CESAR: Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure. A multicenter randomised controlled trial

Peek G. Lancet 2009;374:1351-1363

Clinical Question

  • Is extracorporeal membrane oxygenation (ECMO) a safe, efficacious and cost-effective treatment compared with conventional ventilation for the management of severe adult respiratory failure?


  • Randomised controlled trial
  • 1:1 ratio to receive continued conventional ventilation (CV) or referral to consideration for treatment of ECMO
    • 24% (n = 22) of patients randomised to ECMO never received this treatment
    • 5 patients died either before or during transfer
    • The remaining 17 patients continued with conventional ventilation
  • Non-blinded to clinicians although researchers were blinded at 6 month follow up
  • Power calculations were based on an anticipated 70% mortality in the conventional management group. Assuming a 10% risk of severe disability in the surviving patients in both treatment groups, 120 patients in each group (240 in total) were required to detect deduction by a 1/4 in the proportion of patients achieving the primary outcome from 73% to 55%. This power calculation was reviewed in 2003 (at this point 63 patients (60%) were recruited) and a lower sample size of 180 patients was selected to detect a reduction by a 1/3
  • Intention to treat analysis


  • UK based study.
  • 92 conventional tertiary ICU’s and 11 referral hospitals (which sent patients to conventional centres if allocated to receive conventional management)
  • Glenfield Hospital, Leicester was the only centre where ECMO treatment took place
  • July 2001 – August 2006


  • Inclusion: severe but potentially reversible respiratory failure
    • Murray Score of 3 or more (based on paO2/FiO2 ratio, PEEP, lung compliance, CXR and FiO2 1.0) or
    • uncompensated hypercapnoea with pH of < 7.2 despite optimal conventional ventilation (CV)
    • age 18-65
  • Exclusion: Peak inspiratory pressure > 30cmH2O and/or FiO2 > 0.8 for more than 7 days; signs of intracranial bleeding; contraindications to heparinisation; planned limitation of support
  • 180 patients randomised (90 received CV & 90 received ECMO)
  • Primary cause for severe respiratory failure was pneumonia in about 2/3rd of the patients


  • Transfer to ECMO specialist centre and consideration for ECMO treatment
    • all patients randomised for consideration of ECMO were transferred by Glenfield’s own specialist retrieval team.
    • no patient was transferred on ECMO
  • If patients were haemodynamically stable, a standard protocol was used – pressure restricted   mechanical ventilation at 30cmH2O, titrated optimal PEEP, FiO2 adjusted to maintain SaO2 > 90%, diuresis, PCV 40%, prone positioning
    • Low-volume low-pressure ventilation 84 patients (93%)
  • If patient did not respond within 12 hours they received cannulation and ECMO according to published institutional protocols
  • Non-responders = FiO2 > 0.9 to maintain SpO2 > 90%, respiratory or metabolic acidosis pH < 7.2) or was haemodynamically unstable
  • ECMO:
    • was done in the veno-venous mode with percutaneous cannulation
    • Servo-controlled roller pumps (stockert, Freiburg, Germany) and poly-methyl pentene oxygenators (Medos Medizintechink, Stolbeerg, Germany) were used
    • was continued until lung recovery or apparently irreversible multiorgan failure


  • Continued conventional ventilation (pressure control mode with Siemens Servo 300 ventilators) or high-frequency oscillatory ventilation, or both
    • Low-volume low-pressure ventilation 63 patients (70%)


  • Primary outcome: Death or severe disability (confinement to bed and inability to wash and dress alone) at 6 months after randomisation
    • 63% in ECMO vs. 47% in conventional ventilation group (RR 0.69; 95% CI 0.05-0.97 p = 0.03, NNT 7)
    • RR is not for ECMO but for referral to a single ECMO capable hospital for assessment and management
    • Death < 6 months or before discharge: 57 patients (63%) in ECMO vs. 45 patients (50%) in the conventional ventilation group (p=0.07
    • In the ECMO group, 17 of the 85 patients that arrived in Glenfield continued to receive conventional ventilation rather than ECMO. 14 (82%) of these patients survived
    • Without severe disability: ECMO: 57 patients (63%) vs. CV: 45 patients (50%)
    • Severe disability: 0 (0%) vs. 1 (1)
    • Data for 3 patients was missing from the conventional ventilation group
      • Assuming that these 3 patients had all survived but been severely disabled, or had not been severely disabled, the relative risk of the primary outcome would be 0.67 (95% CI 0.48-0.94, p=0.017), and 0.72 (0.51-1.01, p=0.051), respectively. In the latter comparison, the primary endpoint narrowly misses the threshold for significance
  • Secondary outcomes (ECMO group vs. Conventional management)
    • Use of HFOV of jet ventilation: 6 patients (7%) vs. 13 patients (14%), p=0.21
    • Prone ventilation: 32 patients (4%) vs. 38 (42%), p=0.58 – see weaknesses
    • Use of steroids: 76 (84%) vs. 58 (64%) p=0.001
    • Use of MARS: 15 (17%) vs. 0 p=<0.0001
    • Nitric oxide and CVVHF: No difference
    • ICU LOS (median days): 24 vs. 13
    • Hospital LOS (median days): 35 vs. 17
    • Health status at 6 months
      • Gain of 0.03 quality-adjusted life-years (QALY’s). Lifetime model would predict the cost per QALY of ECMO to be £19,255, (95% C.I.  £7622-£59,200)
    • 5 patient in the ECMO group died – 3 patients died before they could be transferred and 2 died in transit
      • adverse events reported:
      • mechanical failure of the oxygen supply in the ambulance resulting in one death
      • vessel perforation during cannulation which was felt to contribute to the death of the patient
      • catastrophic pulmonary haemorrhage during transfer which was felt to be caused by the underlying disease

Authors’ Conclusions

  • Recommend transferring adult patients with severe but potentially reversible respiratory failure, whose Murray score > 3 or who have a pH < 7.2 on optimum conventional management, to a centre where ECMO-based management is available.


  • Early assignment to treatment groups
  • Incorporation of transport risk into trial design
  • Robust economic analysis
  • 6 month follow up testing was done in patients’ home by trained researchers who were blinded to treatment allocation. Patient and their relatives were instructed not to reveal which treatment had been used. A scarf was used to cover the neck to mask cannulation status


  • This study was more about treatment in a single specialist centre with the potential for ECMO treatment versus standard based UK treatment rather than ECMO vs. conventional ventilation
  • Only 76% in the ECMO group actually received ECMO
  • Perhaps it would have been better if patients were randomised at the point they arrived at Glenfield. The pragmatic decision not to do this is understandable
  • Lack of a management protocol for patients randomised to conventional ventilation. 93% in intervention group vs. 70% in control group were treated with lung protective ventilation, p<0.0001
  • Health evaluation data representation confusing.  On first glance, 33% of patients in the ECMO group reported no problems with mobility vs. 21% of patients in the conventional ventilation group. However the data is misleading. 40 of the 90 patients in the conventional ventilation group had follow up information available = 44%. 19 had no problems with mobility. This is represented as 19 of 90 patients (21%) rather than 19 of 40 (48%). The ECMO group would be 53% (30 of 57 patients). An ARR of 5% looks less impressive than 12% particularly when up to 24% of patients in the ECMO group never received this treatment.
  • Intention to treat analysis is useful only when there is minimal data lost to follow up. Full follow up information was only available in 58% and 36% of the patients in the ECMO group and conventional ventilation group respectively

* Error noted in manuscript: treatment of patients receiving prone position in ECMO group. 32 of 90 patients is 36% and not 4%

The Bottom Line

  • The CESAR study supports ECMO as a valid treatment option for the management of patients with severe respiratory failure. However, it does not show that ECMO is better than conventional ventilation. With incomplete follow up data in nearly half of the patients and 24% of patients in the ECMO group not actually receiving ECMO, we are left with insufficient evidence as to whether ECMO is better or worse than protocolised ARDSnet ventilation in adults patients with severe respiratory failure. The study does highlight the importance of involving specialist units, effective lung protective ventilation and ECMO as an option in refractory respiratory failure


Full text only available with subscription / abstract / doi: 10.1016/SO140-6736(09)6109-2

Editorial, Commentaries or Blogs


Summary author: @stevemathieu75
Summary date: 3rd July 2014
Peer-review editor: @DavidSlessor


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.