Skip to main content

Foundational Trials

Olasveengen

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Intravenous drug administration during out-of-hospital cardiac arrest: a randomized trial
  • Year: 2009
  • Journal published in: JAMA
  • Citation: Olasveengen TM, Sunde K, Brunborg C, Thowsen J, Steen PA, Wik L. Intravenous drug administration during out-of-hospital cardiac arrest: a randomized trial. JAMA. 2009;302(20):2222-2229.

Context & Rationale

  • Background
    • Advanced life support algorithms historically included routine intravenous (IV) access and vasopressor/antiarrhythmic administration during cardiopulmonary resuscitation, despite limited randomised evidence for improved survival to hospital discharge.
    • Physiological benefits (raising coronary/cerebral perfusion pressure) were balanced against plausible harms: impaired microcirculation, increased myocardial oxygen demand, arrhythmogenesis, and potential worsening of neurological injury.
    • Obtaining IV access and administering drugs can compete with high-quality CPR delivery (interruptions, attention diversion), making “real-world net benefit” uncertain even if physiological effects are favourable.
  • Research Question/Hypothesis
    • Does IV drug administration during out-of-hospital cardiac arrest (OHCA) improve survival to hospital discharge compared with a strategy that withholds IV drug administration during resuscitation?
    • Trialists explicitly powered the study assuming survival would be doubled in patients not receiving epinephrine (projected 14% vs 7%).
  • Why This Matters
    • This trial tested a foundational component of advanced life support in a pragmatic prehospital environment where time pressure, competing tasks, and team behaviours shape effectiveness.
    • It directly interrogated whether improvements in short-term endpoints (ROSC, hospital admission) translate into patient-centred outcomes (survival and neurological status).
    • It prefigured later debates about the validity of ROSC as a surrogate endpoint in cardiac arrest trials.

Design & Methods

  • Research Question: In adult, non-traumatic OHCA, does an advanced life support strategy that includes IV access and IV drug administration improve survival to hospital discharge versus withholding IV drug administration during resuscitation?
  • Study Type: Prospective, parallel-group, randomised trial; prehospital (Oslo emergency medical service, Norway); open-label (unblinded); investigator-initiated; patient-level randomisation via sealed envelopes.
  • Population:
    • Setting: single metropolitan EMS system serving Oslo, with transport to 2 university hospitals; ALS delivered by paramedics and physician-staffed emergency units.
    • Inclusion: adults (>18 years) with non-traumatic OHCA in whom resuscitation was attempted by EMS.
    • Key exclusions (predefined): traumatic aetiology; arrest witnessed by ambulance crew; resuscitation initiated/interrupted by a physician outside the resuscitation team; suspected asthma/anaphylactic shock (added in 2006).
    • Recruitment profile: 1183 assessed; 916 randomised; 851 included in primary analysis (418 IV group; 433 no-IV group).
  • Intervention:
    • Advanced life support with IV access and IV drug administration during resuscitation, with drug selection and dosing guided by contemporaneous international guidelines.
    • Drugs actually delivered (primary analysis cohort): epinephrine 330/418 (79%), atropine 194/418 (46%), amiodarone 69/418 (17%).
  • Comparison:
    • Advanced life support without access to IV drug administration during resuscitation (no attempt to establish IV access or administer IV drugs during CPR).
    • IV access could be established ≥5 minutes after ROSC (and drugs then given if clinically indicated).
    • Protocol separation in primary analysis cohort: IV drugs during resuscitation occurred in 42/433 (10%).
  • Blinding: Unblinded (procedural intervention). Outcomes such as ROSC and survival were objective, but intensity/duration of resuscitation and transport/ICU admission decisions were potentially susceptible to performance bias.
  • Statistics: A total of 900 patients were required to detect a doubling of survival to hospital discharge (from 7% to 14%) with 91.4% power at the 5% significance level; analysis was intention-to-treat, but patients randomised and later found to meet predefined exclusion criteria were not included in the intention-to-treat analysis (modified intention-to-treat).
  • Follow-Up Period: To hospital discharge (including cerebral performance category at discharge) and 1 year (national population registry).

Key Results

This trial was not stopped early. Interim analyses were performed annually by an external researcher without disclosure of results to the investigators.

Outcome Intravenous drug administration (n=418) No intravenous drug administration (n=433) Effect p value / 95% CI Notes
Survival to hospital discharge (primary) 44 (10.5%) 40 (9.2%) OR 1.16 95% CI 0.74 to 1.82; P=0.61 Adjusted OR 1.15; 95% CI 0.69 to 1.91; P=0.59
Favourable neurological outcome at discharge (CPC 1–2) 41 (9.8%) 35 (8.1%) OR 1.24 95% CI 0.77 to 1.98; P=0.45 CPC distribution did not differ significantly between groups
Any ROSC during resuscitation 165 (40%) 107 (25%) OR 1.99 95% CI 1.48 to 2.67; P<0.001 Short-term outcome
Admitted to hospital 178 (43%) 126 (29%) OR 1.81 95% CI 1.36 to 2.40; P<0.001 Includes admitted with ROSC and with ongoing CPR
Admitted to hospital with ROSC 133 (32%) 89 (21%) Not reported P<0.001 Subset of hospital admissions
Admitted to ICU 125 (30%) 88 (20%) OR 1.67 95% CI 1.22 to 2.29; P=0.002 Among ICU admissions: therapeutic hypothermia 90/125 (72%) vs 62/88 (70%)
Alive 1 year after cardiac arrest 41 (10%) 36 (8%) Not reported P=0.53 Follow-up via national population registry
  • IV drug administration was associated with substantially higher short-term survival markers (ROSC 165/418 [40%] vs 107/433 [25%]; ICU admission 125/418 [30%] vs 88/433 [20%]).
  • There was no statistically significant difference in survival to discharge (44/418 [10.5%] vs 40/433 [9.2%]) or favourable neurological outcome (CPC 1–2: 41/418 [9.8%] vs 35/433 [8.1%]).
  • Rhythm-stratified outcomes suggested a pronounced ROSC/hospital admission signal in non-shockable arrests (ROSC 80/274 [29%] vs 32/291 [11%]) without a corresponding survival-to-discharge benefit.

Internal Validity

  • Randomisation and Allocation:
    • Allocation by sealed envelopes opened after confirmation of cardiac arrest and initiation of resuscitation.
    • 95 eligible patients were not randomised (e.g., “forgot to randomise”, envelopes unavailable, perceived futility), creating potential for selection bias before randomisation.
  • Drop out or exclusions (post-randomisation):
    • 65/916 (7.1%) randomised patients were excluded from the primary analysis after randomisation because predefined exclusion criteria were identified later (24 IV group; 41 no-IV group).
    • This constitutes a modified intention-to-treat analysis and can introduce bias if post-randomisation exclusions are imbalanced or correlated with prognosis.
  • Performance/Detection Bias:
    • Unblinded intervention; clinicians could plausibly differ in persistence, transport decisions, or threshold for termination.
    • Primary endpoint (survival to discharge) is objective; intermediate endpoints (hospital/ICU admission) can be influenced by clinician behaviour and system practices.
  • Protocol Adherence:
    • Among those randomised to IV drugs (n=442): 344 received IV drug administration as randomised; 74 did not have IV established before end of resuscitation (including 42 ROSC before IV drugs; 12 inability to establish access; 12 judged futile; 8 no explanation).
    • Among those randomised to no-IV strategy (n=474): 388 had no IV drugs as randomised; 45 received IV drugs (27 ROSC followed by rearrest; 13 hospital admission; 5 protocol breach).
  • Baseline Characteristics:
    • Groups were broadly comparable: age 64 (SD 17–18), male 70–72%, presumed cardiac aetiology 70–72%, bystander CPR 62–63%.
    • Initial rhythm distribution showed borderline imbalance (asystole 53% vs 46%; P=0.06; PEA 15% vs 20%; P=0.06), which may influence absolute event rates.
  • Heterogeneity:
    • Single EMS system and standardised protocols likely reduced inter-site variability.
    • However, real-world heterogeneity in arrest aetiology and rhythm remains, reflected in rhythm-stratified analyses.
  • Timing:
    • IV access was frequently not established before end of resuscitation (74/442 randomised to IV strategy).
    • In the no-IV arm, IV access was explicitly deferred until ≥5 minutes after ROSC, meaning that early post-ROSC pharmacology could still occur.
  • Dose:
    • Drug dosing followed contemporaneous international guidelines; specific dosing schedules were not detailed in the main report.
  • Separation of the Variable of Interest:
    • IV drugs during resuscitation: 343/418 (82%) vs 42/433 (10%).
    • Epinephrine: 330/418 (79%) vs 37/433 (9%); atropine: 194/418 (46%) vs 20/433 (5%); amiodarone: 69/418 (17%) vs 17/433 (4%) (all P<0.001).
  • Key Delivery Aspects:
    • CPR quality measures were similar between groups (e.g., hands-off ratio median 0.15 vs 0.14; compressions/min 94 vs 94).
    • Resuscitation intensity differed: CPR duration mean 22 (95% CI 20–23) minutes in the IV group vs 18 (17–19) minutes in the no-IV group (P<0.001), raising concern for differential persistence and “resuscitation dose”.
  • Outcome Assessment:
    • Survival endpoints were objective; neurological outcome was measured using CPC at hospital discharge.
    • Hospital/ICU admission outcomes are clinically meaningful but can be sensitive to system behaviours and unblinded decision-making.
  • Statistical Rigor:
    • Power calculation assumed a very large effect (doubling survival) that was not observed, and the analysed sample (n=851) was smaller than the planned 900 due to post-randomisation exclusions.
    • Observed absolute differences in survival were small (10.5% vs 9.2%), increasing risk of type II error for modest but clinically relevant effects.

Conclusion on Internal Validity: Overall, internal validity is moderate: randomisation and objective primary outcomes support causal inference, but unblinded delivery, post-randomisation exclusions, and meaningful protocol non-adherence (plus longer CPR duration in the IV arm) introduce plausible performance bias and dilute group separation.

External Validity

  • Population Representativeness:
    • Adult, non-traumatic OHCA with high bystander CPR prevalence (~62–63%) in a single Scandinavian urban EMS system.
    • Physician-staffed emergency units and robust CPR-quality monitoring may not reflect paramedic-only or resource-limited systems.
  • Applicability:
    • The “IV drug administration” strategy was a bundle that included atropine (now removed from contemporary adult PEA/asystole algorithms) and amiodarone; modern practice also includes broader use of intraosseous access.
    • Core inference remains relevant: improving ROSC/hospital admission does not necessarily translate into improved survival with good neurological outcome.

Conclusion on External Validity: Generalisability is moderate: the trial reflects real-world prehospital practice but is anchored in a specific EMS configuration and an era of evolving drug algorithms, limiting direct transferability of the bundled drug strategy to contemporary systems.

Strengths & Limitations

  • Strengths:
    • Randomised evaluation of a difficult-to-study prehospital intervention.
    • Prospective capture of CPR-quality metrics, supporting interpretation of mechanistic/process confounding.
    • Patient-centred outcomes (survival to discharge, neurological status) plus 1-year follow-up.
    • Clear protocol distinction (82% vs 10% exposure to IV drugs in the analysed cohort).
  • Limitations:
    • Unblinded design with clinically meaningful endpoints (hospital/ICU admission) susceptible to behavioural/system influence.
    • Modified intention-to-treat with post-randomisation exclusions (65/916), and pre-randomisation non-enrolment of eligible patients (95) for operational reasons.
    • Protocol non-adherence in both arms (IV group often never achieved IV access before resuscitation ended; 10% crossover to IV drugs in the no-IV cohort).
    • Intervention tested a bundled pharmacotherapy approach (epinephrine, atropine, amiodarone), limiting attribution to any single drug.
    • Power assumptions were optimistic (doubling survival), leaving limited power for modest effects and for neurological outcomes.

Interpretation & Why It Matters

  • Signal and meaning
    IV drug administration increased short-term survival markers (ROSC, hospital and ICU admission) but did not improve survival to discharge or favourable neurological outcome, challenging the clinical validity of ROSC-based surrogacy in cardiac arrest care.
  • Mechanistic interpretation
    The combination of higher ROSC rates and longer CPR duration in the IV arm is consistent with either drug-mediated perfusion effects, greater resuscitation persistence, or both; the absence of a discharge survival signal suggests downstream harm, futility, or simply insufficient effect size.
  • Trialist-methodologist lesson
    For resuscitation trials, intermediate outcomes are necessary but insufficient; design features that influence “resuscitation dose” (duration, transport decisions) can materially alter short-term endpoints without improving patient-centred survival.

Controversies & Subsequent Evidence

  • Surrogate endpoints vs patient-centred outcomes: The trial demonstrated dissociation between increased ROSC/hospital admission and unchanged survival to discharge; contemporaneous critical appraisal reinforced that short-term resuscitation metrics do not guarantee meaningful survival.1
  • Bundled pharmacotherapy limits inference: The tested “IV drug administration” strategy included multiple agents (including atropine, later removed from many adult algorithms), constraining attribution of benefit/harm to epinephrine alone and complicating translation to modern practice.
  • Open-label delivery and resuscitation intensity: Longer mean CPR duration in the IV group (22 vs 18 minutes) and more frequent defibrillation raise the possibility that part of the short-term signal reflects differential persistence and decision-making rather than pharmacologic effects alone.
  • Later RCTs refined the signal: PARAMEDIC2 (epinephrine vs placebo) showed increased survival but with more severe neurological impairment among survivors, echoing the pattern of improved short-term outcomes with limited neurological benefit.2
  • Smaller placebo-controlled trial context: A prior double-blind placebo-controlled RCT of adrenaline suggested improved ROSC without definitive survival benefit, limited by sample size—broadly consistent with Olasveengen’s “ROSC without discharge survival” pattern.3
  • Synthesis evidence: Systematic review/meta-analysis and network meta-analysis data support that vasopressors improve ROSC and short-term outcomes, with uncertain or small effects on survival with good neurological outcome, aligning with the mechanistic/clinical dissociation seen here.45
  • Guideline position: International consensus recommendations continue to support vasopressor use in adult cardiac arrest while acknowledging trade-offs in neurological outcome certainty, reflecting the broader evidence pattern to which this trial contributed.6

Summary

  • In 851 analysed adult OHCA patients, IV drug administration increased ROSC (40% vs 25%), hospital admission (43% vs 29%), and ICU admission (30% vs 20%).
  • There was no statistically significant difference in survival to discharge (10.5% vs 9.2%) or favourable neurological outcome at discharge (CPC 1–2: 9.8% vs 8.1%).
  • Protocol non-adherence was clinically important (IV access not established before end of resuscitation in 74/442 randomised to IV strategy; IV drugs administered in 45/474 randomised to no-IV strategy).
  • CPR quality measures were similar, but mean CPR duration was longer in the IV group (22 vs 18 minutes), introducing a “resuscitation dose” confounder consistent with performance bias risk in open-label trials.
  • The trial is landmark for demonstrating that improving short-term resuscitation endpoints does not necessarily improve meaningful survival, shaping later trial design and interpretation in cardiac arrest pharmacotherapy.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Abbreviations: CPC = cerebral performance category; CPR = cardiopulmonary resuscitation; ICU = intensive care unit; IV = intravenous; OHCA = out-of-hospital cardiac arrest; OR = odds ratio; ROSC = return of spontaneous circulation.

Overall Takeaway

Olasveengen’s trial is landmark because it randomised a core element of advanced life support and showed a consistent pattern: IV drug administration improved short-term resuscitation endpoints but did not demonstrably improve survival to discharge or favourable neurological outcome. This dissociation helped reshape how the field values intermediate endpoints and strengthened the call for patient-centred, placebo-controlled resuscitation pharmacotherapy trials.

Overall Summary

  • IV drugs increased ROSC and hospital/ICU admission, but did not improve discharge survival or favourable neurological outcomes, highlighting limits of ROSC as a surrogate endpoint.

Bibliography