Annals of the American Thoracic Society

Rationale: In patients with severe, acute respiratory failure undergoing venovenous extracorporeal membrane oxygenation (VV-ECMO), the optimal strategy for mechanical ventilation is unclear.

Objectives: Our objective was to describe ventilation practices used in centers registered with the Extracorporeal Life Support Organization (ELSO).

Methods: We conducted an international cross-sectional survey of medical directors and ECMO program coordinators from all ELSO-registered centers. The survey was distributed using a commercial website that collected information on center characteristics, the presence of a mechanical ventilator protocol, ventilator settings, and weaning practices. E-mails were sent out to medical directors or coordinators at each ELSO center and their responses were pooled for analysis.

Measurements and Main Results: We analyzed 141 (50%) individual responses from the 283 centers contacted across 28 countries. Only 27% of centers reported having an explicit mechanical ventilation protocol for ECMO patients. The majority of these centers (77%) reported “lung rest” to be the primary goal of mechanical ventilation, whereas 9% reported “lung recruitment” to be their ventilation strategy. A tidal volume of 6 ml/kg or less was targeted by 76% of respondents, and 58% targeted a positive end-expiratory pressure of 6–10 cm H2O while ventilating patients on VV-ECMO. Centers prioritized weaning VV-ECMO before mechanical ventilation.

Conclusions: Although ventilation practices in patients supported by VV-ECMO vary across ELSO centers internationally, the majority of centers used a strategy that targeted lung-protective thresholds and prioritized weaning VV-ECMO over mechanical ventilation.

Among patients with severe, acute respiratory failure undergoing venovenous extracorporeal membrane oxygenation (VV-ECMO), the optimal management strategy for mechanical ventilation is unclear. Although approaches targeting the concept of “lung rest” have been suggested (1, 2), there are no evidence-based guidelines for mechanical ventilation in patients supported by ECMO. As many of these patients likely have acute respiratory distress syndrome (ARDS), the ideal ventilation strategy would promote pulmonary recovery and weaning while minimizing ventilator-associated lung injury (VALI) (3, 4).

With this study, we aimed to describe the ventilation practices used by centers registered with the Extracorporeal Life Support Organization (ELSO), an international group of health care professionals and scientists dedicated to the development of innovations in extracorporeal life support. We further sought to determine the number of centers using a protocol to guide mechanical ventilation practices for patients on VV-ECMO.

We conducted a single, international, cross-sectional survey of ECMO medical directors and ECMO program coordinators from all ELSO-reporting centers. The sampling frame consisted of the publicly available list of ELSO-registered ECMO centers ( Three investigators (T.T., L.D.S., and E.F.) piloted the survey for clarity and relevance of questions. The survey underwent a consensus round, with several authors (E.F., T.T., L.M., L.D.S., and M.D.) contributing to its final form. The survey was distributed via a commercial website (SurveyMonkey; between May 27 and July 30, 2013.

The survey aimed to collect data pertaining to mechanical ventilation practices in patients placed on VV-ECMO for acute respiratory failure. Centers were excluded if they did not use VV-ECMO for acute respiratory failure. In an effort to capture the full breadth of mechanical ventilation practices across the various populations (neonatal, pediatric, and adult) supported on VV-ECMO, we did not restrict centers to a specific definition of acute respiratory failure. The survey collected data organized according to the following domains: (1) patient population and center characteristics, (2) the presence or absence of an explicit mechanical ventilation protocol, (3) mechanical ventilator settings, and (4) weaning practices (see the online supplement). We did not limit centers to a specific definition of “lung rest” (5,6) or “lung recruitment” (1, 7, 8), as we wanted to be as inclusive as possible with the way in which centers characterized these concepts through their individual approach to mechanical ventilation.

Initial e-mails were sent to medical directors at each ELSO center with a link to the survey. If the e-mail delivery failed, the program coordinator for that institution was contacted instead. Follow-up reminder e-mails were sent on a weekly basis for the duration of the study period. If neither the medical director nor program coordinator could be reached by e-mail, the center was excluded. Only one response per institution was accepted. Responses were excluded if they met any of the following criteria: (1) the medical director and/or program coordinator could not be contacted, (2) the same individual submitted data twice, (3) the program coordinator provided a response in addition to the response provided by the medical director from the same site (only the response from the medical director was accepted in this case), (4) two medical directors at the same site submitted separate responses (only the first response was accepted in this situation), or (5) an anonymous response was submitted. All responses were authenticated and reviewed to ensure that duplicates were excluded.

Responses were pooled for exploratory data analysis in which medians and proportions were calculated. If centers had an explicit protocol for mechanical ventilation devoted to patients supported on VV-ECMO, they were encouraged to e-mail their protocol to the investigators. The Research Ethics Board at Mount Sinai Hospital (Toronto, ON, Canada) approved the study protocol.

During the study period, the survey was sent to 288 ELSO-registered ECMO centers in 45 countries across 6 continents. Two centers were excluded because e-mails sent to medical directors failed, and they had no listed coordinator. Two more centers were excluded because e-mails sent out to both the medical director and coordinator failed. Finally, one center was excluded because it does not use VV-ECMO to manage patients with acute respiratory failure. Consequently, 283 centers were successfully contacted.

A total of 152 responses were received. Seven responses, in which the same individual provided more than one response to the survey, were excluded as duplicates. There was one center from which two separate responses were received, from both the medical director and coordinator. More than one medical director provided separate responses to the survey from two centers. For our analysis, we used the response corresponding to the first director to complete the survey. One response came from an anonymous source and was excluded, as we could not confirm it to be a duplicate. Therefore, our analysis included responses corresponding to 141 centers in 28 countries from all 6 continents. This yielded a 50% response rate (Figure 1).

Multiple centers reported the capacity to use VV-ECMO to support multiple patient populations (Table 1). The median (interquartile range) number of patients who could be supported simultaneously at one time across all respondents was 4 (2–5).

Table 1. Patient population and center characteristics

Questionn (%)
ECMO population 
 Adult38 (27)
 Pediatric11 (8)
 Neonatal11 (8)
 Adult–pediatric7 (5)
 Pediatric–neonatal53 (38)
 Adult–neonatal1 (1)
 Adult–pediatric–neonatal20 (14)
ECMO patients that can be supported at one time 
 Median (IQR)4 (2–5)

Definition of abbreviations: ECMO = extracorporeal membrane oxygenation; IQR = interquartile range.

Ventilation Practices across ECMO Centers

Of the 141 centers analyzed, 27% responded that they have an explicit mechanical ventilation protocol for patients supported on VV-ECMO. When asked about the primary goal of mechanical ventilation in VV-ECMO patients, 77% of respondents stated “lung rest” (Table 2) to be the strategy used, whereas 18% reported “lung recruitment” or a combination of both (lung rest and lung recruitment) to be their strategy.

Table 2. Mechanical ventilation practices

QuestionNeonatal/Pediatric: n = 75 [n (%)]Adult: n = 38 [n (%)]Protocol MV: n = 38 [n (%)]Overall: n = 141 [n (%)]
Primary goal of ventilatory support during ECMO    
 Lung rest55 (73)33 (87)33 (87)109 (77)
 Lung recruitment6 (8)1 (3)0 (0)12 (9)
 Varies by physician5 (7)1 (3)4 (11)8 (6)
 Other9 (12)3 (8)1 (3)12 (9)
Preferentially wean patients    
 Off ECMO first71 (95)32 (84)35 (92)127 (90)
 Off MV first4 (5)6 (16)3 (8)14 (10)
During ECMO weaning, preferentially reduce    
 Sweep gas flow rate29 (39)18 (47)17 (45)61 (43)
 FdO215 (20)6 (16)10 (26)26 (18)
 Blood flow rate19 (25)7 (18)6 (16)30 (21)
 Other12 (16)7 (18)5 (13)24 (17)
Primary initial ventilatory mode during ECMO    
 Controlled46 (61)25 (66)26 (68)87 (62)
 Support14 (19)6 (16)6 (16)29 (21)
 Airway pressure release ventilation1 (1)4 (11)1 (3)5 (4)
 Noninvasive0 (0)0 (0)0 (0)0 (0)
 Not intubated/supplemental oxygen0 (0)0 (0)0 (0)1 (1)
 Varies by physician10 (13)1(3)3 (8)12 (9)
 None4 (5)2 (5)2 (5)7 (5)
Tidal volume (per predicted body weight)    
 ≤4 ml/kg21 (28)13 (34)17 (45)44 (31)
 4–6 ml/kg35 (47)18 (47)11 (29)64 (45)
 7–9 ml/kg2 (3)1 (3)0 (0)3 (2)
 ≥10 ml/kg0 (0)0 (0)0 (0)0 (0)
 Not strictly controlled17 (23)6 (16)10 (26)30 (21)
PEEP (during VV-ECMO)    
 ≤5 cm H2O8 (11)0 (0)3 (8)10 (7)
 6–10 cm H2O49 (65)18 (47)19 (50)82 (58)
 11–15 cm H2O8 (11)12 (32)11 (29)30 (21)
 ≥16 cm H2O0 (0)2 (5)0 (0)2 (1)
 Other10 (13)6 (16)5 (13)17 (12)

Definition of abbreviations: ECMO = extracorporeal membrane oxygenation; FdO2 = fraction of delivered oxygen via ECMO; MV = mechanical ventilation; PEEP = positive end-expiratory pressure; VV = venovenous.

The majority (62%) of centers reported using controlled ventilation modes for patients placed on VV-ECMO, while fewer (27%) centers reported using spontaneous breathing modes. Interestingly, 76% of respondents reported ventilating patients using tidal volumes of 6 ml/kg or less (Figure 2). Of note, 21% did not target a particular tidal volume with their ventilation strategy. Of the 30 centers that did not target a specific tidal volume while ventilating patients on VV-ECMO, only 6 endorsed using spontaneous breathing modes (i.e., pressure support); the majority (17 centers) stated that they used controlled modes of ventilation for their patients. Eighty percent of centers adhere to a positive end-expiratory pressure (PEEP) of greater than 5 cm H2O when ventilating patients on VV-ECMO, with the majority (58%) targeting a PEEP between 6 and 10 cm H2O. Data pertaining to peak pressure, plateau pressure, and permissive hypercapnia were not collected with this survey.


When approaching weaning, most centers stated that they would prioritize weaning from the VV-ECMO circuit over the ventilator (Table 2). The majority of responses pertaining to the way in which centers weaned from VV-ECMO was through reduction of the sweep gas flow rates (43%) and the blood flow rates, respectively (21%). ECMO centers using a mechanical ventilation protocol to manage patients on VV-ECMO did not differ regarding target tidal volumes, PEEP, or weaning strategy when compared with centers that did not use a protocol.

In this cross-sectional survey of 141 ELSO-reporting centers across 28 countries, we found that ventilator management policies vary widely between centers, which may reflect uncertainty surrounding the optimal ventilator management strategy in patients requiring VV-ECMO. Although this is supported by the observation that only 27% of centers responding to our survey had an explicit mechanical ventilation protocol, ECMO centers may have employed guidelines for specific ventilation targets that allow for some individual variability, rather than committing them to strict use of a protocol. Despite variability in mechanical ventilation practices across ECMO centers, many interesting themes emerged with respect to PEEP, tidal volumes, and weaning strategy, which may reflect the goals that clinicians attempt to achieve while ventilating patients on VV-ECMO.

Whereas the majority of centers responded that “lung rest” was the primary goal of mechanical ventilation, 22% of centers targeted a PEEP greater than 10 cm H2O. This variability illustrates the lack of data indicating specific PEEP targets for patients on VV-ECMO. In the setting of VV-ECMO, the application of PEEP to improve oxygenation and lung compliance is not required (3, 9), but promoting lung recruitment through higher levels of PEEP might accelerate lung healing (5, 6, 10) through preventing pulmonary vascular leak and inflammatory macrophage activation (79). Conversely, ventilation strategies without the use of PEEP have been employed successfully in VV-ECMO patients, and atelectatic lung has shown the capacity to recover in other clinical conditions such as pneumonia (9, 11, 12). Indeed, it is possible that prepulmonary oxygen delivery in VV-ECMO supports lung healing in the absence of PEEP. The optimal ventilatory strategy in patients with acute respiratory failure on VV-ECMO remains unclear. More information may be available, as the impact of tidal ventilation and the optimal level of PEEP on biotrauma, physiological parameters, and cardiac function will be evaluated in the Strategies for Optimal Lung Ventilation in VV-ECMO for ARDS (SOLVE-ARDS) Study (NCT01990456;

The majority of respondents reported using ventilator settings that met lung-protective thresholds regarding tidal volume and PEEP. This may reflect recognition of the mortality benefit that results from using a lung-protective ventilation strategy in the adult ARDS population (10, 13). Lung-protective ventilation has been suggested to have a mortality benefit in VV-ECMO as well (9, 14). It is important to note, however, that ventilation using conventional tidal volumes exceeding lung-protective thresholds (>6 ml/kg) is more injurious in adult patients than in pediatric or neonatal patients who are ventilated at comparable tidal volumes (11, 12, 15, 16). This is significant considering the fact that most centers (73%) responding to our survey used VV-ECMO to support their pediatric and neonatal populations. These groups did not differ regarding ventilation mode or tidal volume, but a larger proportion of adult patients was managed at a PEEP of 11–15 cm H2O. It is possible that higher levels of PEEP were used in adults for more profound levels of hypoxemia or in an effort to hasten weaning.

Interestingly, 31% of centers reported using “ultraprotective” tidal volumes (<4 ml/kg) while ventilating patients on VV-ECMO. Through augmenting gas exchange, VV-ECMO permits the achievement of lower tidal volumes and reduced airway pressures. This may confer the additional benefit of minimizing lung inflammation and VALI. Centers using VV-ECMO may consider extracorporeal carbon dioxide removal (ECCO2R) as a less invasive alternative to facilitate mechanical ventilation with ultraprotective tidal volumes. Although ECCO2R in combination with ultraprotective mechanical ventilation has not shown survival benefit in ARDS (13), investigators have demonstrated that a reduction in tidal volume below 6 ml/kg reduced lung inflammation (14). Therefore, VV-ECMO and ECCO2R may be used to manage consequent hypercapnic acidosis as clinicians target a lower tidal volume strategy (<6 ml/kg) to minimize VALI. The divergence in tidal volume targets observed in our study demonstrates the uncertainty surrounding the optimal tidal volume in VV-ECMO and how VV-ECMO might be used to facilitate lung protection goals.

Controlled modes of mechanical ventilation were reported to be the most commonly used ventilator mode among ECMO centers. The choice of using controlled ventilation in this setting may represent the level of acuity in patients selected for VV-ECMO, the concomitant use of heavy sedation or neuromuscular blockade, and poor respiratory system compliance. The use of controlled mechanical ventilation may pose an obstacle to weaning because of patient–ventilator asynchrony and the subsequent impairment of diaphragmatic function (15, 16). Indeed, the use of VV-ECMO to facilitate gas exchange in these patients may actually allow a reduction in sedation and cessation of neuromuscular blockade, allowing patients to be more awake, breathing spontaneously, and perhaps even mobile (17, 18). Furthermore, 90% of ECMO centers favored weaning patients from the VV-ECMO circuit before weaning from the ventilator. Accordingly, the work of breathing is transferred to the patient as his/her pulmonary function improves (2, 3). The benefits of this strategy are unclear, as even lung-protective ventilation in patients with severe ARDS may result in VALI (4, 19). Prioritizing weaning from the ventilator may carry the benefit of reducing the risk of VALI, but increase the risk of complications associated with VV-ECMO, such as bleeding (2, 20).

There are several limitations to our study. First, the cross-sectional design of this study fails to capture the way in which ventilator management evolves over time in the management of patients on VV-ECMO. Next, our study had a 50% response rate in the number of ELSO centers contacted. Accordingly, we were unable to comment on the full breadth of ventilator practices among all active centers, and our conclusions were prone to nonresponder bias. To provide a broad description of mechanical ventilation practices across ECMO centers, we framed our survey within the context of “acute respiratory failure.” Recognizing that ARDS likely accounts for the majority of these cases, survey questions were tailored to reflect concepts relevant to ARDS and the prevention of VALI. Although we were able to identify that most centers ventilated patients using tidal volumes and PEEP settings that met lung-protective thresholds, our findings would have been more informative if they had included data pertaining to plateau pressures, fraction of inspired oxygen, and permissive hypercapnia. Finally, our survey did not capture outcome data. Although we would seek to correlate ventilator practices at these centers with clinical outcomes, we recognize that a variety of factors interact in producing these outcomes. In anticipation that a great deal of variability would exist in the way centers approached mechanical ventilation, we collected data with the intent of exploring themes and generating hypotheses for further study.

In conclusion, the majority of centers supporting patients on VV-ECMO for acute respiratory failure used a controlled mode of mechanical ventilation to target lung-protective tidal volumes and moderate levels of PEEP. These centers used VV-ECMO with the intent of providing “lung rest” and preferentially weaned VV-ECMO before weaning mechanical ventilation. Future investigations should focus on establishing optimal PEEP and tidal volume targets to promote pulmonary recovery and weaning while minimizing VALI.

1 . Meade MO, Cook DJ, Guyatt GH, Slutsky AS, Arabi YM, Cooper DJ, Davies AR, Hand LE, Zhou Q, Thabane L, et al.; Lung Open Ventilation Study Investigators. Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:637645.
2 . Peek GJ, Mugford M, Tiruvoipati R, Wilson A, Allen E, Thalanany MM, Hibbert CL, Truesdale A, Clemens F, Cooper N, et al.; CESAR Trial Collaboration. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial. Lancet 2009;374:13511363. [Published erratum appears in Lancet 374:1330.]
3 . Del Sorbo L, Cypel M, Fan E. Extracorporeal life support for adults with severe acute respiratory failure. Lancet Respir Med 2014;2:154164.
4 . Slutsky, AS, Ranieri, VM. Ventilator-induced lung injury. N Engl J Med 2013;369:21262136. [Published erratum appears in N Engl J Med 370:1668–1669.]
5 . Lachmann B. Open up the lung and keep the lung open. Intensive Care Med 1992;18:319321.
6 . Vieillard-Baron A, Jardin F. Right level of positive end-expiratory pressure in acute respiratory distress syndrome. Am J Respir Crit Care Med 2003;167:1576; author reply 15761577.
7 . Carpenter TC, Stenmark KR. Hypoxia decreases lung neprilysin expression and increases pulmonary vascular leak. Am J Physiol Lung Cell Mol Physiol 2001;281:L941L948.
8 . Madjdpour C, Jewell UR, Kneller S, Ziegler U, Schwendener R, Booy C, Kläusli L, Pasch T, Schimmer RC, Beck-Schimmer B. Decreased alveolar oxygen induces lung inflammation. Am J Physiol Lung Cell Mol Physiol 2003;284:L360L367.
9 . Pham T, Combes A, Rozé H, Chevret S, Mercat A, Roch A, Mourvillier B, Ara-Somohano C, Bastien O, Zogheib E, et al. Extracorporeal membrane oxygenation for pandemic influenza A (H1N1)–induced acute respiratory distress syndrome. Am J Respir Crit Care Med 2013;187:276285.
10 . 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;342:13011308.
11 . Kornecki A, Tsuchida S, Ondiveeran HK, Engelberts D, Frndova H, Tanswell AK, Post M, McKerlie C, Belik J, Fox-Robichaud A, et al. Lung development and susceptibility to ventilator-induced lung injury. Am J Respir Crit Care Med 2005;171:743752.
12 . van Kaam A. Lung-protective ventilation in neonatology. Neonatology 2011;99:338341.
13 . Bein T, Weber-Carstens S, Goldmann A, Müller T, Staudinger T, Brederlau J, Muellenbach R, Dembinski R, Graf BM, Wewalka M, et al. Lower tidal volume strategy (≈3 ml/kg) combined with extracorporeal CO2 removal versus “conventional” protective ventilation (6 ml/kg) in severe ARDS. Intensive Care Med 2013;39:847856.
14 . Terragni PP, Del Sorbo L, Mascia L, Urbino R, Martin EL, Birocco A, Faggiano C, Quintel M, Gattinoni L, Ranieri VM. Tidal volume lower than 6 ml/kg enhances lung protection: role of extracorporeal carbon dioxide removal. Anesthesiology 2009;111:826835.
15 . de Wit M, Miller KB, Green DA, Ostman HE, Gennings C, Epstein SK. Ineffective triggering predicts increased duration of mechanical ventilation. Crit Care Med 2009;37:27402745.
16 . Mauri T, Bellani G, Foti G, Grasselli G, Pesenti A. Successful use of neurally adjusted ventilatory assist in a patient with extremely low respiratory system compliance undergoing ECMO. Intensive Care Med 2010;37:166167.
17 . Lee CM, Fan E. ICU-acquired weakness: what is preventing its rehabilitation in critically ill patients? BMC Med 2012;10:115.
18 . Abrams D, Javidfar J, Farrand E, Mongero LB, Agerstrand CL, Ryan P, Zemmel D, Galuskin K, Morrone TM, Boerem P, et al. Early mobilization of patients receiving extracorporeal membrane oxygenation: a retrospective cohort study. Crit Care 2014;18:19.
19 . Terragni PP, Rosboch G, Tealdi A, Corno E, Menaldo E, Davini O, Gandini G, Herrmann P, Mascia L, Quintel M, et al. Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med 2007;175:160166.
20 . Brodie D, Bacchetta M. Extracorporeal membrane oxygenation for ARDS in adults. N Engl J Med 2011;365:19051914.

*These authors contributed equally to this work.

Correspondence and requests for reprints should be addressed to Eddy Fan, M.D., Ph.D., Toronto General Hospital, 585 University Avenue, PMB 11-123, Toronto, ON, M5G 2N2 Canada. E-mail:

This article has an online supplement, which is accessible from this issue’s table of contents online at

Author disclosures are available with the text of this article at


No related items
Comments Post a Comment

New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
Annals of the American Thoracic Society

Click to see any corrections or updates and to confirm this is the authentic version of record