Does ECMO Save Lives in Severe ARDS?

Abstract & Commentary

By David J. Pierson, MD, Editor, Professor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington, Seattle, is Editor for Critical Care Alert.

Synopsis: In a selected population of younger patients early in the course of severe acute respiratory failure, referral to a specialized center for extracorporeal membrane oxygenation did not improve survival but was associated with better 6-month outcomes when severe disability in patients with complete follow-up was included.

Source: Peek GJ, et al. Efficacy and economic assessment of conventional ventilator support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): A multicentre randomised controlled trial. Lancet 2009 Sep 15; Epub ahead of print.

Peek and colleagues have recently published the long-awaited and much-discussed results of the CESAR trial (Conventional ventilatory support vs Extracorporeal membrane oxygenation [ECMO] for Severe Adult Respiratory failure), a British study that has been underway for nearly a decade. The investigators compared venovenous ECMO, in combination with "gentle ventilation" and several other protocol elements, to conventional management in patients with severe acute respiratory failure, primarily looking at mortality and severe disability in survivors but also examining a number of secondary outcomes including cost and quality of life.

The study enrolled patients over a 5-year period from 2001 through 2006, screening patients from hospitals throughout the United Kingdom. Patients meeting entry criteria were randomized to transfer to a single ECMO center (Glenfield Hospital, Leicester) vs continued management at the original hospital, or transfer to a referral center if appropriate. Patients randomized to the conventional management group were managed at the discretion of their physicians, although the use of lung-protective ventilation according to the ARDS Network protocol1 was encouraged. Entry criteria were age 18-65, "potentially reversible acute respiratory failure," and either a Murray lung injury score2(see Table) of 3 or higher, or refractory respiratory acidosis with an arterial pH < 7.20. Patients were excluded if they had a contraindication to anticoagulation (required for ECMO) or if they had been on mechanical ventilation for more than 7 days with inspiratory pressure > 30 cm H2O and/or FIO2 > 0.80.

After arrival at the ECMO center, patients randomized to that treatment group who were hemodynamically stable were managed on pressure-control ventilation with plateau pressure limited to 30 cm H2O, positive end-expiratory pressure (PEEP) titrated to "optimum saturation," maintenance of hematocrit at 40% or higher, and prone positioning if deemed necessary. If the patient did not respond to this protocol within 12 hours, and required FIO2 > 0.90 to maintain SaO2 > 90%, had respiratory or metabolic acidosis with pH < 7.20, or was hemodynamically unstable, cannulation was carried out and ECMO initiated.

A total of 766 potentially eligible patients from 148 centers were evaluated, of whom 180 from 68 centers were enrolled and randomized (90 to ECMO and 90 to conventional management). There were no significant differences between the groups at entry. The patients were young (mean age, 40 years). Respiratory failure was due to pneumonia in 60%, other acute respiratory distress syndrome (ARDS) in 28%, and trauma in 7%. Seventy percent of the patients had 1 or 2 organs failing, while 30% had 3 or more. The median duration of mechanical ventilation prior to study entry was 36 hours, and the median time from study entry to treatment in the ECMO group was 6 hours. Ninety-five percent of all patients qualified by Murray lung injury score (mean PaO2/FIO2 76 mm Hg, PEEP 14 cm H2O, static compliance 27 mL/cm H2O), with only 5% meeting entry criteria because of acidosis.

Of the 90 patients randomized to transfer to the ECMO center, 5 died prior to or during transport, and 16 improved with the center's conventional management protocol, so that only 68 patients (76%) were put on ECMO. Besides ECMO, other differences in management in the 2 study groups included more use of lung-protective ventilation in the ECMO group (only 70% of the conventional management patients ever received a tidal volume of 8 mL/kg or less vs 93% of the ECMO patients; P < 0.0001), and more frequent administration of corticosteroids in the ECMO patients (83% vs 64%; P = 0.001). About 80% of the patients in both groups received continuous venovenous hemofiltration. In the ECMO and conventional management groups, there were no differences in the use of prone positioning (36% vs 42%, respectively), high-frequency ventilation (7% vs 14%), or inhaled nitric oxide (10% vs 7%) for severe hypoxemia. Molecular albumin recirculation, a therapy for liver dysfunction, was used in 17% of the ECMO patients but in none of the patients in the conventional management group.

There was a trend toward lower mortality in the ECMO group as compared to the control group (33 vs 44 deaths), although the difference was not statistically significant (P = 0.07). When both death and severe disability at 6 months were lumped together, the difference (37% vs 53%) was significantly different (P = 0.03) in favor of the patients randomized to be transferred to the ECMO center. A higher proportion of deaths were classified as being due to respiratory failure in the control group (60%) than in the ECMO group (24%). Patients in the ECMO group had longer stays in the ICU (24 vs 13 days) and in the hospital (35 vs 17 days) than patients in the conventional management group, respectively.

With respect to cost and quality of life in survivors, referral to consideration for treatment by ECMO was associated with a gain of 0.03 quality-adjusted life-years (QALYs) at 6-month follow-up. A lifetime model predicted the cost per QALY of ECMO to be $31,112 (95% confidence interval, $12,317-$95,507) in 2005 dollars. Concluding, the authors "recommend transferring of adult patients with severe but potentially reversible respiratory failure, whose Murray score exceeds 3.0 or who have a pH of less than 7.20 on optimum conventional management, to a centre with an ECMO-based management protocol to significantly improve survival without severe disability."


Two previous multicenter randomized controlled trials have evaluated ECMO in the management of severe ARDS. In the 1970s, a study of 90 adult patients with very severe respiratory failure, carried out in 9 medical centers, randomized 42 patients to mechanical ventilation supplemented with partial venoarterial bypass and 48 to conventional mechanical ventilation.3 Only 4 patients in each group survived. Fifteen years later, a second, smaller trial was carried out, in patients similar to those in the initial study but with all study patients transferred to a single institution for uniform management.4 This time ventilator management consisted of pressure-controlled inverse ratio ventilation, adjusted by a comprehensive protocol for mechanical ventilation and other interventions, with and without extracorporeal CO2 removal. Survival in the 21 ECMO patients (33%) was substantially higher than predicted on the basis of other studies, but statistically indistinguishable from that in the 19 non-ECMO patients (42%; P = 0.8).

Since the 1970s trial, ECMO has become standard of care for severe neonatal respiratory failure, and has continued to have strong advocates for the management of adults. According to the CESAR investigators, this new, third trial was needed despite the 2 previous studies showing no survival benefit from ECMO, because "neither of these studies has relevance to modern ECMO because the case selection, ventilation strategies, extracorporeal circuit design, and disease management were completely different from modern protocols."

Although it is the largest study conducted so far, the CESAR trial is unlikely to settle the controversy over whether ECMO belongs in the armamentarium for treating severe acute respiratory failure. As pointed out in the accompanying editorial,5 the new trial's results can be spun in different ways. In their article, the authors present them as positive, emphasizing that the combined outcome of survival and the absence of severe disability was significantly better in the ECMO-referred patients than in those who remained in the community setting for management. They also point out that, according to the statistical methods employed, referral to ECMO is likely to be cost-effective, at least in the British health care system.

The findings can also be presented differently. No significant difference in survival was observed, and it was only when severe disability was added (known to have been present in only 1 of the 180 randomized patients), and the denominator in the conventional management group was adjusted (omitting the 3 patients lost to follow-up), that the primary outcome was significantly better in the ECMO-randomized group. A skeptic might also point out that less than one-quarter of all patients with acute respiratory failure who were referred by their physicians for possible inclusion were accepted into the study; that patients referred for ECMO, whether they received it or not, were managed quite differently than those who remained at their referring institutions, and that these differences might also have contributed to better outcomes.

For example, more ECMO-randomized patients received lung-protective ventilation, the only management approach so far proven to substantially improve outcomes in acute lung injury.1 Other studies of critically ill patients have also demonstrated that high-volume centers, using more standardized, more guideline-congruent management, have better outcomes; as noted by the authors, some of the CESAR-referring centers manage many fewer cases of ARDS than does the ECMO center. And while referral to the ECMO center was found to be cost-effective for the British health care system, this might not translate to the American setting, and the analysis did not include the cost involved in setting up and staffing an ECMO program for an institution not presently offering this modality.

What should the clinician make of this study? It clearly does not pertain to the majority of patients with acute respiratory failure, or even to most patients with acute lung injury or ARDS, who can be managed fine with accepted, evidence-based standard care. Nonetheless, there are some patients with acute hypoxemic respiratory failure whose hypoxemia cannot be corrected sufficiently with standard ventilator adjustments — although what "sufficiently" means in this context is highly clinician-specific and open to much debate. In such cases, after correcting any abnormalities in hemoglobin concentration and cardiac function, the options usually considered are: 1) switching to a different ventilator mode, such as from volume assist-control to pressure control or airway pressure-release ventilation; 2) using higher levels of PEEP, with or without pulmonary artery catheter guidance, colloids, and/or vasopressor support; 3) high-frequency oscillatory ventilation; 4) prone positioning; 5) inhaled nitric oxide; 6) an alternative inhaled vasodilator such as prostacyclin; and/or 7) ECMO. The last, ECMO, is not an option for the great majority of intensivists, simply because it is not available locally. Should it be?

I take the following bottom-line messages from this study: 1) We are still not implementing lung-protective ventilation as often or as effectively as we should be; 2) Patients deemed at very high mortality risk on the basis of earlier studies are more likely to survive today than in the past; and 3) Applied routinely in patients with moderately severe hypoxemic respiratory failure, ECMO does not have a dramatic effect on survival. I have never used ECMO, never having worked in a center that did it. Given that limitation, and acknowledging the possibility that ECMO might save some patients who would otherwise die of hypoxemia, it is hard to muster a lot of enthusiasm for setting up ECMO programs in centers that do not now offer it on the basis of the CESAR trial.


  1. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301-1308.
  2. Murray JF, et al. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720-723.
  3. Zapol W, et al. Extracorporeal membrane oxygenation in severe acute respiratory failure. A randomized prospective study. JAMA 1979;242:2193-2196.
  4. Morris AH, et al. Randomized clinical trial of pressure-controlled inverse ratio ventilation and extracorporeal CO2 removal for adult respiratory distress syndrome. Am J Respir Crit Care 1994;149:295-305.
  5. Zwischenberger JB, Lynch JE. Will CESAR answer the adult ECMO debate? Lancet 2009 Sep 15; Epub ahead of print.