Chapter question: The ventricle is filling faster than it can empty. How do I know, what do I do first, and when does that stop being enough?
Evidence search date: 7 September 2026.
Primary sources: ELSO Red Book 6th edition Ch 4 (surgical venting routes), Ch 33 (cardiac complications), Ch 44 (other mechanical circulatory support, including Table 44-1); ISCCM Manual Ch 30 and Ch 41; Taha Ch 6; ECPR and Resuscitative ECMO Ch 7; ECMO Simulation Ch 22.
External evidence retrieved this session: two randomised trials (EARLY-UNLOAD, EVOLVE-ECMO), five meta-analyses, one international cohort study, one ELSO-registry target-trial emulation, and one physiological meta-analysis.
This chapter contains the sharpest evidence conflict in the book so far: a large and consistent observational literature says unloading saves lives; every randomised and quasi-randomised test of it is null. §15.7.3 sets out why both can be true.
What this chapter covers — and what it does not
This chapter owns | Deferred to |
Why the left ventricle distends and what actually fills it; the energetics of unloading; recognition and its thresholds; the differential diagnosis of pulmonary oedema on VA ECMO; the medical unloading sequence; the comparative menu of mechanical unloading strategies and how to choose; the observational-versus-randomised evidence conflict; timing; failure criteria and escalation | The afterload physiology itself → Chapter 13 §13.3
Pulsatility and what the waveform means → Chapter 13 §13.4
The initiation sequence → Chapter 14
Echocardiographic technique and protocols → Chapter 26
Cannulation routes and technique → Chapter 12
Running an ECMO + Impella circuit → Chapter 80
Running an ECMO + IABP circuit → Chapter 81
Hybrid and triple-cannulation configurations → Chapter 78
ECMO + durable LVAD → Chapter 79
Bridge decisions → Chapters 70–73 |
Clinical pearl — this chapter is where Chapter 13's contradiction becomes a decision
Chapter 13 §13.3 set out two published positions on whether VA ECMO raises left ventricular afterload, and resolved them by noting that they describe hearts with different residual function. That resolution is not re-argued here. This chapter starts from Position B — the severely failing ventricle with a competent aortic valve, pressurised from above — and asks the only question that matters at the bedside: what do I do about it?
It is worth noticing that the Red Book's own physiology chapter, the one that denies the afterload problem, nevertheless instructs: "if there is no LV function, establish left atrial drainage." Even the position that says VA ECMO does not overload the ventricle concedes the ventricle that cannot eject must be drained.
15.1 What actually fills a ventricle that cannot empty
The intuition most clinicians start with is that VA ECMO empties the heart. It drains the right atrium, so less blood reaches the left ventricle, so the left ventricle should be under-filled. That intuition is correct for the largest inflow and wrong overall, because it accounts for only one of four sources.
Physiology — the four inflows to a left ventricle on peripheral VA ECMO
- Residual transpulmonary flow. Venous blood not captured by the drainage cannula continues through the right ventricle and lungs into the left atrium. This is the dominant source, it is the one ECMO flow reduces, and — as Chapter 13 emphasised — a preserved right ventricle keeps delivering it.
- Bronchial venous return. Systemic arterial blood supplied to the airways drains into the pulmonary veins. It bypasses the ECMO drainage cannula entirely and is not reduced by raising flow.
- Thebesian venous return. Blood draining directly from the myocardium into the cardiac chambers. Small, but likewise uncapturable.
- Aortic regurgitation. Retrograde circuit flow pressurises the aortic root through diastole. Any aortic incompetence therefore delivers blood backwards into the ventricle at exactly the moment the pump is at its most effective. The ISCCM manual states that retrograde flow "enhances any prior aortic incompetence," and this is why moderate-to-severe aortic regurgitation is a contraindication to VA ECMO (Chapter 11).
The clinical consequence follows directly: raising ECMO flow reduces only inflow 1, while worsening the outflow obstruction for all four. There is no flow setting at which the left ventricle stops filling.
Danger — the cascade is fast and its end point is not survivable
The Red Book's cardiac complications chapter describes what happens if distension is left unattended: "blood continues to accumulate under pressure until ventricular and systemic pressures equalize. This can result in severe pulmonary edema, intracardiac thrombus, and fatal pulmonary hemorrhage."
The intermediate steps matter clinically because each is separately harmful and separately detectable:
— Rising LV end-diastolic pressure raises wall tension and myocardial oxygen demand while the coronary perfusion gradient falls, producing subendocardial ischaemia in the very ventricle you are trying to recover.
— Rising left atrial and pulmonary venous pressure produces pulmonary oedema. Chapter 13 recorded the ISCCM manual's figure: pulmonary oedema occurs at a left-heart pressure of 20–25 mmHg.
— Stasis in a motionless ventricle, outflow tract and aortic root produces thrombus. The ISCCM manual is blunt about the consequence: "little can be done once LV thrombus is formed," and weaning becomes difficult — though it notes that if the run was always a bridge to transplantation, a left ventricular thrombus matters less.
— Ventricular arrhythmia, which the ECPELLA literature lists as a distension consequence rather than a coincidence.
The failure mode is therefore not "the ventricle gets bigger." It is ischaemia of the organ you are supporting, flooding of the lungs you need for weaning, and thrombus that removes recovery from the list of possible outcomes.
15.2 What "unloading" means — and why the word hides three different things
Clinical writing uses unloading, venting and decompression interchangeably. They are not the same, and the difference determines which device is appropriate.
Physiology — pressure–volume area, and which part each device attacks
Myocardial oxygen consumption tracks the pressure–volume area (PVA), which has two components:
where SW is stroke work — the area inside the pressure–volume loop, the useful external work of ejection — and PE is potential energy, the area to the left of the loop, the energy stored in the wall at end-systole and dissipated as heat.
Peripheral VA ECMO does something specific and unhelpful to this: it reduces stroke work (less is ejected) while increasing potential energy (the ventricle sits at higher end-diastolic and end-systolic pressure and volume). The published account of the net effect is that PVA rises overall, so myocardial oxygen consumption rises, and it rises further as flow and afterload increase.
This is the precise sense in which VA ECMO is a mixed blessing to the myocardium: it guarantees coronary supply by maintaining aortic root pressure, while raising myocardial demand through wall tension.
The three verbs then separate cleanly:
— Volume decompression removes blood from the left heart. It lowers end-diastolic volume and pressure — it attacks potential energy.
— Afterload reduction lowers the pressure the ventricle must generate to eject. It attacks potential energy too, but by a different route, and it also restores stroke work by allowing ejection to resume.
— Preload interruption stops blood reaching the left heart at all. It attacks both stroke work and potential energy, which is why the reported energetic effect of atrial septostomy is the largest — and also why it does the least to preserve the ventricle's remaining useful function.
Certainty: moderate for the direction of these effects (consistent across simulation, animal and small human haemodynamic studies); very low for their clinical importance, since no study has shown that a given change in PVA translates into a change in recovery or survival.
Pitfall — an unloading device that does not reduce afterload does not stop the aortic valve staying shut
Left atrial drainage lowers filling pressures and clears the lungs. It does not lower the pressure the ventricle must overcome to open the aortic valve. A ventricle that was not ejecting before will frequently not be ejecting after — with a smaller cavity, better lungs, and the same stasis in the outflow tract and aortic root.
The authors of the pooled randomised analysis make exactly this point: left atrial venting carries "an increased risk of LV or aortic root thrombus formation in the absence of aortic valve opening."
Aortic valve opening is therefore a separate treatment target from filling pressure, and clearing the chest radiograph does not tell you that you have achieved it.
15.3 Recognition — and the uncomfortable fact that there is no gold standard
The single most important statement in the recent literature on this subject is not about a device. It is the observation, from the group that ran the largest randomised trial in the field, that there is an "absence of validated criteria for diagnosing LV overload." Everything below is therefore a set of signals of differing quality, not a diagnostic standard.
15.3.1 The signals, and what each is worth
Signal | What it means | How much to trust it |
Falling or absent pulse pressure | The aortic valve is opening less often, or not at all | The most continuously available signal, and the most commonly used trigger. But it is ambiguous — Chapter 13 §13.4: a narrowing pulse pressure means either a deteriorating ventricle or an over-supported one |
Aortic valve failing to open on echocardiography | The ventricle is not crossing the ejection threshold | The reference standard for the mechanism. Directly observes the thing that matters |
Rising LV end-diastolic dimension on serial echo | The cavity is enlarging | Specific but insensitive — the pressure–volume relationship is non-linear, so end-diastolic pressure can rise substantially before the dimension changes visibly |
Spontaneous echo contrast in the ventricle, outflow tract or aortic root | Stasis — the pre-thrombotic state | A late and serious sign. Its appearance changes the urgency |
Rising pulmonary artery diastolic pressure and PCWP | Left-heart filling pressure is climbing | Genuinely useful — Chapter 13 §13.5 noted the wedge is the one number on a PA catheter that survives VA ECMO. But see the PCWP–LVEDP caveat below |
Pulmonary oedema on chest radiograph | Filling pressure has exceeded the threshold | Described in the review literature as among "the most frequently encountered and convincing clinical signs of LV overload" — but it is a consequence, and it has a differential (§15.4) |
Frothy or pink secretions in the endotracheal tube | Alveolar flooding, or frank pulmonary haemorrhage | Same standing as the radiograph, and rather harder to ignore. A late sign |
Falling end-tidal CO₂ | Falling transpulmonary blood flow | Free and continuous. Chapter 13 §13.6.1 gives the combined rule: ETCO₂ below 14 mmHg with pulse pressure below 15 mmHg flags a native cardiac output below 1 L/min |
Falling lung compliance and worsening oxygenation | The lungs are wet | Non-specific on a ventilated ECMO patient, but the trend matters — and it has a downstream danger of its own (§15.4) |
New or refractory ventricular arrhythmia | A distended, ischaemic ventricle | Weak. Listed as an indication in reviews on the strength of case reports. Treat as supporting, never as sufficient |
15.3.2 The threshold problem
Four authoritative sources give four different pulse-pressure numbers, for four different purposes. Reading them as one number is the commonest conceptual error in this area.
Number | Source | What question it answers |
Pulse contour 10–15 mmHg | Red Book Ch 5, as a support target | How much of the work should the machine be doing? — a setting |
Pulse pressure 10–20 mmHg | ECPR and Resuscitative ECMO Ch 7: "about 10–20 mmHg of pulse pressure is sufficient to ensure that the LV is ejecting adequately"; its monitoring table adds that lack of pulsatility indicates need for an LV venting strategy | Is the ventricle ejecting at all? — an adequacy check |
Pulse pressure below 15 mmHg (with ETCO₂ below 14) | Red Book Ch 28 | Is native cardiac output below 1 L/min? — a low-output alarm |
Pulse pressure below 10 mmHg | Taha Ch 6 ("struggling to meet an arterial pulsatility of greater than 10 mmHg... LV venting should be utilized"); and the 2024 review literature, which reports that pulse pressure below 10 mmHg early in a VA ECMO run is associated with severe pulmonary oedema and poor prognosis | Should I vent this ventricle? — an intervention trigger |
Pitfall — the target and the trigger are ten millimetres of mercury apart, and the danger zone lies between them
Chapter 13 §13.6.1 flagged the overlap between the recommended target and the low-output warning. Adding the venting trigger completes an uncomfortable picture: the conventional pulse-contour target of 10–15 mmHg sits immediately above the pressure at which the literature says to start venting.
There is no contradiction — a target you are actively maintaining is different from a value you have drifted down to — but there is a practical consequence. A patient resting at the bottom of the recommended target range has no margin. Any further fall in contractility, any increase in flow, any rise in systemic resistance, moves them across the trigger.
Do not treat 10–15 mmHg as a comfortable place to sit and stop looking. Treat it as the range in which the aortic-valve-opening question must be asked on echocardiography, repeatedly.
15.3.3 Three traps in the measurements themselves
Danger — with an intra-aortic balloon pump in place, pulsatility is not evidence of ejection
Chapter 13 §13.4 recorded the Red Book's warning that pulsatility in the presence of an IABP "may provide false reassurance." This chapter must add the practical manoeuvre, because it is the exact circumstance in which the trap springs: a patient who has already been given an IABP to unload the ventricle now has a pulsatile trace that cannot tell you whether the unloading worked.
The ECMO simulation literature specifies the correction: interpretation of a pulsatile waveform in a patient on ECMO plus IABP should prompt the learner to ask for the IABP to be paused, and for simultaneous echocardiography, in order to demonstrate ejection and aortic valve opening.
Pause the balloon. Look at the valve. A device that generates a pulse and a ventricle that generates a pulse are indistinguishable on the arterial line and completely different clinically.
Pitfall — wedge pressure is not end-diastolic pressure on VA ECMO
Some units use PCWP above 15 mmHg as the unloading trigger — the same number the Red Book gives as a haemodynamic target (Chapter 13 §13.6.1), which makes it attractively simple.
The 2024 review literature warns that there is a significant discrepancy between PCWP and LVEDP in VA ECMO patients. Non-pulsatile retrograde flow, altered mitral inflow, changes in pulmonary vascular tone and the catheter position problems described in Chapter 13 §13.5 all degrade the relationship.
Use a rising wedge as a trend that should send you to the echocardiography machine. Do not use a single value as the decision.
Pitfall — a normal-looking left ventricle does not exclude a dangerously loaded one
Serial measurement of LV chamber size is the intuitive echocardiographic approach and is explicitly acknowledged in the review literature as insensitive, because the LVEDP–volume relationship is non-linear. A stiff, ischaemic, previously hypertrophied ventricle can reach a lethal end-diastolic pressure with a cavity that looks unremarkable.
The sensitive echocardiographic sign is not the size of the ventricle. It is whether the aortic valve opens, and on how many beats. Chamber dimension is the confirmatory measure and the tool for following response to treatment.
15.4 Not all pulmonary oedema on VA ECMO is left ventricular distension
Golden Rule 4 requires a differential, and here it is genuinely load-bearing: the three commonest alternatives all present with a wet chest radiograph on a VA ECMO patient, and two of them are made worse by the reflex response to distension.
Alternative | What distinguishes it | Why it matters |
Acute severe mitral regurgitation — ruptured papillary muscle | The simulation literature gives the discriminating picture precisely: disproportionately high wedge pressure and pulmonary oedema, lack of ejection, with or without hypotension, but an echocardiogram showing a well-decompressed left heart | The treatment is surgical, not a vent. A distension diagnosis here delays the operation |
Pulmonary haemorrhage from anticoagulation and thrombocytopenia | Frank blood rather than pink froth; falling haemoglobin; may occur with a decompressed ventricle. The Red Book reports pulmonary haemorrhage in 2.8% of adult VA ECMO patients in the 2017 ELSO Registry report | Adding an unloading device adds another large-bore access site and, for most devices, more anticoagulation demand |
Lung injury from the primary illness, or ventilator-associated | Compatible history; no rise in filling pressures; aortic valve opening preserved | The correct response is ventilator and infection management, not a cannula |
Pulmonary ischaemia from very low transpulmonary flow | High ECMO flow, minimal native output, and a Red Book–described mechanism of its own | The Red Book's stated remedies are to reduce ECMO flow to maintain partial pulmonary perfusion, or convert to a VVA configuration — the opposite of raising flow |
Fluid overload | Positive balance, raised central venous pressure, responsive to ultrafiltration | Treatable without any additional device (§15.5) |
Clinical pearl — the discriminating echo question takes ten seconds
Every alternative in the table above is separated from left ventricular distension by two observations that a competent bedside operator can make immediately: is the left ventricle big and full, and does the aortic valve open?
A large, stagnant ventricle with a closed aortic valve and a wet chest is distension. A small, well-decompressed left heart with a wet chest is not, whatever the wedge pressure says — and in that patient the mitral valve deserves a very careful look before anyone reaches for a vent.
15.5 The medical stage — what to do before anyone is called
The device literature has crowded out a simple point that both the textbooks and the trialists make: most left ventricular loading on VA ECMO is first addressed without an additional cannula. The group that ran EARLY-UNLOAD states its own position plainly: "we do not recommend prophylactic, routine LV unloading at the beginning of VA-ECMO treatment. Instead, we recommend using initial medical therapy."
Step | Why it works | What limits it |
1. Reduce ECMO flow to the lowest level that maintains adequate oxygen delivery | Directly lowers the aortic pressure the ventricle must overcome, and restores transpulmonary flow. Taha: careful titration of flows to the lowest acceptable level "can facilitate LV decompression indirectly through promotion of the heart's native ejection" | The floor is set by oxygen delivery (Chapter 2) and by the mixing point — lowering flow moves the watershed distally and can unmask differential hypoxaemia (Chapter 16) |
2. Reduce systemic vascular resistance — wean vasopressors, consider a vasodilator such as sodium nitroprusside | Chapter 13 §13.6.2: on VA ECMO, flow is the vasopressor. Every unit of resistance you add is afterload the ventricle must cross to open its valve | Genuine vasoplegia. Nitroprusside on a patient with renal or hepatic failure carries its own hazards |
3. Add or increase an inotrope — dobutamine, or an inodilator such as milrinone or levosimendan | Raises the pressure the ventricle can generate, so it crosses the ejection threshold. Both the Red Book and ISCCM list inotropic augmentation as a first-line unloading strategy | The Red Book's own caution: liberal use of inotropes may hamper myocardial recovery and raises myocardial oxygen demand in an ischaemic ventricle. This is a trade, not a free move |
4. Correct volume overload — diuresis or ultrafiltration | Lowers preload from every source at once. The Red Book notes strongly positive fluid balance is associated with poor outcome | Slow. Will not rescue an acutely distending ventricle within the hour |
5. Increase PEEP | Raises intrathoracic pressure, reduces transmural LV pressure and redistributes alveolar oedema. Named explicitly in the 2024 review literature as part of medical LV unloading | Reduces venous return and may reduce circuit drainage. Not a durable answer |
6. Restore an organised, ejecting rhythm | Chapter 13 §13.8 lists any rhythm with absent LV ejection as a direct route to distension and thrombus | Often the arrhythmia is a consequence of distension, which makes this circular. Treat both |
7. Improve venous drainage — an additional drainage cannula (VVA upgrade) | The ISCCM manual states it directly: "LV distension during VA ECMO is also an indication to improve the venous drainage to unload the left ventricle." Better right-heart drainage reduces inflow source 1 at its origin | Addresses only the transpulmonary inflow. Requires a further large-bore venous access placed under live imaging (Chapter 78) |
Danger — the reflex to raise flow is the wrong reflex
A patient on VA ECMO who becomes hypotensive, hypoxaemic or acidotic invites one response above all others: turn the flow up. In the distending ventricle this is precisely wrong, and it is wrong in a way that is initially rewarded — the mean arterial pressure improves while the ventricle is being shut.
Chapter 14 established the rule and this chapter is where it is enforced: when flow and ejection are in conflict, the answer is unloading, not more flow. The Red Book states the cost of higher flows explicitly — impaired cardiac ejection, left ventricular overload and pulmonary congestion.
The discriminating manoeuvre remains Chapter 13 §13.4's: reduce the flow briefly and watch the waveform. If pulsatility returns, the ventricle was being overwhelmed by support, and the treatment direction is down, not up.
15.6 The mechanical menu
When medical measures fail, the choice of device is usually presented as a list. It is more useful as a classification by which inflow or which energetic component the device actually attacks, because that is what predicts both the benefit and the failure mode.
15.6.1 Classified by mechanism
Strategy | Mechanism | Attacks | Key practical facts |
Intra-aortic balloon pump | Counterpulsation displaces aortic volume in diastole, lowering systolic afterload and augmenting diastolic coronary flow | Afterload (potential energy) | Simplest and most available; effect depends on residual native contractility. Smallest measured filling-pressure effect. Invalidates the arterial waveform as a measure of ejection (§15.3.3). Detail in Chapter 81 |
Micro-axial pump across the aortic valve (Impella; with ECMO = ECPELLA / ECMELLA) | Actively draws blood from the LV cavity and expels it into the ascending aorta, continuously and independently of rhythm | Volume decompression and forward flow | Reported sizes: 2.5 (2.5 L/min, 13 Fr), CP (3.5 L/min, 14 Fr), 5.0 (5.0 L/min, 23 Fr, surgical). Works regardless of native function. Also maintains aortic root washout. Highest complication burden. Detail in Chapter 80 |
Transseptal left atrial cannulation (active LA venting; LAVA-ECMO) | A cannula across the interatrial septum, Y-connected into the ECMO drainage limb, so both atria drain into the circuit | Preload only | The intervention tested in both randomised trials. No afterload reduction, so the aortic valve may stay shut (§15.2). The cannula comes out with the circuit — no support during weaning |
Percutaneous atrial septostomy | A balloon-created interatrial defect producing a left-to-right shunt | Preload — both stroke work and potential energy | Largest reported energetic effect; no added prosthetic material; animal data show it maintains LV haemodynamics even as ECMO flow rises. But shunt flow is unpredictable and durability uncertain; human data are small retrospective series |
Transaortic pigtail catheter into the LV | A drain placed retrogradely across the aortic valve | Volume decompression | 7 Fr, placed under fluoroscopic and transoesophageal echocardiographic guidance. Minimally invasive; evidence is case reports only |
Pulmonary artery drainage | A cannula in the pulmonary artery drains flow before it reaches the left heart | Preload (stroke work) and lung protection | Cannulae 10–31 Fr, including dual-lumen designs of 29–31 Fr. Direct lung protection, low thrombus risk; smallest LVEDP effect — one comparison reported 0.5 mmHg versus 5.4 mmHg for a micro-axial pump, a difference that did not reach significance (P=0.143) |
Surgical LV apical vent | Cannula through a small thoracotomy over the LV apex, incorporated into the circuit | Volume decompression, maximally | Largest cannulae, therefore best drainage and best unloading; the Red Book notes the trade directly — higher bleeding risk than minimally invasive strategies |
Surgical left atrial vent | Via the right superior pulmonary vein through a median sternotomy, or via the left atrial appendage through a small left thoracotomy; the Red Book notes the vent "can often be advanced across the mitral valve" | Preload, or preload plus decompression if advanced across the mitral valve | Natural choice when the chest is already open, particularly post-cardiotomy. Same bleeding and infection trade |
Conversion to central cannulation with a vent | Right atrial and aortic cannulation with an LV apex or pulmonary vein vent | Everything — drainage, antegrade flow and decompression | ISCCM: with central cannulation, "LV distension is also not a problem, as LV can be emptied by putting LV vent through LV apex or through the pulmonary veins," and the watershed is abolished. Costs: sternotomy, bleeding, infection, immobility (Chapter 12) |
Evidence — how much do these actually unload? The one direct physiological comparison
A systematic review and meta-analysis of the haemodynamic effect of unloading techniques during veno-arterial extracorporeal life support pooled 8 studies and 149 patients: IABP (1 study), micro-axial pump (2), left ventricular venting (1), and atrial septostomy (4).
Overall unloading effect on preload: standardised mean difference −1.05 (95% CI −1.24 to −0.86); ratio of means 0.60 (0.47 to 0.76) — that is, a 40% reduction in preload.
By technique: micro-axial pump SMD −1.11 (−1.55 to −0.68), ratio of means 0.58 (0.39–0.86); atrial septostomy SMD −1.22 (−1.47 to −0.96), ratio of means 0.54 (0.36–0.83); IABP a non-significant 21% reduction; LV venting a 22% reduction.
Meuwese CL, de Haan M, Zwetsloot PP, Braithwaite S, Ramjankhan F, van der Heijden J, Hermens J, Cremer O, Broomé M, Donker DW. Perfusion. 2020;35(7):664–671. DOI 10.1177/0267659119897478. Retrieved in full.
Certainty: low. The authors' own stated limitations are decisive: none of the included studies had a control group, there were few studies per category, designs were heterogeneous, the literature is dominated by case reports with publication bias favouring success, and — the point that matters most — no dose–response association was demonstrated between the degree of unloading achieved and mortality.
A separate comparison reported in the 2024 review literature gives filling-pressure reductions of IABP −3.9 ± 1.3 mmHg versus atrial septostomy −9.6 ± 2.5 mmHg versus transapical catheterisation 17.2 ± 2.1 mmHg (P<0.001). The third figure is printed without a minus sign in the source; in context it can only be a reduction of 17.2 mmHg, and it is reproduced here with that discrepancy flagged rather than silently corrected.
Clinical pearl — the ranking by unloading power is the inverse of the ranking by safety
Put the two orderings side by side. By measured unloading effect: surgical vent and atrial septostomy and micro-axial pump at the top, left atrial venting and pulmonary artery drainage in the middle, IABP at the bottom. By procedural risk, bleeding, haemolysis and access complications: almost exactly the reverse.
This is why no algorithm can produce a single right answer, and why the device that has the best outcome signal in the observational literature — the IABP — is the one with the weakest physiological effect. §15.7.3 returns to that paradox, because it is the key to the whole evidence base.
Pitfall — device choice is not only about the ventricle
Taha's list of the considerations in choosing a decompression device is worth reproducing because most of it is not physiological: availability of the device and of a qualified operator, the underlying pathology, concomitant pathology, the anticipated duration of recovery, and the management goal — bridge to recovery, to decision, to durable device or to transplantation.
A vent that must be removed at decannulation suits a patient expected to recover in days. A device that can support the circulation after the ECMO circuit is weaned suits a patient facing weeks. The Red Book makes this explicit for ECPELLA: it "provides an attractive option to wean the extracorporeal circuit as soon as practical, and to leave the patient solely on Impella support," and notes that full unloading by a small micro-axial pump is only achievable while part of the cardiac output is being taken by the ECMO circuit.
15.7 The evidence
15.7.1 What the observational literature says
Study | Size | Result |
Meta-analysis of 17 observational studies, J Am Coll Cardiol 2019, DOI 10.1016/j.jacc.2018.10.085 | 3,997 patients; 1,696 (42%) unloaded — IABP 91.7%, percutaneous VAD 5.5%, pulmonary vein or transseptal LA cannulation 2.8% | Mortality 54% with unloading versus 65% without; RR 0.79 (95% CI 0.72–0.87), P<0.00001. Haemolysis higher with unloading; other secondary outcomes not different |
Systematic review and meta-analysis, J Clin Med 2020, DOI 10.3390/jcm9041039 | 62 papers, 7,581 patients; 3,337 (44%) unloaded; overall in-hospital mortality 58.9% | RR 0.88 (0.82–0.93), P<0.0001; weaning RR 1.35 (1.21–1.51); largest benefit in acute myocardial infarction RR 0.75 (0.68–0.83); no difference in complications |
International multicentre cohort with propensity matching, Circulation 2020, DOI 10.1161/CIRCULATIONAHA.120.048792 | 686 patients at 16 centres in 4 countries; 337 (49%) unloaded with a micro-axial pump; 255 matched pairs | 30-day mortality HR 0.79 (0.63–0.98), P=0.03. Complications markedly higher: severe bleeding 38.4% versus 17.9%, access-site ischaemia 21.6% versus 12.3%, abdominal compartment 9.4% versus 3.7%, renal replacement therapy 58.5% versus 39.1% |
Timing analysis, JACC Heart Fail 2022, DOI 10.1016/j.jchf.2022.11.005 | 421 patients at 18 centres in 4 countries, all unloaded; 310 (73.6%) early (within 2 hours) | Early unloading HR 0.64 (0.46–0.88) for 30-day mortality and OR 2.17 (1.19–3.93) for successful weaning from ventilation, without more complications; risk rose and weaning fell almost proportionally with delay |
Meta-analysis restricted to RCTs and propensity-matched studies, Shock 2024, DOI 10.1097/SHK.0000000000002463 | 2 RCTs and 11 propensity-matched studies; 9,858 patients | Mortality RR 0.89 (0.84–0.94), P=0.0001, moderate certainty — confirmed in the IABP studies specifically. Weaning RR 1.15 (1.02–1.29, low certainty); favourable neurological outcome RR 2.45 (1.62–3.69, two studies, low); major bleeding RR 1.27 (1.02–1.59) and haemolysis RR 1.49 (1.10–2.02) |
Meta-analysis by device, Eur Heart J Open 2025, DOI 10.1093/ehjopen/oeaf103 | 26 studies, 22,625 patients | Unloading RR 0.80 (0.73–0.96); IABP RR 0.78 (0.69–0.89). Adverse effects comparable except in micro-axial pump patients — infection RR 1.37 (1.07–1.75) and renal replacement therapy RR 2.02 (1.37–3.00) |
15.7.2 What the randomised literature says
Trial | Design | Result |
EARLY-UNLOAD — Kim MC et al., Circulation 2023, DOI 10.1161/CIRCULATIONAHA.123.066179; NCT04775472 | Single-centre, open-label, 116 patients with cardiogenic shock on VA ECMO, Chonnam National University Hospital, South Korea, March 2021 – September 2022. Early routine transseptal LA cannulation within 12 h of randomisation (n=58, median 1.1 h) versus conventional care permitting rescue cannulation (n=58). Mean age 67.6 ± 13.5; 34 (29.3%) women | 30-day all-cause death 27 (46.6%) versus 26 (44.8%); HR 1.02 (95% CI 0.59–1.74), P=0.942. Crossover to rescue cannulation occurred in 29 of 58 (50%) conventional patients, at a median of 21.8 hours (IQR 12.4–52.2). The only significant secondary difference: faster resolution of pulmonary congestion — median 3 days (IQR 2–6) versus 5 (3–7), P=0.027 |
EARLY-UNLOAD 1-year — Lim Y, Kim MC, Lee SH, Park S, Ahn JH, Hyun DY, Cho KH, Jung YH, Jeong I-S, Ahn Y. Eur Heart J Acute Cardiovasc Care 2025;14(4):203–211, DOI 10.1093/ehjacc/zuae150 | Prespecified 1-year follow-up; 114 of 116 (98.3%) analysed | All-cause death 33/58 (56.9%) versus 32/56 (57.1%); HR 0.97 (0.60–1.58), P=0.887. No difference in cardiac or non-cardiac mortality, heart-failure rehospitalisation HR 1.17 (0.43–3.24), or the composite |
EVOLVE-ECMO — Park H et al., Eur J Heart Fail 2023, DOI 10.1002/ejhf.3014 | Two centres, 60 patients, December 2018 – August 2022. Transseptal LA cannulation at the time of ECMO insertion (n=30) versus conventional (n=30). Unloading actually delivered in 29 (96.7%) and 23 (76.7%); median time to unloading in the conventional arm 48.4 h (IQR 47.8–96.5) | ECMO weaning 70.0% versus 76.7%; RR 0.91 (0.67–1.24), P=0.386. Survival to discharge 53.3% versus 50.0%, P=0.796. Pulmonary congestion score improved significantly only in the early group (2.0 ± 0.7 to 1.7 ± 0.6 at 48 h, P=0.008) |
Pooled analysis of both trials — Ughetto A et al., Eur J Heart Fail 2024, DOI 10.1002/ejhf.3178 | 176 patients (88 versus 88) | No difference in in-hospital mortality, ECMO weaning, or bridge to durable device or transplant. Adverse events similar: stroke 16% versus 10%, limb ischaemia 8% versus 9%, tamponade 3% versus 1%, bleeding 16% versus 20% |
Meta-analysis of early versus bail-out timing, J Cardiothorac Vasc Anesth 2025, DOI 10.1053/j.jvca.2025.01.005 | 6 studies, 1,556 patients; 936 (60%) early | No difference in weaning, in-hospital mortality RR 0.95 (0.86–1.05), 30-day mortality RR 0.75 (0.52–1.10) or ECMO duration; no difference in sepsis, stroke or bridging |
15.7.3 Controversy 1 — Why does unloading work in cohorts and not in trials?
Controversy 1 — Does left ventricular unloading improve survival on VA ECMO?
The question. Six independent observational syntheses, covering more than 30,000 patients, find a consistent 11–21% relative mortality reduction with unloading. Two randomised trials and their pooled analysis find nothing. Which is right?
The case that unloading works. The signal is large, consistent across 26 to 62 studies in different syntheses, survives restriction to propensity-matched designs, is graded moderate certainty in the most methodologically careful meta-analysis, shows an internally coherent dose–response with timing in the largest timing cohort, and is supported by a mechanism that is not in doubt: distended ventricles become ischaemic, flood the lungs and form thrombus.
The case that it does not. Every randomised comparison is null, at 30 days and at one year, for mortality, weaning and bridging. Observational unloading is not a randomised exposure — the patient who is given a device is selected, and the patient who survives long enough to be given one is selected twice. Delayed unloading cannot occur in a patient who died in the first six hours, which inflates the apparent benefit of early unloading by immortal time alone.
What the evidence actually shows — and the resolution this book proposes. Three observations reconcile the two literatures, and none of them requires either to be wrong.
First, the two literatures did not test the same intervention. In the largest observational meta-analysis, 91.7% of the "unloading" was an intra-aortic balloon pump, and the device-stratified 2025 meta-analysis found the mortality signal was carried by IABP (RR 0.78) rather than by micro-axial pumps. Both randomised trials tested transseptal left atrial cannulation — a preload-only intervention that, as §15.2 argued, does not reduce afterload and does not necessarily reopen the aortic valve. A null trial of left atrial venting is not a null trial of afterload reduction.
Second, the two literatures did not enrol the same patients. The 2024 review reports that in the observational meta-analysis the indication for unloading was recorded for only 23% of patients, and of those only 15% were unloaded therapeutically — the rest were pre-emptive. The observational literature is therefore largely a comparison of prophylactic IABP versus nothing in unselected patients, while the trials asked does doing it early beat doing it when needed.
Third, and decisively, the control arms were not untreated. Half of EARLY-UNLOAD's conventional group crossed over to rescue cannulation "according to a clear indication." In EVOLVE-ECMO, 77% of the conventional group were eventually unloaded. These trials did not compare unloading with no unloading. They compared unloading everybody with unloading the people who need it — and found no difference. That is a clinically important finding, and it is not the finding "unloading does not work."
Certainty: moderate that unloading of some kind, in patients with established overload, is beneficial. High that routine prophylactic left atrial venting in unselected patients is not.
Where practice actually sits. Most units unload reactively, on evidence of overload, using whichever device the institution can deploy fastest — and this is exactly the strategy that both trials' control arms represent, and against which nothing has been shown superior.
What would resolve it. Two trials now running are designed for precisely this gap: UNLOAD-ECMO (NCT05577195, 198 patients, micro-axial pump plus VA ECMO versus VA ECMO alone, primary outcome 30-day all-cause death) and the REMAP-ECMO left ventricular unloading domain (NCT05913622, a Bayesian registry-embedded adaptive platform testing routine early IABP as an adjunct, primary outcome successful weaning at 30 days). HERACLES (ISRCTN82431978, 36 patients) compares micro-axial pump against IABP on coronary flow reserve. The REMAP domain is the important one, because it randomises the device the observational signal actually belongs to.
What to take to the bedside. Do not offer routine prophylactic left atrial venting to a patient with no evidence of overload — that has been tested twice and does not help. Do look hard and repeatedly for overload, and treat it when it appears; every trial's control arm did, in half to three-quarters of patients.
15.8 Timing — prophylactic, early, or therapeutic
The three-tier framework used by the EARLY-UNLOAD investigators is the clearest available way to organise the timing question, and each tier now has a different evidentiary standing.
Tier | Definition | Standing |
Prophylactic | Unloading placed before any evidence of overload | Tested and null for left atrial venting (EARLY-UNLOAD, EVOLVE-ECMO, pooled). Untested for IABP and micro-axial pumps |
Early | Mechanical unloading before or instead of a medical trial | Strong observational support (HR 0.64 within 2 hours), no randomised support. The review authors do not recommend it |
Therapeutic | Unloading for medically refractory signs of overload | The recommended approach in the current review literature, and the strategy every trial control arm actually implemented |
Pitfall — the patient who needs rescue unloading early is the sickest patient, not a treatment failure
A post-hoc analysis of EARLY-UNLOAD classified 108 per-protocol patients by unloading timing: prophylactic under 12 h (n=53), early rescue under 30 h (n=16), late rescue over 30 h (n=10), and none (n=29). Rescue unloading was performed in 47% of the conservative arm; median time from pump-on to rescue was 23 hours.
The early-rescue group had the highest in-hospital mortality (87.5%, P=0.004) and 30-day mortality (75%, P=0.029), and rescue within 30 hours carried an odds ratio of 6.87 (95% CI 1.31–35.92) for in-hospital death.
This must not be read as unloading causing harm. The same group had the highest baseline lactate (11.4 versus 4.3–6.3 mmol/L, P=0.011), the highest severity score (SAPS II 68.2 versus 50.6–55.7, P=0.017) and the highest day-1 inotrope requirement (vasoactive-inotropic score 31.2 versus 7.5–13.8, P=0.014). Needing an unloading device within the first day identifies a ventricle that failed to tolerate support at all.
Certainty: very low — a post-hoc subgroup analysis of 16 patients, presented in abstract form.
The bedside meaning is prognostic, not therapeutic. Early refractory overload is a marker of severity that should prompt an explicit conversation about escalation and about ceilings (§15.11), not a reason to withhold the vent.
15.9 ECPR is a separate question, and the two best studies disagree
Controversy 2 — Should the left ventricle be unloaded during extracorporeal CPR?
The question. ECPR patients arrive with a ventricle that has been arrested, is often not ejecting at all, and is being perfused retrogradely from the moment the pump starts. The distension risk is at its maximum. Does unloading help?
Position A — yes. A systematic review and meta-analysis of 13 study records from 32 hospitals, 1,014 ECPR patients (762 VA ECMO alone, 252 ECMELLA), found ECMELLA associated with lower mortality (OR 0.53, 95% CI 0.30–0.91) and higher odds of good neurological outcome (OR 2.22, 1.17–4.22), with numerically but not significantly higher complications. Crit Care Med 2024, DOI 10.1097/CCM.0000000000006157.
Position B — no, and it may harm. A target-trial emulation of the ELSO registry (2020–2023), 3,215 ECPR patients with 621 propensity-matched pairs, found no difference in survival to 90 days (HR 0.92, 0.79–1.08) or survival with favourable function (OR 1.15, 0.67–1.99), concordant across sensitivity analyses — and more complications: renal OR 1.55 (1.16–2.07), cardiovascular OR 1.60 (1.14–2.26), central nervous system bleeding OR 1.75 (1.03–2.96), arrhythmia OR 1.56 (1.04–2.36), haemolysis OR 1.85 (1.10–3.09). Crit Care 2025, DOI 10.1186/s13054-025-05345-3.
What the evidence actually shows. These are not equal. Position A's cohorts are overtly non-comparable on the variables that determine ECPR outcome: the ECMELLA group had more shockable rhythms (58.6% versus 49.3%), far more acute myocardial infarction (79.7% versus 51.5%) and far more percutaneous coronary intervention (79.0% versus 47.5%), while pulmonary embolism — a diagnosis with different prognosis — was concentrated in the ECMO-alone group (9.5% versus 0.7%). A patient with a shockable rhythm and a treatable coronary lesion is both more likely to be offered a micro-axial pump and more likely to survive. Position B's design was built specifically to remove that confounding, and the effect disappeared.
Certainty: low that unloading helps in ECPR; low-to-moderate that it does not and increases complications. Both are observational; the target-trial emulation is the better-designed of the two and it is null.
A single-centre retrospective series of 77 ECPR patients undergoing transseptal unloading pointed the same way, reporting higher 30-day (58.6% versus 33.3%) and 1-year (75.9% versus 52.7%) mortality with unloading within 12 hours. That comparison carries an obvious immortal-time problem — a patient classified as "conventional" had to survive 12 hours to be so classified — and is reported here for completeness rather than as evidence of harm.
What would resolve it. Randomisation within ECPR, stratified by initial rhythm and by whether a coronary lesion was treated. None is currently registered.
What to take to the bedside. Do not extrapolate the cardiogenic-shock unloading literature to the arrested heart. In ECPR, unload for demonstrated overload, expect a higher complication burden than the shock literature reports, and do not add a device on the assumption that the arrest itself is an indication.
15.10 A working framework
No published algorithm has outcome data behind it. What follows is this book's synthesis of the sources above, offered as a sequence of questions rather than a protocol, and labelled as reasoning rather than evidence.
Clinical pearl — six questions that settle most of these decisions
- Is the ventricle actually the problem? Big and stagnant with a shut valve, or small and decompressed with wet lungs (§15.4).
- Have I done the free things? Flow down, pressors down, inotrope up, fluid off, PEEP up (§15.5). Most of these cost nothing and can be done in minutes.
- Am I treating pressure or ejection? Left atrial drainage clears the lungs; only afterload reduction reopens the valve (§15.2).
- How long will this patient need support, and what happens at decannulation? A transseptal cannula leaves with the circuit; a micro-axial pump does not.
- What can this institution actually deliver tonight, and who will place it?
- If this does not work, what is the next step — and is there one? Decide before, not after (§15.11).
15.11 Failure criteria and escalation
Unloading can fail in three distinguishable ways, and they have different answers.
Failure mode | How it presents | Response |
Mechanical failure of the vent | Drainage falls, filling pressures rise again; in small cannulae, clotting. The ISCCM manual describes vent clotting as filling pressures fall with recovery, and a technique of partially occluding the main drainage cannula to preserve vent flow | Confirm cannula position on imaging; check for circuit thrombus (Chapter 31); consider a larger or different route |
Physiological failure — the ventricle is decompressed but still not ejecting | Lungs clear, filling pressures fall, aortic valve still shut, spontaneous echo contrast persists | The intervention addressed preload only. Add afterload reduction or a device that generates forward flow, and re-examine anticoagulation because the thrombotic risk is undiminished |
Failure of the underlying heart | Unloading is adequate, the lungs are dry, and there is still no contractile recovery over days | This is no longer an unloading problem. It is a destination problem — Chapters 70–73 |
Evidence — the escalation criteria the Red Book gives
The Red Book's chapter on other mechanical circulatory support lists the triggers for escalating beyond VA ECMO alone: persistent shock, pulmonary oedema, right ventricular failure, a vasoactive-inotropic score above 20, and inability to wean inotropes. It distinguishes univentricular from biventricular escalation, and compares durable and micro-axial options against VA ECMO — noting that a surgically placed micro-axial pump can deliver full support even without residual left ventricular function, may carry fewer vascular complications, and permits longer support and therefore a genuine bridge-to-decision window.
Certainty: very low. These are expert criteria presented without supporting outcome data, and the vasoactive-inotropic score threshold in particular is a convention rather than a validated cut-off.
Danger — every unloading device is a second bleeding and access problem
The complication figures from the propensity-matched international cohort are the ones to keep in mind before adding a device: severe bleeding 38.4% versus 17.9%, access-site ischaemia 21.6% versus 12.3%, abdominal compartment syndrome 9.4% versus 3.7%, renal replacement therapy 58.5% versus 39.1%. The device-stratified meta-analysis adds infection RR 1.37 and renal replacement RR 2.02 for micro-axial pumps specifically, and the RCT-and-propensity-matched synthesis adds major bleeding RR 1.27 and haemolysis RR 1.49 across all modalities.
Chapter 17's limb-ischaemia arithmetic changes materially when a second arterial device shares the groin. If a micro-axial pump and an ECMO arterial cannula are going into the same limb, the distal perfusion question must be settled at the time of insertion, not afterwards — and axillary insertion exists precisely for this reason.
15.12 The errors that recur
Error | Correction |
Raising ECMO flow because the patient looks worse | It reduces one of four inflows while worsening the outflow obstruction for all of them. Lower the flow and watch the waveform |
Trusting a pulsatile arterial trace in a patient who has an IABP | Pause the balloon and look at the aortic valve on echocardiography |
Using LV chamber size as the screening measure | It is insensitive. Aortic valve opening is the sensitive sign; chamber size is confirmatory |
Treating a single wedge-pressure value as the decision | Wedge and end-diastolic pressure diverge on VA ECMO. Use the trend to trigger an echo |
Assuming a clear chest radiograph after left atrial venting means the ventricle is unloaded | Preload relief without ejection leaves aortic root stasis untouched. Ask about the valve, not the lungs |
Reading the observational meta-analyses as evidence for the device in front of you | Over 90% of that literature is intra-aortic balloon pump. It is not evidence about micro-axial pumps or transseptal cannulae |
Reading the null randomised trials as "unloading does not work" | Half to three-quarters of both control arms were unloaded on indication. The trials tested routine early versus on indication, and found them equivalent |
Extrapolating the cardiogenic-shock unloading data to ECPR | The best-designed ECPR analysis is null and shows more complications (§15.9) |
Diagnosing distension in a patient with pulmonary oedema and a small left ventricle | Look at the mitral valve. Acute severe regurgitation presents exactly this way and needs an operation |
Adding a second arterial device without settling distal perfusion | Chapter 17. The limb has now got two cannulae competing for one artery |
15.13 Key points
- Four things fill a left ventricle on VA ECMO — residual transpulmonary flow, bronchial venous return, thebesian venous return and any aortic regurgitation. Raising flow reduces only the first while worsening the outflow obstruction for all four.
- A preserved right ventricle makes distension worse (Chapter 13). The combination to fear is a strong right ventricle, a weak left ventricle, a competent aortic valve and high flow.
- VA ECMO lowers stroke work but raises potential energy, so the net pressure–volume area — and therefore myocardial oxygen consumption — rises. It guarantees coronary supply while raising myocardial demand.
- Unloading, venting and decompression are three different actions. Preload interruption clears the lungs; only afterload reduction reopens the aortic valve.
- There are no validated diagnostic criteria for left ventricular overload. Every threshold in use is expert convention.
- The published pulse-pressure numbers answer four different questions — 10–15 mmHg is a setting, 10–20 mmHg an adequacy check, below 15 mmHg with a low end-tidal CO₂ a low-output alarm, and below 10 mmHg an intervention trigger. The conventional target sits immediately above the venting trigger.
- Aortic valve opening is the sensitive echocardiographic sign; chamber size is not, because the pressure–volume relationship is non-linear.
- An intra-aortic balloon pump makes the arterial waveform uninterpretable. Pause it and image the valve.
- Not all pulmonary oedema on VA ECMO is distension. A wet chest with a small, well-decompressed left heart should send you to the mitral valve.
- Do the medical steps first — flow down, pressors down, inotrope up, fluid off, PEEP up, rhythm restored, drainage improved. The trialists themselves recommend this.
- Unloading power and safety rank in opposite orders, and the device with the best outcome signal in the observational literature — the balloon pump — has the weakest measured physiological effect.
- Six observational syntheses covering more than 30,000 patients show a mortality benefit; two randomised trials show none. They tested different devices, in different patients, against control arms that were themselves unloaded half to three-quarters of the time.
- Routine prophylactic left atrial venting does not help. That specific question has been asked twice and answered.
- Unloading on indication is the strategy every control arm implemented, and nothing has been shown superior to it.
- In ECPR, the best-designed analysis is null and shows more complications. Do not extrapolate the shock literature to the arrested heart.
- Needing rescue unloading in the first day is a severity marker. It should prompt a conversation about escalation and ceilings, not a reassessment of whether to place the vent.
- Every unloading device adds a bleeding, haemolysis and access-ischaemia burden, and a second arterial device in the same groin makes distal perfusion an immediate question, not a later one.
[VERIFICATION REQUIRED] — open items in this chapter
- The framework in §15.10 is this book's synthesis, not a published algorithm. No unloading algorithm anywhere has outcome data behind it, and this one is offered as a sequence of questions rather than a protocol.
- The reconciliation of the observational and randomised literatures in §15.7.3 — different devices, different indications, unloaded control arms — is this book's reasoning. The three underlying facts are cited; the synthesis is not a published one.
- The 17-study meta-analysis (DOI 10.1016/j.jacc.2018.10.085) is attributed to Russo et al. in the peer-reviewed review used here; a bibliographic index returned the same DOI under Aleksova as first author. The DOI, journal, year, sample size and results agree across both; first authorship is not independently verified and no author list is asserted.
- The filling-pressure comparison quoted in §15.6.1 prints the transapical figure as 17.2 ± 2.1 mmHg without a minus sign where the context requires a reduction. The discrepancy is flagged, not corrected, and the figure should be checked in the primary source before being quoted.
- The class II, level C expert recommendation to consider early unloading, referred to in the pooled randomised analysis, is cited there to a document not retrieved in this session. The issuing body and exact wording are not confirmed and are therefore not reproduced.
- The pulmonary haemorrhage rate of 2.8% is from the 2017 ELSO Registry report as cited by the Red Book; the registry report itself was not retrieved.
- The EARLY-UNLOAD post-hoc timing analysis (§15.8) is a conference abstract describing 16 patients in the critical subgroup. It is reported for its interpretive value and should not be treated as a finding.
- Device specifications (micro-axial pump 2.5 / CP / 5.0 flows and French sizes; 7 Fr pigtail; 10–31 Fr pulmonary artery cannulae; 29–31 Fr dual-lumen) are as reported in the 2024 review literature and were not checked against manufacturer instructions for use. Verify locally before purchase or insertion.
- The vasoactive-inotropic score above 20 escalation threshold is a Red Book expert criterion without cited supporting data.
- The ISCCM Manual, Red Book, Taha, ECPR and Resuscitative ECMO and ECMO Simulation volumes are cited by chapter; editions, editors, years and page numbers are not confirmed and are deliberately not reproduced from memory.
- UNLOAD-ECMO, REMAP-ECMO and HERACLES are described from registry and design-paper information. Enrolment status and any results at the September 2026 search date were not confirmed; UNLOAD-ECMO's listed completion date of 1 December 2025 has passed and no result was retrieved.
Cross-references
- Chapter 11 — VA ECMO: Indications and Patient Selection: aortic regurgitation as a contraindication — inflow source 4 in §15.1
- Chapter 12 — VA ECMO Cannulation: central versus peripheral, and the arterial site decision that determines the distension risk
- Chapter 13 — VA ECMO Haemodynamics: §13.3 the afterload contradiction and Position B; §13.4 pulsatility and the IABP warning; §13.5 the PA catheter; §13.6.1 the targets
- Chapter 14 — Initial VA ECMO Management: the initiation sequence, and the rule that when flow and ejection conflict the answer is unloading
- Chapter 16 — Differential Hypoxaemia: the opposite consequence of the same mixing physics, and the reason lowering flow is not free
- Chapter 17 — Limb Ischaemia and Vascular Complications: a second arterial device in the same limb
- Chapter 18 — VA ECMO Weaning and Decannulation: what an unloading device does, or does not, contribute during weaning
- Chapter 26 — Echocardiography During ECMO: how to make the aortic-valve-opening assessment properly
- Chapter 31 — Circuit Thrombosis: vent clotting and circuit-side thrombus
- Chapters 70–73 — Bridge to Decision, Recovery, Transplant and LVAD: where the failure-of-the-heart pathway in §15.11 leads
- Chapter 78 — Hybrid ECMO Configurations: the VVA upgrade in §15.5
- Chapter 80 — ECMO + Impella and Chapter 81 — ECMO + IABP: the operational detail of the two devices this chapter compares
References
Randomised trials
- Kim MC, et al. Early Left Ventricular Unloading or Conventional Approach After Venoarterial Extracorporeal Membrane Oxygenation: The EARLY-UNLOAD Randomized Clinical Trial. Circulation. 2023. DOI 10.1161/CIRCULATIONAHA.123.066179. NCT04775472.
- Lim Y, Kim MC, Lee SH, Park S, Ahn JH, Hyun DY, Cho KH, Jung YH, Jeong I-S, Ahn Y. Early left ventricular unloading after venoarterial extracorporeal membrane oxygenation: 1-year outcomes of the EARLY-UNLOAD randomized clinical trial. European Heart Journal: Acute Cardiovascular Care. 2025;14(4):203–211. DOI 10.1093/ehjacc/zuae150.
- Park H, et al. Early left atrial venting versus conventional treatment for left ventricular decompression during venoarterial extracorporeal membrane oxygenation support: the EVOLVE-ECMO randomized clinical trial. European Journal of Heart Failure. 2023. DOI 10.1002/ejhf.3014.
Syntheses
- Ughetto A, et al. Early left ventricular unloading via active transseptal left atrial venting in case of cardiogenic shock under veno-arterial extracorporeal membrane oxygenation: a meta-analysis. European Journal of Heart Failure. 2024. DOI 10.1002/ejhf.3178.
- Meuwese CL, de Haan M, Zwetsloot PP, Braithwaite S, Ramjankhan F, van der Heijden J, Hermens J, Cremer O, Broomé M, Donker DW. The hemodynamic effect of different left ventricular unloading techniques during veno-arterial extracorporeal life support: a systematic review and meta-analysis. Perfusion. 2020;35(7):664–671. DOI 10.1177/0267659119897478.
- Left Ventricular Unloading During Extracorporeal Membrane Oxygenation in Patients With Cardiogenic Shock. Journal of the American College of Cardiology. 2019. DOI 10.1016/j.jacc.2018.10.085. [VERIFICATION REQUIRED] — first author disputed between sources.
- Kowalewski M, et al. Left Ventricle Unloading with Veno-Arterial Extracorporeal Membrane Oxygenation for Cardiogenic Shock: Systematic Review and Meta-Analysis. Journal of Clinical Medicine. 2020;9:1039. DOI 10.3390/jcm9041039.
- Kotani Y, et al. Mechanical Left Ventricular Unloading in Cardiogenic Shock Treated with Venoarterial Extracorporeal Membrane Oxygenation: A Systematic Review and Meta-Analysis. Shock. 2024. DOI 10.1097/SHK.0000000000002463.
- Dar B, et al. Left ventricular unloading in patients with cardiogenic shock treated with veno-arterial extracorporeal membrane oxygenation. European Heart Journal Open. 2025. DOI 10.1093/ehjopen/oeaf103.
- Picado-Loaiza S, et al. Early Versus Bail-Out Left Ventricular Unloading During Venoarterial Extracorporeal Membrane Oxygenation: A Systematic Review and Meta-Analysis. Journal of Cardiothoracic and Vascular Anesthesia. 2025. DOI 10.1053/j.jvca.2025.01.005.
- Thevathasan T, et al. Left-Ventricular Unloading With Impella During Refractory Cardiac Arrest Treated With Extracorporeal Cardiopulmonary Resuscitation: A Systematic Review and Meta-Analysis. Critical Care Medicine. 2024. DOI 10.1097/CCM.0000000000006157.
Cohorts
- Schrage B, et al. Left Ventricular Unloading Is Associated With Lower Mortality in Patients With Cardiogenic Shock Treated With Venoarterial Extracorporeal Membrane Oxygenation. Circulation. 2020. DOI 10.1161/CIRCULATIONAHA.120.048792.
- Schrage B, et al. Timing of Active Left Ventricular Unloading in Patients on Venoarterial Extracorporeal Membrane Oxygenation Therapy. JACC: Heart Failure. 2022. DOI 10.1016/j.jchf.2022.11.005.
- Ling R, et al. Left ventricular unloading during extracorporeal cardiopulmonary resuscitation: a target trial emulation of the ELSO registry. Critical Care. 2025;29. DOI 10.1186/s13054-025-05345-3.
- Kim A, et al. Comparison of Early Versus Conventional Left Ventricle Unloading in Patients Undergoing Extracorporeal Cardiopulmonary Resuscitation. Artificial Organs. 2026. DOI 10.1111/aor.70111.
Reviews and design papers
- Lim Y, Kim MC, Jeong I-S. Left ventricle unloading during veno-arterial extracorporeal membrane oxygenation: review with updated evidence. Acute and Critical Care. 2024;39(4):473–487. DOI 10.4266/acc.2024.00801. Source for the three-tier timing framework, device descriptions and specifications, thresholds, and the incidence figures in §15.7.3.
- van Steenwijk MV, et al. A Randomized Embedded Multifactorial Adaptive Platform for Extracorporeal Membrane Oxygenation (REMAP ECMO): design and rationale of the left ventricular unloading trial domain. American Heart Journal. 2024. DOI 10.1016/j.ahj.2024.10.010. NCT05913622.
Textbooks
- ELSO Red Book, 6th edition, Chapter 4 (cannulation — LV apical and left atrial vent routes), Chapter 33 (cardiac complications — the distension cascade, monitoring, pulmonary haemorrhage rate), Chapter 44 (other mechanical circulatory support — ECPELLA, Table 44-1, escalation criteria). [VERIFICATION REQUIRED] — page numbers not confirmed.
- ISCCM Manual of RRT and ECMO in ICU, Chapter 30 (cannulation and the VVA upgrade for distension) and Chapter 41 (cardiac issues — distension mechanism, consequences, the seven-item strategy list, thrombus). [VERIFICATION REQUIRED].
- Taha AR, Caridi-Scheible M, Leiendecker E, et al. ECMO: A Practical Guide to Management, Chapter 6. When venting is not indicated; the 10 mmHg trigger; device-selection considerations. [VERIFICATION REQUIRED].
- Shinar Z, Badulak J, eds. ECPR and Resuscitative ECMO, Chapter 7. The 10–20 mmHg pulse pressure rule and the monitoring table. [VERIFICATION REQUIRED].
- Johnston LC, Su L, eds. Comprehensive Healthcare Simulation: ECMO Simulation, Chapter 22. Pausing the balloon pump to demonstrate ejection; the ruptured papillary muscle picture; ECPELLA complications. [VERIFICATION REQUIRED].
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.
Two contradiction audits were substantial. The first is the observational-versus-randomised conflict in §15.7.3, which is presented in full rather than resolved by preferring one literature — with the three reconciling facts (different devices, different indications, unloaded control arms) cited and the synthesis labelled as this book's reasoning. The second is the ECPR disagreement in §15.9 between a positive meta-analysis and a null target-trial emulation, where the chapter states plainly that the two are not of equal design quality and says why.
The numerical audit found one sign discrepancy in a published figure, which is reproduced with the discrepancy flagged rather than silently corrected, and one authorship discrepancy between two sources for the same DOI, which is recorded rather than resolved.
The redundancy audit removed the afterload argument itself, which belongs to Chapter 13 §13.3, and the operational detail of the two named devices, which belongs to Chapters 80 and 81.
Addendum — 7 September 2026: §15.5 step 1 has a cost this chapter did not price
§15.5 makes "reduce ECMO flow to the lowest level that maintains adequate oxygen delivery" the first medical step for a loaded left ventricle, and names the floor as oxygen delivery and the mixing point, with a forward pointer to Chapter 16. Chapter 16 is now written, and the pointer needs to be sharper than it was.
Lowering the flow does not merely risk unmasking differential hypoxaemia as a side effect. It moves the mixing point distally, which is the exact mechanism of the syndrome — and the syndrome's victims are the coronary and cerebral circulations, in a patient whose lungs are by definition failing.
The one controlled experiment in this area (Rozencwajg S, Heinsar S, Wildi K, et al., Scientific Reports 2023;13:4002, DOI 10.1038/s41598-023-30226-6) randomised sheep with cardiogenic shock and respiratory failure on femoral VA ECMO to 2.5 versus 4.5 L/min. The low-flow group had brain tissue oxygen tension −58% from baseline versus +215%, cerebral NIRS 49 ± 4% versus 67 ± 5% (p=0.003), and a significantly worse global histological brain injury score (p=0.0003) after five hours. The authors also state the counterweight themselves: they did not measure mortality or ventricular injury, so the cerebral benefit of higher flow "may not outweigh LV damage from increased afterload."
What this changes. Nothing in §15.5's ordering — the medical stage is still right, and lowering flow is still the correct first move for a distending ventricle. What it adds is a mandatory second observation. §15.5 step 1 should be read as: reduce the flow, and check the right radial gas as well as the aortic valve before and after.
The unifying statement, set out in full in Chapter 16 §16.8: LV distension and differential hypoxaemia are the same physiology read from opposite ends, both caused by a ventricle ejecting poorly oxygenated blood into a pressurised aorta. The flow dial only trades one organ against the other. Only fixing the lung, or changing the circuit, treats the cause.