The Lancet OCTOBER 17, 2009

Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial

Giles J Peek, Miranda Mugford, Ravindranath Tiruvoipati, Ann Wilson, Elizabeth Allen, Mariamma M Thalanany, Clare L Hibbert, Ann Truesdale, Felicity Clemens, Nicola Cooper, Richard K Firmin, Diana Elbourne

Bottom Line

Referral of adults with severe, potentially reversible respiratory failure to a specialist center for consideration of ECMO significantly improved 6-month survival without severe disability compared to continued conventional ventilatory support.

Key Findings

1. Survival to 6 months without severe disability was significantly higher in the ECMO referral group compared to the conventional management group (63% [57/90] vs. 47% [41/87]; RR 0.69, 95% CI 0.05-0.97, p=0.03).
2. Only 75% (68/90) of patients randomized to the ECMO referral arm actually received ECMO, indicating the trial evaluated a management strategy (referral to a specialist center) rather than just the ECMO device itself.
3. Referral for ECMO consideration was shown to be highly cost-effective, leading to a gain of 0.03 quality-adjusted life-years (QALYs) at 6 months.
4. A lifetime model predicted the cost per QALY of the ECMO referral strategy to be £19,252 (95% CI 7,622-59,200) at a discount rate of 3.5%, well within accepted thresholds for cost-effectiveness.

Study Design

Design
Multicenter RCT
Open-Label
Sample
180
Patients
Duration
6 mo
Median
Setting
Multicenter, UK
Population Adults (aged 18-65 years) with severe but potentially reversible respiratory failure, defined by a Murray score ≥3.0 or uncompensated hypercapnia with pH <7.20 despite optimal conventional ventilation.
Intervention Referral and transfer to a specialized tertiary center for consideration of ECMO and protocolized respiratory management.
Comparator Continued conventional ventilatory management at the referring hospital.
Outcome Composite of death or severe disability at 6 months after randomization.

Study Limitations

The pragmatic trial design evaluated a strategy of transfer to a single specialized center, meaning the outcome differences could be attributed to the specialized critical care and protocolized lung-protective ventilation at the ECMO center rather than the ECMO circuit itself (the 'CESAR effect').
Nearly a quarter (25%) of the patients randomized to the intervention arm never actually required or received ECMO, complicating the assessment of ECMO's direct therapeutic efficacy.
Ventilatory management was not standardized in the control arm; lung-protective low-tidal-volume ventilation was utilized less frequently in the conventional management group (70%) compared to the ECMO center group (93%).
Incomplete long-term follow-up data was noted in a significant portion of patients, and there was a lack of blinding for treating clinicians.

Clinical Significance

The CESAR trial fundamentally altered modern critical care by providing the first robust, contemporary randomized evidence supporting the use of ECMO for severe adult ARDS and respiratory failure. By proving that transferring carefully selected patients (Murray score ≥3.0 or pH <7.20) to highly experienced ECMO centers drastically reduces mortality and severe morbidity, it catalyzed the development of regionalized ECMO retrieval networks and established ECMO as a viable, standard-of-care rescue therapy.

Historical Context

Early randomized trials from the 1970s and 1990s (e.g., the 1979 Zapol trial and 1994 Morris trial) failed to demonstrate a survival benefit for adult ECMO, largely due to severe bleeding complications, outdated circuit technology, and the simultaneous use of highly injurious ventilator settings. By the late 2000s, advances in biocompatible circuits, better anticoagulation, and the establishment of strict lung-protective ventilation principles raised the hypothesis that ECMO combined with 'lung rest' could improve outcomes. CESAR was the first large-scale, modern trial to systematically evaluate and validate this approach, paving the way for the explosive growth of ECMO utilization observed during the subsequent 2009 H1N1 influenza and COVID-19 pandemics.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What are the physiological differences between Veno-Venous (VV) and Veno-Arterial (VA) ECMO, and which modality is primarily indicated for patients like those in the CESAR trial with severe respiratory failure but intact cardiac function?

Key Response

VV-ECMO supports gas exchange (oxygenation and CO2 removal) and facilitates lung protective ventilation by relying on the patient's own cardiac output to perfuse organs. VA-ECMO bypasses the heart and lungs, providing both respiratory and hemodynamic support. For isolated severe ARDS with preserved hemodynamics, VV-ECMO is the standard of care.

Resident
Resident

Based on the inclusion criteria of the CESAR trial, what clinical parameters define 'severe, potentially reversible respiratory failure' that should prompt a resident to initiate a transfer to an ECMO center?

Key Response

The CESAR trial utilized the Murray lung injury score greater than 3.0 or uncompensated hypercapnia (pH less than 7.20) to define severe respiratory failure. Recognizing high FiO2 and PEEP requirements failing to maintain oxygenation, while understanding contraindications like prolonged high-pressure ventilation greater than 7 days, is crucial for timely referral before irreversible lung injury occurs.

Fellow
Fellow

Approximately 25% of the patients randomized to the ECMO arm in the CESAR trial never actually received ECMO. How does this 'intention-to-treat' analysis impact the interpretation of the study's findings regarding the true efficacy of ECMO versus the benefit of management at a highly specialized ARDS center?

Key Response

The trial actually demonstrated the benefit of transferring patients to a specialist center capable of ECMO, not just the machine itself. Because a quarter of patients improved with strict lung-protective ventilation and advanced management at the referral center, the survival benefit reflects the combined effect of expert ARDS management and ECMO availability.

Attending
Attending

The CESAR trial showed a significant survival benefit for patients referred to an ECMO center, but establishing an ECMO program is highly resource-intensive. How should an ICU attending balance the decision to transfer a severe ARDS patient to a regional ECMO hub versus maximizing rescue therapies like proning at their home institution?

Key Response

Attendings must navigate the risk of transport against the benefit of specialized care. The CESAR trial advocates for established early referral networks. If local rescue therapies are failing, delaying transport to an ECMO center can worsen outcomes, especially if the patient passes the 7-day threshold of high-pressure mechanical ventilation, rendering them ineligible for ECMO.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The CESAR trial employed a pragmatic, unblinded, multicenter design where the control group received conventional management at local hospitals without a mandated, standardized mechanical ventilation protocol. How does this lack of a standardized control protocol introduce confounding, and how might a trialist design a subsequent study to isolate the effect of ECMO?

Key Response

The lack of a strict lung-protective ventilation protocol in the control group is a major methodological critique, as control patients may have received injurious ventilation (performance bias). A follow-up study, such as the later EOLIA trial, must mandate optimal, standardized conventional care (e.g., strict low tidal volumes, early proning) in both arms to truly isolate the efficacy of the ECMO intervention itself.

Journal Editor
Journal Editor

From an editorial standpoint, the primary outcome of the CESAR trial was 'survival without severe disability at 6 months', yet the trial lacked standardized care in the control arm. What are the major threats to validity regarding this pragmatic design, and how would a reviewer require authors to frame their conclusions?

Key Response

A tough reviewer would flag the blending of 'referral to an ECMO center' with 'receiving ECMO' and the performance bias introduced by unstandardized control care. Editors would demand clear delineation in the abstract and conclusion that the intervention tested was a 'system-level referral strategy' rather than an assessment of the isolated extracorporeal technology, ensuring readers do not misinterpret the mechanism of benefit.

Guideline Committee
Guideline Committee

When evaluating the CESAR trial alongside subsequent trials like EOLIA for formulating clinical practice guidelines on ARDS, what Level of Evidence does this provide, and how should guidelines structure the recommendation for ECMO referral in severe ARDS?

Key Response

CESAR provides high-quality (Level 1b) evidence for the systemic strategy of referral to an ECMO center. Major guidelines (e.g., ATS/ESICM/SCCM and ELSO) strongly recommend that patients with severe ARDS failing conventional management (including prone positioning) be evaluated for and transferred to ECMO-capable centers. The recommendation relies on CESAR to emphasize that the expertise of an ECMO center improves outcomes, even if the extracorporeal circuit is ultimately withheld.

Clinical Landscape

Noteworthy Related Trials

2000

ARMA Trial (ARDSNet)

n = 861 · NEJM

Tested

Low tidal volume ventilation (6 ml/kg)

Population

Patients with acute lung injury and ARDS

Comparator

Traditional tidal volume ventilation (12 ml/kg)

Endpoint

In-hospital mortality and ventilator-free days

Key result: Low tidal volume ventilation significantly reduced mortality from 39.8% to 31.0% and increased ventilator-free days.
2013

PROSEVA Trial

n = 466 · NEJM

Tested

Prone positioning for at least 16 hours per day

Population

Patients with severe ARDS

Comparator

Supine positioning

Endpoint

28-day all-cause mortality

Key result: Prone positioning significantly decreased 28-day mortality from 32.8% to 16.0% compared to the supine position.
2018

EOLIA Trial

n = 249 · NEJM

Tested

Early venovenous ECMO

Population

Adults with severe ARDS

Comparator

Conventional mechanical ventilation (with optional ECMO crossover)

Endpoint

60-day mortality

Key result: There was no significant difference in 60-day mortality between early ECMO and conventional ventilation, though a high rate of crossover to ECMO occurred in the control group.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis