Limitations
The project has obvious limitations, including those which are typical of any systematic review and meta-analyses. By pooling observational studies, this review cannot overcome the limitations of its primary studies included which were relatively of small numbers ( apart from one (21)) and, still none were based on a randomised allocation. Indeed, the authors believe only meta-analyses of homogeneous well-powered randomised trials should be considered a solid scientific proof of the safety and efficacy of any medical/surgical intervention which is difficult to achieve on the short term period while we are still facing an unresolved pandemic and literature guidance from the available data is needed to support decision on a stretched medical resources setting in most countries The secondary outcome data was missing in numerous studies and the focus by the authors was mortality outcome and hence authors needed to be contacted by emails to fill in missing information. However, systematic reviews and meta-analyses of non-randomized studies (as in the current case) can be meaningful and guide current practice, even if only by emphasizing the limitations of the available clinical evidence (as in the current ECMO use with the COVID 19 pandemic).
Furthermore, the exclusion of 14 reports because including less than 5 cases or being case reports is a call for more collaborative research efforts. This type of collaboration is essential for the present clinical challenges of the COVID 19 crisis. To complement this collaboration, ASAIO has developed a database specific to ECMO use in severe COVID-19 to aid in this effort. Merging and synergizing data between databases such as those obtained by ASAIO, ESLO and “SpecialtyCare” may provide insight about the relevant exposure, demographics, comorbidities, and clinical and laboratory variables that may predict outcome, aid selection of patients or even suggest futility (against the evidence presented here).