Scientific job: CLAIMED. Emergency & Critical Care retains the critically ill patient, the underlying organ failure and overall intensive-care strategy. This article owns the persistent extracorporeal support movement from refractory cardiac/respiratory failure → ECMO candidacy → VV versus VA configuration → cannulation → circuit/oxygenator/flow state → anticoagulation and bleeding/thrombotic surveillance → organ recovery or bridge decision → weaning → decannulation, durable support, transplantation or other destination.

Wait, what? ECMO does not cure the heart or lungs. It buys time while something else changes.

Extracorporeal membrane oxygenation moves blood outside the body through a circuit that can add oxygen, remove carbon dioxide and, depending on configuration, support circulation. It can temporarily replace part of failing lung or heart-lung function while the underlying disease is treated, the organs recover, or another destination becomes possible.

Singapore General Hospital currently describes an intensivist-led VV ECMO service for severe lung failure and an extracorporeal life-support retrieval capability with National Heart Centre Singapore. The architecture is therefore clear: ECMO is not the diagnosis. It is a temporary replacement system.

The ECMO tube

Severe cardiac/respiratory failure → conventional support insufficient → candidacy and goals → VV/VA or other configuration → cannulation → circuit initiation → flow/gas-exchange support → anticoagulation + blood/device surveillance → underlying disease treatment → daily readiness assessment → wean trial → decannulation or bridge to another destination → rehabilitation/recovery.

1. ECMO begins only after defining what is failing

Severe respiratory failure, severe cardiac failure and combined cardiopulmonary failure are different physiological states. Before extracorporeal support is considered, the team needs to understand what conventional ventilation, circulation and disease-specific treatment are failing to achieve.

For eduKateAI, the first fields are organ failure type, reversibility, severity, time course, conventional support already tried, major contraindications and intended bridge.

2. VV and VA ECMO are different support systems

Venovenous ECMO primarily supports gas exchange in severe lung failure while the patient’s own heart continues to provide systemic circulation. Venoarterial ECMO can support both oxygenation and circulation in severe cardiac or combined failure.

For eduKateAI: ECMO ≠ one generic machine state. Configuration determines what physiological job the circuit is doing and what complications become important.

3. Candidacy is a bridge question

ECMO is most useful when there is a plausible destination: recovery of lungs or heart, treatment of a reversible cause, transplantation, durable mechanical support or another defined next state. If there is no realistic bridge, the burden and risk of extracorporeal support can outweigh benefit.

The ethical and clinical question is therefore not simply “Can we put the patient on ECMO?” but what are we bridging toward?

4. Cannulation creates a new vascular-device state

Large cannulae are placed into selected veins and/or arteries to move blood between the patient and the extracorporeal circuit. Cannulation site and configuration affect flow, bleeding, limb perfusion, recirculation and other risks.

The Medical Devices Web owns persistent device identity and safety lineage generally. ECMO owns the active cannula + circuit + patient physiology state during extracorporeal support.

5. The circuit becomes part of the patient’s physiology

The extracorporeal circuit includes tubing, pump and membrane oxygenator. Blood flow, gas flow and circuit pressures interact with the patient’s own cardiac output, ventilation, haemoglobin and metabolic demand.

For eduKateAI, the state should preserve configuration, cannulae, circuit identity, flow, oxygenator state, key gas-exchange parameters and whether the circuit is currently meeting the intended support goal.

6. ECMO does not remove the need for mechanical ventilation decisions

In VV ECMO, mechanical ventilation often continues, but ECMO can allow a more lung-protective strategy while extracorporeal gas exchange carries more of the burden. Respiratory Medicine and Critical Care own the ventilator and lung-disease strategy; ECMO owns how extracorporeal support changes the feasible ventilation state.

For eduKateAI: adequate blood oxygenation on ECMO ≠ lungs recovered.

7. Anticoagulation is a balance, not an on/off switch

Blood contacting artificial surfaces can clot, yet anticoagulation can increase bleeding. The patient may also have surgery, trauma, thrombocytopenia or other reasons for altered coagulation. The system therefore continually balances circuit thrombosis against patient bleeding risk.

Patient Blood Management, Haematology and Laboratory Medicine contribute measurements and blood-product decisions. ECMO owns how those findings interact with the extracorporeal circuit.

8. Bleeding and thrombosis can occur at the same time

ECMO patients can experience cannulation-site bleeding, intracranial or other haemorrhage, thrombosis within the circuit, embolic events or limb ischaemia depending on configuration and clinical state. These are not contradictory events; critical illness and extracorporeal circulation disturb several haemostatic systems simultaneously.

The anti-collapse rule is: bleeding present ≠ clot risk absent.

9. Infection risk follows invasive support

Large vascular cannulae, prolonged intensive care and other devices can create infection risk. Antimicrobial Stewardship & Infection Prevention owns prevention and responsible antimicrobial use; ECMO owns the circuit/cannulation state that modifies that risk.

10. The underlying disease still has to be treated

Pneumonia, ARDS, myocarditis, cardiogenic shock, massive pulmonary embolism and other conditions can lead to extracorporeal support. ECMO does not replace treatment of the cause. Infectious Disease, Cardiology, Respiratory Medicine, Surgery or other specialty owners continue their disease-specific work in parallel.

For eduKateAI: supported physiology ≠ disease controlled.

11. Daily assessment asks whether the patient still needs the circuit

As the lungs or heart improve—or fail to improve—the balance of benefit and risk changes. Teams repeatedly assess gas exchange, haemodynamics, imaging, organ function, ventilator requirements, biomarkers and clinical trajectory.

The ECMO runtime is therefore support → observe → reduce dependence where safe → reassess.

12. Weaning is a test of recovered native function

Before decannulation, support can be reduced in controlled steps to assess whether the patient’s own lungs and/or heart can sustain adequate physiology. A successful weaning trial is not merely turning down a pump; it is evidence that native function can carry the necessary load.

For eduKateAI: stable on high ECMO support ≠ ready for decannulation.

13. Not every ECMO destination is recovery

Some patients recover and are decannulated. Others may require transplantation, ventricular-assist devices or another form of durable support. In some cases, ongoing extracorporeal support no longer offers a realistic benefit and goals-of-care decisions become central.

Transplantation, Cardiac Surgery, Palliative Care and Clinical Ethics may become the next owners. ECMO’s job is to make the bridge state explicit.

14. Retrieval makes geography part of the support state

SGH describes an extracorporeal life-support retrieval capability able to initiate support at referring hospitals and transport selected patients. That demonstrates a HealthOS truth: the usefulness of an advanced therapy depends on whether specialist teams, equipment and transfer pathways can reach the patient in time.

15. Rehabilitation begins while the patient is still critically ill

Prolonged critical illness can produce profound weakness, cognitive change, swallowing problems and psychological effects. When safe, mobilisation and rehabilitation planning can begin before extracorporeal support ends.

The final human receipt is not “off ECMO”. It is native organ support sufficient + complications addressed + function and recovery route established.

Characteristic failure modes

• ECMO = treatment of disease error: temporary support substitutes for treating the cause.
• VV = VA error: respiratory and circulatory support configurations collapse together.
• Stable = recovered error: good physiology on full extracorporeal support is mistaken for native organ recovery.
• Bleeding = no clot risk error: competing haemostatic risks are simplified.
• Circuit-only error: patient physiology and human goals disappear behind pump numbers.
• No-destination error: support continues without an explicit bridge target.
• Decannulated = recovered error: rehabilitation and post-critical-care burden disappear from outcomes.

The eduKateAI routing contract

• Canonical public owner: ECMO & Extracorporeal Life Support Web.
• Input state: severe respiratory/cardiac failure being assessed for or managed with extracorporeal support.
• Primary job: preserve support configuration, cannulation, circuit state, anticoagulation, complications, native-organ recovery and bridge destination.
• Do not collapse: ECMO ≠ disease treatment; VV ≠ VA; stable on support ≠ recovered; circuit flow ≠ human outcome.
• Handoffs: Emergency/Critical Care, Respiratory Medicine, Cardiovascular Medicine, Cardiac Surgery, Haematology/PBM, Laboratory Medicine, Infection Prevention, Medical Devices, Transplantation, Rehabilitation, Clinical Ethics and Palliative Care.
• Return receipt: candidacy accepted/rejected, configuration active, circuit stable/complicated, native function improving/not improving, wean passed/failed, decannulated/bridged, post-ECMO receiver confirmed.

Authoritative routes

• Singapore General Hospital — MICU and VV ECMO Service
• Singapore General Hospital — Respiratory & Critical Care Medicine
• Extracorporeal Life Support Organization

Educational boundary: this article explains ECMO and extracorporeal-life-support information architecture. It does not determine ECMO candidacy, select VV/VA configuration, direct anticoagulation, interpret circuit parameters or make weaning/decannulation decisions for an individual patient.