Remember me
Abstract
Venoarterial extracorporeal membrane oxygenation (VA ECMO) is a form of temporary mechanical circulatory support and simultaneous extracorporeal gas exchange for acute cardiorespiratory failure, including refractory cardiogenic shock (CS) and cardiac arrest (CA). Few studies have assessed predictors of successful weaning (SW) from VA ECMO. This systematic review and meta-analysis aimed to identify a multiparameter strategy associated with SW from VA ECMO. PubMed and the Cochrane Library and the International Clinical Trials Registry Platform were searched. Studies reporting adult patients with CS or CA treated with VA ECMO published from the year 2000 onwards were included. Primary outcomes were hemodynamic, laboratory, and echocardiography parameters associated with a VA ECMO SW. A total of 11 studies (n=653) were included in this review. Pooled VA ECMO SW was 45% (95%CI: 39–50%, I2 7%) and in-hospital mortality rate was 46.6% (95%CI: 33–60%; I2 36%). In the SW group, pulse pressure [MD 12.7 (95%CI: 7.3–18) I2 = 0%] and mean blood pressure [MD 20.15 (95%CI: 13.8–26.4 I2 = 0) were higher. They also had lower values of creatinine [MD –0.59 (95%CI: –0.9 to –0.2) I2 = 7%], lactate [MD –3.1 (95%CI: –5.4 to –0.7) I2 = 89%], and creatine kinase [–2779.5 (95%CI: –5387 to –171) I2 = 38%]. And higher left and right ventricular ejection fraction, MD 17.9% (95%CI: –0.2–36.2) I2 = 91%, and MD 15.9% (95%CI 11.9–20) I2 = 0%, respectively. Different hemodynamic, laboratory, and echocardiographic parameters were associated with successful device removal. This systematic review demonstrated the relationship of multiparametric assessment on VA ECMO SW.
Keywords: Cardiac arrest, cardiogenic shock, echocardiogram, extracorporeal membrane oxygenation, mortality, weaning
How to cite this article: Introduction 
Despite advances in management techniques and medical therapy, refractory cardiogenic shock (CS) remains a life-threatening condition with high-mortality rates.[1] Veno-arterial extracorporeal membrane oxygenation (VA ECMO) should be considered in patients with profound CS and in the setting of cardiac arrest (CA).[2],[3],[4],[5] VA ECMO may be used in patients with Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profiles 1 or 2 as a bridge to decision, bridge to recovery, and bridge to bridge for either long-term mechanic circulatory support or urgent heart transplantation.[6],[7]
At the crux of the decision to wean support is the demonstration of adequate myocardial recovery to provide sufficient blood and oxygen delivery to end organs to meet metabolic demands. Therefore, end-organ dysfunction, particularly liver and pulmonary failure, should be recovered before decannulation.[8]
This device can sometimes be successfully removed if the patient has partially or fully recovered from the condition that necessitated the use of ECMO. However, to date, only a few studies have reported strategies for weaning from VA ECMO.[9]
Prematurely weaning when the patient is not ready exposes the already compromised heart to stressors from high-dose inotropes, hemodynamic instability, and emergent recannulation and re-initiation of ECMO. On the other hand, delayed withdrawal can unnecessarily prolong the exposure to risks of ECMO-related complications and increase morbidity and mortality.[10]
Clinical practice guidelines provide limited information on the weaning process.[11] To date, there are limited studies investigating the association of different parameters on successful weaning (SW) from VA ECMO.
Therefore, this study aimed to conduct a systematic review and meta-analysis to identify a multiparameter strategy associated with a SW from VA ECMO in adult patients with CA or CS.
Methods This study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement[12] in conducting and reporting this systematic review. The review protocol was registered on the PROSPERO database.
Search methods for identification of studies
Systematic searches were conducted to identify studies using PubMed and the Cochrane Library, and the International Clinical Trials Registry Platform published from the year 2000 onwards was included. The terms “weaning,” “wean,” “removal,” “cardiogenic shock,” “cardiac arrest,” “extracorporeal membrane oxygenation,” “ECMO,” and “ECLS” were used as either keywords or MeSH headings in title or abstract.
The reference list of all relevant publications (retrieved full texts of the key articles and identified reviews) were hand searched.
Selection of studies for inclusion
Those studies with inclusion criteria[1] that provide data on patients who required VA ECMO for CS or CA[2], include patients aged 18 years or older[3], report at least one primary outcome of interest[4], be a prospective or retrospective observational investigation[5], include at least 10 patients[6],[5], published in English or Spanish language[7], and published since 2000 were included for this analysis.
Primary outcome was hemodynamic, laboratory, or echocardiography parameters associated with VA ECMO weaning. Secondary outcomes included all-cause mortality at the longest follow-up available. This study excluded the following types of studies: narrative or systematic reviews, case reports, or case series.
Titles and abstracts were independently screened by two reviewers (Lucrecia María Burgos and Leonardo Seoane) to identify potentially relevant articles. Discrepancies in judgment were resolved after discussion. Full-text articles were included in this analysis if they fulfilled the inclusion criteria.
Data extraction and management
Data extraction was performed independently by two reviewers (Lucrecia María Burgos and Leonardo Seoane) and all disagreements were resolved through discussion or arbitration. Data extraction included first author, year of publication, region, study design, number of patients, study period, characteristics of the study population, outcome data, and methodological quality items.
Bias assessment
Randomized controlled trials were assessed for evidence of bias with the Cochrane risk-of-bias tool.[13] A study's overall risk of bias was judged to be high if any domain was at high risk of bias, with the exception of caregiver blinding for which this study accepted standardization of mechanical ventilation, sedation, and weaning in both study groups to mitigate performance bias in these necessarily unblinded trials. The Newcastle–Ottawa Scale (NOS) was used to assess methodological strength in non-randomized studies.[14] A “star system” has been developed in which a study is judged on three broad perspectives: the selection of the study groups, the comparability of the groups, and the ascertainment of the outcome of observational studies. In this system, nine stars represent the highest level and those studies that get six stars are of high quality.
Statistical analysis
For outcomes reported by at least two studies, meta-analysis was performed. This study pooled risk ratio (RR) and 95% confidence interval (CI) for binary data classification, and continuous outcomes as mean differences (MD) and 95% CI. This study estimated the pooled prevalence according to the method described by Neyeloff et al.[15]
Heterogeneity among studies was quantified with the I2 metric, which is independent of the number of studies in a meta-analysis. I2 > 50% indicates significant heterogeneity between the studies.[16] Based on the test of heterogeneity, the pooled RR was calculated using the fixed-effects model when lacking heterogeneity, whereas random-effects modeling was adopted when heterogeneity existed. All P values were two tailed and P < 0.05 was considered statistically significant.
Median and range values were converted to means and SDs according to the actual suggested statistical method.[17]
Two-sided P values < 0.05 were considered statistically significant.
Publication bias was estimated in case more than 10 studies were found by the visual inspection of funnel plot. Egger's regression test was used to examine the asymmetry of the funnel plot.[18]
Statistical analyses were conducted using Review Manager (RevMan version 5.3, Cochrane Collaboration, London, United Kingdom) and Meta-Essentials 1.5. The screening process was performed with the reference manager Rayyan QCRI.[19]
Results Search results
Three hundred sixty-eight studies were identified through a computerized literature search among which 75 were duplicates and 262 were excluded after an initial review of titles and abstracts. The remaining 31 publications were reviewed in full-text and assessed against inclusion criteria. Finally, 11 were selected for the systematic review and meta-analysis.[20],[21],[22],[23],[24],[25],[26],[27],[28],[29],[30],[31] The search and selection process is depicted in a PRISMA flow diagram [Figure 1].
Risk of bias in included studies
Risk of bias evaluation according to NOS assessment tool for cohort studies is reported in [Table S1] in the supplementary appendix.
Study and patients characteristics
The study and patients characteristics are reported [Table 1] and [Table 2], respectively. Of the studies included for quantitative analysis, five were from European centers and the rest of them from Asia, with recruitment periods between 1995 and 2016. Most of them were unicentric except the Sugiura et al. study.[30] The majority of the included study had a retrospective design.[20],[23],[24],[26],[28],[29],[30] Only the study conducted by Sugiura et al.[30] included patients with CA, the rest were patients in CS.
The definition of SW varied in the reported studies and was described in eight studies.[21],[22],[23],[24],[25],[26],[28] Most authors considered weaning successful if the patient survives for longer than 48 h after ECMO explantation.[23],[26],[28]
Of note, Aissaoui et al. study from 2016[25] was a substudy from the original one of 2011,[21] but they added hemodynamic and echocardiographic parameters not described in the first one.
This study could not pool the microvascular outcomes of Akin et al.[26] study and biomarkers outcomes and Luyt et al.[22] study, because their results were only reported in one study. As well as multiple echocardiographic parameters evaluated in the Huang et al. study.[28]
Effects of interventions
A total of 11 studies assessing 652 patients were included. The VA ECMO SW was reported in all studies,[20],[21],[22],[23],[24],[25],[26],[27],[28],[29],[30] with a pooled prevalence of 45% (95%CI 39–50%, I2 7%). In-hospital mortality rate was 46.6% (95%CI 33–60%; I2 36%).
Hemodynamic parameters
In the SW group from VA ECMO, there was a significant higher pulse pressure, MD 12.72 (95%CI 7.39–18.05) P for effect < 0.0001, I2 = 0% [Figure 2]a, and there was also an increase in mean artery pressure (MAP) with MD 20.15 (95%CI 13.87–26.43, P for effect < 0.0001, I2 = 0%) [Figure 2]b. No statistical difference regarding the systolic blood pressure was found between the weaned and the non-weaned groups. The pooled results from the random effect models are shown in [Figure 2]c.
Figure 2: Forest plot of the comparison of different hemodynamic parameters on VA ECMO weaning (a) Pulse pressure (mmHg) (b) Mean blood pressure (mmHg) and (c) Systolic blood pressure (mmHg)Laboratory parameters
In the SW group, patients had decreased values of creatinine with a MD –0.59 (95%CI –0.94 to –0.24) P for effect = 0.001, I2 = 7% [Figure 3]a, lactate with MD –3.16 (95%CI –5.28 to –1.04) P for effect = 0.003, I2 = 86% [Figure 3]b, and creatine kinase (CK) with a MD –2779.53 (95%CI –5387 to –171) P for effect = 0.04, I2 = 38%. The pooled results from the fixed and random effect models are shown in [Figure 3].
Figure 3: Forest plot of the comparison of different laboratory parameters on VA ECMO weaning (a) Creatinine (mg/dl) (b) Lactate (mmol/l) (c) Creatine kinase (UI/l) and (d) Creatine kinase-MB. Lee et al. admission measure and Matsumoto et al. and Sugiura et al. peak creatinine. Park et al. did not specify the time. Sugiura et al. and Matsumoto et al. were lactate at 24 h, Li et al. at 12 h, Zhang et al. at 48 h, and Lee et al. did not specify the time. Matsumoto et al. and Sugiura et al. were the peak value and Zhang et al. measure at 48 h. Matsumoto et al. was the peak value, Zhang et al. measure at 48 h, and Lee et al. at admission.Echocardiographic parameters
In the VA ECMO SW group, patients had higher left ventricular ejection fraction (LVEF) expressed in %, MD 17,98 (95%CI –0,224–36,2) P for effect = 0.05, I2 = 91%, but it did not reach the significance [Figure 4]a. Interestingly, the weaned patients had a significant higher right ventricular ejection fraction (RVEF) in %, MD 15.99 (95%CI 11,9–20,08) P for effect < 0.0001, I2 = 0% [Figure 4]b.
Figure 4: Forest plot of the comparison of different echocardiographic parameters on VA ECMO weaning (a) Left ventricular ejection fraction (b) Right ventricular ejection fraction Discussion This is the first systematic review and meta-analysis assessing the association of multiparameters to predict a successful VA ECMO weaning in adult patients presenting with CA and CS to the knowledge of the authors.
By pooling data from observational studies, this study found that hemodynamic parameters, such as pulse pressure and MAP, laboratory findings (lactate, creatinine, and CK), and echo parameters (RVEF) were associated with VA ECMO SW.
In this meta-analysis, the pooled prevalence of VA ECMO SW was 45% (95%CI 40–51%), with low heterogeneity. There is a wide variation in the percentage of patients with refractory CS who are successfully weaned from ECMO varies from 31 to 76%, depending on the underlying cause of CS and the definition of SW[31],[32],[33],[34] Indeed, the definition of SW varied in the reported studies. Some consider weaning successful if the patient survives for longer than 48 h after ECMO explantation[23],[26],[28] and others if the patient does not require further mechanical support because of recurring CS during the following 30 days after the ECMO removal.[21] The pooled in-hospital mortality rate was 46.6% and similar to the rate reported by the ELSO registry.[35]
The importance of this study lies in using different hemodynamic, laboratory, and echocardiographic parameters to identify those patients who could potentially initiate a weaning trial from VA ECMO. But SW also depends on multiple other factors, such as the reversibility of the underlying cause of CS, comorbidities, and severity of organ dysfunction at ECMO initiation.[36]
The biological parameters may reflect the severity of organ dysfunction at ECMO initiation. Thus, weaned patients had decreased values of creatinine, lactate, and CK.
Interestingly, weaned patients had higher values of pulse pressure and MAP. The MAP is determined by the cardiac output (CO), systemic vascular resistance (SVR), and central venous pressure.[37] Thus, MAP and also pulsed pressure reflect CO, clinical parameters may be able to estimate cardiac function.
However, comprehensive and dynamic echocardiographic measurements are mandatory to manage patients with CS assisted by ECMO.
In this meta-analysis, only LVEF and RVEF were analyzed and the weaned patients had a significant higher RVEF. The right ventricular, in minuscula function is an important determinant of VA ECMO weaning success.[25],[28] But at bedside, it is difficult to determine the functioning of the right ventricle in maximal ECMO flow because the ECMO circuit creates negative pressure by draining venous blood from the right atrium.[38] This study found that in the VA ECMO SW group, patients had higher LVEF but it did not reach the significance. The low number of studies included and how left ventricular function were assessed may explain why we did not find significant differences in LVEF between those weaned and not weaned patients.
Weaning trials are essential to assess the behavior of ventricles during increases in preload and to determine whether the ECMO can be safely removed.[36],[37]
Multiple strategies for weaning from VA ECMO are largely unknown and most of them based on empirical evidence. Both fast and slow weaning protocols have been described.[39],[40] And strategies involving transthoracic echocardiography[21] and hemodynamic transesophageal echocardiography[41] have been reported.
Based on evidence, Aissaoui et al.,[36],[9] proposed an integrated weaning trial strategy that includes the etiology of cardiac failure, hemodynamic stability taking into account the MAP among others, tolerate a full weaning trial and echocardiographic assessment when the patient is under minimal ECMO support, evaluating LVEF, velocity time integral, lateral mitral annulus peak systolic velocity, and 3D RVEF if it is feasible.
This meta-analysis should be interpreted within the context of its limitations. This study presents data from cohorts with retrospective data collection. The differences between the patient characteristics, the institutions in terms of patient selection, volume and expertise could explain the significant heterogeneity of the outcomes between studies. Another explanation is the variation in the definition of SW. The comparability between the two groups in most of the observational studies was not well established, which might have affected the pooled results. As another limitation, this study can mention the different moments in which the laboratory parameters were measured, although the variation is small between 12 and 24 h, these patients are characterized by presenting clinical liability and that is reflected in these parameters. This study could not assess the publication bias because of the low number of studies included in the meta-analysis.
In spite of these limitations, the mentioned parameters could be taken into account to identify the patients most likely to be able to remove the device successfully. Future research with larger prospective studies is necessary to confirm these findings.
Conclusion This systematic review and meta-analysis is the first to quantitatively demonstrate the relationship of multiparametric assessment on VA ECMO SW. Hemodynamic, laboratory, and echocardiographic parameters were associated to successful device removal.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Correspondence Address:
Lucrecia María Burgos
Instituto Cardiovascular de Buenos Aires, Blanco Encalada 1543, CABA. CP1428
Argentina
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/aca.aca_79_22
Comments (0)