-care arm (71/413 [17.2%] transferred to the CAC) reduced separation of the system-level exposure.
  • Baseline comparability (selected prehospital metrics):
    • Age (mean [SD]): 63.8 (16.0) vs 63.2 (16.0); male sex: 286/414 (69%) vs 274/413 (67%).
    • Witnessed arrest: 331/414 (80%) vs 331/413 (80%); bystander CPR: 331/414 (80%) vs 331/413 (80%).
    • Time from collapse to ROSC (median [IQR]): 25 min (17–37) vs 25 min (17–37).
  • Heterogeneity:
    • Standard care occurred across multiple receiving hospitals with potentially variable access to cath labs, ICU expertise, and post-arrest bundles.
    • Final diagnoses included a substantial non-cardiac aetiology despite prehospital screening (86/411 [20.9%] vs 74/411 [18.0%]), which plausibly dilutes the effect of a coronary-capability intervention.
  • Timing:
    • Time from arrest to arrival at first hospital (median [IQR]) was longer with CAC destination: 84 min (70–103) vs 77 min (64–96).
    • Time to coronary angiography (median [IQR]) was shorter with CAC strategy: 2.3 h (1.6–3.4) vs 5.7 h (3.0–9.8).
  • Dose / intensity of the intervention:
    • Coronary angiography occurred in 231/412 (56%) vs 153/410 (37%).
    • Revascularisation occurred in 83/413 (20.1%) vs 65/409 (15.9%), implying that the incremental angiography yield did not translate into a large absolute increase in definitive coronary intervention.
  • Separation of the variable of interest:
    • Destination separation was attenuated by inter-hospital transfer (standard care to CAC: 71/413 [17.2%]).
    • Coronary angiography rate separation: 56% vs 37% (absolute difference 19%).
    • Time-to-angiography separation: 2.3 h vs 5.7 h (median difference 3.4 h).
  • Outcome assessment and statistical rigour:
    • Primary analysis estimated a risk ratio for survival using intention-to-treat principles; adjusted analyses were also reported and did not change inference.
    • Exclusion of participants with unknown mortality status from the primary analysis introduces a small risk of bias, but event counts were balanced and the effect estimate was near-null.

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomised design and objective primary outcome support causal inference, but open-label care, post-randomisation opt-out/unknown status, and meaningful contamination (standard-care transfers to CAC) plausibly attenuated any true destination effect.

External Validity

  • Population representativeness:
    • Represents adults with ROSC after OHCA in a high-resource urban EMS system, specifically excluding STEMI (where immediate PCI is already standard).
    • High rates of witnessed arrest and bystander CPR (~80% each) reflect a relatively optimised chain-of-survival environment, which may not extrapolate to systems with lower baseline performance.
  • Applicability across systems:
    • Requires EMS capability for immediate post-ROSC ECG triage and access to a designated CAC with 24/7 cath-lab and specialist ICU pathways.
    • In regions with longer transport distances, the observed increase in time-to-hospital arrival (84 vs 77 min) could be larger and potentially harmful.
    • In healthcare systems where “standard” receiving hospitals already provide high-quality post-arrest ICU and readily available angiography, incremental benefit of CAC destination is likely smaller.

Conclusion on External Validity: Generalisability is moderate: findings are most applicable to mature urban EMS networks with mixed receiving hospital capabilities and feasible bypass times, and less applicable to rural/remote systems or those with markedly different baseline post-arrest care infrastructure.

Strengths & Limitations

  • Strengths:
    • Pragmatic prehospital randomisation directly tested a system-level destination strategy (high policy relevance).
    • Large sample (n=862) with objective, clinically meaningful primary endpoint (30-day all-cause mortality).
    • Captures real-world delivery constraints (transport time trade-offs, inter-hospital transfers, and variable hospital capabilities).
  • Limitations:
    • Open-label design with potential for differential co-interventions and withdrawal-of-life-sustaining therapy practices.
    • Post-randomisation withdrawals/unknown 30-day status reduced the primary analysis set (411 vs 412) and secondary outcomes had substantial missingness (notably EQ-5D-5L).
    • Contamination reduced contrast between groups (17.2% of standard-care patients transferred to the CAC; 37% of standard-care patients received angiography).
    • Prehospital “presumed cardiac” screening was imperfect (≈18–21% non-cardiac cause by final diagnosis), plausibly diluting any coronary-capability benefit.

Interpretation & Why It Matters

  • Clinical policy signal
    • Routine bypass to a CAC for ROSC after non-ST-elevation OHCA did not improve 30-day survival, neurological recovery (mRS), or 3-month quality of life in this UK EMS system.
    • The strategy delivered more and faster coronary angiography, yet the downstream increase in revascularisation was modest and did not translate into improved outcomes.
  • Mechanistic implication
    • If benefit exists for “CAC care”, it is likely concentrated in clinically identifiable subgroups (eg, STEMI, cardiogenic shock, refractory ventricular arrhythmia, need for advanced mechanical circulatory support), rather than a universal ROSC-without-STEMI population.
    • Transport time increases may offset any marginal in-hospital advantages when baseline post-arrest care in standard hospitals is already strong.

Controversies & Subsequent Evidence

  • Contemporary commentary emphasised that ARREST’s null result challenges assumptions underpinning universal “cardiac arrest centre” bypass policies, and re-emphasises the primacy of early chain-of-survival interventions over destination changes.1
  • ARREST is directionally consistent with randomised evidence that routine immediate coronary angiography in comatose OHCA survivors without ST elevation does not improve survival, supporting a selective rather than universal invasive strategy in this phenotype.23
  • Prespecified subgroup analysis showed an interaction by age tertile for 30-day mortality (Pinteraction=0.003): <57 years 55/134 (41.0%) vs 90/156 (57.7%) RR 0.71; 95% CI 0.55 to 0.93; 57–71 years 103/153 (67.3%) vs 70/133 (52.6%) RR 1.28; 95% CI 1.01 to 1.61; ≥72 years 99/121 (81.8%) vs 93/121 (77.3%) RR 1.06; 95% CI 0.92 to 1.23.
  • Recent international guideline updates incorporate the accumulating randomised evidence base and recommend selective early coronary angiography based on ECG and clinical instability, while recognising uncertainty about universal CAC destination for non-ST-elevation OHCA populations.4

Summary

  • ARREST randomised 862 adults with ROSC after non-ST-elevation OHCA to expedited CAC transport versus standard destination in a pragmatic UK EMS system.
  • There was no difference in 30-day mortality: 258/411 (63%) vs 258/412 (63%); RR for survival 1.00; 95% CI 0.90 to 1.11; P=0.96.
  • There was no improvement in functional outcome (mRS) at discharge or 3 months, and EQ-5D-5L at 3 months was similar (mean difference −0.04; 95% CI −0.12 to 0.05).
  • The CAC strategy increased coronary angiography (56% vs 37%) and shortened time to angiography (2.3 h vs 5.7 h) but did not translate into improved survival or neurological recovery.
  • Interpretation must account for contamination (17.2% of standard-care patients transferred to the CAC) and post-randomisation withdrawals/unknown outcomes, which likely reduced exposure separation and power for secondary outcomes.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Overall Takeaway

ARREST is a landmark pragmatic EMS-randomised destination trial showing that routine expedited transport to a cardiac arrest centre for ROSC after non-ST-elevation OHCA did not improve 30-day survival or patient-centred neurological and quality-of-life outcomes. The findings support selective, physiology-driven triage rather than universal bypass policies, especially where standard receiving hospitals already deliver strong post–cardiac arrest ICU care and timely coronary evaluation.

Overall Summary

  • Universal CAC bypass for non-ST-elevation OHCA with ROSC did not reduce 30-day mortality (RR for survival 1.00; 95% CI 0.90–1.11).
  • More and faster angiography at the CAC did not translate into improved survival or neurological outcomes.
  • System design decisions should prioritise transport-time trade-offs and selective identification of patients most likely to benefit from specialist centre resources.

Bibliography