Introduction
H1N1 influenza has been the focus of substantial research given its higher case fatality among younger subjects and the potential for fulminant acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) [
1]. In light of observational and randomized trials in support of extracorporeal membrane oxygenation (ECMO), this approach has been advocated for and employed in several cases of complicated H1N1 infection [
2]. The premises for use of ECMO in these patients, including ECMO-supported inter-hospital transfer, have been mainly the young age, the relatively low prevalence of comorbidities, and the likelihood for reversible lung failure typical of these patients [
3]. Yet, there is uncertainty as to the risk-benefit balance of ECMO in patients with H1N1 infection, given variability in selection, procedural and logistic features involved in the implementation of such a technically demanding treatment [
4].
Whereas H1N1 influenza virus is currently in the post-pandemic period, regional outbreaks are still ongoing and thus defining more accurately the role of ECMO in this condition is clinically relevant [
4]. Moreover, lessons learned from the experience with Influenza A H1N1 infection may prove informative and beneficial for the management of similar instances of ALI.
Systematic reviews based on explicit and sound methods can increase statistical power, appraise quality of clinical evidence, and inform current clinical practice and future research efforts [
5]. We thus performed a systematic review focusing on the use of ECMO in patients with H1N1 influenza.
Discussion
This comprehensive systematic review, pooling data on the outlook of ECMO in 266 patients with suspected or confirmed H1N1 complicated with ARDS, have several implications. First, ECMO appears feasible in young patients with H1N1 or suspected H1N1 patients with severe ALI. Accordingly, this treatment may represent a promising alternative to standard protective ventilation. Nonetheless, short-term survival can be further improved, most likely by refinements in ancillary therapy and means of mechanical ventilation.
The H1N1 influenza pandemic has generated a plethora of research efforts, focusing on prevention, diagnosis and treatment [
19]. Indeed, such efforts have been largely justified by the substantial risk of death even in young and apparently healthy subjects developing fulminant ARDS [
1,
20]. Given these premises, the application of current state-of-the-art ECMO technologies has been proposed as a promising means to reduce the morbidity and mortality of ARDS, complicating suspected or confirmed H1N1 influenza [
2,
21]. Despite the availability of several recent studies focusing on the risk-benefit balance of ECMO in this setting, the vast majority of such publications are case reports or very small series. Even larger studies have usually a single center or single country setting, and thus have limited external validity.
Given the superior statistical power and external validity of systematic reviews, our work provides important insights on the clinical role of ECMO in the context of H1N1 infection complicated by ALI. Indeed, ECMO usage is feasible in this clinical setting, as testified by the now large number of patients who have successfully received this treatment, have been maintained on ECMO for several days and have obtained a final satisfactory clinical outcome. Thus, ECMO implementation in this type of patients can be recommended in selected centers provided training, logistics and resources are adequate. Apparently, the positive results reported by the included studies stemming from several international tertiary care centers may be due to the short period (median of two days) occurring between the start of mechanical ventilation and the start of ECMO, the elevated standards of the clinical centers that performed ECMO, as suggested by the high referral rate, and the young age of the treated patients (median 35 years). However, further studies are required to confirm or disprove the importance of these patient and procedural factors to achieve favorable clinical results with ECMO in ALI due to H1N1 infection. Notably, as many as 50% of patients were transported under ECMO, a significant undertaking which provides further evidence in support of the safety of this modality of life support.
Quoting verbatim the Extracorporeal Life Support Organization, the main indications for ECMO are acute severe heart or lung failure with high mortality risk despite optimal conventional therapy [
22]. Thus, extracorporeal life support (ECLS) is considered when a 50% mortality risk is predicted, whereas ECLS is patently indicated in most circumstances at 80% mortality risk. Severity of illness and mortality risk should be appraised as precisely as possible using measurements for the appropriate age group and organ failure. Most contraindications are relative, balancing the risks of the procedure (including the risk of using valuable resources which could be employed for others) versus the potential benefits. The relative contraindications are: 1) conditions incompatible with normal life if the patient recovers; 2) preexisting conditions which affect the quality of life (central nervous system status, end stage malignancy, risk of systemic bleeding with anticoagulation); 3) age and size of the patient; and 4) futility: patients who are too sick, have been on conventional therapy too long or have a fatal diagnosis. The application of these guidelines in the context of ARDS is jeopardized by the lack of clear principles for selecting the venous-venous versus the venous-arterial configuration [
22].
Data in the literature are scarce, but a careful revision of the patient population included in this systematic review depicts a cohort of ARDS patients suffering from severe circulatory failure and organ dysfunction or in need of supramaximal inotropic support to maintain hemodynamic stability. Accordingly, the care of these patients still requires significant improvement. There are also institutional issues that need to be addressed as the treatment of these patients is variable in terms of allocation (general versus cardiothoracic intensive care units) and specialists who are caring for them (intensive care unit specialists versus cardiac surgeons). These differences necessarily bias the trends and attitudes in clinical management.
We formally tried to explore the impact of veno-arterial versus veno-venous ECMO by means of meta-regression, given that the latter is usually associated with fewer vascular complications, but may provide inadequate hemodynamic support and less blood oxygenation in comparison to veno-arterial ECMO. On the other side, veno-arterial ECMO is much more invasive in terms of vascular access, risk of bleeding and may produce harlequin syndrome (that is, loco-regional and asymmetric discrepancies in blood flow distribution appearing as differences in skin color, sweating and temperature). However, we did not find statistical evidence for a difference between these two means to establish ECMO.
Despite the small number of patients, which has prevented its inclusion in our meta-analysis, Roch and co-workers reported a series of nine patients treated with ECMO at a single institution for H1N1 ARDS. Among them, six were treated with veno-venous and three with veno-arterial ECMO. Baseline respiratory parameters were not different between the two groups and both benefited immediately from the extracorporeal support in terms of gas exchange; whereas there was a striking difference among survivors and non-survivors in terms of the hemodynamic effects of ECMO: patients who died had no improvement of circulatory function, as defined as increased requirements or inability to wean from inotropic agents or increased lactates [
23].
The issue of cardiac failure in ARDS was also appraised in detail by Brogan
et al. in a retrospective revision of the Extracorporeal Life Support Organization (ELSO) database of adults treated for any respiratory failure [
24]. They reported a steady increase in the utilization of the veno-venous mode (from 44% in the period 1986 to 1991 to 72% in 2002 to 2006). However, more patients in the most recent years required inotropic agents/vasopressors, intra-aortic balloon pump (IABP) or had suffered preimplant cardiac arrest. Furthermore, Stöhr
et al., in a series of 30 patients, described a 40% need for primary veno-arterial cannulation in patients suffering from ARDS and an additional need for a change of configuration in 11 patients, mainly for insufficient oxygenation [
25]. Interestingly, the authors showed a configuration related mortality: patients who received a veno-veno/arterial ECMO displayed a decreased mortality when compared to patients on veno-venous or veno-arterial mode (27% vs 63% vs 75%;
P = 0.05).
The most recent case series included in this systematic review is very promising [
16]. The authors not only confirmed the low mortality rate warranted by ECMO but suggested, by a propensity score matching, that referral and transfer to an ECMO center is associated with 50% reduction in hospital mortality when compared with matched non-ECMO-referred patients. Our data, thus, confirm these and other authors' results showing the beneficial effects of ECMO for the treatment of H1N1 ARDS.
Limitations
This work has several limitations, including all those typical of systematic reviews and meta-analyses. Moreover, pooling observational studies, this review cannot overcome the limitations of primary studies, which were of relatively high quality, but still none was based on randomized allocation [
26]. Indeed, only meta-analyses of homogeneous randomized trials should be considered the final scientific proof of the efficacy and safety of any medical intervention. However, systematic reviews and meta-analyses of non-randomized studies can be meaningful and guide clinical research and practice, even if only by emphasizing the limitations of the available clinical evidence.
This meta-analysis is also not powerful enough to define the exact role of different types of ECMO. Furthermore, the exclusion of more than 70 reports because including less than 10 cases clearly calls for more collaborative research efforts. This type of collaboration is essential for present and future clinical challenges resembling the H1N1 pandemic. If only a fraction of these separate case series were to be combined, we would have achieved much greater statistical power and precision. In addition, we did not formally appraise agreement between reviewers before final consensus for study search, selection, abstraction or appraisal. Finally, no cases of H1N1 outbreaks have been reported recently, and thus the main strength of the present work is to prove that the experience with ECMO for ALI due to H1N1 infection, given its favorable effects on younger and previously healthier patients, may be helpful in the future in similar situations.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ZA, LG, BZG and PF participated in conception, design of the study, analysis of data and in drafting of the article. FG, PN and PA helped in interpretation of data and performed a critical revision of revision of the manuscript for important intellectual content. All authors approved the final version of the manuscript.