Chapter question: The cannulae are in and the pump is running. What happens in the next few hours?
Evidence search date: 7 September 2026. This chapter contains the first randomised evidence in Part III about how to set VA ECMO rather than whether to use it — the BLENDER trial (Burrell et al., Intensive Care Medicine 2024), retrieved in full from the publisher. Other sources: ISCCM Manual Ch 30 (initiation sequence and safety checklist); ELSO Red Book Ch 32 (immediate post-cannulation actions); ECMO in the Adult Patient Ch 6 (cannulation checklists).
What this chapter covers — and what it does not
This chapter owns | Deferred to |
Starting the circuit; the first flow titration and the hypotension that follows it; oxygen targets after BLENDER; CO₂ and sweep in the first hours; the immediate post-cannulation checklist; the distal perfusion deadline; ventilator and sedation in the first 24–72 hours; what must be established before the team disperses | Who to cannulate → Chapter 11
Cannulation technique → Chapter 12
The physiology behind every setting here → Chapter 13
LV distension and unloading → Chapter 15
Differential hypoxaemia → Chapter 16
Limb ischaemia management → Chapter 17
Weaning → Chapter 18
ECPR-specific post-arrest care → Chapters 19–20
Anticoagulation in depth → Part VII
The daily round, once established → Chapter 8 |
14.1 Starting the circuit
The ISCCM manual gives an explicit initiation sequence. The order matters, because opening the wrong clamp first can drive blood backwards through a pump that is not yet generating forward flow.
Before starting: work through the initiation safety checklist (§14.6) and ensure the activated clotting time is above 250 seconds.
- Turn on the pump and increase speed to about 500 RPM — enough to provide forward flow before any clamp comes off.
- Open the clamp on the venous drainage line, and increase pump speed gradually to about 1,500 RPM.
- Remove the clamp from the blood outflow line, and increase pump speed further.
- Start sweep gas at 5–6 L/min, keeping the gas-flow to blood-flow ratio at 1:1. Monitor gas inlet pressure — ideally around 40–45 psi.
Danger — never let the pump sit stopped with the circuit unclamped
A centrifugal pump generates no forward flow at rest and does not occlude the circuit. With the pump stopped and both clamps off, blood flows backwards — from the pressurised arterial side, through the pump head, and back into the venous system. On VA this is a massive left-to-right shunt (Chapter 9).
This is why the sequence above brings the pump up to speed first and takes the clamps off after, and why it opens the drainage clamp before the return clamp. The same logic runs in reverse at every emergency: clamp, then stop.
14.2 The first flow titration — and the hypotension that follows
The target. The ISCCM manual sets the initial VA target as the flow needed to maintain a cardiac index of 2.0–2.4 L/min/m². Chapter 13 gives the alternative formulation from the Red Book — go to maximum flow first to establish drainage capacity, then back off until the pulse contour is 10–15 mmHg.
Clinical pearl — the two published approaches answer different questions, and you need both
The cardiac index target answers "is oxygen delivery adequate?" The pulse contour target answers "is the ventricle still ejecting?" Neither alone is sufficient, because the flow that satisfies the first can abolish the second.
In practice: set flow initially for the cardiac index, then look at the arterial waveform and the aortic valve on echocardiography and ask whether you have bought that index at the cost of ejection (Chapter 13 §13.6). If you have, the answer is not more flow — it is an unloading strategy (Chapter 15).
Physiology — why the blood pressure falls as you increase flow, and what to do
The ISCCM manual describes it plainly: as pump speed is increased, venous drainage may fall and hypotension results. This surprises people, because more flow ought to mean more pressure.
The mechanism is preload. The circuit drains the venous reservoir faster than it refills; the drainage cannula begins to collapse the vein around it, and drainage — and therefore total flow — falls. What looks like a pump problem is a volume problem.
The source's sequence: exclude cannula malposition first, then give volume — packed red cells, albumin or crystalloid — as boluses of 5–10 mL/kg, with explicit care to avoid fluid overload.
Note how this differs from VV (Chapter 7), where hypotension at initiation always has another cause because VV ECMO is haemodynamically neutral. On VA, hypotension during the first titration usually is the circuit — specifically, its appetite for preload.
Pitfall — the 5–10 mL/kg bolus is a diagnostic, not a strategy
The warning to avoid fluid overload is not decorative. Chapter 11 recorded that strongly positive fluid balances are associated with poor outcome on VA support, and Chapter 9 recorded the same argument on the VV side. A patient who needs repeated boluses to hold flow is telling you something — about drainage cannula size, position, or the possibility that the flow target itself is too ambitious for this patient's venous return.
Give volume to establish flow. Do not keep giving volume to defend a number.
14.3 Oxygen targets — what BLENDER changed
Evidence — BLENDER: the first randomised trial of how to set VA ECMO
Burrell and colleagues (Intensive Care Medicine 2024;50(9):1470–1483) randomised 300 adults on VA ECMO — 149 conservative, 151 liberal — in a registry-embedded, open-label, multicentre trial across 12 hospitals, drawn from 934 VA ECMO patients reported to the registry across 26 hospitals. Randomisation had to occur within 6 hours of ECMO initiation.
The two strategies:
— Conservative: SaO₂ 92–96%, oxygenator FbO₂ 0.5–1.0, ventilator FiO₂ minimum 0.21.
— Liberal: SaO₂ 97–100%, oxygenator FbO₂ 1.0, ventilator FiO₂ minimum 0.5.
Primary outcome — ICU-free days to day 28: median 0 days (IQR 0–13.7) conservative versus 0 days (IQR 0–13.3) liberal; median treatment effect 0 days (95% CI −3.1 to 3.1).
Mortality: day 28 39.6% versus 39.1%; day 60 43% versus 41.1%.
Everything else was similar too — duration of ECMO and mechanical ventilation, ICU and hospital length of stay, and 6-month functional outcomes. Subgroups were consistent, including cardiogenic shock, ECPR, age above and below 50, SAVE score above and below the median, and timing of randomisation. Critical hypoxaemia episodes, cardiopulmonary resuscitation and seizures were similar.
Authors' conclusion: "In adults receiving VA-ECMO in ICU, a conservative compared to a liberal oxygen strategy, did not affect the number of ICU-free days to day 28."
Certainty: moderate. A genuine randomised trial with a clean null result across every outcome measured, but open-label, modest in size for a mortality question, and with the deliverability problem below.
Pitfall — the most instructive number in BLENDER is a protocol deviation rate
The conservative group had 44 major protocol deviations (29.5%), against 2 (1.3%) in the liberal group — p less than 0.001.
Almost a third of the conservative arm could not be kept in its assigned band. That single figure carries two lessons. First, it means separation between the arms was imperfect, so a true difference would have been harder to detect — the null result is a little weaker than it looks. Second, and more usefully at the bedside: holding a VA ECMO patient at SaO₂ 92–96% is genuinely difficult, because the oxygenator delivers blood at near-maximal saturation and the mixing physiology of Chapter 13 does not offer fine control.
A target that a trial team could not deliver in 30% of patients is not a target to enforce rigidly on a ward round.
What this means for practice. BLENDER does not license hyperoxia — it found no benefit to conservative targeting, not a benefit to liberal targeting. What it does is remove the obligation to chase a tight low-normal saturation band in a patient where doing so is difficult and where the effort has no measurable payoff. Avoid gratuitous hyperoxia where it is easy to do so; do not destabilise a patient to achieve 92–96%.
A correction to Chapter 13. That chapter states that no randomised evidence supports any haemodynamic or gas-exchange target on VA ECMO, because the trials of Chapter 11 tested whether to use VA ECMO rather than how to set it. BLENDER is the exception, and Chapter 13 now carries a dated addendum saying so. The rest of that chapter's targets — flow, MAP, pulse contour, DO₂:VO₂ — remain untested by randomisation.
14.4 Carbon dioxide in the first hours
Sweep gas starts at 5–6 L/min at a 1:1 ratio with blood flow (§14.1), then is titrated. The specific hazard in the first hours is not hypercapnia but its opposite.
Danger — aim to avoid hypocarbia
The Red Book's immediate post-cannulation guidance is explicit that sweep gas and mechanical ventilation should be titrated with the aim of avoiding hypocarbia.
The reason is that many of these patients arrive profoundly acidotic — from shock, from arrest, or both — and a fully efficient membrane lung at a 1:1 sweep ratio will clear CO₂ far faster than any lung the patient has ever had. Rapid normalisation of a chronically or acutely elevated PaCO₂ causes a large fall in cerebral blood flow at precisely the moment the brain is most vulnerable, and the same caution appears in Chapter 8 for VV.
Correct the CO₂ deliberately and gradually. The membrane is capable of doing it far too fast.
14.5 The immediate post-cannulation checklist
The ELSO Red Book gives a minute-by-minute action list for the moments after cannulation. It appears in the ECPR chapter, and the ECPR-specific elements belong to Chapters 19–20 — but the actions below generalise to any VA initiation and are reproduced here on that basis.
Issue | Action | Note |
Access cannula | Confirm position by echocardiography and/or fluoroscopy; secure; dress | Drainage cannula tip in the right atrium, at the SVC–RA junction |
Mean arterial pressure | Measure from a right upper limb arterial line | No optimal MAP has been demonstrated; aim 60–80 mmHg. May need to rapidly down-titrate or hold pressors and inotropes |
PaO₂ | Measure from the right-sided arterial line | Avoid hyperoxia where possible. If hypoxaemic, consider differential hypoxia (Chapter 16) |
Arterial blood gases | Titrate sweep gas flow and mechanical ventilation | Aim to avoid hypocarbia (§14.4) |
Drainage and flows | If drainage is insufficient despite correct placement, consider fluids or transfusion | If total flow — VA plus native — is excessive, consider reducing VA flow |
Cardiac pathology and LV decompression | Echocardiography and 12-lead ECG | A comprehensive study by an experienced imager. Ask specifically about failure of LV decompression (Chapter 15) |
General measures | Endotracheal tube position and end-tidal CO₂; establish central access if not already present; review sedation and analgesia; chest radiograph | ETCO₂ assesses the native cardiopulmonary circulation (Chapter 13 §13.5) |
Laboratory investigations | Standard set including cardiac biomarkers | Anticoagulation targets must account for increased bleeding risk after arrest and after PCI |
Adapted from Table 32-3, ELSO Red Book 6th ed. The ECPR-specific content of that table is deferred to Chapters 19–20.
Clinical pearl — "may need to rapidly down-titrate or hold pressors"
This is easy to read past and is one of the most important instructions in the table. A patient cannulated in extremis is usually on substantial vasopressor and inotrope support. The moment the pump takes over the circulation, that support becomes both unnecessary and harmful — it raises the afterload the failing ventricle must overcome and lowers pump flow (Chapter 13 §13.6).
The Red Book's word is rapidly. Do not wean these drugs over hours out of caution; the physiology changed in one step and the prescription should follow.
Pitfall — "no optimal MAP has been demonstrated"
The Red Book states the 60–80 mmHg aim and, in the same breath, that no optimal mean arterial pressure has been shown. Chapter 13 records a slightly different published figure — Red Book Ch 28's MAP above 65 mmHg, and Ch 5's 60–70 mmHg controlled by flow.
These are not in conflict so much as jointly uncertain. Treat 60–80 mmHg as a reasonable band rather than a defended target, and remember that on VA the way to raise the pressure is usually flow, not a pressor.
14.6 The initiation safety checklist
The ISCCM manual's pre-initiation checklist is longer than most teams will run from memory, which is the point — it is designed to be printed and read aloud at sign-out.
Patient and plan: consent reconfirmed · name and hospital ID · height, weight, body surface area recorded · assessment of condition and anticipated problems · mode decided (VA or VV).
Baseline state: vital signs including heart rate and rhythm, arterial pressure, central venous pressure, temperature · ventilator settings · detailed neurological assessment including GCS, pupillary size and reaction, and any focal deficit · current infusions, fluids, feeds and sedation.
Investigations: arterial and venous or mixed venous blood gas · full blood count · renal and liver function · lactate · glucose · group and crossmatch · prothrombin time, APTT, ACT · chest radiograph · echocardiogram.
Readiness: blood products available · IV access patent · urinary catheter · crash cart including defibrillator at the bedside.
Circuit and cannulae: cannula position reconfirmed by transthoracic or transoesophageal echocardiography, properly secured, dressing in place · cannulation site inspected for infection, bleeding or haematoma · peripheral pulses checked in the cannulated limbs · tube clamps at the bedside · full length of tubing inspected for cracks or fibrin · oxygenator inspected for clot, fibrin or air · all connections tight · pump correctly seated in the cradle · emergency hand crank available.
Clinical pearl — the neurological examination is a baseline you cannot obtain later
The checklist asks for GCS, pupils and focal deficits before initiation. In a patient who will spend the next fortnight sedated and anticoagulated, and in whom intracranial haemorrhage and ischaemic stroke are among the commonest catastrophic complications, this is the only clean neurological baseline anyone will ever have.
The same argument appears in Chapter 8 for the daily examination. Here it is stronger, because there is no way back to this moment.
Cross-check against the procedural checklists in ECMO in the Adult Patient Ch 6, which add: two units of red cells available, platelets above 100,000 or a plan in place, cannula size agreed and available, a stated back-up plan for failed insertion, invasive blood pressure, CVP and end-tidal CO₂ monitoring, antibiotics, and anticoagulation — with a sign-out covering dressing, line securement, instrument count and safe disposal of guidewires and sharps.
14.7 The distal perfusion deadline
Chapter 12 established that the distal perfusion wire should go in before the arterial cannula, because distal flow is much lower afterwards. In emergency cannulation and ECPR that is often not possible, and this chapter owns what happens next.
Danger — four hours
The ELSO Red Book is specific: where femoral cannulation has been used and a distal perfusion cannula has not already been positioned, this should be undertaken ideally within 4 hours to reduce the risk of limb ischaemia.
This converts Chapter 12's technical preference into a time-bound obligation with an owner. In the chaos after an emergency cannulation it is exactly the task that gets deferred to "the morning" — by which point the limb may be beyond rescue.
Write the deadline in the notes at the time of cannulation, and name who will do it.
14.8 Ventilator, sedation and the first 24–72 hours
Evidence — the first phase has a defined task
The Red Book frames the early period explicitly: during the first 24–72 hours of ECLS support, ventilator-induced lung injury must be minimised and patient self-inflicted lung injury (P-SILI) must be avoided. In later phases, sedatives and neuromuscular blockers are titrated down and withdrawn, and the intensity of breathing is monitored — the source suggests oesophageal pressure monitoring — with some centres using low-dose neuromuscular blockade to limit excessive respiratory drive.
Certainty: low — expert practice, and the specific claims about P-SILI on ECMO are extrapolated from the ARDS literature rather than tested on VA support.
Two VA-specific points modify the general picture:
- The lung is not usually the problem, but it is still at risk. On VA support for cardiogenic shock the lungs may be relatively healthy. They are nonetheless receiving little pulmonary blood flow, are exposed to whatever ventilation is set, and — if the left ventricle distends — will be the first organ to declare it, as pulmonary oedema and then pulmonary haemorrhage (Chapter 15).
- Sedation depth cannot be traded against respiratory drive the way it can on VV. Chapter 8's argument — that sweep gas controls drive, so sedation can be lightened — depends on the circuit clearing the patient's CO₂. That remains true on VA, but the competing consideration is different: an agitated patient with femoral arterial and venous cannulae is at risk of decannulation and of the arterial catastrophe that follows.
14.9 Before the team disperses
The cannulation team, the perfusionist, the echocardiographer and the intensivist are all present at the moment of initiation and will not be again. Six things should be settled while they are.
- Cannula positions confirmed and documented, with insertion depths and skin markings recorded.
- A right upper limb arterial line sited — for MAP, for PaO₂, and because it samples what the coronary and cerebral circulations receive (Chapters 12, 13, 16).
- A baseline echocardiogram by an experienced imager, specifically documenting aortic valve opening, LV size and end-diastolic dimension, and RV function — the reference points against which every later study will be read (Chapter 13).
- The distal perfusion decision made and, if deferred, its four-hour deadline written down with a named owner (§14.7).
- Vasoactive drugs reviewed and reduced, not left running because nobody wanted to touch them (§14.5).
- The destination stated in the notes — bridge to recovery, procedure, device, transplant, or a named decision with a named decision-maker and a date (Chapter 11 §11.6). This is the item most often skipped and the one Chapter 8's destination review depends on.
14.10 Controversies
Controversy 1 — What oxygen target should be set on VA ECMO after BLENDER?
The question. BLENDER found no benefit from a conservative (SaO₂ 92–96%) over a liberal (97–100%) strategy. Should units now stop targeting conservative oxygenation?
The case for abandoning the conservative target. It was tested properly and it did not help — not on ICU-free days, not on mortality at 28 or 60 days, not on ECMO or ventilation duration, not on 6-month function, and not in any prespecified subgroup. And it was hard to deliver: 29.5% major protocol deviations against 1.3%. Enforcing a target that a trial team could hold in only 70% of patients, for no measurable benefit, spends attention that could go elsewhere.
The case for keeping it. BLENDER shows no benefit; it does not show safety of hyperoxia. Its liberal arm targeted 97–100% saturation, not deliberate hyperoxia, so the trial does not license running the oxygenator at FbO₂ 1.0 and ignoring the resulting PaO₂. The observational literature associating hyperoxia with harm after cardiac arrest and on ECMO is not overturned by a 300-patient trial with imperfect separation, and the Red Book's own post-cannulation guidance still says to avoid hyperoxia where possible.
What the evidence actually shows. A clean null result on every measured outcome, with moderate certainty, limited by open-label design, modest size for a mortality endpoint, and the deviation rate that narrowed the achieved separation between arms.
Where practice actually sits. Most units read BLENDER as removing the obligation to chase 92–96% rather than as an endorsement of high targets. The pragmatic position is to reduce oxygenator FbO₂ when it is easy and the patient is stable, and not to destabilise anyone in pursuit of a saturation band.
What would resolve it. A trial powered for mortality, or one comparing a genuinely liberal strategy against a deliberately hyperoxic one, with achieved separation reported. BLENDER's own investigators have written separately on hyperoxia during VA ECMO, which suggests the question is regarded as open.
This book's position. Avoid gratuitous hyperoxia; do not fight for 92–96%. Set the oxygenator to what the patient needs, review it on the round, and spend the attention saved on the aortic valve and the limb.
Controversy 2 — Should flow be set to a cardiac index or to a pulse contour?
The question. Two published approaches to the initial flow titration point in opposite directions. ISCCM: titrate flow to achieve a cardiac index of 2.0–2.4 L/min/m². Red Book Ch 5: go to maximum flow, then back off until the pulse contour is 10–15 mmHg. Which governs?
The case for the cardiac index target. It is anchored to the thing ECMO exists to provide — oxygen delivery — and it is the familiar currency of shock management. It gives a defensible number for a junior operator at 3 a.m.
The case for the pulse contour target. It is anchored to the thing that kills these patients — loss of ejection, distension, pulmonary oedema and intracardiac thrombosis. Cardiac index says nothing about whether the ventricle is still working, and on VA the index can be met entirely by the machine.
What the evidence actually shows. Neither target has randomised support. Certainty: very low for both. Chapter 13 records the additional complication that the recommended pulse-contour band (10–15 mmHg) overlaps the pulse-pressure threshold used to warn of near-absent native output.
Where practice actually sits. Both, sequentially — flow set for delivery, then interrogated for its effect on ejection. That is what §14.2 recommends.
What would resolve it. A study relating initial flow strategy to LV distension, unloading requirement and survival. None is known here.
The transferable point. These are not competing targets so much as a target and a constraint. The cardiac index tells you how much flow you need; the pulse contour and the aortic valve tell you how much flow this ventricle can tolerate. When they conflict, the answer is not more flow — it is unloading (Chapter 15).
14.11 The errors that recur
Error | Correction |
Unclamping before the pump is generating forward flow | Pump to ~500 RPM first, then drainage clamp, then return clamp. Backwards flow on VA is a massive left-to-right shunt |
Responding to hypotension during the first titration by increasing RPM | It is usually preload. Exclude malposition, then volume in 5–10 mL/kg aliquots |
Giving repeated boluses to defend a flow target | A bolus is a diagnostic. Persistent need points to cannula size, position, or an over-ambitious target |
Leaving vasopressors and inotropes running after the pump takes over | Down-titrate rapidly — they raise ventricular afterload and lower pump flow |
Normalising the PaCO₂ quickly | The membrane can do it far too fast. Aim to avoid hypocarbia and correct gradually |
Deferring the distal perfusion cannula to the morning | Four hours. Write the deadline and name the owner at cannulation |
Not documenting a neurological examination before initiation | It is the only clean baseline that will ever exist for this patient |
Fighting to hold SaO₂ 92–96% | BLENDER found no benefit, and 29.5% of its conservative arm could not be held there |
Letting the team disperse without a documented destination | Chapter 8's destination review has nothing to review |
14.12 Key points
- Pump first, clamps second — drainage before return. A stopped centrifugal pump on an unclamped VA circuit shunts blood backwards.
- Ensure ACT above 250 seconds before initiating; start sweep at 5–6 L/min at a 1:1 ratio with blood flow.
- Initial flow targets a cardiac index of 2.0–2.4 L/min/m² — then immediately ask what that flow did to the pulse contour and the aortic valve.
- Hypotension during the first titration is usually preload, not pump failure. Exclude malposition, then give 5–10 mL/kg aliquots — as a diagnostic, not a strategy.
- BLENDER is the first randomised trial of how to set VA ECMO: conservative (92–96%) versus liberal (97–100%) oxygen targets made no difference to ICU-free days, mortality at 28 or 60 days, ECMO or ventilation duration, or 6-month function.
- The conservative target was hard to deliver — 29.5% major protocol deviations versus 1.3%. Avoid gratuitous hyperoxia; do not destabilise a patient chasing a saturation band.
- Aim to avoid hypocarbia. The membrane clears CO₂ far faster than any native lung, and rapid correction drops cerebral blood flow.
- Down-titrate vasopressors and inotropes rapidly once the pump takes over — the Red Book's word is rapidly.
- 60–80 mmHg is a reasonable MAP band, not a defended target — no optimal MAP has been demonstrated, and on VA the lever is flow, not a pressor.
- Measure MAP and PaO₂ from a right upper limb arterial line. Hypoxaemia there means differential hypoxia until proven otherwise.
- If a distal perfusion cannula was not placed at cannulation, place it within 4 hours — with a named owner and a written deadline.
- Document a full neurological examination before initiation. There is no way back to that baseline.
- Settle six things before the team disperses: cannula positions, the right radial line, a baseline echo documenting aortic valve opening, the distal perfusion decision, the vasoactive review, and the destination.
[VERIFICATION REQUIRED] — open items in this chapter
- BLENDER was retrieved in full from the publisher and is the best-verified trial in Part III. Its bibliographic details, targets, primary and secondary results, subgroups, safety outcomes and protocol deviation rates are all as reported in that article. Not independently checked: the full author list beyond the first three, and the precise definition used for "major protocol deviation".
- The initiation sequence (500 RPM, drainage clamp, 1,500 RPM, return clamp, sweep 5–6 L/min at 1:1, gas inlet 40–45 psi) is from the ISCCM Manual and is stated without supporting evidence. Device instructions for use govern.
- The cardiac index target of 2.0–2.4 L/min/m², the ACT above 250 seconds initiation threshold and the 5–10 mL/kg bolus aliquot are all ISCCM figures without citations retrieved.
- The 4-hour distal perfusion window is from the Red Book, cited there to a reference not retrieved. It is a stated ideal, not a validated threshold.
- The 60–80 mmHg MAP aim appears in the Red Book's ECPR chapter alongside an explicit statement that no optimal MAP has been demonstrated. Chapter 13 records two other published figures. All are consensus.
- The immediate post-cannulation table is adapted from the Red Book's ECPR chapter (Table 32-3) and applied here to VA initiation generally. That extension is this book's judgement; the ECPR-specific content is deferred to Chapters 19–20.
- The 24–72 hour early-phase framing and the P-SILI caution are expert practice, extrapolated from the ARDS and VV literature rather than tested on VA support.
- The pre-cannulation coagulation targets seen in the project library (haemoglobin above 7.0 g/dL, platelets above 100,000, INR below 2.0) come from one transport service's referral checklist and are reproduced nowhere in this chapter as a general standard.
- SAVE score appears as a BLENDER subgroup variable. It is a VA ECMO prognostic score not otherwise covered in this book; it has not been described or verified here and is a gap to close in Chapter 11 or Chapter 30.
Cross-references
- Chapter 7 — Initial VV ECMO Management: the parallel chapter, and the contrast — on VV, hypotension at initiation always has another cause
- Chapter 8 — Daily VV ECMO Management: where the routine takes over, and the destination review that §14.9 sets up
- Chapter 11 — VA ECMO: Indications and Patient Selection: the destination that must be named, and the fluid-balance warning
- Chapter 12 — VA ECMO Cannulation: the distal perfusion wire that should already be in, and the right radial line
- Chapter 13 — VA ECMO Haemodynamics: the physiology behind every setting here — flow as the vasopressor, the pulse contour, the mixing point. Carries a dated addendum recording BLENDER
- Chapter 15 — LV Distension and LV Unloading: where the answer goes when flow and ejection conflict
- Chapter 16 — Differential Hypoxaemia / Harlequin Syndrome: why the right-sided PaO₂ matters from the first hour
- Chapter 17 — Limb Ischaemia and Vascular Complications: what the four-hour deadline is protecting against
- Chapters 19–20 — ECPR: the arrest-specific post-resuscitation care that Table 32-3 also covers
- Part VII — Anticoagulation: the ACT threshold here, and the post-arrest and post-PCI bleeding-risk caveat
References
- Burrell A, Bailey MJ, Bellomo R, et al. Conservative or liberal oxygen targets in patients on venoarterial extracorporeal membrane oxygenation (BLENDER). Intensive Care Medicine. 2024;50(9):1470–1483. DOI: 10.1007/s00134-024-07564-8. Retrieved in full from the publisher.
- ISCCM Manual of RRT and ECMO in ICU, Chapter 30, Cannulation, Priming and Initiation of ECMO. The initiation sequence; ACT threshold; sweep gas starting settings and gas:blood ratio; gas inlet pressure; the cardiac index target; hypotension from reduced venous drainage and its management; the ECMO initiation safety checklist. [VERIFICATION REQUIRED] — edition, editors, year and page numbers not confirmed.
- Extracorporeal Life Support: The ELSO Red Book, 6th edition, Chapter 32 (Table 32-3, immediate post-arrest management) and the early-phase ventilation and sedation guidance. Post-cannulation actions; right upper limb arterial monitoring; MAP aim with the explicit absence of a demonstrated optimum; avoiding hyperoxia and hypocarbia; the 4-hour distal perfusion window; the 24–72 hour lung-protection phase.
- ECMO in the Adult Patient (Core Critical Care series), 2017, Chapter 6. Cannulation insertion and removal checklists, including blood product availability, back-up plan for failed insertion, monitoring requirements, and sign-out items. [VERIFICATION REQUIRED] — editors, publisher and page numbers deliberately not stated.
Chapter status
Drafted and audited 7 September 2026. Ten-pass quality control completed: clinical, physiology, evidence, citation, numerical, safety, contradiction, redundancy, bedside utility and literature-currency passes.
This is the best-evidenced chapter in Part III so far, because BLENDER is a properly conducted randomised trial retrieved in full rather than through a secondary summary. Its most useful finding for the bedside is arguably not the null primary outcome but the 29.5% protocol deviation rate in the conservative arm — evidence that the target is hard to deliver, which is exactly the sort of practical fact that trial reports usually bury. Everything else in the chapter remains expert practice, and the verification callout says which figures rest on nothing more.