Practical referral criteria for respiratory and cardiac ECMO and ECPR, the scores used (Murray, RESP, SAVE), what the ECMO centre needs on the phone, and how to prepare for retrieval.
At the bedside
Timed steps- 1
Recognise the trigger
EarlyRespiratory: P/F <150 despite optimisation is the time to call (not necessarily cannulate); P/F <80, or pH <7.25 with hypercapnia, is the time to act. Cardiac: SCAI stage C not improving, or any stage D/E. ECPR: witnessed refractory arrest — call within the first 10–15 min of CPR.
- 2
Phone the ECMO centre with an ISBAR summary
CallAge, weight, height, diagnosis and day of illness, days on ventilator, current settings (mode, FiO₂, PEEP, Pplat, RR, VT), ABG with P/F, prone/NMB done, vasopressor doses, echo, lactate, renal function, platelets/INR, neuro status, comorbidities, CT head if relevant, family wishes.
- 3
Calculate Murray and RESP (VV) or SAVE (VA)
ScoreMurray lung injury score >3 was the CESAR referral trigger. RESP and SAVE estimate survival if ECMO is given; they support but do not replace clinical judgement.
- 4
Prepare for cannulation or retrieval
PrepareCross-match blood (at least 4 units PRBC or as the centre asks), check platelets, fibrinogen, INR; ask which vessels to keep free (often right IJ and both femoral veins); arterial line in the right radial for VA (upper-body oxygenation).
- 5
If the patient is too unstable to move
TransferRequest mobile ECMO retrieval: the team cannulates at your bedside and transports on ECMO. Conventional transfer of a very hypoxaemic patient is high risk.
- 6
Inform and consent
FamilyECMO is a bridge, not a cure: explain likely duration, bleeding and stroke risk, cost, and the possibility of withdrawal if there is no recovery.
Referral triggers (call the ECMO centre)
| Situation | Call when | Usual action threshold |
|---|---|---|
| Severe ARDS | P/F <150 on FiO₂ ≥0.6 despite optimisation, or proning needed | P/F <80 (ELSO/EOLIA) or Murray >3 (CESAR) |
| Hypercapnic failure | pH <7.30 with Pplat approaching 30 | pH <7.25 with PaCO₂ ≥60 for >6 h despite RR 35 |
| Cardiogenic shock | SCAI C with rising lactate or escalating drugs | SCAI D–E: refractory hypotension, worsening lactate, end-organ failure |
| Massive PE | Arrest or refractory shock despite (or contraindication to) thrombolysis | Immediate |
| Refractory cardiac arrest (ECPR) | Witnessed arrest, shockable rhythm, good CPR | No ROSC after 10–15 min of ALS in an eligible patient |
Thresholds summarise ELSO, EOLIA and CESAR entry criteria; local centres may differ.
Typical ECPR inclusion criteria (centre-specific)
| Criterion | Typical requirement |
|---|---|
| Age | Usually <65–70 years (biological, not only chronological) |
| Arrest | Witnessed, with immediate bystander CPR (no-flow <5 min) |
| Rhythm | Initial shockable rhythm favoured; PE, hypothermia and poisoning also considered |
| Low-flow time | Expected arrest-to-ECMO-flow <60 min |
| Signs of life | ETCO₂ >10 mmHg, gasping or pupillary reaction during CPR favour candidacy |
| Exclusions | Unwitnessed arrest, end-stage disease, known DNAR, aortic dissection, severe AR, uncontrolled bleeding |
RESP score risk classes (survival on VV ECMO)
| RESP score | Class | Hospital survival in derivation cohort |
|---|---|---|
| ≥6 | I | 92% |
| 3 to 5 | II | 76% |
| −1 to 2 | III | 57% |
| −5 to −2 | IV | 33% |
| ≤−6 | V | 18% |
Pre-ECMO variables: age, immunocompromise, days of ventilation, diagnosis group, CNS dysfunction, non-pulmonary infection, NMB, nitric oxide, bicarbonate infusion, cardiac arrest, PaCO₂ and peak pressure.
Calculators for this topic
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
ECMO referral pathway
Step 1 of ~4
From deterioration to referral, retrieval or ECPR.
Decision: Refractory cardiac arrest now?
Based on Labib A, August E, Agerstrand C, et al; Richardson ASC, Tonna JE, Nanjayya V, et al. Educational use; follow local protocol.
Pathophysiology
Reversibility is the key concept: ECMO buys time for the lung or heart to recover. Time on injurious ventilation causes fibroproliferation and reduces reversibility; each minute of low-flow CPR worsens hypoxic brain injury.
Evidence
- ECPR trials: ARREST (2020, n=30, refractory VF OHCA) stopped early: survival to discharge 43% vs 7% with ECMO-facilitated resuscitation in an expert system. Prague OHCA (2022) — invasive strategy including ECPR: 180-day good neurological survival 31.5% vs 22.0% (p=0.09). INCEPTION (2023, multicentre): 30-day good neurological survival 20% vs 16% (not significant).
- AHA 2025 ALS: ECPR may be reasonable for selected patients with refractory arrest when delivered within a skilled, organised system of care.
- Scores: RESP (VV) and SAVE (VA) were derived from ELSO registry data; discrimination is moderate and they were not designed to select or exclude individual patients.
Controversies
- How far to transport a severely hypoxaemic patient conventionally vs waiting for retrieval.
- ECPR in Indian cities: long EMS response times and low bystander CPR rates mean most out-of-hospital arrests are not ECPR candidates; in-hospital witnessed arrests (cath lab, ICU, ED) are the realistic population.
🇮🇳 India notes
- Keep an updated list of regional ECMO centres and retrieval teams (phone numbers) on the ICU wall.
- Road transfer times in Indian cities are unpredictable; factor traffic into the choice between conventional transfer and mobile retrieval.
- Families often need time to arrange funds — start the conversation at the 'call' threshold, not the 'act' threshold.
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
Test yourself
Module quiz · score 70% to complete the module
Educational reference for trained clinicians. Check doses against your hospital protocol and the patient in front of you.