Extracorporeal membrane oxygenation technology for adults: an evidence mapping based on systematic reviews

We searched four electronic databases including Pubmed, EMBASE, Cochrane Library, Web of Science from their inception through January 2023. Search strategies were constructed using combinations of words describing the intervention of interest (“Extracorporeal Membrane Oxygenation” or “Venovenous ECMOs” or “Venovenous Extracorporeal Membrane Oxygenation” or “Venovenous ECMO” or “Venoarterial ECMO” or “Venoarterial ECMOs” or “Venoarterial Extracorporeal Membrane Oxygenation” or “Extracorporeal Membrane Oxygenations” or “ECMO” or “ECMO Treatment” or “ECMO Treatments” or “Extracorporeal Life Support” or “ECMO Extracorporeal Membrane Oxygenation” or “ECLS” or “ECLS Treatment” or “ECLS Treatments” or “Extracorporeal Life Supports” or “Extracorporeal Gas Exchange”), and the studied type (“meta-analysis” or “systematic review”). Depending on characteristics of the database, medical subject headings (MeSH) and free vocabulary words were combined. No language restrictions were imposed.

EndNote X9 (Clarivate Analytics, Spring Garden, Pennsylvania, USA) software managed search results and deduplication. Two reviewers independently screened all potentially relevant studies based on recorded titles and abstracts and then cross-checked them. In the event of disagreement, the study was provisionally included to obtain additional information. After an initial selection decision was made, the full texts of the chosen studies were downloaded for further review. Two independent reviewers then conducted a new selection process based on a full-text analysis of eligible SRs. An all-researcher meeting was convened to reach a definitive decision on disparate related studies. We extracted information regarding the population, intervention, comparison, and outcomes (PICO) process, certainty of evidence statements, and the number of studies included in each SR.

The included SRs were classified according to the topic of the investigation. If multiple SRs on similar clinical topics were identified, we chose the most relevant and best-performed SR for each topic based on the results of the GRADE assessment. Besides, SRs for each topic were depicted only once on the bubble plot. The results of the evidence mapping were presented using characteristic tables of the included SRs and a bubble plot display. Each bubble in the graph represents the evidence evaluated by the SRs investigating the efficacy of ECMO for clinical topics. The visual representation or evidence mapping displays information on four dimensions using a bubble plot: X-axis, Y-axis, bubble size, and color. This enabled us to provide the following forms of information regarding each included SR.

Four electronic databases searches yielded a total of 1933 records from inception to January 2023. After removing duplicates, 975 records were screened based on their titles and abstracts. The initial screening identified 338 potentially relevant studies evaluated against eligibility criteria. A total of 321 articles were deemed ineligible after thoroughly examining of their full text due to non-compliance with the established eligibility criteria. Ultimately, a total of 17 SRs with or without a meta-analysis on ECMO were included for systematic scoping review and evidence synthesis [22‐38]. Figure 1 displays all comprehensive review processes, exclusion numbers, and reasons for full-text exclusions.

In terms of methodological quality, the overall quality was rated as “Moderate” for six [23, 27, 28, 35‐37], seven SRs scored “Low” [25, 26, 29, 31‐34], four SRs scored “Critically Low” [22, 24, 30, 38], and “High” for none SRs according to AMSTAR-2 criteria. The most frequent flaws were as follows: lack of a reasonable explanation for the selection of study design type for inclusion, the absence of a report on sources of funding for included studies, a lack of a statement regarding potential sources of conflict of interest, and the absence of a protocol. Figure 3 depicts the methodological quality of the 17 SRs included in the analysis.

The evidence quality of included one SR was considered high, 5 SRs were moderate and low quality respectively, while 6 SRs were assessed as having very low quality. The results of this evaluation can be found in Table 3.

For diseases that overlap, the overall evidence quality was considered. Individual SRs reflected the conclusions, which were confirmed by an internal review. The evidence mapping on ECMO for adults is presented in Fig. 4.

There are several SRs for ECMO based on available evidence, with only 17 SRs centered on various diseases meeting the criteria. Most SRs contain a small number of primary studies, indicating limited evidence for this issue. RCT is the most reliable evidence to evaluate the efficacy and safety of interventions [40]. However, the majority of the main studies used to support the efficacy and safety of ECMO were not RCT, according to the SRs included in this evidence mapping, which might be a phenomenon with significant ethical implications.

SR, an essential component of evidence-based medicine, has become the highest level of evidence as it synthesizes all available evidence on a given topic [41]. However, if the quality and criteria of SR differ widely, the findings of reviews may be exaggerated. Although methodological quality assessment is not a core task of evidence mapping, it has been recommended that any review should include this process to assess the consistency of their conclusions [42]. The AMSTAR-2 tool has been used extensively as an effective method to evaluate the methodological quality of SR [43]. Increasing numbers of SR have been conducted on ECMO, but we used the AMSTAR-2 tool to assess methodological quality to ascertain the validity of their conclusions. Unfortunately, we discovered no “High” methodological quality of SRs, but rather six “Moderate” SRs, seven “Low” SRs, and four “Critically Low” SRs. The most frequent shortcomings were as follows: lack of a reasonable explanation for the selection of study design type for inclusion, the absence of a report on sources of funding for included studies, a lack of a statement regarding potential sources of conflict of interest, and the absence of a protocol, all of which would require the attention of future researchers.

Our evidence mapping emphasizes the areas where SRs have reported “beneficial”, “probably beneficial”, “inconclusive”, “no effect”, or “harmful” effects while simultaneously displaying the research concentration and volume. ECMO was beneficial for some clinical topics, such as severe ARDS and ALI due to H1N1 influenza infection. It is probably beneficial for certain clinical topics, such as ARDS, ALF or ACLF, and cardiac arrest. Conclusions regarding dependent ARDS, ARF, ARF due to the H1N1 influenza pandemic, and cardiac arrest of cardiac origin were inconclusive. Significantly, we found no evidence of a harmful association between ECMO and various clinical topics, which may be due to the fact that few RCTs with negative conclusions have been published [44]. The fact that the efficacy and safety outcome of ECMO in treating severe ARDS is not only concluded as having a “beneficial” effect but also supported by “high” quality evidence, indicating that ECMO is a potentially promising support technique for severe ARDS and is also consistent with the ARDS management guidelines, is particularly noteworthy. According to the formal management guidelines of ARDS [45], severe cases of ARDS with PaO₂/FiO₂ < 80 mmHg and/or dangerous MV, despite optimization of ARDS management including high PEEP, neuromuscular blocking agents and prone positioning, should probably be considered for venovenous ECMO. The decision to use ECMO should be evaluated early by contacting an expert center with the strong agreement. Being distinct from severe ARDS, ECMO demonstrated a probably beneficial effect in another SR that did not differentiate the severity of ARDS, suggesting that ECMO may have a more effective therapeutic effect in severe ARDS, which requires further research to confirm. The current estimated mortality rate for ARDS is approximately 30–40%, with severe forms of ARDS having higher mortality rates than mild or moderate forms of ARDS [46]. It is highly promising to assert that using ECMO technology will significantly contribute to reducing mortality rates, especially associated with severe ARDS. The continuous advancements in ECMO technology are expected to achieve such a favorable effect.

An additionally interesting finding was that ECMO outcome indicators tend to focus on survival, mortality, and a favorable neurological outcome. However, none of the studies evaluated the quality of life as an outcome or conducted an economic evaluation. In the past few decades, quality of life has emerged as a significant concept and objective for research and practice in the fields of health and medicine, which can inform clinicians and policymakers about how to prioritize and allocate healthcare resources when assessing the benefits of different treatment options [47]. Similarly, economic evaluation contributes to the most efficient allocation of societal resources [48]. ECMO is an essential technology for critically ill patients with cardiopulmonary failure; consequently, evaluations of the quality of life and economic impact are particularly crucial.

Cardiopulmonary failure is a condition that the majority of critically ill patients may experience, indicating that ECMO may be increasingly required for a prolonged time in the future. According to searched studies, high-quality design research was absent on the adverse effects of long-lasting ECMO use. Four of the 17 included SRs reported potential adverse effects of prolonged ECMO use, including bleeding, barotrauma, sepsis, and circuit-related complications. Specifically, two SRs [24, 31] suggested that ECMO may increase the risk of bleeding in ARDS or ARF patients, and ALI due to H1N1 infection patients with severe comorbidities or multiorgan failure remained at high risk of in-hospital death if prolonged support (over one week) was required in most cases [23]. Having said that, only a few studies on the adverse effects of ECMO were included in this mapping, and most primary studies are not RCTs, so the findings were not entirely accurate. Thus, related RCTs ought to concentrate on developing high-quality studies to evaluate the adverse effects of ECMO.

This evidence mapping has described the research focus, reported in the existing SRs, and identified the gaps in evidence to identify clinical topics that should be prioritized for future research [49]. However, it can not answer more specific questions, such as the optimal parameter selection, application timing and duration for ECMO in a particular specific health topic. To advance our evidence-based understanding of ECMO, we should acquire additional data on the efficacy and safety of ECMO across and within each clinical condition and patient population using meta-analyses of primary studies. In addition, the large number of clinical topics classified as lacking conclusive evidence calls for additional primary research. In some of the clinical topics included in the category of the inconclusive evidence, additional studies have been published, necessitating an update of the present SRs.

This evidence mapping, unlike previous ones, provides a comprehensive summary of the current evidence for all categories of clinical topics associated with ECMO without restrictions. Moreover, our study conducted a systematic and exhaustive search of four databases and utilized a relatively dependable study design, SR. Then, we used the AMSTAR-2 tool to assess the methodological quality of inclusion in SRs, and the GRADE system to assess the quality of evidence for inclusion in SR outcomes, visually presenting the results of the existing evidence in the form of a bubble plot based on multiple significant dimensions. In addition, we determine the rating of conclusions based on the description of research results and conclusions, which may avoid the uncertainty caused by policy recommendations determined solely based on research results or conclusions in a certain sense [50], which are not only instructive for future research and important for preventing the waste of academic resources, but also essential for policymakers. Nonetheless, it is important to note that this research does have a few limitations. First, we excluded other study designs (such as RCT, case report, cohort study, or cross-sectional study, etc.), even though the fact that SRs generally could provide the highest quality evidence for decision-making. Second, only four frequent literature databases were searched, but literature from other sources, such as clinical trial registration websites, was not focused on, so literature omission was inevitable. Especially since I was unable to find any information related to ECMO and pregnancy. Third, most of the included SRs were based on observational or cohort studies with poor methodological quality, which might have led to bias and affected the intrinsic authenticity of SR to some extent.