2.1 Search strategy

Three experienced and independent investigators conducted a comprehensive, unbiased search on PubMed/MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials databases from their inception to identify studies (up to May 10th, 2024, without inception limits) pertinent to the research question.

Details regarding the search strategy are made available in the Supplementary Appendix (Supplementary Material, Search Strategy).

Duplicate publications were removed using EndNote X9 (Clarivate Analytics), and the resulting citations were uploaded to Rayyan for screening.¹⁹

Notably, both backward and forward snowballing techniques were applied to scrutinize the references of selected articles, aiming to identify additional studies for potential inclusion in the systematic review.

No additional language restrictions were imposed.

2.2 Study selection

Following removal of duplicate records from multiple databases using Zotero duplicate identification and then manually checking deleted records, every reference identified through the database search and literature review underwent independent assessment by the three investigators, at both title and abstract levels. In cases where concerns or disagreements arose, full-text articles were consulted, and any disagreements were resolved through discussion ultimately involving a third, senior investigator.

2.3 Data extraction and quality assessment

Two independent investigators conducted data extraction, aided by standardized forms for each of the included trials. All available data outlined in the research protocol, including study characteristics (such as first author, year of publication, and country), setting, sample size, details on ECMO support, and outcomes, were extracted.

2.4 Primary outcome

The primary outcome of our study was the rate of barotrauma development or progression. Development of barotrauma was defined as development of PNX, PMD, or subcutaneous emphysema while on ECMO support. Progression of barotrauma was defined according to the authors of each individual study. If no definition was reported, barotrauma progression was defined as the need for additional therapeutic interventions to treat barotrauma (e.g. chest drain), or enlargement of original barotrauma (e.g. worsening PNX, development of bilateral PNX in a patient with unilateral PNX, development of PNX in addition to PMD, etc.).

Additional outcomes included all-cause longest follow-up mortality, successful weaning from ECMO/achievement of lung transplantation, and rate of intubation for patients receiving ECMO without invasive ventilation.

2.5 Statistical analysis

We presented the results from individual studies, typically encompassing predictive performance for predefined outcomes. Provided the heterogeneity in the literature and considering that most of the retrieved studies were case reports or case series with <5 patients, quantitative data synthesis or analysis were not performed.

3.1 Characteristics of the included studies

Details on study characteristics are presented in Table 1. All but three studies were published after 2017. The remaining three articles were published in 2009,²⁶ 2015,²⁹ and 2016,³⁶ respectively. Four studies were performed in Japan,^{23, 35, 38, 40} three in the United States,^{22, 30, 37} two in China,^{27, 39} two in Germany^{29, 34} two in Italy,^{17, 33} and the others were published in Portugal,³² Saudi Arabia,¹⁶ Qatar,³⁶ Ireland,²⁴ Norway,²⁶ UK,²⁸ and Canada,²⁵ respectively. Six were retrospective observational studies^{17, 22, 25, 31, 32, 37} and the remaining 15 were case reports. Only one study compared patients managed with an “ECMO-first (invasive ventilation as rescue)” approach to patients managed with “invasive ventilation first (ECMO as rescue)” approach.³¹ Eleven studies investigated patients with Coronavirus Disease 2019 (COVID-19) pneumonia/ARDS,^{16, 17, 23, 25, 30, 31, 33-35, 38, 40} three studies investigated patients with P. jirovecii pneumonia/ARDS,^{32, 36, 39} and two studies investigated patients with autoimmune-related interstitial lung disease (i.e., dermatomyositis).^{24, 27} The remaining studies examined patients with mixed-etiology ARDS,³⁷ chest trauma-related bronchopleural fistula,²² Legionella pneumonia/ARDS,²⁶ Leptospirosis infection (with pulmonary hemorrhage),²⁸ and non-COVID-19 pneumonia/ARDS.²⁹

3.2 Extracorporeal membrane oxygenation and mechanical ventilation settings

Details on ECMO settings are presented in Table 2. All patients were treated with V-V-ECMO. Of these, 20 patients (44.4%) underwent ECMO implantation before receiving invasive ventilation. The most common cannulation configuration was femoro-femoral, while heparin was the most commonly reported anticoagulant administered. Thirteen studies reported details on ventilation/respiratory support settings before and after ECMO implantation,^{16, 23, 28-31, 33, 35-40} and in all but one³¹ cases, ventilation settings were adjusted after ECMO implantation. In particular, patients were switched from conventional to ultraprotective ventilation in four studies (seven patients),^{33, 35-37} while lower peak inspiratory pressure (PIP) and positive end-expiratory pressure (PEEP) were used in five studies (five patients).^{23, 28, 38-40} One patient was extubated while on ECMO.³⁰ In one case,²⁹ separate two-lungs protective ventilation was used.

3.3 Primary and secondary outcomes

4.1 Key findings

In this scoping review, we found that ECMO implantation with the goal of limiting barotrauma progression is feasible and is also generally associated with good outcomes, although data remains scarce and generally limited to individual case reports. We also found that colleagues generally consider ECMO implementation to allow ultraprotective and/or very low-pressure ventilation, while in almost half of the reported cases an “ECMO without invasive ventilation” approach was selected. Our data mirror our original hypothesis.

4.2 Relationship to previous studies

Our scoping review aimed to systematically assess the current practice on ECMO use for preventing barotrauma occurrence or limiting its progression. Previous reviews either investigated the effect of ECMO on survival, the feasibility and safety of ECMO without invasive ventilation, or the feasibility and safety of physiotherapy on ECMO.^{9, 13, 45, 46} Compared with these reviews, our study focused on a very specific patient population. We found a greater rate of successful weaning from ECMO and survival than the one reported for the general ARDS population on ECMO,⁹ as well as for ARDS patients with barotrauma.² However, this is likely explained by the fact that studies included in our review present data of a highly selected population treated in experienced centers. Furthermore, studies reporting unsuccessful outcomes are less likely to be published. Nevertheless, we cannot exclude that the low mortality rate observed in our study may at least in part be related to the efficacy of the investigated strategy. Notably, our rate of awake ECMO failure is in line with what has already been reported in the published literature for patients with ARDS.¹³

Previous randomized controlled trials comparing ultraprotective ventilation with standard protective ventilation strategies in patients with extracorporeal support did not report data on barotrauma,^{47, 48} or found no difference in its occurrence rate between ultraprotective and standard protective ventilation.⁴⁹ Compared with these studies, our review focused on patients with or at high risk for barotrauma, therefore focusing on a highly selected population representing 5 to 15% of patients generally enrolled in ARDS trials.^{2, 3} Furthermore, a relevant proportion of our patients were COVID-19 patients, who are considered to be at higher risk for barotrauma as compared with non-COVID-19 ARDS patients.^{2, 50, 51}

Previous systematic reviews on the management of air leaks during mechanical ventilation confirmed that the general approach of critical care clinicians includes ventilation strategies aimed at reducing airway pressures, a finding also confirmed by our study.^{4, 5} Compared with these studies, which only briefly mentioned the possibility of using ECMO, we specifically focused on the possibility of ECMO implantation to facilitate either ultraprotective invasive ventilation with very low airway pressure or avoidance of positive pressure ventilation at all.

4.3 Implication of study findings

Our study provides baseline data on the current practice and patient outcomes on use of ECMO to prevent barotrauma development and progression in patients with respiratory failure. Our data suggest that ECMO implantation in this setting is feasible and potentially associated with good outcomes. Our data showed that the general approach of clinicians is to implant ECMO in order to allow for ultraprotective and/or low-pressure ventilation. Both results are in line with our original hypothesis. Notably, in about half of the reported cases, clinicians chose to avoid invasive ventilation at all while on ECMO, suggesting that some colleagues begin to consider this as a viable alternative approach to ultraprotective ventilation. In one additional case report, the patient was extubated while on ECMO.³⁰ These strategies were generally associated with either avoidance of barotrauma progression or development in retrieved studies. Only one before/after retrospective study compared an “ECMO-first” to an “invasive ventilation first” approach for COVID-19 ARDS patients with barotrauma and found that the “ECMO first” approach might be associated with improved survival.³¹ Notably, in this study, all patients requiring escalation of support died, confirming the high mortality associated with barotrauma development and/or failure of awake ECMO in ARDS patients.^{2, 13}

One additional study used a well-known radiological sign (the Macklin-like radiological sign or Macklin effect) to identify patients with severe COVID-19 ARDS at high-risk for barotrauma and candidate these patients to the “ECMO first” approach while avoiding invasive ventilation.¹⁷ The Macklin effect has been associated with a very high risk of the development of barotrauma in COVID-19 ARDS patients,^52-56 and some authors suggested applying ECMO without invasive ventilation to prevent barotrauma in these high-risk patients,^{15, 17} either using an “ECMO first” approach or extubating patients while on ECMO.⁵⁷

The use of ECMO without invasive ventilation is a well-established practice in patients awaiting lung transplantation,^{13, 58} and became increasingly popular also for adult and pediatric patients with COVID-19.^{13, 59, 60} The principal advantages of awake ECMO include prevention of issues associated with sedation and immobilization, improved communication with relatives and staff, and avoidance of complications related to invasive ventilation such as ventilator-associated pneumonia.^{14, 45, 46, 61} The present study offers preliminary evidence to support the hypothesis that awake ECMO may also be effective in the treatment or prevention of barotrauma, supporting the hypotheses of some authors.^{15, 17, 62}

It is noteworthy that some authors have also reported the complete avoidance of ventilation while on ECMO to prevent ventilation-associated lung injury in patients with such severely depressed lung compliance that even ultraprotective ventilation becomes unfeasible.⁶³ This approach may prove an interesting alternative for the management of such extreme conditions.

Of note, most of the studies included in our review focused on COVID-19 patients. The pathophysiology of COVID-19 ARDS is different from non-COVID-19 ARDS,^64-67 and therefore our results may not apply to non-COVID-19 patients.

Collectively, our data suggested that the use of ECMO to prevent or limit barotrauma progression may indeed warrant further investigations, and we provide some baseline data to plan future studies. In particular, our study highlighted that “awake” ECMO without invasive ventilation is a relatively common approach in this setting, the other being ECMO alongside ultraprotective ventilation. Future studies should compare these strategies with current standard care to assess feasibility, safety, and efficacy on a wider scale of each approach and investigate different populations.

4.4 Study limitations

Our study has some limitations. The limited number of patients enrolled contributes to the heterogeneity of the findings; hence, our investigation needs to be considered hypothesis-generating only. However, this remains the largest review on the topic available to date.

The fact that the majority of included investigations are in the form of case reports underscores that, at present, the use of ECMO for barotrauma prevention remains anecdotal. However, management of an air leak in the context of severe respiratory failure is challenging, and very few data are available to guide its therapeutic management.

Most studies investigated patients with COVID-19 ARDS; therefore, our findings may not be generalized to reflect other populations of critically ill patients.

Only one study included a control group undergoing invasive ventilation without ECMO; therefore, there is very limited data on direct comparison with other approaches.