Skip to main content
Download PDF

Extracorporeal cardiopulmonary resuscitation versus standard treatment for refractory out-of-hospital cardiac arrest: a Bayesian meta-analysis

Critical Care volume 28, Article number: 217 (2024) Cite this article

Abstract

Background

The outcomes of several randomized trials on extracorporeal cardiopulmonary resuscitation (ECPR) in patients with refractory out-of-hospital cardiac arrest were examined using frequentist methods, resulting in a dichotomous interpretation of results based on p-values rather than in the probability of clinically relevant treatment effects. To determine such a probability of a clinically relevant ECPR-based treatment effect on neurological outcomes, the authors of these trials performed a Bayesian meta-analysis of the totality of randomized ECPR evidence.

Methods

A systematic search was applied to three electronic databases. Randomized trials that compared ECPR-based treatment with conventional CPR for refractory out-of-hospital cardiac arrest were included. The study was preregistered in INPLASY (INPLASY2023120060). The primary Bayesian hierarchical meta-analysis estimated the difference in 6-month neurologically favorable survival in patients with all rhythms, and a secondary analysis assessed this difference in patients with shockable rhythms (Bayesian hierarchical random-effects model). Primary Bayesian analyses were performed under vague priors. Outcomes were formulated as estimated median relative risks, mean absolute risk differences, and numbers needed to treat with corresponding 95% credible intervals (CrIs). The posterior probabilities of various clinically relevant absolute risk difference thresholds were estimated.

Results

Three randomized trials were included in the analysis (ECPR, n = 209 patients; conventional CPR, n = 211 patients). The estimated median relative risk of ECPR for 6-month neurologically favorable survival was 1.47 (95%CrI 0.73–3.32) with a mean absolute risk difference of 8.7% (− 5.0; 42.7%) in patients with all rhythms, and the median relative risk was 1.54 (95%CrI 0.79–3.71) with a mean absolute risk difference of 10.8% (95%CrI − 4.2; 73.9%) in patients with shockable rhythms. The posterior probabilities of an absolute risk difference > 0% and > 5% were 91.0% and 71.1% in patients with all rhythms and 92.4% and 75.8% in patients with shockable rhythms, respectively.

Conclusion

The current Bayesian meta-analysis found a 71.1% and 75.8% posterior probability of a clinically relevant ECPR-based treatment effect on 6-month neurologically favorable survival in patients with all rhythms and shockable rhythms. These results must be interpreted within the context of the reported credible intervals and varying designs of the randomized trials.

Registration

INPLASY (INPLASY2023120060, December 14th, 2023, https://doi.org/10.37766/inplasy2023.12.0060).

Background

Following encouraging results derived from observational studies [1, 2], several randomized controlled trials (RCTs) comparing extracorporeal cardiopulmonary resuscitation (ECPR) to conventional cardiopulmonary resuscitation (CCPR) have been conducted [3,4,5]. A patient-level pooled analysis of two of these trials [6]—and several other meta-analyses [7,8,9,10,11,12]—were published following the publication of these RCTs, focusing on the potential presence of a statistically significant difference in (neurologically favorable) survival. Still, such analyses are heavily dependent on the sample size and event rate, and may be underpowered to detect a clinically relevant treatment effect. To augment the pooled sample size, some of these meta-analyses have also included non-randomized data [11,12,13,14]. However, the addition of such studies increases the risk of various forms of bias that are associated with a non-randomized trial design (including treatment selection bias and allocation bias). Moreover, by expanding the sample size, a potentially established statistically significant difference (based on the p-value) may no longer be clinically relevant.

All of the previously performed pooled analyses have applied the frequentist statistical framework, which relies on well-known concepts such as null hypothesis significance testing and p-values [15]. In this paradigm, the treatment of interest is considered a ‘fixed parameter’ or ‘fixed points estimate’, and the data is considered ‘unknown’. In that light, the p-value denotes the probability of observing the trial data—or more extreme—in future experiments, conditional on the assumption that the null hypothesis is true (i.e., analogous to the probability of a test result, given a disease [16, 17]). Still, the frequentist approach presents several cognitive difficulties, such as, the (in)correct interpretation of the p-value [15, 17,18,19]. Indeed, a p-value below the threshold of statistical significance (i.e., < 0.05) does not imply that a treatment effect is definitely clinically relevant, while a p-value above the threshold of statistical significance (i.e., > 0.05) does not mean a treatment is necessarily clinically ineffective [16]. Some of these difficulties can be addressed by the application of Bayesian statistical inference. In Bayesian statistical inference, the treatment effect (or hypothesis) is considered 'unknown' and has a probability distribution reflecting its uncertainty. Instead, the data is observed, and therefore 'known'. Using this posterior probability distribution, the Bayesian statistical paradigm allows the estimation of the probability of a treatment effect (or hypothesis) in the light of the observed data. Inherently, this inverse probability thinking resembles clinical reasoning (i.e., analogous to the probability of the disease, given the test result). Therefore, Bayesian statistical inference can be used to estimate the probability of clinically relevant treatment effects, particularly in trials evaluating a resource-intensive treatment such as ECPR, which may be subjected to small sample sizes and low statistical power.

Thus, the researchers behind three recent RCTs conducted a joint systematic review and Bayesian meta-analysis to determine whether there is a clinically relevant effect of ECPR on neurologically favorable survival in patients with refractory OHCA.

Methods

Protocol registration

This systematic review and Bayesian meta-analysis was prospectively registered on the international platform of registered systematic reviews and meta-analyses (INPLASY) [20, 21] (registration number: INPLASY2023120060; date of registration: December 14th 2023; https://doi.org/10.37766/inplasy2023.12.0060) and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 statement [22].

Search strategy and study inclusion

A systematic search was performed in MEDLINE, PubMed Central, EMBASE, and the Cochrane Library, including abstract presentations during meetings and preprints (see Supplementary Material 1 for the full search strategy).

Randomized studies were included when comparing ECPR-based treatment to CCPR in patients with refractory OHCA and reporting the primary outcome (neurologically favorable survival at six months). Studies were excluded if they applied a non-randomized design or did not report results on the outcome of neurologically favorable survival. The search was performed by two experienced reviewers (SH and MJK).

Data extraction

Data extraction was performed by two reviewers using a predefined worksheet, which can be found in Supplementary Material 2 (SH and JFHU).

Risk of bias assessment

Risk of bias was independently assessed using the Risk of Bias 2.0 tool (RoB 2.0) [23] by two reviewers (SH and JFHU). The final judgment ranged from ‘high’ to ‘low’ risk of bias, based on the RoB 2.0 tool.

Outcomes

The primary outcome of the current study was 6-month neurologically favorable survival, preferably categorized by the Cerebral Performance Category (CPC) [24]. CPC scores of 1 and 2 were deemed neurologically favorable, and CPC of 3–5 were considered neurologically unfavorable. Alternatively, neurological performance could be assessed using the modified Rankin Scale (mRS) [25], in which an mRS of 0–3 indicated neurologically favorable survival. Given the importance of long-term outcomes and the potential of improvement in neurological status during the first months after refractory OHCA [3], the 6-month time point was considered the primary outcome.

Primary analyses were performed in patients with all rhythms (i.e. refractory OHCA of presumed cardiac origin). Secondary analyses were conducted on patients with shockable rhythms.

The pre-specified outcomes were expressed as mean natural logarithmic (log) relative risks (RRs), median RRs, and mean absolute risk differences (ARDs), accompanied by 95% credible intervals (CrIs), assuming a normal distribution for Bayesian analyses. To enhance clinical interpretation, posterior probabilities of various treatment effect sizes and numbers needed to treat (NNT) were also calculated, with credible intervals denoted as proposed by Altman [26].

Priors

The primary and secondary analyses were performed under vague priors (i.e. a prior assuming no difference, with a normal and wide distribution capturing plausible treatment effects; please see the statistical analysis section for the prior rationale and specification).

Minimal clinically important differences

As proposed previously in an expert consensus statement, a 5% absolute risk difference in survival was deemed clinically important in patients with refractory OHCA due to ventricular fibrillation [27, 28]. Thus, the posterior probability of the MCID was evaluated in the primary and secondary analyses. However, as the referenced expert consensus statement by Nichol et al. did not specifically address the question of the MCID in ECPR research [27], the posterior probabilities of the 10%, 15%, and 20% absolute risk difference thresholds were also studied. Notably, these ARDs also captured the range of the trial protocols’ predefined expected treatment effects [29, 30]. ARDs were converted to NNTs (or numbers needed to harm [NNH]) to facilitate clinical interpretation.

Statistical analysis

Categorical variables were pooled using a random-effects model (using an inverse-variance weighting approach [31]). When continuous variables warranted conversion, Wan’s method was applied [32], and the data were pooled using a random-effects model (inverse-variance weighting) as well. The control group’s (CCPR) risk of survival was considered the baseline risk, and the RR of 6-month neurologically favorable survival in the ECPR group was calculated with CCPR as the reference group. The survival risk of the various trial CCPR groups, also defined as the assumed control risk (ACR), was calculated from a random effects model (expressed as a mean % with corresponding 95% confidence intervals [CIs]). This ACR subsequently facilitated further estimations of ARDs using the approach previously proposed by the Cochrane Collaboration [33, 34]. The rationale and methods for calculating these endpoints are outlined in Supplementary Material 3.

Under a vague prior, no difference between groups is assumed, with a wide distribution of possible treatment effects. As such, the mean RR (mean or μ) of this prior was set to 0 on the log RR scale, and the SD (σ) was set to 2 on the log RR scale. The rationale for this prior selection and prior specification can be found in Supplementary Material 4. All Bayesian hierarchical meta-analyses were performed using a random-effects model. In addition, between-study heterogeneity was assessed using the I2 metric and τ2 [35].

Posterior probabilities were estimated by use of Markov Chain Monte Carlo (MCMC) sampling algorithms using dedicated openly available software (JASP, version 0.17.1 for Mac) [36], and R statistics using the ‘brms’ package (R Statistics, version 4.2.2, Vienna, Austria), by setting four chains and 10,000 saved iterations per chain.

Publication bias assessment

Publication bias was assessed visually by inspection of the funnel plots and statistically evaluated by Egger’s test, in which a p-value of < 0.05 was considered to indicate the presence of statistically significant publication bias (using Meta-Essentials [37]).

Results

Study inclusion

Our systematic search yielded 2081 studies across three electronic databases. After removing duplicate studies, 766 studies were screened based on their titles and abstracts. Eventually, 13 studies were retrieved and assessed for eligibility based on full-text assessment. Finally, three randomized trials were included in the primary analysis [3,4,5] (PRISMA flowchart is presented in Fig. 1).

Fig. 1
figure 1

2020 PRISMA flowchart for the study inclusion

Full size image

The ARREST and INCEPTION-trial included patients with shockable refractory OHCA [3, 5], while Prague OHCA included patients with refractory OHCA of presumed cardiac origin (including non-shockable rhythms) [4]. The results of the initial Prague OHCA report were used in the primary analysis, whereas the results of a post-hoc study evaluating the outcomes of patients with shockable rhythms were used in the secondary analysis [38].

Study and patient characteristics

The three included randomized studies were conducted in the United States [3], Czech Republic [4], and the Netherlands [5] and published between 2020 and 2023. Two trials had a single-center design [3, 4] and one trial had a multi-center design [5].

A total of 420 patients were included in the study (ECPR-based treatment, n = 209 patients; CCPR, n = 211 patients). The mean ages of the patients in the sample were 57 years (SD, 13 years; ECPR group) and 56 years (SD, 12 years; CCPR group). Further studies and patient characteristics are shown in Table 1.

Table 1 Study and patient characteristics of the included trials
Full size table

Risk of bias

The risk of bias assessment showed some concerns regarding the presence of the risk of bias, as reported in Supplementary Material 5. These risks mainly arose from bias due to the randomization process or bias due to deviations from the intended intervention.

Assumed control risk for the CCPR group

For the primary analysis, the ACR was calculated in patients with all rhythms (OHCA of presumed cardiac origin), which yielded a CCPR risk of 6-month neurologically favorable survival of 18.4% (95% CI 11.8–27.7%, I2 = 33%, p = 0.228, τ2 = 0.077). In the secondary analysis (shockable OHCA), the calculated ACR was 19.9% (95% CI 8.3–40.7%, I2 = 76%, p = 0.015, τ2 = 0.513) in the CCPR group. The calculated ACR facilitated further estimation of the absolute risk differences between the groups.

Primary analysis: all rhythms

The primary analysis for 6-month neurologically favorable survival in all patients under a vague prior produced a mean log RR of 0.40 (95%CrI − 0.27;1.20) and translated median RR of 1.47 (95%CrI 0.76–3.32, Table 2, Fig. 2A). This resulted in a mean ARD of 8.7% (95%CrI − 5.0; 42.7%), in favor of ECPR-based treatment (NNT 11, 95%CrI 20 ∞ 2, denoted according to Altman [26]). The posterior probability of an ARD of > 0% was 91.0%.

Table 2 Primary and secondary analyses under a vague prior
Full size table
Fig. 2
figure 2

Primary (A) and secondary (B) Bayesian meta-analyses of primary outcomes under a vague prior. The black horizontal line denotes 95% credible interval

Full size image

The posterior probabilities of an ARD > 5%, > 10%, > 15%, and > 20% were 71.1%, 43.7%, 23.5%, and 13.4%, respectively. The full posterior probability distribution of the absolute risk difference in favor of the ECPR-based treatment is presented in Fig. 3A.

Fig. 3
figure 3

Full posterior probability distribution of the primary (A) and secondary (B) Bayesian meta-analyses of the primary outcome under a vague prior. The black horizontal line denotes the 95% credible interval. ARD: absolute risk difference, CCPR: conventional cardiopulmonary resuscitation, ECPR: extracorporeal cardiopulmonary resuscitation

Full size image

Secondary analysis under the vague prior (shockable rhythms)

The secondary analysis for 6-month neurologically favorable survival in patients with shockable rhythms under a vague prior estimated a slightly more positive mean log RR of 0.43 (95%CrI -0.23; 1.31) and translated median RR of 1.54 (95%CrI 0.79–3.71) (Table 2, Fig. 2B). This resulted in a mean ARD of 10.8% (95%CrI − 4.2; 73.9%) in favor of ECPR-based treatment (NNT 9, 95%CrI 24 ∞ 1). The posterior probability of an ARD of > 0% was 92.4%. The posterior probabilities of an ARD > 5%, > 10%, > 15%, and > 20% were 75.8%, 50.4%, 29.8%, and 17.6%, respectively (Fig. 3B).

Publication bias

There was no statistically significant evidence of publication bias (Egger’s test—p = 0.239, Supplementary Material 6).

Discussion

The current Bayesian meta-analysis estimated a 91.0–92.4% posterior probability of any effect of ECPR-based treatment on neurologically favorable survival in patients with refractory OHCA, and a 71.1–75.8% posterior probability of a clinically relevant effect of ECPR-based treatment, in patients with all rhythms and shockable rhythms, respectively. We observed an estimated mean absolute risk difference in 6-month neurologically favorable survival ranging between 8.7% and 10.8% in favor of ECPR-based treatment in these patient groups. Due to the use of the applicable random-effects models and the relatively low number of included patients, credible intervals were wide, illustrating the uncertainty of the estimated effects.

A recent patient-level meta-analysis conducted by the authors of the ARREST and Prague OHCA trials—including 286 patients with refractory OHCA—demonstrated a statistically significant 6-month neurological survival benefit in favor of an ECPR-based treatment (ARD 12.7%, 95% CI 2.6–22.7%, p = 0.015) [6]. These results are encouraging and seem to favor the use of ECPR in this patient population, although it must be noted that the two single-center trials were performed at two acknowledged centers with extensive expertise in ECPR. As such, the results of this individual patient data analysis may be less generalizable to ‘real-world’ settings. Simultaneously with ARREST and Prague OHCA, the INCEPTION trial was conducted, a multi-center RCT performed in 10 Dutch centers with established cardiac surgery programs [5]. This trial failed to demonstrate a statistically significant effect of the ECPR-based treatment. Given the multi-center and pragmatic design of the INCEPTION trial, pooling these additional data may enhance the external validity and generalizability of randomized ECPR outcomes, as presented in the current analysis. However, it should be kept in mind that the effectiveness of ECPR-based treatment for OHCA may vary among different systems for the chains of survival.

During the past years, several study-level meta-analyses of these randomized trials have been conducted under a frequentist framework, with or without incorporation of in-hospital cardiac arrest patients [13], and with or without incorporation of non-randomized studies to augment the analysis’ power [7,8,9,10, 12,13,14]. When focused on OHCA, the majority of pooled analyses did not observe a statistically significant ECPR treatment effect [7,8,9, 13]. Still Scquizzato and colleagues reported a statistically significant favorable neurological survival benefit at longest follow-up (odds ratio 1.90, 95% CI 1.16–3.13, p = 0.02). These diverging results seem to be the consequence of heterogeneity in statistical approaches, as the latter study applied a fixed effects model to the pooled analysis [10]. Particularly the use of fixed versus random effects models under the frequentist framework seems disputable since it may not be realistic to assume one true underlying effect size in ECPR research, given the RCTs’ varying study designs [39, 40]. Furthermore, in their commendable updated meta-analysis, Low et al. report a statistically significant benefit of ECPR in refractory OHCA (OR 0.67, 95% CI 0.51–0.88) [12] after the addition of two studies, as compared to their 2023-analysis [13]. Although well-performed, this analysis heavily relies on the incorporation of non-randomized data, which subjects the meta-analysis to all biases associated with such designs. Consequently, results may become less reliable. These observations underline the limitations of the frequentist statistical framework, which reduces the interpretation of randomized trials and meta-analyses to a mere a positive or negative result, based on the p-value and the significance level (usually 0.05). In turn, the p-value is determined by the sample size and treatment effect. Indeed, even in case of a small effect, an infinitely large sample size can lead to a statistically significant result (the p-value crossing the 0.05-boundary), while a large treatment effect may not statistically surface in studies or pooled analyses comprising small sample sizes. This is also illustrated by the interpretation of Prague OHCA as a negative trial based on the p-value [4]. Therefore, particularly in a cost- and resource-consuming research-field as refractory OHCA and ECPR, the Bayesian approach may be applicable as it facilitates a more intuitive clinical interpretation of the trials’ results in terms of posterior probabilities and clinically relevant treatment effects.

The Bayesian methodological approach relies on the combination of the ‘prior’ and ‘likelihood’ (i.e., the current data), to obtain the ‘posterior’ [41]. To ensure objectivity, we have opted to perform all our primary analyses under a vague prior, which assumes no difference between treatments and considers all possible treatment effects equally likely. As such, the analyses performed in this study are a reflection of the totality of randomized evidence in ECPR literature, and are not influenced by the biases that are associated with non-randomized study designs.

The ARREST trial applied Bayesian inference as the primary statistical approach, declaring the superiority of ECPR if the posterior probability of an ECPR survival benefit exceeded 98.6% during interim analyses [3]. Furthermore, the Prague OHCA and INCEPTION trial reported a post-hoc Bayesian reanalysis of their results [28, 42], following the primary report under the frequentist framework [4, 5]. The Bayesian reanalysis of the Prague OHCA trial demonstrated a posterior probability of any effect of ECPR-based treatment (ARD > 0%) of 96.1% for 6-month neurologically favorable survival under a weakly informative prior, but did not study further treatment effect thresholds such as the MCID [42]. The post-hoc analysis of the INCEPTION trial reported a 71.6% posterior probability of any effect of ECPR-based treatment and a 42.1% posterior probability of the MCID (ARD > 5%) under a minimally informative prior [28].

Although the credible intervals were wide in the current pooled analysis, our findings seem to support the presence of a clinically relevant effect of ECPR-based treatment in refractory OHCA, with a posterior probability of an ARD of > 5% for 6-month neurologically favorable survival of 71.1% and 75.8% in patients with all rhythms and shockable rhythms, respectively. The MCID of an ARD > 5% was based on a published expert consensus statement by Nichol et al., who evaluated experts’ interpretation of a clinically relevant treatment effect for any intervention applied to patients with refractory shockable OHCA regarding a good neurological outcome at discharge [27]. However, there was high variability in the judgment of this threshold, and this survey was not focused on ECPR, an exceptionally cost- and resource-consuming intervention. Consequently, we also estimated the posterior probabilities of various more extreme treatment effect size ranges, including > 10%, 15%, and > 20% ARD.

Our findings must be interpreted with caution, as the varying designs of the trials and experience of the centers performing ECPR should be considered, as well as the uncertainty of the analyses, as reflected by the relatively wide credible intervals. Consequently, we believe that ECPR can be highly effective in ideal circumstances, which comprises dedicated treatment teams and a sufficient case load [40]. In such instances, the probability of clinically relevant results is highly likely. Instead, in less ideal circumstances, such as in inexperienced centers with rather low case volumes, ECPR may provide less benefit, and its implementation should perhaps be reconsidered. Clinical guidelines could aid in these processes by appointing centers of expertise and defining thresholds for case volumes.

Finally, we advocate for the use of a predefined Bayesian approach in randomized clinical trials studying diseases such as refractory OHCA and interventions like ECPR, given their relative infrequency, and the time- and cost-consuming nature of the intervention. It may indeed be implausible to achieve a sufficient sample size to detect a clinically relevant treatment effect under the frequentist statistical framework. For future trials, the Bayesian approach could facilitate the estimation of clinically relevant treatment effects, while the results of the current pooled analysis of RCTs can be used as an informed prior for such studies.

Strengths and limitations

First, this meta-analysis only included RCTs, intended to reduce various forms of risk bias associated with non-randomized study designs. In addition, the study was performed by the authors of the three RCTs. Nevertheless, a potential limitation is related to the facts that the included RCTs had varying designs (single- or multi-center), varied markedly in sample size (i.e., ARREST comprised 30 patients), applied different randomization protocols, and Prague OHCA and INCEPTION were subjected to deviations from the intended treatment [4, 5].

Not all patients in the ECPR groups of the Prague OHCA and INCEPTION trials’ intention-to-treat analyses actually received ECPR [4, 5]. This could be a consequence of ECPR being part of a bundle of therapies (i.e. a hyperinvasive approach including ECPR in Prague OHCA [4]), the randomization process (i.e. pre-hospital randomization), the pre-hospital return of spontaneous circulation, or cross-overs. Although per-protocol data is available [43], the inclusion of such analyses exposes the current study to new forms of bias. Therefore, the term ‘ECPR-based treatment’ was used in this study.

The choice and specification of an informed prior within the Bayesian framework is heavily debated in the literature and may be exposed to criticism in terms of the relative subjectivity of its elicitation. Therefore, all analyses were conducted under a vague prior, assuming no difference between treatments with a wide distribution of possible treatment effects, ensuring optimal objectivity. In addition, there is wide variability in the determination of the MCID in ECPR research. We attempted to mitigate for this limitation by studying several other treatment effect sizes. Finally, some advocate for using fixed-effects models in meta-analyses with a relatively low number of trials. However, this artificially decreases the uncertainty of the observed effects. Therefore, we employed the appropriate random-effects model, which does not assume one true effect size to underlie all studies in the meta-analysis.

Conclusion

The current Bayesian meta-analysis found a 71.1% and 75.8% probability of a clinically relevant ECPR-based treatment effect on 6-month neurologically favorable survival in refractory OHCA patients with all rhythms and shockable rhythms, respectively. The observed differences may be considered clinically relevant but must be interpreted within the context of the reported credible intervals and varying trial designs.

Availability of data and materials

All coding and analyses are openly shared and accessible through the corresponding author’s GitHub repository (https://github.com/samuelheuts/Bayesian_analysis_ECPRtrials) upon publication of this study.

Abbreviations

ACR:

Assumed control risk

ARD:

Absolute risk difference

CCPR:

Conventional cardiopulmonary resuscitation

CPC:

Cerebral performance category

CI:

Confidence interval

CrI:

Credible interval

ECPR:

Extracorporeal cardiopulmonary resuscitation

MCID:

Minimal clinically important difference

MCMC:

Markov Chain Monte Carlo

mRS:

Modified Rankin Scale

NNT:

Number needed to treat

OHCA:

Out-of-hospital cardiac arrest

RCT:

Randomized controlled trial

RoB:

Risk of bias

RR:

Relative risk

V-A ECMO:

Veno-arterial extracorporeal membrane oxygenation

References

  1. Yannopoulos D, Bartos JA, Raveendran G, Conterato M, Frascone RJ, Trembley A, John R, Connett J, Benditt DG, Lurie KG, et al. Coronary artery disease in patients with out-of-hospital refractory ventricular fibrillation cardiac arrest. J Am Coll Cardiol. 2017;70(9):1109–17.

    Article  PubMed  Google Scholar 

  2. Choi DS, Kim T, Ro YS, Ahn KO, Lee EJ, Hwang SS, Song SW, Song KJ, Shin SD. Extracorporeal life support and survival after out-of-hospital cardiac arrest in a nationwide registry: a propensity score-matched analysis. Resuscitation. 2016;99:26–32.

    Article  PubMed  Google Scholar 

  3. Yannopoulos D, Bartos J, Raveendran G, Walser E, Connett J, Murray TA, Collins G, Zhang L, Kalra R, Kosmopoulos M, et al. Advanced reperfusion strategies for patients with out-of-hospital cardiac arrest and refractory ventricular fibrillation (ARREST): a phase 2, single centre, open-label, randomised controlled trial. Lancet. 2020;396(10265):1807–16.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Belohlavek J, Smalcova J, Rob D, Franek O, Smid O, Pokorna M, Horak J, Mrazek V, Kovarnik T, Zemanek D, et al. Effect of intra-arrest transport, extracorporeal cardiopulmonary resuscitation, and immediate invasive assessment and treatment on functional neurologic outcome in refractory out-of-hospital cardiac arrest: a randomized clinical trial. JAMA. 2022;327(8):737–47.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Suverein MM, Delnoij TSR, Lorusso R, Brandon Bravo Bruinsma GJ, Otterspoor L, Elzo Kraemer CV, Vlaar APJ, van der Heijden JJ, Scholten E, den Uil C, et al. early extracorporeal cpr for refractory out-of-hospital cardiac arrest. N Engl J Med. 2023;388(4):299–309.

    Article  PubMed  Google Scholar 

  6. Belohlavek J, Yannopoulos D, Smalcova J, Rob D, Bartos J, Huptych M, Kavalkova P, Kalra R, Grunau B, Taccone FS, Aufderheide TP. Intraarrest transport, extracorporeal cardiopulmonary resuscitation, and early invasive management in refractory out-of-hospital cardiac arrest: an individual patient data pooled analysis of two randomised trials. EClinicalMedicine. 2023;59:101988.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Gomes DA, Presume J, Ferreira J, Oliveira AF, Miranda T, Brizido C, Strong C, Tralhao A. Extracorporeal cardiopulmonary resuscitation for refractory out-of-hospital cardiac arrest: a systematic review and meta-analysis of randomized clinical trials. Intern Emerg Med. 2023;18(7):2113–20.

    Article  PubMed  Google Scholar 

  8. Cheema HA, Shafiee A, Jafarabady K, Seighali N, Shahid A, Ahmad A, Ahmad I, Ahmad S, Pahuja M, Dani SS. Extracorporeal cardiopulmonary resuscitation for out-of-hospital cardiac arrest: a meta-analysis of randomized controlled trials. Pacing Clin Electrophysiol. 2023;46(10):1246–50.

    Article  PubMed  Google Scholar 

  9. Kiyohara Y, Kampaktsis PN, Briasoulis A, Kuno T. Extracorporeal membrane oxygenation-facilitated resuscitation in out-of-hospital cardiac arrest: a meta-analysis of randomized controlled trials. J Cardiovasc Med (Hagerstown). 2023;24(7):414–9.

    Article  CAS  PubMed  Google Scholar 

  10. Scquizzato T, Bonaccorso A, Swol J, Gamberini L, Scandroglio AM, Landoni G, Zangrillo A. Refractory out-of-hospital cardiac arrest and extracorporeal cardiopulmonary resuscitation: a meta-analysis of randomized trials. Artif Organs. 2023;47(5):806–16.

    Article  PubMed  Google Scholar 

  11. Wang JY, Chen Y, Dong R, Li S, Peng JM, Hu XY, Jiang W, Wang CY, Weng L, Du B. China critical care clinical trials G: extracorporeal vs. Conventional CPR for out-of-hospital cardiac arrest: a systematic review and meta-analysis. Am J Emerg Med. 2024;80:185–93.

    Article  PubMed  Google Scholar 

  12. Low CJW, Ling RR, Ramanathan K, Chen Y, Rochwerg B, Kitamura T, Iwami T, Ong MEH, Okada Y. Extracorporeal cardiopulmonary resuscitation versus conventional CPR in cardiac arrest: an updated meta-analysis and trial sequential analysis. Crit Care. 2024;28(1):57.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Low CJW, Ramanathan K, Ling RR, Ho MJC, Chen Y, Lorusso R, MacLaren G, Shekar K, Brodie D. Extracorporeal cardiopulmonary resuscitation versus conventional cardiopulmonary resuscitation in adults with cardiac arrest: a comparative meta-analysis and trial sequential analysis. Lancet Respir Med. 2023;11(10):883–93.

    Article  PubMed  Google Scholar 

  14. Scquizzato T, Bonaccorso A, Consonni M, Scandroglio AM, Swol J, Landoni G, Zangrillo A. Extracorporeal cardiopulmonary resuscitation for out-of-hospital cardiac arrest: A systematic review and meta-analysis of randomized and propensity score-matched studies. Artif Organs. 2022;46(5):755–62.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Brophy JM. Key issues in the statistical interpretation of randomized clinical trials. Can J Cardiol. 2021;37(9):1312–21.

    Article  PubMed  Google Scholar 

  16. Goodman S. A dirty dozen: twelve p-value misconceptions. Semin Hematol. 2008;45(3):135–40.

    Article  PubMed  Google Scholar 

  17. Wasserstein RL, Lazar NA. The ASA statement on p-values: context, process, and purpose. Am Stat. 2016;70(2):129–33.

    Article  Google Scholar 

  18. Kaul S, Diamond GA. Trial and error: How to avoid commonly encountered limitations of published clinical trials. J Am Coll Cardiol. 2010;55(5):415–27.

    Article  PubMed  Google Scholar 

  19. Wasserstein RL, Schirm AL, Lazar NA. Moving to a world beyond “p < 005.” Am Stat. 2019;73:1–19.

    Article  Google Scholar 

  20. Pieper D, Rombey T. Where to prospectively register a systematic review. Syst Rev. 2022;11(1):8.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Canellas J, Ritto FG, Rodolico A, Aguglia E, Fernandes GVO, Figueredo C, Vettore MV. The international platform of registered systematic review and meta-analysis protocols (INPLASY) at 3 years: an analysis of 4,658 registered protocols on inplasy.com, platform features, and website statistics. Front Res Metr Anal. 2023;8:1135853.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Sterne JAC, Savovic J, Page MJ, Elbers RG, Blencowe NS, Boutron I, Cates CJ, Cheng HY, Corbett MS, Eldridge SM, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.

    Article  PubMed  Google Scholar 

  24. Becker LB, Aufderheide TP, Geocadin RG, Callaway CW, Lazar RM, Donnino MW, Nadkarni VM, Abella BS, Adrie C, Berg RA, et al. Primary outcomes for resuscitation science studies: a consensus statement from the American Heart Association. Circulation. 2011;124(19):2158–77.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Rittenberger JC, Raina K, Holm MB, Kim YJ, Callaway CW. Association between cerebral performance category, modified rankin scale, and discharge disposition after cardiac arrest. Resuscitation. 2011;82(8):1036–40.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Altman DG. Confidence intervals for the number needed to treat. BMJ. 1998;317(7168):1309–12.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Nichol G, Brown SP, Perkins GD, Kim F, Sterz F, Broeckel Elrod JA, Mentzelopoulos S, Lyon R, Arabi Y, Castren M, et al. What change in outcomes after cardiac arrest is necessary to change practice? Results Int Surv Resuscit. 2016;107:115–20.

    Google Scholar 

  28. Heuts S, van de Koolwijk AF, Gabrio A, Ubben JFH, van der Horst ICC, Delnoij TSR, Suverein MM, Maessen JG, Lorusso R, van de Poll MCG: A Bayesian re-analysis of the INCEPTION-trial. Eur Heart J Acute Cardiovasc Care 2023.

  29. Belohlavek J, Kucera K, Jarkovsky J, Franek O, Pokorna M, Danda J, Skripsky R, Kandrnal V, Balik M, Kunstyr J, et al. Hyperinvasive approach to out-of hospital cardiac arrest using mechanical chest compression device, prehospital intraarrest cooling, extracorporeal life support and early invasive assessment compared to standard of care. A randomized parallel groups comparative study proposal. “Prague OHCA study.” J Transl Med. 2012;10:163.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Bol ME, Suverein MM, Lorusso R, Delnoij TSR, Brandon Bravo Bruinsma GJ, Otterspoor L, Kuijpers M, Lam KY, Vlaar APJ, Elzo Kraemer CV, et al. Early initiation of extracorporeal life support in refractory out-of-hospital cardiac arrest: Design and rationale of the INCEPTION trial. Am Heart J. 2019;210:58–68.

    Article  PubMed  Google Scholar 

  31. Marín-Martínez F, Sánchez-Meca J. Weighting by inverse variance or by sample size in meta-analysis: a simulation study. Edu Psychol Measur. 2009;70(1):56.

    Article  Google Scholar 

  32. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol. 2014;14:135.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Schünemann HJ, Vist GE, Higgins JPT, Santesso N, Deeks JJ, Glasziou P, Akl EA, Guyatt GH. Interpreting results and drawing conclusions. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editors. Cochrane handbook for systematic reviews of interventions. Wiley; 2019.

    Google Scholar 

  34. Heuts S, Gabrio A, Veenstra L, Maesen B, Kats S, Maessen JG, Walton AS, Nanayakkara S, Lansky AJ. van ’t Hof AWJ, Vriesendorp PA: Stroke reduction by cerebral embolic protection devices in transcatheter aortic valve implantation: a systematic review and Bayesian meta-analysis. Heart. 2023. https://doi.org/10.1136/heartjnl-2023-323359.

    Article  Google Scholar 

  35. Borenstein M. Avoiding common mistakes in meta-analysis: Understanding the distinct roles of Q, I-squared, tau-squared, and the prediction interval in reporting heterogeneity. Res Synth Methods. 2023. https://doi.org/10.1002/jrsm.16785.

    Article  PubMed  Google Scholar 

  36. Berkhout S, Haaf J, Gronau Q, Heck D, Wagenmakers E. a tutorial on bayesian model-averaged meta-analysis in JASP. 2021.

  37. Suurmond R, van Rhee H, Hak T. Introduction, comparison, and validation of Meta-Essentials: A free and simple tool for meta-analysis. Res Synth Methods. 2017;8(4):537–53.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Rob D, Smalcova J, Smid O, Kral A, Kovarnik T, Zemanek D, Kavalkova P, Huptych M, Komarek A, Franek O, et al. Extracorporeal versus conventional cardiopulmonary resuscitation for refractory out-of-hospital cardiac arrest: a secondary analysis of the Prague OHCA trial. Crit Care. 2022;26(1):330.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Wetterslev J, Jakobsen JC, Gluud C. Trial sequential analysis in systematic reviews with meta-analysis. BMC Med Res Methodol. 2017;17(1):39.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Ubben JFH, Heuts S, Delnoij TSR, Suverein MM, van de Koolwijk AF, van der Horst ICC, Maessen JG, Bartos J, Kavalkova P, Rob D, et al. Extracorporeal cardiopulmonary resuscitation for refractory OHCA: lessons from three randomized controlled trials-the trialists’ view. Eur Heart J Acute Cardiovasc Care. 2023;12(8):540–7.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Yarnell CJ, Abrams D, Baldwin MR, Brodie D, Fan E, Ferguson ND, Hua M, Madahar P, McAuley DF, Munshi L, et al. Clinical trials in critical care: can a Bayesian approach enhance clinical and scientific decision making? Lancet Respir Med. 2021;9(2):207–16.

    Article  PubMed  Google Scholar 

  42. Rob D, Komarek A, Smalcova J, Belohlavek J. Effect of intraarrest transport, extracorporeal cardiopulmonary resuscitation, and invasive treatment: a post hoc bayesian reanalysis of a randomized clinical trial. Chest. 2023. https://doi.org/10.1016/j.chest.2023.07.030.

    Article  PubMed  Google Scholar 

  43. Ubben JFH, Suverein MM, Delnoij TSR, Heuts S, Winkens B, Gabrio A, van der Horst ICC, Maessen JG, Lorusso R, van de Poll MCG. Early extracorporeal CPR for refractory out-of-hospital cardiac arrest: a pre-planned per-protocol analysis of the INCEPTION-trial. Resuscitation. 2024;194:110033. https://doi.org/10.1016/j.resuscitation.2023.110033.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

None.

Funding

No funding was received for this study.

Author information

Author notes

  1. Samuel Heuts and Johannes F. H. Ubben contributed equally as first authors.

  2. Demetris Yannopoulos, Jan Bělohlávek, Roberto Lorusso and Marcel C. G. van de Poll share senior authorship.

Authors and Affiliations

  1. Department of Cardiothoracic Surgery, Maastricht University Medical Center, MUMC+), P. Debyelaan 25, 6229HX, Maastricht, The Netherlands

    Samuel Heuts, Michal J. Kawczynski, Jos G. Maessen & Roberto Lorusso

  2. Cardiovascular Research Institute Maastricht (CARIM), University Maastricht, Maastricht, The Netherlands

    Samuel Heuts, Michal J. Kawczynski, Iwan C. C. van der Horst, Jos G. Maessen & Roberto Lorusso

  3. Department of Anesthesiology and Pain Medicine, Maastricht University Medical Center (MUMC+), Maastricht, The Netherlands

    Johannes F. H. Ubben

  4. Department of Methodology and Statistics, University Maastricht, Maastricht, The Netherlands

    Andrea Gabrio

  5. Care and Public Health Research Institute (CAPHRI), University Maastricht, Maastricht, The Netherlands

    Andrea Gabrio

  6. Department of Intensive Care Medicine, Maastricht University Medical Center (MUMC+), Maastricht, The Netherlands

    Martje M. Suverein, Thijs S. R. Delnoij, Iwan C. C. van der Horst & Marcel C. G. van de Poll

  7. 2nd Department of Medicine-Department of Cardiovascular Medicine, First Medical School, General University Hospital and Charles University in Prague, Prague, Czech Republic

    Petra Kavalkova, Daniel Rob & Jan Bělohlávek

  8. Department of Probability and Mathematical Statistics, Faculty of Mathematics and Statistics, Charles University in Prague, Prague, Czech Republic

    Arnošt Komárek

  9. Center for Resuscitation Medicine, University of Minnesota Medical School, Minneapolis, MN, USA

    Demetris Yannopoulos

  10. School of Nutrition and Translational Research in Metabolism, University Maastricht, Maastricht, The Netherlands

    Marcel C. G. van de Poll

Authors
  1. Samuel Heuts
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Johannes F. H. Ubben
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Michal J. Kawczynski
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Andrea Gabrio
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Martje M. Suverein
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Thijs S. R. Delnoij
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Petra Kavalkova
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Daniel Rob
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Arnošt Komárek
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Iwan C. C. van der Horst
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Jos G. Maessen
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Demetris Yannopoulos
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Jan Bělohlávek
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Roberto Lorusso
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Marcel C. G. van de Poll
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

SH: conceptualization, data curation, formal analysis, investigation, methodology, software, visualization, writing original draft. JFHU: conceptualization, data curation, methodology, validation, visualization, writing original draft. MJK: data curation, formal analysis, investigation, methodology, software, visualization, writing—reviewing and editing. AG: data curation, formal analysis, investigation, methodology, software, supervision, validation, writing—reviewing and editing. MMS: conceptualization, data curation, supervision, validation, writing—reviewing and editing. TSRD: conceptualization, data curation, supervision, validation, writing—reviewing and editing. PK: conceptualization, data curation, supervision, validation, writing—reviewing and editing. DR: conceptualization, data curation, supervision, validation, writing—reviewing and editing. AK: conceptualization, data curation, supervision, validation, writing—reviewing and editing. ICCvdH: conceptualization, data curation, supervision, validation, writing original draft. JGM: conceptualization, data curation, supervision, validation, writing—reviewing and editing. DY: conceptualization, data curation, supervision, validation, writing—reviewing and editing. JB: conceptualization, data curation, supervision, validation, writing—reviewing and editing. RL: conceptualization, data curation, supervision, validation, writing—reviewing and editing. MCGvdP: conceptualization, data curation, formal analysis, investigation, methodology, writing original draft.

Corresponding author

Correspondence to Samuel Heuts.

Ethics declarations

Ethical approval and consent to participate

Not applicable.

Consent for publication

No informed consent from patients was required for this analysis. All authors have consented to the publication of this study.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Heuts, S., Ubben, J.F.H., Kawczynski, M.J. et al. Extracorporeal cardiopulmonary resuscitation versus standard treatment for refractory out-of-hospital cardiac arrest: a Bayesian meta-analysis. Crit Care 28, 217 (2024). https://doi.org/10.1186/s13054-024-05008-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13054-024-05008-9

Keywords

Critical Care

ISSN: 1364-8535