<aside> 🧭

Chapter question: The cannulae are in and the pump is running. What happens in the next few hours?

</aside>

<aside> 📅

Evidence search date: 7 September 2026. This chapter contains the first randomised evidence in Part III about how to set VA ECMO rather than whether to use it — the BLENDER trial (Burrell et al., Intensive Care Medicine 2024), retrieved in full from the publisher. Other sources: ISCCM Manual Ch 30 (initiation sequence and safety checklist); ELSO Red Book Ch 32 (immediate post-cannulation actions); ECMO in the Adult Patient Ch 6 (cannulation checklists).

</aside>

What this chapter covers — and what it does not

This chapter owns Deferred to
Starting the circuit; the first flow titration and the hypotension that follows it; oxygen targets after BLENDER; CO₂ and sweep in the first hours; the immediate post-cannulation checklist; the distal perfusion deadline; ventilator and sedation in the first 24–72 hours; what must be established before the team disperses Who to cannulate → Chapter 11
Cannulation technique → Chapter 12
The physiology behind every setting here → Chapter 13
LV distension and unloading → Chapter 15
Differential hypoxaemia → Chapter 16
Limb ischaemia management → Chapter 17
Weaning → Chapter 18
ECPR-specific post-arrest care → Chapters 19–20
Anticoagulation in depth → Part VII
The daily round, once established → Chapter 8

14.1 Starting the circuit

The ISCCM manual gives an explicit initiation sequence. The order matters, because opening the wrong clamp first can drive blood backwards through a pump that is not yet generating forward flow.

Before starting: work through the initiation safety checklist (§14.6) and ensure the activated clotting time is above 250 seconds.

  1. Turn on the pump and increase speed to about 500 RPM — enough to provide forward flow before any clamp comes off.
  2. Open the clamp on the venous drainage line, and increase pump speed gradually to about 1,500 RPM.
  3. Remove the clamp from the blood outflow line, and increase pump speed further.
  4. Start sweep gas at 5–6 L/min, keeping the gas-flow to blood-flow ratio at 1:1. Monitor gas inlet pressure — ideally around 40–45 psi.

<aside> ⚠️

Danger — never let the pump sit stopped with the circuit unclamped

A centrifugal pump generates no forward flow at rest and does not occlude the circuit. With the pump stopped and both clamps off, blood flows backwards — from the pressurised arterial side, through the pump head, and back into the venous system. On VA this is a massive left-to-right shunt (Chapter 9).

This is why the sequence above brings the pump up to speed first and takes the clamps off after, and why it opens the drainage clamp before the return clamp. The same logic runs in reverse at every emergency: clamp, then stop.

</aside>

14.2 The first flow titration — and the hypotension that follows

The target. The ISCCM manual sets the initial VA target as the flow needed to maintain a cardiac index of 2.0–2.4 L/min/m². Chapter 13 gives the alternative formulation from the Red Book — go to maximum flow first to establish drainage capacity, then back off until the pulse contour is 10–15 mmHg.

<aside> 💡

Clinical pearl — the two published approaches answer different questions, and you need both

The cardiac index target answers "is oxygen delivery adequate?" The pulse contour target answers "is the ventricle still ejecting?" Neither alone is sufficient, because the flow that satisfies the first can abolish the second.

In practice: set flow initially for the cardiac index, then look at the arterial waveform and the aortic valve on echocardiography and ask whether you have bought that index at the cost of ejection (Chapter 13 §13.6). If you have, the answer is not more flow — it is an unloading strategy (Chapter 15).

</aside>

<aside> 🫀

Physiology — why the blood pressure falls as you increase flow, and what to do

The ISCCM manual describes it plainly: as pump speed is increased, venous drainage may fall and hypotension results. This surprises people, because more flow ought to mean more pressure.

The mechanism is preload. The circuit drains the venous reservoir faster than it refills; the drainage cannula begins to collapse the vein around it, and drainage — and therefore total flow — falls. What looks like a pump problem is a volume problem.

The source's sequence: exclude cannula malposition first, then give volume — packed red cells, albumin or crystalloid — as boluses of 5–10 mL/kg, with explicit care to avoid fluid overload.

Note how this differs from VV (Chapter 7), where hypotension at initiation always has another cause because VV ECMO is haemodynamically neutral. On VA, hypotension during the first titration usually is the circuit — specifically, its appetite for preload.

</aside>

<aside> 🕳️

Pitfall — the 5–10 mL/kg bolus is a diagnostic, not a strategy

The warning to avoid fluid overload is not decorative. Chapter 11 recorded that strongly positive fluid balances are associated with poor outcome on VA support, and Chapter 9 recorded the same argument on the VV side. A patient who needs repeated boluses to hold flow is telling you something — about drainage cannula size, position, or the possibility that the flow target itself is too ambitious for this patient's venous return.

Give volume to establish flow. Do not keep giving volume to defend a number.

</aside>