Chapter question: What is the circulation actually doing when a pump is pushing blood backwards up the aorta — and what should I measure to find out?
Evidence search date: 6 September 2026. Sources: ELSO Red Book 6th edition Ch 5 (The Physiology of Extracorporeal Life Support), Ch 27–28 (adult cardiac failure); ISCCM Manual Ch 26 (physiology) and Ch 33 (monitoring); Taha Ch 6 (Physiology II: Venoarterial ECMO); ECMO in the Adult Patient Ch 4.
This chapter contains a direct contradiction between two chapters of the same textbook on whether VA ECMO raises left ventricular afterload. It is set out in §13.3 and resolved rather than concealed.
What this chapter covers — and what it does not
This chapter owns | Deferred to |
The three circulations and the mixing point quantitatively; the afterload question and its resolution; pulsatility — what the arterial waveform means and what to target; estimating native cardiac output; why the pulmonary artery catheter misleads on VA; the relationship between flow, pressure and oxygen delivery; the haemodynamic monitoring dataset | Who to cannulate → Chapter 11
Where the cannulae go → Chapter 12
The hour-by-hour initiation sequence → Chapter 14
LV distension — recognition and unloading → Chapter 15
Differential hypoxaemia in depth → Chapter 16
Limb ischaemia → Chapter 17
Weaning studies and echo criteria → Chapter 18
Fick, DO₂–VO₂ and the critical threshold → Chapter 2 |
13.1 Three circulations in one patient
On VV ECMO there is one circulation, and the circuit adds oxygen to it. On VA ECMO there are effectively three things to think about at once:
- The native circulation — whatever the left ventricle still ejects, having passed through the lungs, carrying whatever oxygen those lungs managed to add.
- The extracorporeal circulation — fully oxygenated blood delivered by the pump, at a flow you set.
- The mixed field — the territory where the two meet, whose location moves with the balance between them.
Physiology — total flow is conserved; the pulse is not
The ISCCM manual states the core relation plainly: as venous blood is drained from the right atrium and perfused into the aorta, the total flow remains constant, but the pulse contour decreases, because less blood is ejected from the left ventricle.
Push this to its limit and the picture becomes clear. When extracorporeal flow reaches 100% of venous return, the systemic pulse contour is flat — the patient is on full cardiopulmonary bypass. In VA ECMO the flow is conventionally maintained at about 80% of venous return, so roughly 20% still passes through the heart and lungs, and the pulse pressure is around 10 mmHg.
This is the single most important idea in VA haemodynamics: the pulse pressure is a readout of how much blood is still going through the heart. It is not a measure of blood pressure adequacy. It is a measure of the division of labour between the patient and the machine.
Two consequences follow immediately, and they are the reason Chapters 15 and 16 exist:
- Blood that does not pass through the lungs does not get oxygenated by them. If the lungs are failing and the heart is ejecting, that poorly oxygenated blood goes to whatever the native circulation reaches first — the coronary and cerebral arteries (Chapter 16).
- Blood that does not leave the left ventricle accumulates in it. The ISCCM manual gives the mechanism and a number: the left side fills from bronchial and thebesian venous flow, and "when that pressure reaches 20–25 mmHg, pulmonary edema occurs and the left ventricle becomes overdistended" (Chapter 15).
13.2 The mixing point, and why it moves
Chapter 12 established that femoral return creates a watershed and upper-body return does not. The quantitative question is where the watershed sits, and the answer is that it is not a fixed anatomical location — it is wherever the two opposing flows balance.
Change | Effect on the mixing point | Clinical meaning |
Native cardiac output improves | Moves distally — further down the aorta | More of the body is perfused by native blood. Good news for the heart, dangerous for the brain if the lungs are still failing |
Native cardiac output falls | Moves proximally — towards the arch | More of the body is perfused by the circuit. Differential hypoxaemia risk falls; LV distension risk rises |
ECMO flow increased | Moves proximally | The standard manoeuvre for differential hypoxaemia — at the cost of more afterload and less ejection |
Lung function improves | Position unchanged, but the consequence disappears | The watershed only matters when native blood is poorly oxygenated |
Pitfall — the patient who looks pink and is having a myocardial infarction
ECMO in the Adult Patient gives the clearest description of this trap: a patient can "appear to be very well oxygenated (i.e. be pink and have a high measured PaO₂) but have an ischaemic electrocardiogram."
The reason is that the very well oxygenated ECMO blood is distributed to every vascular bed except the coronaries, while the blood entering the coronary arteries came out of a failing lung. A femoral or left-sided arterial sample will read reassuringly. The heart is being starved.
This is why Chapter 12 sites a right radial arterial line at cannulation, and why a new ischaemic ECG on VA ECMO is a differential-hypoxaemia diagnosis until proven otherwise (Chapter 16).
Danger — never interrupt the sweep gas on a VA circuit
ECMO in the Adult Patient states it in capitals — interrupting sweep gas in a veno-arterial circuit is "something that should NEVER be done" — because it injects blood with a low oxygen content directly into the arterial tree, producing what it calls a reverse harlequin effect: the circuit-perfused territory becomes the hypoxaemic one.
This independently corroborates the warning in Chapters 10 and 11 that the sweep gas off trial is a VV-only manoeuvre. In VV, turning off the sweep makes the circuit inert. In VA, it turns the circuit into a device that actively pumps deoxygenated blood into the aorta.
13.3 Does VA ECMO raise or lower LV afterload? The two answers, and the resolution
Two chapters of the ELSO Red Book — the most authoritative single text in this field — say opposite things about the central haemodynamic consequence of VA ECMO. Both are quoted here in full, because a reader who has met only one of them will misunderstand the other.
Controversy 1 — The afterload contradiction
The question. Does peripheral VA ECMO increase or decrease left ventricular afterload?
Position A — it decreases it. ELSO Red Book Chapter 5: "Mixing of cardiac and retrograde ECLS flow does not add significantly to arterial pressure or systemic resistance. VA ECMO does not increase left ventricular afterload above normal. In fact, it decreases afterload by maintaining non pulsatile perfusion." The same chapter adds that once the heart recovers pulsatile flow, work can be minimised by unloading systolic pressure with an intra-aortic balloon pump or synchronised pulsatile flow.
Position B — it increases it. ELSO Red Book Chapter 27: "With a competent aortic valve and poor LV function, peripheral VA ECMO pressurizes the aorta resulting in proportionally excessive LV afterload, more so when right ventricular contractility is preserved. Without adequate LV ejection into the aorta, the LV overdistends, which may rapidly culminate in severe pulmonary edema, pulmonary hemorrhage, and increased risk of LV cavity and aortic root thrombosis."
What the evidence actually shows. Neither statement is supported by a trial; both are physiological argument in a textbook. Certainty: low for either as stated. But they are not equally applicable, and the disagreement is not really about physics — it is about which patient is being described.
The resolution. They are answering different questions about different hearts.
— Position A is about the systemic circuit as a whole, in a heart that still ejects. At the level of "does the arterial tree see a much higher resistance because flow arrives retrogradely?", the answer is no: total flow is conserved (§13.1), and non-pulsatile perfusion at a given mean pressure presents a lower time-averaged impedance than pulsatile perfusion does. Nothing about the direction of flow in the aorta inherently raises systemic vascular resistance.
— Position B is about the left ventricle specifically, in a heart that barely ejects. The failing LV must generate enough pressure to open the aortic valve against an aortic root that the pump is actively pressurising. That is a threshold, not an average. If the ventricle cannot cross it, it does not eject at all — end-diastolic volume and pressure climb, and the distension cascade begins.
So the operative variable is residual LV function, and the two positions describe the two ends of it. In a heart with reasonable contractility, raising flow modestly redistributes the work and Position A holds. In a severely failing ventricle with a competent aortic valve, raising flow can stop ejection altogether and Position B holds — which is precisely why the Red Book's own cardiac chapters, dealing with the sickest hearts, describe the afterload problem while its physiology chapter, reasoning about the general case, does not.
Where practice actually sits. Every experienced unit manages as though Position B is true, because the failure mode it describes is catastrophic and the one Position A describes is benign. The whole of Chapter 15 exists because of Position B.
What would resolve it. Pressure–volume loop studies across a range of residual LV function on VA support, relating flow to aortic valve opening and end-diastolic pressure. Taha's Chapter 6 presents the P–V loop framework this would use.
What to take to the bedside. Do not ask "does VA ECMO raise afterload?" Ask "is this ventricle still ejecting, and what happens to that when I change the flow?" The instrument for the first question is the arterial waveform; for the second, echocardiography.
Physiology — why right ventricular function makes the left ventricle worse
The Red Book's phrase "more so when right ventricular contractility is preserved" rewards attention, because it inverts an intuition.
Blood not captured by the drainage cannula continues as native venous return, through a working right ventricle, across the lungs, and into the left atrium. A good RV therefore keeps delivering volume to a left ventricle that cannot eject it. The better the right heart, the faster the left heart fills against a pressurised aorta.
So in biventricular failure the LV distension problem is often mild, and in isolated left ventricular failure with a preserved RV it can be severe and fast. The combination to fear is a strong right ventricle, a weak left ventricle, a competent aortic valve, and high circuit flow.
13.4 Pulsatility — reading the arterial waveform
The arterial line on VA ECMO is not primarily a blood-pressure monitor. It is a continuous readout of native cardiac ejection, and it should be read that way.
Decreased pulsatility | Increased pulsatility |
Increased pump flow
Decreased heart contractility
Pericardial collection
Hypovolaemia
Pneumothorax
Aortic valve thrombosis | Decreased pump flow
Increased heart contractility |
Source: Table 4.1, ECMO in the Adult Patient (Core Critical Care, 2017).
The ISCCM manual adds the consequences of losing pulsatility — thrombus, myocardial ischaemia and pulmonary oedema — and lists its causes as poor myocardial function, excessive VA ECMO support, inadequate preload, and right ventricular failure.
Clinical pearl — falling pulsatility has two opposite meanings, and you must not guess
A narrowing pulse pressure means either the heart is getting worse or the machine is taking over more of the work. Those demand opposite responses, and the waveform alone cannot distinguish them.
The discriminating manoeuvre is simple: reduce the flow slightly and watch. If pulsatility returns promptly, the ventricle was being overwhelmed by support. If it does not, the ventricle has deteriorated. The Red Book uses exactly this test in reverse as the sign of recovery — "pulse contour increases when flow is decreased."
Do this deliberately and briefly, with echocardiography available, not as an experiment on an unstable patient.
Danger — an intra-aortic balloon pump makes the waveform lie
The Red Book warns that pulsatility on the arterial waveform in the presence of an IABP may provide false reassurance. The balloon generates a pulse contour whether or not the ventricle is ejecting.
In any patient with both devices, pulsatility is no longer a valid readout of native ejection, and the question must be answered by assessing aortic valve opening on echocardiography — which the Red Book states is important precisely for this reason.
13.5 Estimating native cardiac output — and why the PA catheter misleads
Pitfall — thermodilution measures the wrong circulation
The Red Book is explicit on two counts. First, a pulmonary artery catheter measures only the native pulmonary circulation — it tells you nothing about the 60–80% of systemic flow arriving through the cannula. A cardiac index of 1.4 L/min/m² on a patient with 4 L/min of circuit flow is not a shock state; it is the expected reading.
Second, the measurement itself is unreliable here: correct thermodilution of transpulmonary flow depends on catheter position and "may interfere with dynamically changing venous flows around the drainage cannula."
A PA catheter still has uses on VA ECMO — pulmonary capillary wedge pressure is a genuine signal of left-heart loading (§13.6, Chapter 15). But its cardiac output number must never be read as the patient's cardiac output.
Evidence — a two-variable bedside estimate of native output
The Red Book offers a usable surrogate: end-tidal CO₂ monitoring as an estimate of transpulmonary blood flow, since CO₂ can only reach the alveoli via blood the right ventricle actually pushes through the lungs.
The specific rule: an end-tidal CO₂ below 14 mmHg together with a pulse pressure below 15 mmHg may predict a native cardiac output below 1 L/min "with good accuracy."
Certainty: low. This is a single threshold pair cited by the Red Book to one reference, not retrieved here; no sensitivity, specificity or validation cohort is given. Treat it as a flag that native output is very low, not as a measurement.
Its value is that both variables are already on the monitor, continuously, for free — and that they fail in the same direction for the same reason, which is what makes the combination more informative than either alone.
The reference standard remains echocardiography. The ISCCM manual states it directly: "the best way to assess heart function in VA ECMO is by echocardiography." Specifically — does the aortic valve open, on how many beats, and what is the left ventricular size and end-diastolic dimension doing over time.
13.6 Flow, pressure and the targets
Chapter 14 covers the initiation sequence. This section covers the physiological principles that sequence implements — what flow and pressure each control, and what the published targets are.
13.6.1 Published haemodynamic targets
Target | Value | Source |
Mean arterial pressure | Above 65 mmHg (Red Book Ch 28); 60–70 mmHg controlled by ECLS flow (Red Book Ch 5) | Widely accepted |
Systolic blood pressure | Above 90 mmHg | Red Book Ch 28 |
Cardiac index (where a PA catheter is used — see §13.5) | Above 2.2 L/min/m² | Red Book Ch 28 |
Pulmonary capillary wedge pressure | Below 15 mmHg | Red Book Ch 28 |
Pulse contour / pulse pressure | 10–15 mmHg (Red Book Ch 5); around 10 mmHg at 80% of venous return (ISCCM Ch 26) | Both |
Oxygen delivery to consumption ratio | DO₂:VO₂ of 3:1 or higher | Red Book Ch 5 |
Venous saturation during weaning | Above 70% | Red Book Ch 5 |
Pitfall — the pulse-pressure target and the low-output alarm overlap
Put two of the numbers above side by side. The Red Book's target pulse contour is 10–15 mmHg. The Red Book's warning rule is that a pulse pressure below 15 mmHg, with an end-tidal CO₂ below 14 mmHg, predicts a native cardiac output below 1 L/min.
These are not contradictory — one is a single variable used as a support target, the other a two-variable rule for a different purpose. But the overlap is clinically meaningful: a patient sitting comfortably at the recommended pulse contour is, by the other rule's reckoning, close to the zone of minimal native ejection. That is exactly the zone in which LV distension develops (§13.3, Chapter 15).
The practical implication is that the pulse-pressure target is a floor to be respected, not a comfortable middle. At 10–15 mmHg you should already be asking the aortic-valve-opening question on echocardiography, not waiting for pulsatility to disappear entirely.
13.6.2 The Red Book's flow-setting principle
Red Book Chapter 5 gives an explicit sequence, and its logic is worth extracting because it explains why the numbers above are what they are:
- Plan the circuit from the estimated metabolic rate — for adults, 3–4 mL/kg/min for both oxygen and carbon dioxide. The worked example given: a septic 80 kg adult needs 5 L/min of flow and an oxygenator with rated flow over 5 L/min to supply 300 mL O₂/min.
- Go to the highest flow first, to establish the maximum drainage capacity, following the pulse contour. If the drainage cannula is large enough, total bypass with non-pulsatile flow results.
- Then decrease the flow until the pulse contour is 10–15 mmHg.
- Reduce vasoactive drugs to low or absent levels, and use ECLS flow — not vasopressors — to control arterial pressure at 60–70 mmHg.
- If there is no LV function, establish left atrial drainage (Chapter 15).
- Reassess oxygen kinetics once stable, usually at 6–12 hours. If DO₂:VO₂ is below 3 and the patient is anaemic, transfuse to a haemoglobin of 12–14 g/dL, which should produce an arterial saturation around 95% and a venous saturation around 65%.
Clinical pearl — on VA ECMO, flow is the vasopressor
The instruction to "use ECLS flow to control arterial pressure" and to reduce vasoactive drugs to minimal levels is a genuine reversal of ordinary ICU practice, and it follows from MAP = cardiac output × systemic vascular resistance.
On VA support you own the cardiac output term directly — you can turn it up. Reaching for noradrenaline instead raises resistance, which increases the afterload the failing ventricle must overcome and, on the circuit side, increases pump afterload and reduces flow (§13.7). The Red Book notes separately that high-dose vasopressors impede the microcirculation and that liberal inotrope use may hamper myocardial recovery.
The caveat is real: many of these patients have a genuinely vasoplegic component, and some vasopressor is appropriate. The point is the order of reach — flow first, pressor second — not that pressors are forbidden.
Danger — higher flow is not a free good
The Red Book is explicit that it may be necessary to increase flow above 4 L/min, and equally explicit about the cost: higher VA ECMO flows increase impaired cardiac ejection, left ventricular overload and pulmonary congestion.
So flow sits between two walls, exactly as it did in VV (Chapter 8) but for different reasons. Too little flow and oxygen delivery is inadequate. Too much flow and the ventricle stops ejecting, distends, and floods the lungs. The window is found with the arterial waveform and echocardiography, not with a formula.
13.7 Why flow falls when the blood pressure rises
A phenomenon with no VV equivalent, and a common source of confusion.
Physiology — the pump is afterload-sensitive
A centrifugal pump is not a fixed-displacement device. It generates a pressure difference, and the flow it achieves depends on what it is pumping against. Raise the pressure in the aorta and, at unchanged revolutions per minute, the flow will fall.
ECMO in the Adult Patient describes this happening spontaneously as the heart recovers: "a better cardiac function will increase systemic pressure and this in turn will decrease ECMO flow if the pump energy is not increased."
So a falling circuit flow on VA ECMO has an entirely benign explanation that does not exist in VV — the patient's own heart is getting better and raising the pressure the pump works against. The ISCCM manual lists the alternatives to exclude: inadequate preload (hypovolaemia, with chatter and haemolysis), mechanical obstruction, excessive afterload from thrombus or kink, and inadequate RPM.
Chapter 9's falling-flow algorithm applies here with one addition at the top: on VA, ask what the blood pressure did. Rising MAP with falling flow and rising pulsatility is recovery. Falling flow with unchanged pressure is the Chapter 9 problem.
13.8 The haemodynamic monitoring dataset
Parameter | What to look for | Why it matters |
Rhythm | Ventricular fibrillation or any rhythm with absent LV ejection | Leads to LV distension and thrombus formation |
MAP | Hypotension — MAP = cardiac output × SVR | Either insufficient ECMO flow or low SVR. Distinguish before treating |
Pulsatility | Loss of pulsatility from poor myocardial function, excessive support, inadequate preload or RV failure | Thrombus, myocardial ischaemia, pulmonary oedema |
Flow | Low flow from inadequate preload, excessive afterload (thrombus, kink, SVR) or inadequate RPM | Hypovolaemia may show as chatter and haemolysis (Chapter 9) |
Gas exchange | Circuit settings, oxygenator function, and upper-body hypoxaemia in femoro-femoral cannulation | Chapter 16 |
Tissue oxygen delivery | Falling SvO₂ and rising lactate; consider flow, haemoglobin and SaO₂. Also excessive consumption — fever, shivering | DO₂ = cardiac output × CaO₂. Shivering is treatable |
Distal limb | Absent dorsalis pedis and posterior tibial pulses; cyanosis; cold limb | Chapter 17 |
Echocardiography | Aortic valve opening, LV size and end-diastolic dimension, RV function, and — with an IABP — the only valid assessment of ejection | The reference standard for all of the above |
Adapted from Table 1, ISCCM Manual Ch 33, with the echocardiography row added from Red Book Ch 27 and ISCCM Ch 26.
13.9 Controversy 2 — What pulse pressure should be targeted?
Controversy 2 — Is 10–15 mmHg the right pulse-pressure target?
The question. The published target is a pulse contour of 10–15 mmHg, obtained by backing the flow off from full support. Is that the right place to sit?
The case for a low target (maximal support). The purpose of VA ECMO is to rest a failing heart and guarantee systemic oxygen delivery. Higher flow does both. A ventricle doing very little work is a ventricle consuming little oxygen, which is the rationale for mechanical support in the first place.
The case for a higher target (more native ejection). Every increment of ejection reduces the risk of the complications that actually kill these patients — LV distension, pulmonary oedema, intracardiac and aortic root thrombosis. Keeping the aortic valve opening also keeps blood moving through the left heart, and pulsatility may matter for the microcirculation and for renal perfusion.
What the evidence actually shows. The 10–15 mmHg figure and the 80%-of-venous-return convention are expert practice, reproduced consistently across texts but with no supporting outcome data and no comparative study of pulse-pressure targets. The overlap with the low-native-output warning threshold (§13.6.1) suggests the conventional target may sit closer to the danger zone than its confident presentation implies. Certainty: very low.
Where practice actually sits. Most units aim for some pulsatility and treat its complete loss as an event requiring action, rather than defending a specific number. The real target in practice is an aortic valve that opens, assessed on echocardiography, with the pulse pressure used as the continuous proxy between studies.
What would resolve it. A study relating pulse-pressure or aortic-valve-opening targets to LV distension, thrombosis and survival. The instrument exists — echocardiography — but no such trial is known here.
What to take to the bedside. Target the aortic valve opening, not the number. Use 10–15 mmHg as the floor that prompts you to look, and treat sustained loss of pulsatility as a reason to reduce flow, add an unloading strategy, or both (Chapter 15).
13.10 The errors that recur
Error | Correction |
Reading the PA catheter's cardiac output as the patient's cardiac output | It measures only the native pulmonary circulation, and unreliably. The wedge pressure is the useful number |
Treating hypotension with a vasopressor first | On VA, flow is the primary lever. Pressors raise the afterload the ventricle must overcome and reduce pump flow |
Reading falling pulsatility as one thing | It means either a worsening heart or an over-supported one. Reduce flow briefly and watch |
Trusting the arterial waveform in a patient with an IABP | The balloon generates a pulse regardless of ejection. Use echocardiography |
Reading a falling circuit flow as a circuit problem | On VA it may be recovery — rising MAP raises pump afterload and lowers flow. Check the pressure first |
Increasing flow to fix differential hypoxaemia without considering the ventricle | It moves the mixing point proximally and also increases afterload and distension risk (Chapters 15 and 16) |
Sampling arterial gases from a femoral or left-sided line | Right radial samples what the coronaries and brain receive |
Interrupting sweep gas on a VA circuit | It pumps deoxygenated blood into the aorta — the reverse harlequin effect. Never do it |
13.11 Key points
- Total flow is conserved; the pulse is not. Pulse pressure is a readout of how much blood still goes through the heart — the division of labour between patient and machine, not a measure of blood-pressure adequacy.
- VA ECMO conventionally runs at about 80% of venous return, leaving roughly 20% through the heart and lungs and a pulse pressure around 10 mmHg. At 100% the trace is flat and the patient is on full bypass.
- The mixing point moves. Improving native output pushes it distally — good for the heart, dangerous for the brain if the lungs are still failing.
- The afterload question has two published answers and they describe two different hearts. In a ventricle that still ejects, VA ECMO does not meaningfully raise afterload. In a severely failing ventricle with a competent aortic valve, it raises the pressure threshold for ejection and can stop it altogether. Manage as though the second is true.
- A preserved right ventricle makes left ventricular distension worse, because it keeps delivering volume the left ventricle cannot eject.
- Falling pulsatility has two opposite meanings. Reduce flow briefly and watch: prompt return means over-support; no return means deterioration.
- An IABP invalidates pulsatility as a measure of ejection. Assess aortic valve opening on echocardiography instead.
- The PA catheter measures the wrong circulation on VA, and measures it unreliably. Its wedge pressure remains useful.
- End-tidal CO₂ below 14 mmHg with pulse pressure below 15 mmHg flags a native cardiac output below 1 L/min — a free, continuous, two-variable warning, not a measurement.
- On VA ECMO, flow is the vasopressor. Use flow to control arterial pressure at 60–70 mmHg and keep vasoactive drugs minimal — reach for flow first, pressor second.
- Flow sits between two walls: too little and oxygen delivery fails; too much and the ventricle stops ejecting, distends, and floods the lungs.
- Falling circuit flow with rising blood pressure is recovery, not a fault — the centrifugal pump is afterload-sensitive. Check the pressure before working through Chapter 9's algorithm.
- Target the aortic valve opening, not the pulse-pressure number. Echocardiography is the reference standard for every question in this chapter.
[VERIFICATION REQUIRED] — open items in this chapter
- The afterload contradiction is real and is presented as found. Red Book Ch 5 and Ch 27 make opposing statements; neither cites supporting outcome data. The resolution offered in §13.3 — that the two describe hearts with different residual function — is this book's reasoning, not a published reconciliation, and is labelled as such. A reader who disagrees with the resolution still has both quotations in full.
- The ETCO₂ under 14 mmHg plus pulse pressure under 15 mmHg rule is cited by Red Book Ch 28 to a single reference not retrieved; no sensitivity, specificity, cohort or validation is given. "Good accuracy" is the source's phrase, not a measured performance.
- The 80% of venous return convention and the 10 mmHg resulting pulse pressure are from ISCCM Ch 26 and are stated without supporting evidence.
- The 10–15 mmHg pulse contour target, the 60–70 mmHg flow-controlled pressure, the DO₂:VO₂ ≥ 3 ratio, the 12–14 g/dL transfusion target and the venous saturation above 70% weaning threshold are all from Red Book Ch 5, presented there as a management scheme without trial support.
- The MAP above 65, SBP above 90, cardiac index above 2.2, PCWP below 15 targets are described in Red Book Ch 28 as "widely accepted" — a consensus claim, not a cited one.
- The left-heart pressure of 20–25 mmHg at which pulmonary oedema occurs is from ISCCM Ch 26, without a citation retrieved.
- The adult metabolic rate of 3–4 mL/kg/min for oxygen and carbon dioxide is a planning estimate from Red Book Ch 5.
- No randomised evidence supports any haemodynamic target in this chapter. The relevant trials (Chapter 11) tested whether to use VA ECMO at all, not how to set it.
- Taha's Chapter 6 pressure–volume loop figures were described from surrounding text, as the figures themselves could not be read.
Cross-references
- Chapter 2 — Cardiopulmonary Physiology for ECMO: Fick, DO₂–VO₂ and the critical delivery threshold that §13.6 applies
- Chapter 8 — Daily VV ECMO Management: the two-wall framing of flow, here reappearing with different walls
- Chapter 9 — Persistent Hypoxaemia on VV ECMO: the falling-flow algorithm, which applies on VA with the blood-pressure question added at the top
- Chapter 11 — VA ECMO: Indications and Patient Selection: severe aortic regurgitation as a contraindication — the same retrograde-pressurisation physics as §13.3
- Chapter 12 — VA ECMO Cannulation: the arterial site as the decision that sets up everything in this chapter, and the right radial line
- Chapter 14 — Initial VA ECMO Management: the hour-by-hour sequence that implements §13.6
- Chapter 15 — LV Distension and LV Unloading: the clinical problem that Position B in §13.3 describes
- Chapter 16 — Differential Hypoxaemia / Harlequin Syndrome: the consequence of the mixing point in §13.2
- Chapter 17 — Limb Ischaemia and Vascular Complications
- Chapter 18 — VA ECMO Weaning and Decannulation: where "pulse contour increases when flow is decreased" becomes a weaning criterion
- Chapter 30 — ECMO Data Interpretation
References
- Extracorporeal Life Support: The ELSO Red Book, 6th edition, Chapter 5, The Physiology of Extracorporeal Life Support. The afterload position A; the flow-setting sequence; metabolic rate planning; pulse contour target; DO₂:VO₂ ≥ 3; transfusion and venous saturation targets; recovery signalled by rising pulse contour at reduced flow.
- ELSO Red Book, 6th edition, Chapters 27 and 28, Initiating ECLS for Adult Cardiac Failure and Management of Adult Patients with Cardiac Failure. The afterload position B and the distension cascade; haemodynamic targets; PA catheter limitations; the ETCO₂ and pulse pressure rule; IABP false reassurance; cautions on vasopressors and inotropes.
- ISCCM Manual of RRT and ECMO in ICU, Chapter 26 (Physiology during ECMO Support) and Chapter 33 (Monitoring during ECMO). Total flow conserved with falling pulse contour; the 80% convention and ~10 mmHg pulse pressure; bronchial and thebesian filling with pulmonary oedema at 20–25 mmHg; the haemodynamic monitoring checklist; echocardiography as the best assessment of heart function. [VERIFICATION REQUIRED] — edition, editors, year and page numbers not confirmed.
- ECMO in the Adult Patient (Core Critical Care series), 2017, Chapter 4. Table 4.1, causes of changed arterial pressure waveform; the pink patient with an ischaemic ECG; falling ECMO flow as cardiac function improves; the warning never to interrupt sweep gas on a VA circuit and the reverse harlequin effect. [VERIFICATION REQUIRED] — editors, publisher and page numbers deliberately not stated; not verified in this session.
- Taha AR, Caridi-Scheible M, Leiendecker E, et al. ECMO: A Practical Guide to Management, Chapter 6, Physiology II: Venoarterial ECMO. Pressure–volume loop effects of afterload and inotropy; DO₂ = cardiac output × oxygen content and its modifiable variables. [VERIFICATION REQUIRED] — full editor list, publisher, year and page numbers not confirmed; figures not readable.
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.
The contradiction audit was the substantial one here. Two chapters of the ELSO Red Book state opposite positions on whether VA ECMO raises left ventricular afterload; both are quoted in full in §13.3, the disagreement is attributed to the different hearts they describe, and the reconciliation is explicitly labelled as this book's reasoning rather than a published one. A second, quieter tension — between the recommended pulse-pressure target and the pulse-pressure threshold used to warn of minimal native output — is surfaced in §13.6.1 rather than smoothed over. No haemodynamic target in this chapter has randomised support, and the chapter says so.
Addendum — 7 September 2026: one target in this chapter is no longer untested
§13.6 and the verification callout above state that no randomised evidence supports any target on VA ECMO, because the trials in Chapter 11 tested whether to use VA ECMO rather than how to set it. That is now partly wrong, and the correction is worth making explicitly.
The BLENDER trial (Burrell A, Bailey MJ, Bellomo R, et al., Intensive Care Medicine 2024;50(9):1470–1483, DOI 10.1007/s00134-024-07564-8) randomised 300 adults on VA ECMO to a conservative oxygen target (SaO₂ 92–96%) or a liberal one (97–100%). It found no difference in ICU-free days to day 28 (median treatment effect 0 days, 95% CI −3.1 to 3.1), in mortality at day 28 (39.6% versus 39.1%) or day 60, in duration of ECMO or ventilation, or in 6-month functional outcomes — and consistency across every prespecified subgroup.
It also found that the conservative target was difficult to deliver: 29.5% major protocol deviations versus 1.3%.
What this changes: the oxygenation target on VA ECMO now has randomised evidence, and that evidence is null. What it does not change: every other target in §13.6.1 — flow, MAP, pulse contour, DO₂:VO₂, transfusion threshold, venous saturation — remains untested by randomisation, and the statement in this chapter's verification callout stands for all of them.
Full treatment is in Chapter 14 §14.3, and the trial is entered in the Landmark Trial Database.