Diesen Beitrag gibt es auch auf Deutsch.
A 44-year-old man collapses in the street, ventricular fibrillation, bystanders start CPR at once. After 20 minutes and several shocks there is still no return of spontaneous circulation. Extracorporeal cardiopulmonary resuscitation (ECPR) with venoarterial ECMO is the last option for such patients. So far, it almost always means transport with ongoing chest compressions to an ECMO centre and cannulation there.
The Dutch OnScene study brought ECMO to the patient instead, by helicopter.
Why the question was open
Three randomised trials tested in-hospital ECPR. ARREST showed better survival to discharge in a single, highly experienced centre. The Prague Hyperinvasive trial found no higher rate of good neurological outcome at six months. INCEPTION, multicentre and also Dutch, showed no survival benefit. All three shared long arrest times: a median of 59, 60 and 74 minutes. Since arrest duration strongly determines survival, the obvious question was whether ECPR simply came too late.
Study design
OnScene was a nationwide, nonrandomised study with a stepped-wedge design. All four Dutch helicopter stations started between October 2021 and June 2022 without ECMO on board and switched to ECMO on board one after another. The analysis followed intention to treat: what counted was whether the team could offer ECPR at the time of the call, not whether it actually cannulated.
Eligible patients were 18 to 50 years old with an initial shockable rhythm or pulseless electrical activity with suspected pulmonary embolism. The arrest had to be witnessed or show signs of life and last more than 20 minutes after helicopter dispatch. Exclusions included end-tidal CO2 below 1.2 kPa and an expected arrival at an ECMO centre within 30 minutes. Cannulation was performed by helicopter physicians, anaesthesiologists and trauma surgeons, after a structured training programme.
Of 3,362 dispatches, 231 patients remained: 144 in the ECPR phase and 87 in the conventional phase. Median age was 44 years in both groups.
Results
57 of 144 patients (40%) in the ECPR phase and 17 of 87 (20%) in the conventional phase survived to discharge. The odds ratio was 2.72 (95% CI 1.59 to 4.66, p = 0.002), adjusted 2.66 (1.56 to 4.55). A good neurological outcome (CPC 1 or 2) at six months was seen in 36% versus 13% (OR 4.22, 95% CI 2.82 to 6.32).
The time gain was large. Median time to ECMO flow was 45 minutes with prehospital cannulation and 77 minutes with in-hospital cannulation. From helicopter arrival to flow took 27 instead of 64 minutes.
Not every patient in the ECPR phase was cannulated. An attempt was made in 98 of 144 (68%). 32 patients regained spontaneous circulation first, and in 11 the femoral vessels were not visible on ultrasound. In the conventional phase, 24 patients received in-hospital ECPR.
Safety
Cannulation failed in 15 of 98 prehospital attempts (15%) and in 4 of 24 in-hospital attempts (17%), in line with published in-hospital ECPR data. Bleeding and cannula-site infections occurred at similar rates. In total, 32 adverse events were reported in the ECPR phase and 11 in the conventional phase.
Limitations
The main limitation is the design. Without randomisation, selection effects remain possible. The rhythm distribution shows this: 89% had ventricular fibrillation or tachycardia in the ECPR phase versus 75% in the conventional phase. Pulmonary embolism as the cause was more frequent in the conventional phase (23% versus 11%). Adjusting for rhythm did not materially change the result, and the difference persisted among patients with shockable rhythms (42% versus 25%, post hoc).
The order of the helicopter stations was not randomised. 390 patients were planned and 231 enrolled, because recruitment was slower than expected. Neurological outcome was assessed by a single unblinded intensivist. The direct comparison of prehospital with in-hospital ECPR (35% versus 25% survival) rests on small numbers and no random allocation.
The findings apply only to comparable systems: young patients, a helicopter network reaching almost the entire country within 30 minutes, and physicians with a dedicated training programme. The study was funded by the Dutch Ministry of Health. Author disclosures are available online from the journal and are not part of the article text.
What this means for practice
OnScene provides the first prospective comparison for prehospital ECPR, not randomised proof. The message is still clear: ECPR is time-critical, and the time gain depends on the system, not on a single centre.
- In refractory arrest with a shockable rhythm, consider ECPR early, not after 40 minutes.
- Know your own times: how long does it take in your region from collapse to ECMO flow?
- Plan an ECPR programme as a chain, from dispatch through cannulation to intensive care.
I am happy to discuss how ECPR pathways are organised in your region on LinkedIn.
Source: Reis Miranda D, Moors X, Ali S, et al., Gommers D. Prehospital Extracorporeal Life Support in Refractory Out-of-Hospital Cardiac Arrest. N Engl J Med, online 7 Nov 2026. doi:10.1056/NEJMoa2602919. ClinicalTrials.gov NCT04620070.
Source: DOI: 10.1056/NEJMoa2602919 · Original article