📋

Chapter 8 — Daily VV ECMO Management

🧭

Chapter question: What do I check, target and titrate every day?

Chapter 7 covered the first hours. This chapter covers every day after that: the structure of the ECMO ward round, what the circuit is telling you, what the patient's targets should be, and the daily discipline of asking whether support can come down.

What this chapter covers — and what it does not

This chapter owns	Deferred to
The daily round as a repeatable structure; routine circuit surveillance and its thresholds; daily patient targets; trend-based monitoring; daily down-titration; fluid balance and de-resuscitation; sedation holds and neurological assessment; mobilisation; infection surveillance; the recurring destination review	Acute troubleshooting of hypoxaemia, low flow and alarms → Chapter 9 Weaning and decannulation → Chapter 10 Oxygenator failure and circuit exchange technique → Chapter 32 Haemolysis → Chapter 37 Anticoagulation monitoring in depth → Part VII Ventilation during ECMO in depth → Chapter 53 The printable daily checklist and round template → Chapters 87–88

8.1 Why the ECMO round is a different round

An ICU ward round asks one question of one system: is the patient getting better, and what is stopping that? An ECMO round asks it of two systems that are physically joined and that fail in different ways.

The patient deteriorates over hours to days, announces itself through symptoms and physiology, and is monitored continuously. The circuit deteriorates over days to weeks, is largely silent until it is not, and is monitored only if someone deliberately looks. A membrane lung that will fail on day 9 is usually already declaring itself on day 6 — in a slowly rising transmembrane pressure, a slowly falling post-oxygenator oxygen tension, a rising lactate dehydrogenase — to a team that is measuring those things and plotting them.

🫀

Physiology — why the circuit fails quietly

The membrane lung starts with reserve: an adult oxygenator's rated flow (Chapter 6) normally exceeds the blood flow the patient actually requires, so the device begins with gas-exchange capacity in hand. Fibrin deposition and clot can therefore consume a substantial fraction of the exchange surface before arterial saturation moves at all. The first thing that changes is not the patient's numbers but the circuit's numbers: resistance rises before transfer falls, and transfer falls before the patient desaturates.

This is why circuit surveillance is trend-based, not threshold-based, and why a single normal set of values is close to meaningless.

The practical consequence is the organising principle of this chapter:

The ECMO round is two rounds done in a fixed order — the circuit round, then the patient round — followed by one question asked out loud: can support come down today, and if not, why not?

8.2 The circuit round

Do this first, at the bedside, with your hands on the circuit and your eyes on the trend chart — not from the notes.

8.2.1 Flow, speed and the relationship between them

Record blood flow (L/min) and pump speed (RPM) together, every day, and compare them with yesterday's pair. Flow alone is uninterpretable; the ratio is the information.

Same RPM, falling flow → rising resistance somewhere in the circuit, or falling preload to the drainage cannula. This is the single most useful daily observation in ECMO.
Rising RPM needed to hold the same flow → the same finding, expressed the other way.
Flow stable, RPM stable, patient worsening → the circuit is fine; the problem is the patient or the membrane's transfer, not its resistance.

⚠️

Danger — very low flows are not "gentle"

Blood flow through any limb of the circuit should not be allowed to fall below approximately 1.5 L/min, and several authorities advocate keeping it at or above 2–2.5 L/min. Below this, stagnation promotes thrombus formation in the circuit and oxygenator. If the patient's requirement has genuinely fallen this far, that is an argument for a weaning trial and decannulation (Chapter 10) — not for running a circuit indefinitely at a flow that will clot it.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.2.2 Circuit pressures

Most modern adult circuits allow pressure measurement at three points. Naming conventions vary between units; what matters is that you know which port is which on your circuit.

Site	Normally	A change means
Pre-pump / access line (often "P3")	Negative	More negative → drainage is being obstructed: hypovolaemia, cannula malposition, kinking, tamponade, raised intra-abdominal pressure, or the patient coughing/straining
Post-pump / pre-oxygenator ("P1")	Highest pressure in the circuit	Isolated rise → resistance within or after the oxygenator
Post-oxygenator / return ("P2")	Lower than P1	Rise in P1 and P2 together → obstruction downstream of the oxygenator: kinked return limb, return-cannula malposition or clot

⚠️

Danger — the negative-pressure threshold

Pre-pump (access line) pressure is ideally kept around −60 mmHg. Pressures more negative than −100 mmHg increase blood trauma and free haemoglobin release, and are a mechanical cause of haemolysis that no amount of transfusion will fix.

If you are running more negative than this, the answer is to address the cause of poor drainage (volume, position, cannula), not to accept it as the price of flow.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

🕳️

Pitfall — the pressure port itself is a hazard

The pre-pump limb is under negative pressure, so any breach at the measuring port can entrain air. Some centres deliberately omit access-line pressure monitoring for this reason. Know which convention your unit follows; if the port exists, it must be luer-locked, checked daily and never left open.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.2.3 Transmembrane pressure (ΔP) — the single best early-warning number

\Delta P = P_{\text{pre-oxygenator}} - P_{\text{post-oxygenator}}

ΔP is the pressure cost of pushing blood through the fibre bundle. It is a resistance measurement, and it rises as the bundle is progressively occupied by fibrin and thrombus.

Record it daily, at a documented blood flow. ΔP is flow-dependent, so a value without its flow is uninterpretable. A ΔP of 40 mmHg at 5 L/min and a ΔP of 40 mmHg at 2.5 L/min describe very different membranes.
Plot it. The absolute number varies by oxygenator model, so the trend within one circuit is what carries the information.
A rising ΔP predicts oxygenator failure before gas exchange deteriorates. Its purpose is to convert an emergency circuit change into a planned one.

💡

Clinical pearl — plan the change, do not be surprised by it

The reason to trend ΔP alongside post-oxygenator gases is precisely so that oxygenator replacement becomes an elective, daytime, fully-staffed, blood-available procedure rather than a 3 a.m. emergency in a hypoxaemic patient. That single organisational consequence is the main clinical value of daily circuit surveillance.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.2.4 Gas transfer across the membrane

Pre- and post-oxygenator blood gases assess whether the membrane is still doing its job, as distinct from whether it is still easy to push blood through. The two decline separately, and both should be tracked.

🕳️

Pitfall — circuit gases are only comparable if the sweep is standardised

Post-oxygenator gases must be drawn with the sweep gas set at 100% oxygen, otherwise today's value cannot be compared with yesterday's. A "falling post-membrane PO₂" in a patient whose sweep gas oxygen concentration (FdO₂) was weaned yesterday is an artefact, not a failing membrane.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

Two distinct failure modes are described:

Shunt effect — blood passes the bundle without contacting gas-exchange surface, so oxygen transfer falls.
Dead-space effect — clot around fibres leaves ventilated surface that no longer sees blood, so carbon dioxide clearance falls disproportionately.

A membrane that is losing CO₂ clearance (rising sweep requirement for the same PaCO₂) while oxygenation is still adequate is a common and under-recognised early pattern.

8.2.5 Haemolysis markers

Haemolysis is the circuit injuring the patient, and it is both a marker of circuit stress and a cause of harm (renal injury, and possibly the unexplained hypertension attributed to nitric-oxide scavenging by free haemoglobin).

Monitor with plasma-free haemoglobin, lactate dehydrogenase, haptoglobin and a blood film. Interpret them against the mechanical findings above.

If haemolysis is rising, look for	Because
Excessively negative access pressure	Free haemoglobin rises as suction pressure becomes more negative
High blood flow relative to body size	Free haemoglobin falls when blood flow is reduced; flow rates greater than approximately 147 mL/kg/min have been associated with haemolysis (Butt 2024)
Clot burden in the oxygenator or pump head	Oxygenator replacement can resolve otherwise unexplained haemolysis
Cannula position (tip against a wall, partial obstruction)	Local high shear

📊

Evidence — what the numbers do and do not tell you

A 2024 review of oxygenator monitoring and change-out (Butt et al., Journal of Extra-Corporeal Technology) reports that plasma-free haemoglobin greater than 50 mg/dL after 24 hours has been associated with mortality, that rising D-dimer over the days preceding an exchange correlates with clot burden, and that fibrinogen outside a range of roughly 2–6 g/L carries thrombotic (high) and bleeding (low) associations.

Its central and most honest conclusion is that these are not individually indications to change a circuit: "While individually they are not a direct indication to change the ECMO circuit, but collectively they can provide a holistic picture." The same review provides no validated transmembrane-pressure threshold, no oxygen-transfer threshold, and no standardised change-out criteria — because none exist.

Certainty: low. These are associations from observational series, not decision thresholds.

🕳️

Pitfall — free haemoglobin breaks the laboratory

High plasma-free haemoglobin interferes with optical laboratory assays. A grossly haemolysed sample can therefore produce spurious results across a whole biochemistry panel — including the ones you are about to act on.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.2.6 Look at it and listen to it

The whole system should be inspected at least once daily, deliberately, by someone who knows what a new circuit looks like.

Oxygenator — inspect the fibre bundle and both headers for clot and for fibrin deposition. Condensed water in the gas phase can be cleared with a brief maximal sweep flush of less than one second; do not confuse condensation with plasma leak.
Pump head — listen. A clicking or frictional sound suggests fibrin on the impeller.
Tubing — the whole length, including the parts under the patient and behind the pump, for clot, kinks, and for the connectors.
Cannula sites — dressing, bleeding, position markings, and the depth marking against the recorded insertion depth. A cannula that has migrated 2 cm is a recirculation problem waiting to happen.

⚠️

Danger — do not use alcohol-based solutions on the tubing

Alcohol-based agents cause cracking of polyvinyl chloride circuit tubing. This applies to hand-rub carried on gloves, not only to deliberate cleaning.

Source: ISCCM Manual of RRT and ECMO in ICU.

8.2.7 The heat exchanger and the water bath

💡

Clinical pearl — the temperature problem that is not the patient

Heat-exchanger failure occurs in approximately 2% of ECMO patients and is frequently misattributed to the patient's own condition — an unexplained fall in temperature blamed on sepsis or on sedation. Check the water-bath temperature and the exchanger daily as a distinct item. The water bath can also harbour organisms and is a recognised infection source.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.2.8 Daily safety-equipment check

These are the items that are only ever needed in an emergency, and are therefore the items that are found missing during one.

Battery back-up — confirm by disconnecting external mains power briefly and observing the console switch to battery.
Hand crank — present on the cart, and every member of staff on the shift knows where it is and how to use it.
Clamps — the correct number, on the cart.
Flow probe — coupling gel applied, to prevent false low-flow alarms.
Emergency circuit / primed backup — per unit policy, and its location known.

Source: ISCCM Manual of RRT and ECMO in ICU.

8.2.9 The daily circuit dataset

Parameter	Frequency	What a change means
Blood flow and pump RPM (as a pair)	Hourly; reviewed daily	Falling flow at fixed RPM = rising resistance or falling preload
Sweep gas flow and FdO₂	Hourly; reviewed daily	Rising sweep for the same PaCO₂ = falling CO₂ clearance
Pre- and post-oxygenator pressure, and ΔP	Daily minimum, at a documented flow	Rising ΔP = fibrin/thrombus load in the bundle
Access (pre-pump) pressure	Continuous where used	More negative than −100 mmHg = haemolysis risk
Post-oxygenator blood gas (sweep at 100% O₂)	Daily, or when transfer is questioned	Falling post-membrane PO₂ = failing gas transfer
Plasma-free haemoglobin, LDH, haptoglobin, film	Daily to alternate-daily per unit policy	Rising = circuit-related red-cell trauma
Platelets, fibrinogen, D-dimer	Daily	Consumption pattern suggests circuit clot burden
Point-of-care anticoagulation test (per unit protocol)	Per protocol, at least daily reviewed	See Part VII
Water-bath temperature and heat exchanger	Daily	Failure in ~2%; mimics patient pathology
Visual and auditory inspection of circuit, oxygenator, pump head, cannula sites	Daily minimum	Clot, kink, migration, bleeding

Composite of the monitoring tables in ECMO in the Adult Patient (Core Critical Care, 2017) and the ISCCM Manual of RRT and ECMO in ICU.

8.3 The patient round

The circuit round tells you whether your support is intact. The patient round tells you whether the disease is resolving, and whether the support is doing harm.

8.3.1 Oxygenation — accept the number the patient can live with

Established VV ECMO does not aim for normal arterial saturation. Recirculation, native shunt and the mixing of oxygenated return with desaturated venous blood mean that arterial saturations in the high 80s to low 90s are ordinary and, in a patient with adequate cardiac output, haemoglobin and lactate clearance, are not a problem to be corrected.

The judgement is not "what is the SaO₂?" but "is oxygen delivery meeting oxygen consumption?" — assessed from lactate, central or mixed venous saturation, capillary refill, mentation where assessable, and urine output. A patient with SaO₂ 86%, ScvO₂ 72%, lactate 1.1 mmol/L and good peripheries is adequately oxygenated. A patient with SaO₂ 94% and a rising lactate is not.

🕳️

Pitfall — chasing saturation with flow

The reflex response to a saturation of 87% is to increase blood flow. Before doing so, ask what changed: fever, agitation, shivering and sepsis raise oxygen consumption; a falling haemoglobin lowers delivery; recirculation rises with hypovolaemia and cannula migration. Treating the cause (cooling, sedation, transfusion, position, volume) fixes the physiology; raising flow into a partially clotted circuit at high negative pressure adds haemolysis to the problem.

The systematic approach to hypoxaemia on VV ECMO is Chapter 9.

8.3.2 Carbon dioxide and pH

Carbon dioxide is controlled almost entirely by sweep gas flow, and this is the most powerful and most abused control on the console.

Titrate sweep to the pH and PaCO₂ you intend, not to normality for its own sake. In a patient with pre-existing hypercapnia, rapid normalisation risks a large fall in cerebral blood flow.
Change sweep in small steps and re-check. In an established patient a sudden large increase in sweep can drop PaCO₂ precipitously.
A rising sweep requirement to hold the same PaCO₂ is a circuit finding, not a patient finding — go back to §8.2.4.

8.3.3 The ventilator — protect the lung, and keep protecting it

📊

Evidence — ELSO ventilator settings during ECLS

As summarised by Szuldrzyński and colleagues (Perfusion, 2024), ELSO recommends during ECLS: plateau pressure no greater than 30 cmH₂O (with less than 25 cmH₂O recommended), PEEP at least 10 cmH₂O, respiratory rate 4–15/min, and the minimum FiO₂ compatible with adequate oxygenation.

The operating principle that follows is explicit in the same source: "increase ECLS support rather than ventilator settings if gas exchange deteriorates."

Certainty: low to moderate — guideline recommendations, largely expert consensus rather than randomised evidence. Ventilation during ECMO is covered in depth in Chapter 53.

The daily ventilator question on ECMO therefore inverts the usual one. It is not "can I improve the gas exchange with the ventilator?" — the membrane does that. It is:

"Are these settings still the least injurious settings that keep the lung open?"

Record and trend plateau pressure, driving pressure, PEEP, respiratory rate and FiO₂ as a set. Driving pressure is the variable most consistently associated with outcome in ARDS, and respiratory rate contributes to mechanical power; a lung protected by low driving pressure but ventilated at 25 breaths per minute is not fully protected.

8.3.4 Haemodynamics, fluid balance and de-resuscitation

Most VV ECMO patients arrive volume-overloaded from their pre-ECMO resuscitation. The circuit then makes overload visible in a way it was not before: hypovolaemia and overload both present as drainage problems, and the access line tells you which.

Chattering / line shaking with very negative access pressure → the drainage cannula is intermittently obstructed. This is most often relative hypovolaemia, but is also caused by cannula migration, raised intra-abdominal pressure and coughing.
Rising body weight, positive cumulative balance, worsening oedema, rising oxygen requirement on stable circuit parameters → overload.

Once shock has resolved and the circuit is providing stable support, the daily goal in most patients becomes negative fluid balance — by diuresis where renal function permits, and by ultrafiltration where it does not. Set a target for the day and review yesterday's achievement against it; a cumulative balance that no one owns will not become negative by itself.

💡

Clinical pearl — weigh the patient

Cumulative fluid balance charts on ECMO drift, because circuit priming, flushes, blood products, dialysis effluent and continuous infusions are recorded inconsistently. Where a bed scale exists, daily weight is the more honest number.

8.3.5 Renal function and renal replacement

Acute kidney injury is common on ECMO, and roughly half of patients require renal replacement therapy. Contributing mechanisms include the underlying critical illness, inflammation from the blood–biomaterial interface, haemolysis and free-haemoglobin load, transfusion, and (in VA support) non-pulsatile flow.

Daily: review creatinine and urine output trend, cumulative balance, and whether the reason for a rising creatinine is renal, haemodynamic, or a laboratory artefact from haemolysis (§8.2.5). CRRT delivered in series with the ECMO circuit is covered in Chapter 55.

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

8.3.6 Sedation, the daily neurological assessment, and mobilisation

💡

Clinical pearl — sedation holds are easier on ECMO, not harder

This is counter-intuitive and it is the single most under-used advantage of extracorporeal support. In a conventionally ventilated ARDS patient, lightening sedation risks respiratory drive that the lung cannot tolerate. On ECMO, respiratory drive is controlled by the sweep gas: raising the sweep lowers PaCO₂ and suppresses drive, so sedation can be lightened while drive is managed independently.

The source is blunt about why this matters — interrupting sedation allows neurological assessment, which is described as "a key step in ensuring that ECMO is not futile."

Source: ECMO in the Adult Patient (Core Critical Care, 2017).

Every day, therefore, ask three questions in order:

Is there a reason this patient must stay deeply sedated today? Acceptable reasons exist — refractory hypoxaemia requiring paralysis, agitation endangering cannulae, raised intracranial pressure. "They are on ECMO" is not one of them.
Has a neurological examination been performed and documented? Intracranial haemorrhage and ischaemic stroke are among the most consequential complications of ECMO and are frequently discovered late in a continuously sedated patient. A daily documented examination is the surveillance tool.
Can this patient sit, stand, or walk today? Awake, mobilising ECMO is feasible in selected patients, and mobilisation is easier to achieve in a patient whose sedation has never been allowed to accumulate than in one being woken on day 14.

⚠️

Danger — the sedation hold is also a cannula risk

Waking a patient with femoral cannulae introduces a real risk of agitation, cannula migration and decannulation. The hold must be planned, staffed, and abandoned early if the patient becomes agitated — not attempted by a single nurse during a busy shift. Cannula position markings should be checked before and after.

8.3.7 Infection surveillance

ECMO patients have multiple large-bore intravascular devices, are frequently immunoparalysed, and often lose the usual signs. Two specific points:

Temperature is unreliable. The heat exchanger holds the patient at whatever temperature the water bath is set to. Fever is therefore masked, and a patient can be profoundly septic and normothermic. Trend the water-bath temperature setting — if it is being lowered progressively to hold the patient's temperature, the patient is generating heat.
The water bath itself can be a source of organisms (§8.2.7).

Daily: review cannula sites, culture results, antimicrobial day-count and de-escalation opportunity, and inflammatory trend interpreted with the caveat that the circuit itself generates an inflammatory response.

8.3.8 Nutrition, pressure areas and the ordinary ICU round

Nothing about ECMO removes the standard daily ICU items, and being on ECMO makes several of them harder and more important — enteral nutrition (usually feasible; ECMO is not in itself a contraindication), glycaemic control, pressure-area care in a patient who cannot be turned freely, eye and mouth care, bowel care, thromboprophylaxis decisions in the context of systemic anticoagulation, and physiotherapy.

🕳️

Pitfall — the circuit eats the round

The commonest failure mode of an ECMO round is that the team spends twenty minutes on ΔP and sweep and then leaves without discussing the pressure sore, the feed that has been held for four days, or the family who have not been updated. The circuit is the interesting part; it is rarely the part that decides the outcome.

8.3.9 Transfusion and anticoagulation review

Anticoagulation is covered in Part VII; the daily-round obligations are narrower:

Review the anticoagulation target and whether it is being met, using your unit's chosen assay consistently.
Review the bleeding and thrombotic events of the last 24 hours, at cannula sites and elsewhere.
Review the transfusion threshold and whether it is being applied. Transfusion contributes to inflammation, to renal injury and to circuit load; a haemoglobin target on ECMO should be a decision, not a habit.

📊

Evidence — how common these events are, and how thin the evidence is

In a review by Levy, Staudinger and Steiner (Intensive Care Medicine, 2022), 40.2% of 7,579 venovenous ECMO patients experienced a bleeding or thrombotic event. Commonly cited targets include ACT 180–200 s, aPTT 40–50 s initially and 60–80 s thereafter, and anti-Xa 0.3–0.7 IU/mL.

The same review states plainly that no evidence-based consensus exists for anticoagulation management in ECMO. These are conventions, not thresholds derived from outcome trials.

Certainty: very low for the specific targets; moderate for the event frequency (large registry-based denominator).

8.4 The daily target set

Targets must be written down, because a target that lives only in the consultant's head changes with the consultant.

Domain	What to set	Why it is set as a range, not a number
Oxygenation	A minimum acceptable SaO₂ and a delivery marker (lactate, ScvO₂)	Adequacy is a delivery question, not a saturation question
PaCO₂ / pH	A target range, with an explicit statement if permissive hypercapnia is intended	Rate of change matters as much as the value
Ventilator	Plateau, driving pressure, PEEP, rate, FiO₂ ceiling	Lung rest is defined by the whole set, not by tidal volume alone
ECMO blood flow	A target range, with the minimum acceptable flow stated	Both extremes are harmful (§8.2.1)
Fluid balance	A numeric 24-hour target	Otherwise no one owns it
Sedation	A target sedation score and a stated plan for the daily hold	Drive is controllable by sweep — depth should be a choice
Anticoagulation	Assay, target range, and today's decision if bleeding	Unit-specific; consistency matters more than the choice
Transfusion	Haemoglobin and platelet thresholds	Must be a decision, not a reflex
Mobilisation	Today's level (passive / sitting / standing / walking)	Deconditioning starts on day 1

8.5 The down-titration discipline

Support that is never reduced is support that is never tested. The default trajectory of an ECMO patient who is improving should be less support today than yesterday, and the round should force that question rather than leave it to be noticed.

Ask, in this order, every day:

Can FdO₂ (the oxygen concentration in the sweep gas) come down? This costs nothing and is usually the first thing that can be reduced. Note that once FdO₂ is below 100%, the post-oxygenator gas is no longer directly comparable with previous ones unless a standardised sample is taken (§8.2.4).
Can sweep gas flow come down? Falling sweep requirement is the most sensitive sign that the native lung is recovering CO₂ clearance — usually the first function to return.
Can blood flow come down? Only after gas exchange has demonstrably improved, and never below the minimum flow at which the circuit is safe (§8.2.1).
Is the patient now a candidate for a formal weaning trial? If sweep can be reduced substantially with acceptable gas exchange, go to Chapter 10.
Can the ventilator settings come down — FiO₂ first, then the rest, keeping the lung-rest principle of §8.3.3 intact?
Can sedation come down? (§8.3.6)
Can any line, drug or device come out today?

💡

Clinical pearl — CO₂ recovers before oxygenation

Because carbon dioxide is far more diffusible than oxygen, the native lung typically regains the ability to clear CO₂ well before it regains the ability to oxygenate. Practically: falling sweep requirement is the earliest bedside marker of lung recovery, and it usually appears days before the patient could tolerate a trial off support. Watching the sweep trend is a better recovery monitor than watching the chest radiograph.

8.6 The destination review

Every ECMO round should end by restating, out loud, the answer to a question asked at cannulation: what is this patient's bridge to?

Bridge to recovery — is the underlying lung injury actually reversible, and is there evidence today that it is reversing? Compliance, sweep requirement and imaging over days, not hours.
Bridge to transplantation — is the patient still a candidate, and has the assessment progressed today?
Bridge to decision — what decision, by when, and what information is still outstanding?

When the answer becomes "none of these", that is a finding, and it belongs in the round rather than being deferred to a crisis. Ethical and end-of-life management of the patient who cannot be weaned and cannot be transplanted is Part XI.

⚠️

Danger — the bridge to nowhere is reached gradually

No single day is the day a patient becomes non-recoverable. The transition happens across a fortnight of individually reasonable daily decisions, each of which continues support. The only defence is a destination question asked every day, documented, with the date on which the current answer was last genuinely reviewed. A neurological examination that has not been possible for ten days (§8.3.6) is not a neutral finding in that review.

8.7 Controversies

⚖️

Controversy 1 — Should oxygenator change-out be triggered by protocol or by judgement?

The question. Should units define numeric change-out criteria (a ΔP threshold, a plasma-free haemoglobin level, a post-membrane PO₂ floor) and act on them, or make each decision on the whole clinical picture?

The case for protocolised triggers. Explicit criteria convert emergency exchanges into planned ones, reduce variation between operators and shifts, and give junior staff a defensible action. Exchanging a membrane electively is a low-risk procedure in experienced hands; exchanging one during acute hypoxaemia is not.

The case against. No validated threshold exists for any of these variables. Absolute values are oxygenator-model dependent and flow dependent, so a threshold that is appropriate for one device is arbitrary for another. Protocolised triggers would cause unnecessary exchanges — each of which carries transfusion, circulatory interruption and cost.

What the evidence actually shows. The most recent dedicated review (Butt et al., 2024) identified the relevant monitoring parameters — fibrinogen, pre- and post-membrane gases, plasma-free haemoglobin, D-dimer, platelets, flows, pressures — but explicitly declined to convert them into criteria, concluding that individually they do not indicate change-out and that only collectively do they provide a picture. Certainty: very low. There is no comparative study of a protocolised versus a judgement-based change-out strategy.

Where practice actually sits. Most experienced units run a hybrid: mandatory daily measurement and plotting of a defined parameter set (which is not controversial), with the decision made by an experienced clinician on the combination and, critically, on the trajectory.

What would resolve it. A multicentre registry linking serially recorded ΔP, gas-transfer and haemolysis trajectories to hard oxygenator-failure events — sufficient to derive and then prospectively validate a trigger. Nothing of this kind currently exists.

⚖️

Controversy 2 — How low an arterial saturation is acceptable on VV ECMO?

The question. Established VV ECMO frequently delivers SaO₂ in the mid-80s. Should this be accepted, or should flow, haemoglobin and configuration be escalated to achieve a higher number?

The case for permissive hypoxaemia. Oxygen delivery, not saturation, determines tissue oxygenation. Raising flow to correct saturation demands more negative drainage pressures and higher shear, causing haemolysis; raising haemoglobin means transfusion, with its own inflammatory, renal and immunological costs. A patient with normal lactate and normal venous saturation is, by definition, not oxygen-limited.

The case for a higher target. The tolerance of the brain to sustained modest hypoxaemia over days to weeks is not well characterised, and neurological outcome is difficult to assess in a sedated patient. Chronic hypoxaemia may contribute to the cognitive impairment reported after ECMO. Absence of a raised lactate is a crude assurance.

What the evidence actually shows. There is no randomised comparison of oxygenation targets in VV ECMO, and no observational dataset adequate to separate the effect of a saturation target from the severity of illness that produced it. Certainty: very low.

Where practice actually sits. Most units accept SaO₂ in the high 80s when delivery markers are satisfactory, escalate investigation (rather than flow) when saturation falls further, and are more cautious in patients with cerebrovascular disease or raised intracranial pressure.

What would resolve it. A randomised trial of conservative versus liberal oxygenation targets in VV ECMO with neurocognitive outcomes at 6–12 months. This does not exist and there is no signal that it is planned. [VERIFICATION REQUIRED] — trial registries were not searched for this chapter.

8.8 The five errors that recur on ECMO rounds

Error	Correction
Recording circuit values without plotting them	A number without its trend cannot predict membrane failure. Plot ΔP and post-membrane gases
Comparing post-oxygenator gases drawn at different FdO₂	Standardise: sweep at 100% oxygen for the comparison sample
Recording flow without RPM	The pair is the finding; either alone is not
Treating a low saturation by raising flow	Establish whether the problem is consumption, delivery, recirculation or the membrane (Chapter 9) before changing flow
Never lightening sedation "because they are on ECMO"	Sweep controls respiratory drive; a daily documented neurological examination is a core safety measure, not a luxury

8.9 Key points

The ECMO round is two rounds in a fixed order — circuit, then patient — ending with an explicit down-titration and destination question.
Resistance rises before transfer falls, and transfer falls before the patient desaturates. Circuit surveillance exists to exploit that sequence and to make oxygenator change elective rather than emergent.
ΔP is only interpretable with its blood flow, and only useful as a trend. No validated absolute threshold exists.
Post-oxygenator gases are comparable only if drawn with sweep at 100% oxygen.
Access-line pressure more negative than about −100 mmHg causes haemolysis; blood flow below about 1.5 L/min causes stagnation and thrombosis. Both extremes are unsafe.
Adequacy of oxygenation is a delivery question, not a saturation question. SaO₂ in the high 80s with normal lactate and venous saturation is acceptable.
On ECMO, the ventilator is not the tool for gas exchange — the guidance is to increase extracorporeal support rather than ventilator settings when gas exchange deteriorates.
Sedation holds are easier on ECMO than off it, because sweep gas controls respiratory drive. A daily neurological examination is the principal means of detecting the complications that most often change the outcome.
Temperature is masked by the heat exchanger; heat-exchanger failure occurs in about 2% and mimics patient pathology.
Falling sweep requirement is the earliest marker of native lung recovery — it precedes improvement in oxygenation.
The destination question — bridge to what? — is asked daily, or it will be asked for the first time in a crisis.

🔍

[VERIFICATION REQUIRED] — open items in this chapter

Butt et al. 2024 (J Extra Corpor Technol 2024;56:20–29; DOI 10.1051/ject/2023047): the D-dimer figures reported in the retrieved text carry inconsistent units between passages (mg/dL in one, mg/L in another). Only the direction of change has been used here. The exact values and units require checking against the published article. The PMID has not been independently confirmed in this session.
Plasma-free haemoglobin greater than 50 mg/dL at 24 h and the 147 mL/kg/min flow threshold are cited by Butt et al. from primary sources that have not been retrieved and read here. They are reported as associations, not thresholds for action.
The fibrinogen ranges (2–6 g/L) are quoted from the same review; the primary sources behind them have not been read.
The ELSO ventilator settings in §8.3.3 are quoted as summarised by Szuldrzyński et al. 2024, not read from the ELSO source document directly. The ELSO general guidelines remain at v1.4 (August 2017).
Anticoagulation target ranges in §8.3.9 are quoted from Levy et al. 2022; the underlying sources are not primary outcome trials, and the same review states that no evidence-based consensus exists.
No monitoring frequency in this chapter is evidence-based. "Daily", "hourly" and "alternate-daily" reflect the composite practice described in the cited texts, not tested intervals.
The statement that approximately half of ECMO patients require RRT is taken from a 2017 textbook and may not reflect current practice; the primary registry data have not been retrieved.

Cross-references

Chapter 6 — VV ECMO Physiology: recirculation, mixing, and why arterial saturation behaves as it does
Chapter 7 — Initial VV ECMO Management: the first 24 hours, initial settings, initial ventilator strategy
Chapter 9 — VV ECMO Troubleshooting: the systematic approach to hypoxaemia, low flow, chattering and alarms
Chapter 10 — VV ECMO Weaning and Decannulation: when down-titration becomes a weaning trial
Chapter 32 — Oxygenator Failure and Circuit Exchange: the procedure, and failure modes in detail
Chapter 37 — Haemolysis: mechanisms, quantification and management
Chapter 53 — Mechanical Ventilation During ECMO: lung rest, driving pressure, mechanical power, proning
Chapter 55 — CRRT During ECMO: configurations and in-series delivery
Part VII — Anticoagulation: assays, targets, bleeding and thrombosis
Part XI — Ethics and Palliative Care: the bridge to nowhere
Chapters 87–88 — Practical Handbook: the printable daily checklist and round template

References

ECMO in the Adult Patient (Core Critical Care series), 2017. Circuit pressure monitoring and thresholds; transmembrane pressure; minimum safe flow; standardised circuit gas sampling; haemolysis monitoring; heat-exchanger failure rate; sedation interruption; renal replacement frequency. [VERIFICATION REQUIRED] — editors, publisher and page numbers are deliberately not stated: they were not verified in this session and are not reproduced from memory.
ISCCM Manual of RRT and ECMO in ICU. Indian Society of Critical Care Medicine. Daily whole-system assessment; pressure-port sites; interpretation of pre- versus post-membrane pressure rises; pump-head auscultation; sweep flush; battery, hand-crank and clamp checks; tubing care. [VERIFICATION REQUIRED] — edition, editors, year and page numbers not confirmed in this session.
Butt SP, Razzaq N, Saleem Y, Cook B, Abdulaziz S. Improving ECMO therapy: monitoring oxygenator functionality and identifying key indicators, factors, and considerations for changeout. J Extra Corpor Technol. 2024;56:20–29. DOI: 10.1051/ject/2023047.
Szuldrzyński et al. Mechanical ventilation during extracorporeal membrane oxygenation support: new trends and perspectives. Perfusion. 2024;39(1S):107S–114S. DOI: 10.1177/02676591241232270. [VERIFICATION REQUIRED] — full author list and initials not confirmed in this session.
Levy, Staudinger, Steiner. How to manage anticoagulation during extracorporeal membrane oxygenation. Intensive Care Med. 2022. DOI: 10.1007/s00134-022-06723-z. [VERIFICATION REQUIRED] — full author initials, volume, issue and pages not confirmed in this session.

📅

Chapter status

Drafted and audited 6 September 2026. Ten-pass quality control completed: clinical, physiology, evidence, citation, numerical, safety, contradiction, redundancy, bedside utility and literature-currency passes.

Open verification items are listed above rather than resolved silently. This chapter contains no fabricated citations, thresholds, doses or guideline statements; every unretrieved item is declared.

🔗

Addendum — 6 September 2026, added while drafting Chapter 9

Controversy 1 above states that no validated numeric oxygenator change-out threshold exists. That remains accurate, but it should be read alongside one figure found subsequently in a major source.

The ELSO Red Book, 6th edition (Chapter 7) states that when oxygen transfer is below 150 mL/min, membrane lung exchange should be considered. This is a published consensus prompt for consideration, not a validated trigger, and it does not contradict Butt et al. 2024 — that review's point is that no threshold has been tested against outcomes, which is still true of this one.

The practical position: there is a number available to prompt a decision, and no evidence that the number is correct. Chapter 9, §9.5 and Controversy 2 develop this in full.