Background
Out-of-hospital cardiac arrest (OHCA) survival has been stuck for so long. Despite improvements in bystander CPR, early defibrillation and advanced life support, patients who remain in refractory cardiac arrest have a dismal prognosis which has stayed mostly the same for many years.
Extracorporeal cardiopulmonary resuscitation (ECPR) offers something fundamentally different. Rather than relying on chest compressions to generate a fraction of normal cardiac output, veno-arterial extracorporeal membrane oxygenation (V-A ECMO) can restore systemic perfusion, provide oxygen delivery and allow time to identify and treat the underlying cause of cardiac arrest.
The ARREST trial demonstrated an extraordinary improvement in survival to hospital discharge, from 7% with conventional resuscitation to 43% with ECPR. However, this was a small, single-centre study involving a highly selected population and an exceptionally experienced ECPR service. Subsequently, Prague OHCA and INCEPTION failed to demonstrate statistically significant improvements in their primary outcomes, raising questions about whether the benefits of ECPR could be reproduced outside specialist centres. All three trials relied on transporting patients to hospital before initiating ECPR. We know that low-flow duration is a major determinant of neurological outcome following cardiac arrest. Is it therefore surprising that ECPR becomes less effective when we spend an hour getting patients onto ECMO?
What if, instead of taking patients to ECMO, we took ECMO to the patients? The Dutch investigators behind OnScene have attempted to answer this question.
The Abstract
BACKGROUND
In patients with refractory out-of-hospital cardiac arrest, extracorporeal cardiopulmonary resuscitation (ECPR) with the use of venoarterial extracorporeal membrane oxygenation can be performed to establish reperfusion. Data comparing prehospital ECPR with conventional care regarding survival until hospital discharge among patients who are 50 years of age or younger with refractory out-of-hospital cardiac arrest are limited.
METHODS
In a nationwide, nonrandomized study conducted in the Netherlands, we evaluated the outcomes in young adults who were attended by Helicopter Emergency Medical Services (HEMS) that were equipped to deliver prehospital ECPR (intervention group) as compared with those attended by HEMS without ECPR capability (conventional group). The primary outcome was survival until hospital discharge.
RESULTS
A total of 144 patients were included in the intervention group, and 87 were included in the conventional group. In the intervention group, 57 of 144 patients (40%) survived until hospital discharge, as compared with 17 of 87 patients (20%) in the conventional group (odds ratio, 2.72; 95% confidence interval [CI], 1.59 to 4.66; P=0.002). At 6 months, a favorable neurologic outcome was reported in 51 of 141 patients (36%) in the intervention group and in 11 of 85 patients (13%) in the conventional group (odds ratio, 4.22; 95% CI, 2.82 to 6.32). There were 32 adverse events in the intervention group and 11 in the conventional group.
CONCLUSIONS
Among patients who were 50 years of age or younger with out-of-hospital cardiac arrest, incorporating ECPR into HEMS was associated with improved survival until hospital discharge, as compared with conventional care.
What kind of study is this?
OnScene was a prospective, non-randomised study using a stepped-wedge design.
The Netherlands has four HEMS stations, all of which initially operated without pre-hospital ECPR capability. Between 2022 and 2025, the stations progressively introduced ECPR following a structured training programme. Patients treated before implementation formed the conventional group, while those treated after implementation formed the intervention group.
This allowed the investigators to evaluate a complex intervention across an entire national HEMS system, rather than within a single specialist centre.
Importantly, it wasn’t a randomised trial. Treatment allocation depended on the HEMS station and the period during which the patient was treated. We’ll return to that later, but it is worth recognising that this was an evaluation of a national system of care.
Tell me about the patients
The investigators selected patients in whom ECPR was most likely to offer a meaningful benefit. Patients had to be aged between 18 and 50 years, have a witnessed cardiac arrest or signs of life, and present with either an initial shockable rhythm or PEA where pulmonary embolism was suspected. The arrest had to remain refractory for more than 20 minutes after HEMS dispatch, and the team had to be ready for vascular puncture within 45 minutes of dispatch. Patients were excluded if their end-tidal carbon dioxide was below 1.2 kPa, or if they could be transported to an ECMO centre within an estimated 30 minutes of collapse.
Over almost four years, HEMS received 3,362 study-related dispatches. Of these, 231 patients ultimately met the study criteria, representing just 6.9% of the original dispatches. This was therefore a highly selected population, but that is entirely appropriate. ECPR is not a treatment for every cardiac arrest, and identifying patients with potentially reversible pathology and a realistic prospect of neurological recovery is fundamental to its success. The median age was 44 years in both groups, and almost all arrests were witnessed.
There was one notable difference between groups. An initial shockable rhythm was present in 89% of the intervention group compared with 75% of the conventional group. This is important because presenting rhythm is strongly associated with outcome, although the investigators accounted for this difference in an adjusted analysis.
What were the measured outcomes in this study?
The primary outcome was survival to hospital discharge. The principal secondary outcome was favourable neurological function at six months, defined as a Cerebral Performance Category (CPC) score of 1 or 2. These are exactly the outcomes we should be interested in. Establishing ECMO flow is technically impressive, but what matters is whether patients survive and return to meaningful neurological function.
The investigators also examined time to extracorporeal circulation, procedural complications and outcomes among patients who actually underwent pre-hospital or hospital-based ECPR.
What are the main results?
Before you read this, ask yourself two questions. What would the percentage survival be for this cohort in the system I work in, and what would be a ‘good’ increase in that survival; one worth paying for? Survival to hospital discharge occurred in 57 of 144 patients (39.6%) in the intervention group, compared with 17 of 87 patients (19.5%) receiving conventional care. That is an absolute survival difference of 20%, an approximate doubling of survival, and an unadjusted number needed to treat of just five.
At six months, 51 of 141 patients (36.2%) in the intervention group had a favourable neurological outcome, compared with 11 of 85 (12.9%) in the conventional group. Again, an extraordinary difference. To put this into context, we are discussing patients with refractory cardiac arrest, not an unselected population in whom many would have achieved ROSC following early defibrillation. These are patients who have already failed conventional resuscitation and would ordinarily face a very poor prognosis.
The only comparative study of refractory OHCA reporting outcomes of this magnitude is ARREST. They achieved 43% survival with ECPR, but within a highly experienced single-centre programme. OnScene has demonstrated comparable survival following implementation across a national HEMS system.
How quickly did they establish ECMO flow?
Median time to extracorporeal blood flow was 45 minutes (IQR 37–50) in patients receiving pre-hospital ECPR, compared with 77 minutes (IQR 65–90) among those receiving hospital-based ECPR (in the conventional group). That’s a 32-minute difference between the reported medians, and a substantial reduction in the time patients spend receiving conventional CPR before extracorporeal circulation can be established.
The investigators also demonstrated that pre-hospital ECPR could be delivered by HEMS physicians who had undergone a structured training programme, despite the vast majority having no previous ECMO experience. This study basically shows that they can.
What about the limitations?
The most significant is the non-randomised design. The intervention and conventional groups were treated during different periods, and the greater proportion of patients with an initial shockable rhythm in the intervention group may have contributed to the observed difference in survival. Although the investigators adjusted for presenting rhythm and study period, residual confounding remains possible.
The study also recruited fewer patients than originally planned, and the neurological outcome assessment was unblinded. Furthermore, the results apply to a carefully selected group of relatively young patients with refractory OHCA, rather than to all patients experiencing cardiac arrest.
Cannulation failure occurred in 15% of attempted pre-hospital procedures. The supplementary appendix describes serious complications, including vascular injury and incorrect cannulation including 12 occurrences of inadvertent V-V cannulation. ECPR is an invasive intervention with substantial potential for harm, and harm clearly occurred in this study as a result of procedural complications. That’s important for anyone acting on these findings – training, governance and QA will be needed.
Despite that, this is a study demonstrating a very large difference in patient-centred outcomes.
Is there anything else to look at?
There have been several ECPR trials in recent years, and the three major randomised trials of ECPR have produced mixed results.
ARREST demonstrated survival to hospital discharge of 43% with ECPR compared with 7% with conventional resuscitation. Prague OHCA reported favourable neurological survival at 180 days of 31.5% versus 22%, while INCEPTION reported favourable neurological survival at 30 days of 20% versus 16%. The populations, interventions and endpoints differed considerably, so these figures cannot be compared directly. Nevertheless, OnScene’s results are among the most impressive reported in comparative ECPR research.
There is an important difference in how these systems delivered treatment. ARREST, Prague OHCA and INCEPTION all relied on transporting patients to hospital for extracorporeal support. OnScene moved the capability into the pre-hospital environment, more like the French SAMU model which is arguably the most well known established system in the world (maybe Minnesota has a claim on that too).
This suggests that the effectiveness of ECPR is determined not simply by whether extracorporeal circulation is established, but by how quickly it can be established, and that’a about systems and geography as much as the technology itself.
And what about the French experience?
OnScene isn’t the only paper published in 2026 to suggest that taking ECMO to the patient might be better than taking the patient to ECMO. Soumagnac, Lamhaut and colleagues have published an analysis of the Greater Paris OHCA registry in Critical Care. Over 13 years, they identified 1,031 patients who underwent attempted ECPR for refractory cardiac arrest, comparing 256 patients managed with on-scene ECPR with 775 who underwent hospital-based ECPR. The results in terms of the differences between the groups are consistent with OnScene.
Favourable neurological survival was 13.3% with on-scene ECPR compared with 5.5% following hospital-based ECPR, with an adjusted odds ratio of 2.3 (95% CI 1.4–3.9). Survival to hospital discharge was 15.6% versus 7.6%, while median time from CPR initiation to ECMO flow was 70 versus 93 minutes respectively. Once again, earlier extracorporeal circulation was associated with approximately twice the likelihood of a favourable outcome.
The absolute survival figures were considerably lower than those reported in OnScene, but the populations were very different. The French cohort was older, included a broader range of presenting rhythms and had substantially longer times to extracorporeal circulation. Also, the Paris study included only patients in whom ECPR was attempted, whereas OnScene looked at a wider ECPR-capable HEMS strategy, including patients who achieved ROSC without cannulation.
The association between on-scene ECPR and improved neurological outcome persisted even after adjustment for time to extracorporeal circulation. That adjustment is not straightforward, because time to flow may itself mediate the benefit of pre-hospital ECPR. Nevertheless, the French investigators suggest that there may be a wider benefit from the specialist teams, patient selection and coordinated resuscitation pathways associated with pre-hospital ECPR.
Perhaps the lesson is that we’re not simply trying to move a cannulation procedure out of hospital. We’re trying to deliver an entire resuscitation system earlier and more effectively.
Experience also seems to matter as a favourable neurological survival with on-scene ECPR increased from 2% in 2011–2014 to 19% in 2021–2024, whereas hospital-based outcomes remained broadly unchanged. Although observational and potentially influenced by changes in patient selection and practice, this suggests that an ECPR programme may improve as it matures.
Taken together, the Dutch and French studies provide observational evidence that well-organised pre-hospital ECPR strategies can improve outcomes in refractory OHCA.
The destination is reperfusion
In the NEJM the commentary from Demetris Yannopoulos, whose Minnesota group delivered the ARREST trial is worth a read. Yannopoulos describes how ECPR has evolved from a late rescue intervention towards a strategy for achieving earlier reperfusion. He recognises the results of OnScene, but argues that the real achievement is not simply demonstrating that ECMO can be established in a field, an ambulance or a patient’s living room. It is demonstrating that an entire regional system can be organised to deliver extracorporeal circulation to patients who would otherwise be unlikely to survive.
There is an important distinction between the location of cannulation and the time to effective circulation. In OnScene, median time to extracorporeal flow was 45 minutes following pre-hospital cannulation, compared with 77 minutes among patients cannulated in hospital within the conventional-care group. However, this was not a randomised comparison of cannulation locations, and we should be cautious about attributing the difference in outcomes solely to where the cannulae were inserted.
Yannopoulos also highlights the technical challenges. Cannulation failed in 15 of 98 pre-hospital attempts, and in another 11 patients the vessels could not be adequately visualised. These are important reminders that pre-hospital ECPR is a complex intervention, and one that requires much more than simply putting an ECMO circuit into a helicopter or car. Notably in London and in other services looking at ECMO there is a lot of training and oversight and this would clearly need to be built in to any future service developments. Rather than establishing isolated ECPR teams everywhere, we should build coordinated regional networks that bring patients and experienced teams together by whichever route is fastest. Sometimes that will mean transporting the patient to an ECMO centre. Sometimes it will mean bringing the ECMO team to the patient. And sometimes it might involve meeting somewhere in between.
The question is no longer whether pre-hospital ECPR is feasible. We know that it is. The question is how we design systems capable of delivering extracorporeal circulation to more patients, earlier, without compromising expertise or the quality of conventional resuscitation. As Yannopoulos concludes, “The destination is timely reperfusion; the route should follow the patient.”
Should we change practice based on this study?
We need to ask questions as to whether we can and/or should look at ECPR capable systems. We now have evidence from a national HEMS system that pre-hospital ECPR can be delivered by appropriately trained teams, can establish extracorporeal circulation substantially earlier than hospital-based pathways, and is associated with a dramatic improvement in survival. Almost 40% of patients survived to hospital discharge, with favourable neurological outcomes in more than a third. In the conventional group, only 20% survived. These are impressive outcomes in patients with refractory OHCA.
Together with the French experience, this provides further evidence that pre-hospital ECPR is associated with dramatically improved neurological survival, while ARREST has already demonstrated what can be achieved by a highly developed hospital-based ECPR system. That said, OnScene isn’t a randomised trial, and we cannot conclude that ECPR was responsible for the entire difference.
The other big signal in this and other studies is that time matters. Pre-hospital cannulation makes a lot of sense, but it’s not without complexity and in some systems, more efficient transfer pathways or faster hospital-based initiation may be the right approach. For HEMS systems considering ECPR, OnScene demonstrates that this capability need not be restricted to a single highly specialised centre. With appropriate training, governance and integration into a wider ECMO network, pre-hospital ECPR can be delivered at scale.
And for the wider resuscitation community, perhaps the most important message is that refractory cardiac arrest does not necessarily mean that further resuscitation is futile.
What does this mean for the UK?
With kudos and thanks to my friends Dinis Miranda, Xavier Moors, the OnScene team, and ECPR colleagues around the world.
Best wishes, Hutch
“ECPR saves lives, just do it!” (Lionel Lamhaut, Paris)
Additional thoughts from across the pond – Zaf Qasim
Kudos to the Dutch group on this extraordinary endeavour – no doubt, as Hutch says, this reiterates again the importance of thinking about options beyond what is conventional in terms of resuscitating the refractory OOHCA patient.
What should our readers in the US take away from this paper? I cannot emphasize enough that it is not the ECPR that saves the patient, it is the SYSTEM – one that has a high enough volume amongst a relatively small number of operators that can care for these patients from scene to ICU and onto rehabilitation. The reps are important. INCEPTION, for example, suffered from having wide distribution of ECPR resulting in low individual volumes. This must be taken into account when attempting to introduce this into our systems.
The US faces some more unique challenges in delivering this prehospital, including a lack of a unified prehospital system as seen in the Netherlands, variable non-physician staffing models, significant geographical distances to traverse, competing financial considerations amongst hospital systems, and sometimes a varying acceptance of the evidence by individuals in leadership positions that may result in efforts never getting off the ground. One could argue these barriers exist in other countries also, just with different labels. For a long time, people in the US looked at the Paris model and said “yeah, we can’t do that here.” However, the Minneapolis team proved that those barriers can (and should) be overcome.
But in as diverse an environment as the US prehospital sphere, this will need to take different forms to achieve similar results. And that’s ok. What I see as a much more feasible solution (at least for now) is for individual regional centers to develop excellence in ECPR delivery, and partner with EMS agencies to identify likely patients early, utilizing HEMS when appropriate to deliver the patient to that single center. Then rinse, re-evaluate the outcomes, adjust the protocols and procedures, and repeat until you get it right FOR YOUR SYSTEM. Perhaps it is not as sexy as having people being cannulated in the Louvre, but it might just leverage our existing resources to develop a system of care that can provide better outcomes for the subset of patients in refractory OHCA – that 1% marginal gain we often talk about.
References
1. Reis Miranda D, Moors X, Ali S, et al. Pre-hospital Extracorporeal Life Support in Refractory Out-of-Hospital Cardiac Arrest. *N Engl J Med*. 2026. [DOI: 10.1056/NEJMoa2602919](https://doi.org/10.1056/NEJMoa2602919).
2. Yannopoulos D, Bartos J, Raveendran G, 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;396:1807–1816.
3. Belohlavek J, Smalcova J, Rob 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;327:737–747.
4. Suverein MM, Delnoij TSR, Lorusso R, et al. Early extracorporeal CPR for refractory out-of-hospital cardiac arrest. *N Engl J Med*. 2023;388:299–309.
5. Soumagnac T, et al. Extracorporeal cardiopulmonary resuscitation strategies for refractory out-of-hospital cardiac arrest. *Critical Care*. 2026. [DOI: 10.1186/s13054-026-06421-y](https://doi.org/10.1186/s13054-026-06421-y).
6. Yannopoulos D. Bringing Rapid Reperfusion to More Patients with Out-of-Hospital Cardiac Arrest. *N Engl J Med*. 2026. [DOI: 10.1056/NEJMe2612650](https://doi.org/10.1056/NEJMe2612650).
7. Patterson T, Perkins GD, Perkins A, et al. Expedited transfer to a cardiac arrest centre for non-ST-elevation out-of-hospital cardiac arrest (ARREST): a UK prospective, multicentre, parallel, randomised clinical trial. The Lancet. Published online October 2023:1329-1337. doi:10.1016/s0140-6736(23)01351-x
8. Prague OHCA trial https://www.thebottomline.org.uk/summaries/icm/prague-ohca/https://www.thebottomline.org.uk/summaries/icm/prague-ohca/
9. INCEPTION trial https://www.nejm.org/doi/full/10.1056/NEJMoa2204511
