🎯

Chapter 4 — VV ECMO: Indications and Patient Selection

Part II — VV ECMO · Chapter 4

🗓️

Evidence search date: 6 September 2026. This cycle identified a new GRADE-based guideline (Australia and New Zealand, 2026) that post-dates all ELSO respiratory guidance and is compared against it below.

Clinical Question

A patient with severe respiratory failure is deteriorating despite optimal conventional management. Who should be cannulated, who should not, and how is that distinction actually made?

Why This Matters

Chapter 1 established the finding that governs this chapter: no trial has prospectively validated a selection rule, and selection remains the single largest determinant of outcome. Everything downstream — cannulation technique, flow targets, anticoagulation, weaning — is refinement. The decision made here sets the ceiling on what any of it can achieve.

The decision also has two distinct failure modes, and they are not symmetrical. Cannulating a patient who would have recovered exposes them to bleeding, stroke and weeks of ICU. Cannulating a patient who cannot recover commits a family, a team and a bed to a course with no exit — the situation Chapter 1 named as the most dangerous moment in an ECMO programme. Declining to cannulate a patient who would have survived is invisible, and therefore the error least likely to be reviewed.

Before ECMO: What Must Already Have Been Done

ECMO is considered when conventional therapy has been optimised and has failed — not when it is merely difficult. Every published indication threshold in this chapter is explicitly conditional on the following having been delivered first.

Intervention	What "optimised" means	Why it precedes ECMO
Treat the cause	Source control, appropriate antimicrobials, management of sepsis, pancreatitis, trauma or shock	ECMO does not treat the underlying disease — Chapter 1
Lung-protective ventilation	Tidal volume ~6 mL/kg predicted body weight; plateau pressure kept at or below 30–32 cmH₂O	Prevents ventilator-induced lung injury; the trial thresholds assume it
PEEP trial	PEEP at least 10 cmH₂O, titrated	Recruitment may resolve the shunt without a circuit
FiO₂ escalation	FiO₂ at or above 0.8–0.9 when the threshold is being assessed	Thresholds are defined at high FiO₂, not at any FiO₂
Neuromuscular blockade	Considered where ventilator dyssynchrony or high drive persists	Reduces VO₂ and self-inflicted lung injury — Chapter 2
Prone positioning	A genuine trial, unless contraindicated	The single most important omission in practice — see below
Adjuncts as appropriate	Recruitment manoeuvres, inhaled pulmonary vasodilators, conservative fluid strategy	May obviate cannulation entirely

🚨

Danger — ECMO is not a substitute for proning

The 2021 ELSO VV guideline is unambiguous: "ECMO should not be an alternative to proning; proning is a complement that should be performed before ECMO." The same document notes that in 2017 only 11% of US ECMO patients had received prone positioning.

A patient cannulated without a proning trial has been exposed to the risks of an extracorporeal circuit before a cheaper, safer and independently effective intervention was tried. If proning was genuinely contraindicated or genuinely failed, record which. (ELSO VV ECMO Guideline, ASAIO J 2021)

The Indication Thresholds

The three sets of criteria in common use are not identical, and the differences matter. CESAR randomised to referral; EOLIA defined trial entry; ELSO defines clinical consideration.

	CESAR (2009)	EOLIA (2018)	ELSO (2021)
Inclusion / consideration	Age 18–65; potentially reversible severe respiratory failure; Murray score ≥ 3, or pH < 7.20 after ventilator optimisation	Age ≥ 18; PaO₂/FiO₂ < 50 mmHg for > 3 h, or PaO₂/FiO₂ < 80 mmHg for > 6 h, or pH < 7.25 with PaCO₂ ≥ 60 mmHg	PaO₂/FiO₂ < 80 mmHg, or pH < 7.25 (with PaCO₂ ≥ 60 mmHg), or ventilatory support as a bridge to lung transplantation
Exclusions / contraindications	Mechanical ventilation > 7 days at peak inspiratory pressure > 30 cmH₂O and FiO₂ > 80%; CNS haemorrhage; contraindication to anticoagulation; planned limitation of life-sustaining treatment	Mechanical ventilation ≥ 7 days; pregnancy; BMI > 45 kg/m²; chronic lung disease on oxygen or NIV; HIT; malignancy with fatal prognosis within 5 years; need for VA ECMO; SAPS II > 90; non-drug-induced coma after cardiac arrest; irreversible neurological injury; no vascular access; planned limitation of treatment	Absolute: irreversible underlying condition with anticipated non-recovery. Relative: CNS haemorrhage or significant CNS injury; systemic bleeding; contraindication to anticoagulation; immunosuppression; older age; mechanical ventilation > 7 days at plateau pressure > 30 cmH₂O and FiO₂ > 90%

Conditions attached to those thresholds. The oxygenation criteria apply only despite optimisation of mechanical ventilation — FiO₂ ≥ 80%, tidal volume 6 mL/kg, a PEEP trial of ≥ 10 cmH₂O, and a trial of proning where feasible. The pH criterion applies only at a respiratory rate of at least 35 breaths per minute with settings adjusted to hold plateau pressure at or below 32 cmH₂O (EOLIA) or 30 cmH₂O (ELSO). (Comparison as tabulated in the ELSO Red Book, 6th ed.)

💡

Clinical Pearl

The EOLIA thresholds describe a patient who is already very sick: a PaO₂/FiO₂ below 80 on FiO₂ 0.8 means a PaO₂ of about 64 mmHg on nearly pure oxygen. If you find yourself asking whether a patient is sick enough, they usually are not — and the more useful question is whether the referral should be made now, before they reach the threshold. Referral and cannulation are different decisions with different time constants.

⚠️

Pitfall — EOLIA's exclusions are not a contraindication list

The middle column above is a trial eligibility list, designed to produce a clean study population. It is not a statement about who can be helped. Pregnancy, BMI above 45 and immunosuppression appear there because the investigators chose a conservative population — not because ECMO is contraindicated in those patients, and each now has its own chapter (61, 67, 66).

The ELSO Red Book makes this explicit: there are no agreed absolute contraindications except end-stage respiratory failure where transplantation will not be considered, and the EOLIA exclusions "can be taken as a conservative approach." Using them as a barrier list will deny ECMO to patients who would benefit.

Measuring Severity: the Murray Score

The Murray lung injury score predates ARDS as we now define it, but survives because CESAR used it and because it captures mechanics and radiology, not only oxygenation. Each component scores 0–4; the total is divided by the number of components used.

Component	0	1	2	3	4
PaO₂/FiO₂ (mmHg)	≥300	225–299	175–224	100–174	<100
Chest radiograph (quadrants infiltrated)	Normal	1	2	3	4
PEEP (cmH₂O)	≤5	6–8	9–11	12–14	≥15
Compliance (mL/cmH₂O)	≥80	60–79	40–59	20–39	≤19

A score of 0 indicates no lung injury; above 2.5 indicates ARDS; ≥ 3 is the conventional ECMO consideration threshold.

⚠️

Pitfall — severity is not benefit

A high Murray score says this patient is likely to die without ECMO. It says nothing about whether they will survive with it. Severity scores and survival scores answer different questions, and the ELSO guideline uses them that way: the Murray score estimates mortality without ECMO, while the RESP score predicts survival on ECMO. Both are needed, and neither identifies benefit.

Contraindications, Honestly

📊

Evidence — there is only one absolute contraindication

The 2021 ELSO VV ECMO guideline recognises a single absolute contraindication: anticipated non-recovery without a plan for viable decannulation — whether from irreversible disease, or multi-organ failure with no transplant option. Everything else is relative.

It explicitly permits initiation as a bridge to decision where there is ongoing multidisciplinary discussion and a clear time limit. On age it states there is "increasing risk of death with increasing age, but no threshold is established." (Tonna JE, Abrams D, Brodie D, et al. ASAIO J 2021)

Consideration	Absolute or relative	Rationale	Exceptions
Anticipated non-recovery with no decannulation plan	Absolute	The circuit has no destination; the run cannot end well	Bridge to decision, time-limited, with explicit multidisciplinary review
Intracranial haemorrhage or significant CNS injury	Relative	Systemic anticoagulation risks extension	Low- or no-anticoagulation VV runs are now feasible for periods; patients with intracranial bleeding have been supported to recovery. Individualised, and covered in Chapters 36 and 45
Contraindication to anticoagulation / active systemic bleeding	Relative	Circuit thrombosis versus haemorrhage	Source control may make it manageable; modern circuits tolerate reduced anticoagulation
Prolonged mechanical ventilation (> 7 days at high pressure and FiO₂)	Relative	Associated with worse pre- and post-ECMO mortality	Association, not a rule — and now contested (see Timing). The trajectory matters more than the day count
Immunosuppression	Relative	Historically associated with poor outcome	Increasingly challenged; Chapter 66
Advanced age	Relative	Continuous relationship with mortality	No threshold is established. Physiological reserve and frailty are more informative than chronological age
Obesity	Relative at most	Historically excluded (EOLIA used BMI > 45)	Contemporary data describe an "obesity paradox" with comparable or better survival; Chapter 67
Multi-organ failure	Relative	Cumulative organ failure worsens outcome	Depends entirely on reversibility of the drivers

💡

Clinical Pearl

Almost every "contraindication" in the list above is a risk modifier, not a gate. The one true gate is the reversibility-and-destination question. When declining a patient, be able to say which of those two failed — irreversible disease, or no viable destination. "Too old", "too big" and "immunosuppressed" are not answers.

Where the Guidelines Now Disagree

📊

Evidence — a GRADE-based guideline has arrived

Chapter 1 noted that, as of the 2026 ELSO roadmap, no ELSO guideline uses GRADE methodology. That gap is now partly filled from outside ELSO. Newman, Burrell, Buscher and colleagues have published evidence-based VV ECMO guidelines for Australia and New Zealand using GRADE methodology, developed by a multidisciplinary panel with a methodologist and patient representatives, in accordance with NHMRC 2016 standards. (Newman S, et al. Crit Care Resusc 2026)

Its headline positions: a conditional recommendation for VV ECMO in addition to standard care in severe hypoxaemic respiratory failure when conventional treatment fails, citing CESAR and EOLIA but noting that certainty of evidence remains limited because of uncertainty about long-term outcomes; insufficient evidence to recommend ECMO routinely in hypercapnic respiratory failure from severe asthma or COPD exacerbation, in the absence of randomised trials; typical mechanical ventilation of fewer than seven days before initiation; a conditional recommendation against routine proning during ECMO, citing limited evidence and cannula displacement risk; and minimisation of tidal volume, driving pressure and mechanical power while supported.

Question	ELSO 2021	ANZ GRADE 2026	Practical implication
Severe hypoxaemic failure	Consider at PaO₂/FiO₂ < 80 after optimisation including proning	Conditional recommendation for, certainty limited	Agreement on direction; the newer document is more explicit that certainty is low
Hypercapnic failure	Listed as an indication (pH < 7.25 with PaCO₂ ≥ 60)	Insufficient evidence to recommend routinely in asthma or COPD	A genuine divergence. Hypercapnic indications rest on physiology and observation, not trials — see Controversies
Proning before ECMO	Mandatory — "not an alternative to proning"	Patients should have optimised ventilation strategies first	No disagreement
Proning during ECMO	Not addressed in this guideline	Conditional recommendation against routine use	Distinct question from proning before ECMO. Do not conflate the two
Duration of ventilation before ECMO	> 7 days at high pressure and FiO₂ is a relative contraindication; no safe threshold defined	Typically fewer than 7 days	Broad agreement, with the evidence now weaker than the convention (see Timing)
Methodology	Narrative / consensus, ungraded	GRADE	Weight the ANZ document as a graded appraisal and ELSO as expert synthesis — Chapter 1

Predicting Survival: the RESP Score

📊

Evidence — RESP

Schmidt and colleagues derived the Respiratory ECMO Survival Prediction score from 2,355 patients in the ELSO Registry supported for severe acute respiratory failure between 2000 and 2012, of whom 1,338 (57%) were discharged alive. Discrimination in the derivation cohort was c = 0.74 (95% CI 0.72–0.76); in an external validation cohort of 140 patients it was c = 0.92 (95% CI 0.89–0.97). (Schmidt M, et al. Am J Respir Crit Care Med 2014;189:1374–1382)

Certainty: moderate for discrimination; low for individual prediction.

The score is built from variables a clinician already knows before cannulation, and their direction is instructive in itself:

Domain	Variables (direction of effect)
Patient	Age (worse with each band above 50); immunocompromised status (worse)
Timing	Duration of mechanical ventilation before ECMO — shorter is better, with the largest positive weight for under 48 hours
Diagnosis	Asthma carries by far the largest favourable weight, then aspiration, then viral or bacterial pneumonia and trauma or burn; non-respiratory diagnoses score lowest
Comorbid state	CNS dysfunction (large negative weight); acute non-pulmonary infection (negative); cardiac arrest before ECMO (negative)
Pre-ECMO treatment	Neuromuscular blockade (slightly favourable); nitric oxide and bicarbonate infusion (unfavourable — markers of severity)
Physiology	PaCO₂ ≥ 75 mmHg and peak inspiratory pressure ≥ 42 cmH₂O (both unfavourable)

⚠️

Pitfall — what a prognostic score cannot do

RESP discriminates: it separates groups likely to survive from groups likely to die. It does not identify who benefits from ECMO, because it was derived entirely in patients who received it — there is no untreated comparator anywhere in the model. A low RESP score means "this patient is likely to die"; it does not mean "ECMO will not help this patient."

Discrimination is also not calibration, and the derivation cohort is registry data from 2000–2012 with a case mix that no longer matches current practice. Use RESP to structure a conversation and to inform referral, never as a threshold for offering or withholding support.

Timing

The conventional rule — cannulate within seven days of intubation — comes from consistent observational associations between prolonged pre-ECMO ventilation and worse outcome, and it is embedded in CESAR, EOLIA and ELSO alike. The 2021 ELSO guideline states that "increasing duration of mechanical ventilation before ECMO is associated with worsening mortality," while declining to define a safe threshold.

That convention is now less secure than it looks.

📊

Evidence — the seven-day rule is contested

A 2026 review of outcome improvement in VV ECMO notes that while earlier retrospective analyses associated prolonged pre-ECMO ventilation with worse outcomes, recent Japanese cohort data found no independent association. The same review argues for individualised risk–benefit assessment rather than rigid exclusion criteria: age relates continuously to mortality without a clear threshold; obesity appears protective ("obesity paradox"), with obese patients showing comparable or superior survival; and biological frailty may be more informative than chronological age, though recent Australian and New Zealand data on frailty and outcome have been inconsistent. On centre volume, the relationship with survival "remains controversial," with some registry analyses showing benefit at higher volume and others showing none or the reverse. (Supady A. Perfusion 2026)

Certainty: low. Conflicting observational datasets, no randomised evidence on timing.

💡

Clinical Pearl

Treat day seven as a prompt to think harder, not a door closing. What is actually being measured by "days ventilated" is accumulated ventilator-induced lung injury and the probability of irreversibility — and a patient ventilated for nine days on 6 mL/kg with a driving pressure of 12 is in a different position from one ventilated for four days at a plateau of 35. Ask about the trajectory and the mechanics, not the calendar.

The Three Questions

Thresholds and scores inform the decision; they do not make it. Three questions do, and all three must be answered before cannulation.

Is the lung injury reversible — or is there a destination if it is not? Pneumonia, aspiration, trauma and status asthmaticus recover. Established fibrosis does not. If it does not recover, is this patient a transplant candidate? If neither, the absolute contraindication applies.
Has conventional therapy genuinely been optimised and failed? Specifically: lung-protective settings, a PEEP trial, neuromuscular blockade where indicated, and a proning trial or a documented reason there was none.
Will this patient survive the therapy? Not the disease — the therapy. Weeks of anticoagulation, bleeding risk, immobility, delirium and deconditioning. This is where physiological reserve, frailty and comorbidity belong, and where RESP structures the conversation.

If the answer to any of these is unclear, the correct action is often referral and discussion, not a decision. Referral costs a phone call.

Special Populations at a Glance

Historic exclusions that are now individualised rather than automatic:

Population	Historic position	Current position	Chapter
Pregnancy	EOLIA exclusion	Supported; ELSO published narrative guidance in October 2025	61
Obesity	EOLIA excluded BMI > 45	Comparable or better survival reported; access and flow are the practical problems, not candidacy	67
Immunocompromise	Relative contraindication	Increasingly offered; depends on the specific immune defect and its reversibility	66
Trauma	Bleeding risk considered prohibitive	Feasible once the bleeding source is controlled	62
Asthma / severe obstructive disease	Rare indication	Carries the most favourable diagnostic weight in RESP; but ANZ 2026 finds insufficient evidence to recommend routinely	53, 63
Bridge to lung transplantation	—	An explicit ELSO indication; requires an assessed and listed candidate	82

The Selection Algorithm

Clinical Pearls

Refer before you need to cannulate. Referral and cannulation are separate decisions. CESAR randomised patients to a referral pathway, and that pathway was what showed benefit.
Ask "what was the plateau pressure and driving pressure", not "how many days ventilated". The day count is a proxy; the mechanics are the thing.
Record the proning trial — done, contraindicated, or failed. It is the most consequential omission in the pre-ECMO pathway.
Name the destination on day zero and put it in the notes: recovery, decision, or transplant. Chapter 1's bridge taxonomy is a selection instrument, not an afterthought.
A patient who is "not sick enough yet" is a patient to discuss now, because the window between meeting the threshold and being too injured to benefit can be short.

Pitfalls

Using EOLIA's exclusion list as a contraindication list.
Treating a low RESP score as a reason to withhold ECMO, when the score contains no untreated comparator.
Cannulating without a proning trial and without recording why.
Reading the seven-day rule as a hard cut-off when the underlying evidence is observational and now inconsistent.
Declining on the basis of age, BMI or immunosuppression alone rather than on reversibility and destination.
Confusing "conditional recommendation against routine proning during ECMO" with any statement about proning before it.

Controversies

⚖️

Controversy — do the severity thresholds identify the patients who benefit?

Clinical question: Should the EOLIA thresholds be treated as the criteria for cannulation?

Evidence supporting yes: They are the only prospectively applied entry criteria from a randomised trial of early ECMO in severe ARDS, they define a population with roughly 46% mortality on conventional care, and both ELSO and the 2026 ANZ guideline are built around them.

Evidence supporting no: Thresholds define severity, not responsiveness. EOLIA's primary result did not reach conventional significance, and 28% of its control arm crossed over to rescue ECMO — so the trial compared early against late ECMO within an already-selected group. A post hoc Bayesian reinterpretation of EOLIA subsequently argued that the probability of a mortality benefit was substantially higher than the frequentist result implied, depending on the prior adopted. No analysis has identified which patients within the threshold group actually benefit.

Current consensus: Use the thresholds as the trigger for a structured decision, not as the decision itself.

Practical approach: Meeting a threshold opens the conversation about reversibility, destination and tolerance of the therapy. It does not close it.

Knowledge gap: No prospectively validated selection rule exists — the finding from Chapter 1 that this chapter cannot resolve.

See also, in this workspace: ECMO in Severe ARDS — EOLIA and the Bayesian Reinterpretation

⚖️

Controversy — is hypercapnic respiratory failure an indication?

Clinical question: Should extracorporeal support be offered for refractory hypercapnia in asthma or COPD exacerbation?

Evidence supporting yes: ELSO lists pH < 7.25 with PaCO₂ ≥ 60 mmHg as an indication. The physiology is favourable — Chapter 2 showed CO₂ clearance is achievable at low blood flow, so support can be delivered with smaller cannulae and lower risk. Asthma carries the largest favourable diagnostic weight in the RESP score, and these patients are typically young with reversible disease.

Evidence supporting no: The 2026 ANZ GRADE guideline finds insufficient evidence to recommend ECMO routinely in severe asthma or COPD exacerbation, because no randomised trials exist. The favourable RESP weight reflects survival among those selected for ECMO, not evidence that ECMO caused it.

Current consensus: Reasonable in refractory cases with a clearly reversible obstructive process; not a routine indication.

Practical approach: Distinguish the physiological case (strong) from the outcome evidence (absent), and be explicit with families about which one is being relied on.

Knowledge gap: No randomised data. Chapter 53 covers the related ECCO₂R question.

Evidence Summary

Statement	Certainty	Basis
ECMO is considered only after conventional therapy, including a proning trial, has been optimised and failed	Consensus, near-universal	ELSO 2021; ANZ 2026; ELSO Red Book algorithm
Prone positioning should precede ECMO and is not replaced by it	Consensus, strongly stated	ELSO VV guideline 2021, which also reports only 11% of US ECMO patients proned in 2017
VV ECMO in addition to standard care in severe hypoxaemic failure	Conditional recommendation, limited certainty	ANZ GRADE guideline 2026, citing CESAR and EOLIA
The only absolute contraindication is anticipated non-recovery without a viable decannulation plan	Consensus	ELSO 2021; ELSO Red Book states there are no agreed absolute contraindications beyond this
RESP discriminates survival on ECMO (c = 0.74 derivation; 0.92 in a small external cohort)	Moderate for discrimination	Schmidt 2014, 2,355 registry patients 2000–2012
RESP or any score identifies who benefits from ECMO	Not established	Derived without an untreated comparator; benefit cannot be estimated from it
Mechanical ventilation beyond 7 days before ECMO worsens outcome	Low — inconsistent	Consistent older observational associations; a recent Japanese cohort found no independent association (Supady 2026)
Obesity worsens outcome on VV ECMO	Contradicted	Contemporary data describe comparable or better survival — an "obesity paradox"
ECMO for hypercapnic failure in asthma or COPD	Insufficient evidence	ANZ GRADE 2026; no randomised trials
Higher centre volume improves survival	Uncertain	Conflicting registry and cohort analyses; no agreed volume threshold

Key Takeaways

Selection sets the ceiling on everything that follows, and no prospectively validated selection rule exists.
Every published threshold is conditional on conventional therapy having been optimised first — most importantly, a proning trial.
CESAR, EOLIA and ELSO criteria answer different questions: referral, trial entry, and clinical consideration. Do not use them interchangeably.
EOLIA's exclusions are trial eligibility criteria, not contraindications. Reading them as a barrier list denies ECMO to patients who would benefit.
There is one absolute contraindication: anticipated non-recovery with no viable decannulation plan.
Murray estimates mortality without ECMO; RESP predicts survival on it. Neither identifies benefit.
The seven-day rule is a prompt, not a gate, and the evidence behind it is now inconsistent.
Age, obesity and immunosuppression are risk modifiers, not exclusions. Obesity may even be protective.
A 2026 GRADE-based guideline now exists from Australia and New Zealand, and it disagrees with ELSO on the hypercapnic indication.
When the answer is unclear, refer and discuss. Referral is cheap; a late referral is not.

Key References

Tonna JE, Abrams D, Brodie D, et al. Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): guideline from the Extracorporeal Life Support Organization. ASAIO J. 2021;67(6):601–610. DOI: 10.1097/MAT.0000000000001432
Newman S, Burrell A, Buscher H, et al. (Nair P, senior author). Evidence-based guidelines for the use of extracorporeal membrane oxygenation in Australia and New Zealand using GRADE methodology, series part 1: venovenous ECMO indications and management. Crit Care Resusc. 2026. DOI: 10.1016/j.ccrj.2026.100163
Schmidt M, Bailey M, Sheldrake J, et al. Predicting survival after extracorporeal membrane oxygenation for severe acute respiratory failure: the Respiratory Extracorporeal Membrane Oxygenation Survival Prediction (RESP) score. Am J Respir Crit Care Med. 2014;189(11):1374–1382. DOI: 10.1164/rccm.201311-2023OC. PMID: 24693864
Supady A. Improving outcomes of patients with venovenous extracorporeal membrane oxygenation. Perfusion. 2026;41(1_suppl). DOI: 10.1177/02676591261424665
Combes A, Hajage D, Capellier G, et al. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome (EOLIA). N Engl J Med. 2018. DOI: 10.1056/NEJMoa1800385. PMID: 29791822
Peek GJ, Mugford M, Tiruvoipati R, et al. CESAR trial. Lancet. 2009. [DOI/PMID — VERIFICATION REQUIRED]
Goligher EC, Tomlinson G, Hajage D, et al. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome and posterior probability of mortality benefit in a post hoc Bayesian analysis of a randomized clinical trial. JAMA. 2018. DOI: 10.1001/jama.2018.14276. [Posterior probability values — VERIFICATION REQUIRED]
Brodie D, Peek G, MacLaren G, et al. (eds). Extracorporeal Life Support: The ELSO Red Book, 6th edition — Chapter 23, Initiating ECLS for Adult Respiratory Failure (CESAR/EOLIA/ELSO criteria comparison table; ARDS management algorithm)
Taha AR, Caridi-Scheible M, Leiendecker E, et al. ECMO: A Practical Guide to Management — Chapter 3, Indications for ECMO (Murray score; oxygenation index)
Indian Society of Critical Care Medicine. ISCCM Manual of RRT and ECMO in ICU — pre-ECMO optimisation and mortality-risk framing
Cross-references: Chapter 1 (selection as the largest determinant of outcome; the bridge taxonomy), Chapter 2 (CO₂ clearance at low flow), Chapter 3 (configuration choice), Chapter 5 (cannulation), Chapter 7 (initial management), Chapters 61, 62, 66, 67 (special populations), Chapter 82 (bridge to transplant)

🔍

Citation verification status. The ELSO 2021 VV guideline (citation, DOI, indication thresholds, the single absolute contraindication, the proning statement and the 11% figure), the ANZ 2026 GRADE guideline (authors, journal, DOI, and its stated recommendations), the RESP score (citation, DOI, PMID, cohort size, survival and both c-statistics), and Supady 2026 (citation, DOI, and its statements on timing, obesity, frailty and centre volume) were all verified against source records during the 6 September 2026 search cycle. The CESAR/EOLIA/ELSO comparison table and the Murray score were transcribed from the named project texts and cross-checked between two independent sources within the library.

Not verified this cycle: the CESAR bibliographic identifiers (carried forward from Chapter 1), and the specific posterior probability values from the Goligher Bayesian reanalysis — the paper is cited with its verified DOI, and the reanalysis is described directionally only, with no numbers quoted. The ANZ 2026 recommendations were read from a guideline summary rather than the full text; the individual GRADE certainty ratings attached to each recommendation should be confirmed against the primary publication before they are quoted.

Deliberately not quoted: the RESP score's risk classes and their predicted survival percentages, which were not obtainable from the primary record in this cycle.

⚠️

Educational use only. This chapter does not replace institutional ECMO protocols, local policy, specialist consultation, current guidelines or patient-specific clinical judgement.