Submitted by Dr Arthur CW Lau, ICU, Pamela Youde Nethersole Eastern Hospital, Hong Kong; on 19 Sep 2009
CRITICAL APPRAISAL: Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial. (HKSCCM link here)
Authors: Giles J Peek MD a b , Prof Miranda Mugford DPhil c, Ravindranath Tiruvoipati FRCSEd b, Prof Andrew Wilson MD d, Elizabeth Allen PhD e, Mariamma M Thalanany MSc c, Clare L Hibbert PhD f, Ann Truesdale BSc e, Felicity Clemens MSc e, Nicola Cooper PhD d, Richard K Firmin MBBS a b, Prof Diana Elbourne PhD e, for the CESAR trial collaboration. The Lancet, Early Online Publication, 16 September 2009
Using a standard method for critical appraisal from Users' Guides to the Medical Literature: II. How to Use an Article About Therapy or Prevention A. Are the Results of the Study Valid?

Gordon H. Guyatt, MD, MSc; David L. Sackett, MD, MSc; Deborah J. Cook, MD, MSc; the Evidence-Based Medicine Working Group; Gordon Guyatt, MD, MSc; Eric Bass, MD, MPH; Patrick Brill-Edwards, MD; George Browman, MD, MSc; Deborah Cook, MD, MSc; Michael Farkouh, MD; Hertzel Gerstein, MD, MSc; Brian Haynes, D, MSc, PhD; Robert Hayward, MD, MPH; Anne Holbrook, MD, PharmD, MSc; Roman Jaeschke, MD, MSc; Elizabeth Juniper, MCSP, MSc; Andreas Laupacis, MD, MSc; Hui Lee, MD, MSc; Mitchell Levine, MD, MSc; Virginia Moyer, MD, MPH; Jim Nishikawa, MD; Andrew Oxman, MD, MSc, FACPM; Ameen Patel, MD; John Philbrick, MD; W. Scott Richardson, MD; Stephane Sauve, MD, MSc; David Sackett, MD, MSc; Jack Sinclair, MD; K. S. Trout, FRCE; Peter Tugwell, MD, MSc; Sean Tunis, MD, MSc; Stephen Walter, PhD; John Williams, Jr, MD, MHS; Mark Wilson, MD, MPH. JAMA. 1993;270(21):2598-2601.

I. Are the results of the study valid?
A. Primary Guides:
Was the assignment of patients to treatments randomized?
Yes. The intensivist in the originating hospital contacted the advisory team at Glenfield Hospital to confirm eligibility and bed availability. After obtaining informed assent from relatives on behalf of the patient, the adviser then telephoned the independent central randomization service. Patients were randomly allocated by minimization in a 1:1 ratio to conventional management by intermittent positive-pressure ventilation or high frequency oscillatory ventilation, or both, or consideration for treatment by ECMO.
Minimisation factors were type of centre; age; hours of high-pressure or high FiO2 ventilation; presence of hypoxia or hypercarbia; diagnostic group; and number of organs failed. They did not stratify patients according to pulmonary and extra pulmonary acute respiratory distress syndrome because their previous experience of treating patients with ECMO indicated that this stratification did not have an effect on outcome, whereas the other minimisation criteria did affect outcome.
See Figure 1 of original paper

Were all patients who entered the trial properly accounted for and attributed at its conclusion?
Was followup complete?
Yes for the primary outcome, No for some patients assessed at 6M.
ECMO group: 52 assessed at 6M (33 died before 6M, 5 had restricted information about status at 6M from GP or hospital data)
Conventional: 32 assessed at 6M (44 died before 6M, 3 withdrew from the study and had no information about severe disability at 6M, 11 had restricted information about status at 6M from GP or hospital data)
Refer to Figure 1
Refer to Table 5 (Top)
Were patients analyzed in the groups to which they were randomized?
Yes, intention-to-treat analysis of patients after screening for eligibility (90 ECMO vs 90 conventional):
ECMO: 90 patients (68 (75%) actually received ECMO, 17 had conventional treatment, 3 died before transport, 2 died in transit)
Conventional: 90 patients
Those randomized to ECMO but not received it or died are also included in the final analysis, i.e. this is a true intention-to-treat analysis.
B. Secondary Guides:
Were patients, health workers, and study personnel "blind" to treatment?
No, because could not be blinded, but subsequent assessors are blinded:
At 6M FU, at the 6-month follow-up, testing was done in the patients’ homes by trained researchers who were masked to treatment allocation. Patients and their relatives were instructed not to reveal which treatment had been used. A scarf was used to cover the neck, thereby masking cannulation status.
The data monitoring committee reviewed interim analyses in strict confidence on seven occasions. Except for individuals who supplied confidential information, all study personnel (including the steering committee, funders, collaborators, and administrative staff ) were masked from results of the interim analysis.

Were the groups similar at the start of the trial?
-         Similar age, primary diagnosis, number of organs failed, duration of IPPV on entry, duration of high pressure ventilation and high FiO2 or both, disorder leading to study entry (Hypoxia assessed by Murray Score and hypercapnoea), APACHE II score and most other parameters
-         Refer to Table 1
Aside from the experimental intervention, were the groups treated equally?
Not exactly. Patients randomly allocated to consideration for treatment by ECMO were transferred to Glenfield Hospital and treated using a standardised protocol for disease management
There was absence of a standardized treatment protocol in the conventional management group, which was largely caused by the inability of participating units to reach a consensus on the constituents of best treatment, but they were advised to follow a low-volume low-pressure ventilation strategy.
No crossover to receive ECMO was allowed.
-         84% had steroid in ECMO, 64% in conventional (p = 0.001) – N.B. from previous study by the ARDS Network, N Engl J Med. 2006 Apr; 354(16):1671-84, steroid given after 2 weeks into ARDS actually increased mortality. So more steroid in the ECMO does not favour survival, if not increase mortality.
-         17% MARS in ECMO, 0% in conventional (p < 0.0001)
-         93% treatment by low-volume low-pressure ventilation strategy at any time; 70% in conventional (p < 0.0001); i.e. 20.4 days under above ventilatory treatment strategy vs 15 days in conventional group (p < 0.0001)
Refer to Table 2

II. What were the results?
How large was the treatment effect?
The primary outcome measure was death or severe disability at 6 months after randomisation (defined as death by 6 months or before discharge from hospital at any time to the end of data collection). Severe disability was defined as confinement to bed and inability to wash or dress alone; according to this definition, all patients were severely disabled at randomisation, but no patients were disabled before they became ill and entered the study.
Measures of primary outcome:

Absolute risk reduction
16% (NNT 6.25)
Relative risk
Relative risk reduction (RRR)
95% CI for the RRR
3% to 95% (p = 0.03)

Refer to Table 3
How precise was the estimate of the treatment effect?
Intention-to-treat analysis of primary outcome:
RR 0.69 (0.05 to 0.97), p = 0.03; RRR 0.31 (0.03 to 0.95)
For other outcomes, non significant:
1.          Death <= 6M or before discharge
2.          Severe disability
3.          Cause of death
III. Will the results help me in caring for my patients?
Can the results be applied to my patient care?
For patients with respiratory tract diseases with severe acute respiratory failure, inclusion and exclusion critieria are very general.
1.      Adult patients (18-65 years in UK)
2.      With severe, but potentially reversible respiratory failure. Severe respiratory failure will be defined as a Murray score >3.0, or uncompensated hypercapnoea with a pH <7.20
1. Duration of high pressure and/or high FiO2 ventilation >7 days
2. Intra-cranial bleeding
3. Any other contra-indication to limited heparinisation
4. Patients who are moribund and have any contra-indication to continuation of active treatment
-         Single centre with expertise in ECMO in UK: limits generalisability
-         Dedicated team and the associated expense in training
-         Capital investment in equipment and its maintenance required
-         Cost benefits is only directly applicable to the UK health-care system
-         Inward transport done by ECMO team, including transfer of patients from referral hospitals to conventional treatment centres.
-         Long-distance air travel available from the Royal Air Force (RAF) – expensive and unrealistic in normal situation if not for a clinical trial
Were all clinically important outcomes considered?
Yes, death and severe disability as the primary outcome.
Critical care days, hospital days, HRQOL by EQ-5D, QALY, SGRQ, MMSE, HADS depression score, HADS anxiety score, Sleep problem scores, Caregiver strain index, Restrictions to UL movements, Lung capacity (FEV1, FVC, FER, PEFR),
Refer to Figure 2
Table 4
Refer to Table 5
Are the likely treatment benefits worth the potential harms and costs?
NNT 6.25. Note that:
In ECMO group, only 68 (75%) actually received ECMO, 17 had conventional treatment, 3 died before transport, 2 died in transit), so the benefit is not for ECMO treatment but for referral to a single ECMO-capable hospital for ECMO assessment and management if criteria are met. The good point is that this is a truly intention-to-treat analysis. Per protocol analysis may further favour the ECMO group.
ECMO cases: twice the LOS and twice the cost.
Cost-effectiveness and Cost-utility analyses also done: cost effective in UK
Refer to Tables 6 and 7