Ave, CESAR, morituri te salutant! (Hail, CESAR, those who are about to die salute you!)

Article details 
Evidence-Based Medicine Journal Club 
 
Edited by: Eric B Milbrandt. University of Pittsburgh Department of Critical Care Medicine


Commentary
Th e use of ECMO for the treatment of acute respiratory failure in adults has been debated since the mid-1970s. Prior to the publication of the Conventional ventilation or ECMO for Severe Adult Respiratory failure (CESAR) trial results, there were two negative randomized controlled trials [2,3] in contradistinction to a number of positive institutional experiences [4][5][6][7]. Th e relevancy of these randomized trials to modern ECMO has been questioned due to issues of case selection, ventilation strategies, extracorporeal circuit design, and disease management that were completely diff erent from modern protocols.
CESAR is the fi rst contemporary randomized controlled trial of ECMO referral for respiratory failure in adults compared to conventional supportive critical care. Importantly, the intervention in CESAR was referral to an ECMO center not treatment with ECMO. In fact, only 75% of ECMO-referred patients actually received ECMO. Despite this limited application, the two major eff ects of the intervention were impressive. First, management of adults with severe respiratory failure at a center that has ECMO capability resulted in increased 6-month survival without severe disability compared to conventional management. Second, referral to a center that has ECMO capability was cost-eff ective from the perspective of the UK National Health Service. Th e absolute risk reduction for the primary outcome was 16%, which translates into a number-needed-to-treat of 6.2 patients. Put another way, the intervention will result in one additional life saved for every 6.2 in whom it is attempted, compared to conventional management.
Strengths of the trial were an early assignment to treatment groups, intention-to-treat analysis, incorporation of transport risk into trial design, and a robust economic analysis. Th e forethought of their design allows the fi ndings to be considered pragmatically and reconciles some unanswered questions regarding ECMO use. Importantly, the study shows that ECMO referral is benefi cial -rather than the narrower question of only ECMO use. Th is distinction allows a broader take on the study fi ndings. Th e overwhelming majority of hospitals responsible for the management of adults with severe respiratory failure do not have ECMO capabilities, though they are responsible for the decision to refer patients to a center that does.
Despite the strengths of this study, there are several limitations that challenge both the generalizability and validity of the fi ndings. As the management of patients randomized to ECMO-consideration was performed at an expert high case volume center, it bears questioning whether the results would be similar in smaller or less experienced centers [8]. Furthermore, the argument can be made that the fi ndings are specifi c to the United Kingdom's health care system and not generalizable to other health care networks. In fact, the translation of currency into US dollars should really only be interpreted for scale, rather than as a refl ection of cost-eff ectiveness from a US perspective. Th ree patients in the conventional group who were known to be alive at 6 months but who asked to be withdrawn from the study were excluded from the calculation of the primary endpoint due to missing information about severe disability. As the authors point out, assuming that these three patients had all 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 signifi cance.
A more concerning aspect of the study was the lack of a management protocol for patients randomized to conven tional treatment, leaving the reader to wonder if the ECMO referral group was compared to an appropriate standard of care. Th e authors indicate there was a diff erence of 23% between treatment groups with respect to the use of a lung protective ventilation strategy at any time. Could lower adherence to this strategy in the conven tional management group account for the mortality diff erence observed or was it universally attempted but not possible in the sickest patients due to the severity of their underlying disease? We wonder.
Th e CESAR trial clearly informs our understanding of the role of ECMO referral in a modern health care network, but will likely not represent the fi nal referendum on this technology. Further study is needed to show that the results of CESAR are not merely specifi c to the single ECMO center in the study or to the United Kingdom, but that they apply to all adults with severe respiratory failure. Th e cost-eff ectiveness analysis is encouraging, but modeling in other health care environments would be needed prior to wholesale adoption. Ultimately, ECMO will likely remain a luxury commodity without highvolume use, and as such will continue to have a place in the management of severe respiratory failure at referral centers -independent of cost-eff ectiveness. Will new challenges such as infl uenza H1N1 force us to reconsider the economic burden of ECMO [9]? If so, the optimal positioning of centers with this capability will need to be determined as will protocols for initiating referrals and transfers. Time and circumstance will tell.

Recommendation
Referral of adult patients with severe respiratory failure to an ECMO-capable facility results in improved 6-month survival without disability and is cost-eff ective from the standpoint of the UK National Health Service. Replication of the CESAR fi ndings will establish whether the trial describes a limited institutional experience or off ers a preferred management strategy for patients with severe respiratory failure.