Author: Adam Levin
Peer reviewers: Sarah Yong, Aidan Burrell, Arne Diehl, Chris Nickson

Everything ECMO 016

A 52-year-old man was commenced on VA ECMO for cardiogenic shock following a large anterior STEMI. He is intubated and ventilated, and requiring low dose inotropes. He is anticoagulated on an heparin infusion.

On day 4 of his admission, his free Hb and D-dimer have increased. You notice the ECMO flows have fallen from 3.5 L/min to 2.5L/min.

Figure 1. Low (2.52 L/min) circuit flows at 4000 rpm.

Q1. What is your approach to this problem?

This scenario is highly suggestive of severe haemolysis.

  • Haemolysis in an ECMO patient occurs due to circuit or patient-related problems.

Important differentials for ECMO-related haemolysis include:

  • Pump head thrombosis
  • Oxygenator thrombosis
  • Access insufficiency

An approach to this includes:

  • Assess for access insufficiency.
  • Assess the circuit to look for any clot burden in pump, oxygenator or tubing
  • Consider ultrasound to exclude clot at cannula sites
  • X-rays to confirm optimal cannula positioning
  • Consider other non ECMO cause of haemolysis

You look closer and note multiple clots in the oxygenator.

His APTT is 64 and stable.

See also Everything ECMO 004 for more information about ECMO-related haemoylsis. A succinct review of haemolytic anaemia can be found here: LITFL CCC – Haemolytic Anaemia.

Q2. What would you do now?

You change the circuit (see Everything ECMO 012 for more information on ECMO circuit changes).

His ECMO flows immediately return to normal and his biochemistry improves over the following two days.

4.4 L/min ECMO flow

Figure 2. Higher (4.44 L/min) circuit flows at 4000 rpm following ECMO circuit change.

ECMO oxygenator clot

Figure 3. Evidence of oxygenator thrombosis, image taken following replacement with a new oxygenator.

The patient remains stable on VA ECMO support over the next few days.

However on day 7, ECMO blood flows start to decrease again. Further clots are noted in the oxygenator. Plasma free Hb, LDH and d-dimer once again start to climb. The patient is now also thrombocytopaenic, with a nadir of 23 x 109/L platelets.

You decide to rapidly exchange a second circuit.

Q3. What are the potential causes of this?

The patient has developed recurrent clot formation, leading to haemolysis and circuit loss. In addition there is now thrombocytopenia.

Differentials for thrombocytopaenia can be grouped into decreased production, increased destruction, increased aggregation, dilution, sequestration or spurious causes.

Given this patient’s known clinical and laboratory findings, the most likely causes include:

  • ECMO-induced circuit thrombosis and thrombocytopenia
  • Sepsis
  • Heparin induced thrombotic thrombocytopaenia syndrome (HITTS)
  • Disseminated intravascular coagulation
  • Drug-induced thrombocytopaenia
  • Thrombotic microangiopathy

Q4. You are concerned about HITTS. What is the underlying pathophysiology of HITTS, and why are ECMO patients at risk?

Heparin induced thrombocytopaenia (HIT type 2) is an immune-mediated condition whereby auto-antibodies to complexes of platelet factor 4 bound to heparin cause a thrombocytopaenia through platelet aggregation. Systemic thrombosis can also occur through platelet activation and endothelial injury resulting in HITTS. Note that HIT type 1 is rarely clinically significant, does not cause thrombosis, and is due to non-immune mediated thrombocytopaenia caused by a direct effect of heparin.

90% of patients with HIT have thrombocytopaenia and 50% have thrombosis which can lead to infarction of any end organ including limbs and skin. The thromboses may be arterial or venous.

HIT occurs in about 5% of people exposed to heparin. Those at increased risk include females, patients receiving unfractionated as opposed to low molecular weight heparin (LMWH), and those on high doses of heparin.

The incidence of HIT in ECMO has been reported to be between 1-4% (Glick et al, 2015; Laverdure et al, 2016). A high suspicion of HIT is necessary in ECMO patients given an often prolonged exposure to heparin. Furthermore many ECMO circuit components are heparin coated and could potentiate HIT even after parenteral heparin is ceased (Murphy et al, 2015).

Q5. How can the diagnosis of HITTS be confirmed in this patient?

If suspicion of HITTS exists, the 4Ts score’ calculator should be used, taking into account:

  • Thrombocytopaenia (Fall in platelets ≥50%, level usually <60 x 109/L, but not below 20 x 109/L)
  • Timing of thrombocytopaenia (Timing: days 5-10)
  • Thromobosis
  • Absence of oTher potential causes of thrombocytopaenia

Note – If the clinical suspicion is intermediate to high, it is important to stop heparin and commence alternative anticoagulation while awaiting the result.

The ELISA immunoassay measures levels of antibodies in the circulation against PF4 antigen. HITS is excluded if <0.4 optical density units or confirmed if >2.0 optical density units. Overall the assay is rapid (1-3 days depending on your lab) but unreliable due to a lack of specificity – it has been shown to positive in 30-39% of patients on ECMO despite HITTS having only a 1-4% incidence in this population (Glick et al, 2015; Laverdure et al, 2016).

Also, the PF4 ELISA immunoassay often provides an indeterminate result. In this scenario, a serotonin release assay can be performed which helps to either confirm or exclude HITTS. This is a functional test of platelet factor 4 antibodies (it measures the ability of HIT antibody from patient serum to activate test platelets). However it usually takes much longer to get a result and rarely guides clinical management.

Q6. How is HITTS managed?

Key steps include:

  • Stop heparin immediately.
  • Provide anticoagulation (if needed) with a non-heparin anticoagulant (bivalirudin is most commonly used at the Alfred ICU) should be used. Titrate APTT to 60-80 s.
  • Use a non-heparin-bonded ECMO circuit.
  • The patient should avoid heparin lifelong.

Argobatran, fondaparinux and danaparoid are other alternatives to heparin described in the literature.

Q7. What is the likely outcome for this patient?

HITTS is a potentially fatal condition, however if promptly recognized and appropriate anticoagulation selection is made it is a completely reversible process.

HITTS antibodies can persist for two to three months post HITTS, however, despite this platelet count generally resolves within seven days of cessation of heparin. If platelet counts do not improve within this timeframe an alternate cause of thrombocytopaenia needs to be considered.


Key References

The post Holy HITTS! appeared first on INTENSIVE.

Can you interpret this ICU chest radiograph?

Labs and Lytes 033

Author: David Humphreys
Reviewer: Sarah Yong, Chris Nickson

A 59-year-old female patient has had a prolonged stay in your ICU.

This her most recent chest radiograph:

L&L 033 CXR

Figure 1. Chest radiograph

Q1. Describe and interpret the chest radiograph.

This is a supine mobile CXR with extensive pathology and numerous hardware in situ.

  • The most striking abnormality is complete opacification of the right lung field, consistent with diffuse consolidation
  • Airspace opacities also involve the left lower and upper lobes
  • VV ECMO cannulae appropriately placed in the right internal jugular vein (IJV) (tip at junction between superior vena cava (SVC) and right atrium (RA)) and right femoral vein (RFV) (tip is in inferior vena cava (IVC))
  • Tracheostomy tube, left internal jugular central venous line, mediastinal surgical clips, and a nasogastric tube are also visible

The presence of extensive pulmonary pathology, tracheostomy and VV ECMO cannulae indicate a patient with severe pulmonary disease and critically impaired gas exchange. The number of small metallic vascular clips in the centre of the image (posterior to the mediastinal shadow) are the result of a recent bilateral sequential lung transplant for idiopathic pulmonary hypertension.

In an immune-suppressed transplant patient, the possible causes of the consolidation are numerous. Fungal infection is an important possibility.


All case-based scenarios on INTENSIVE are fictional. They may include realistic non-identifiable clinical data and are derived from learning points taken from clinical practice. Clinical details are not those of any particular person; they are created to add educational value to the scenarios.

The post Can you interpret this ICU chest radiograph? appeared first on INTENSIVE.

Cardiac arrest on ECMO… Does it still count?

Author: Ken Hoffman
Peer reviewers: Chris Nickson, Arne Diehl

Everything ECMO 015

The emergency alarm goes off in the ICU and you are the first doctor to arrive.

The patient is a 45-year-old man transferred to the unit and started on extracorporeal membrane oxygenation (ECMO) two hours prior.

Q1. What is the first thing you should determine if a patient arrests on ECMO?

The first thing to establish is whether the patient is on VA or VV ECMO.

  • VV ECMO – Cardiac arrest is associated with a complete loss of cardiac output as the native heart is acting as the pump
  • VA ECMO – The ECMO circuit likely provides at least partial circulatory support despite the arrhythmia as the ECMO machine is acting as the pump

Q2. You establish that the patient is on VA ECMO and has had a cardiac arrest, what should you check next?

The next thing to check is the ECMO circuit flow and the blood pressure.

If the circuit flow is able to maintain systemic perfusion (e.g. >4L/min) then you have time to think!

If the circuit is not flowing, assess urgently for circuit factors:

  • Pump failure – the ECMO machine will have stopped, clamp the circuit, transfer to the emergency drive unit (the “hand crank”). Check the power supply; you may need to exchange the ECMO unit.
  • Decannulation or circuit rupture – this will be self evident and messy, clamp the circuit and turn off the pump to prevent further air embolism or exsanguination, apply pressure to any bleeding sites, then manage the patient as per ALS – emergent recannulation.
  • Air embolism – clamp the circuit on the return line and turn off the pump, position the patient head down and manage as per ALS, consider aspirating the right atrium and/or right ventricle.
  • Fulminant circuit or pump head thrombosis, clamp the circuit and manage the patient.

Q3. Should you do chest compressions for cardiac arrest in a patient on VA ECMO?

This depends on whether the ECMO circuit is still perfusing the patient.

  • If the circuit is maintaining systemic perfusion then you do NOT need to perform chest compressions. Even brief periods of low flow can usually be tolerated.
  • If there is pump failure, switch to the hand crank immediately.
  • If there is no circuit flow (see above), then you should manage as per ALS guidelines until ECMO flow is re-established. Adrenaline boluses in the minutes before ECMO support should be avoided.

Q4. What will the arterial line show in a VA ECMO arrest?

There will be a loss of arterial pulsatility due to a loss of native cardiac output.

This will result in an arterial “flat line” trace and the mean arterial pressure (MAP) will likely drop depending on amount of ECMO support and the patient’s total peripheral resistance.

Q5. What rhythms would you expect to see?

The rhythm may be a non-shockable or shockable rhythm.

  • Non-shockable rhythms are asystole and pulseless electrical activity (PEA).
  • Shockable rhythms are ventricular fibrillation (VF) and ventricular tachycardia (VT).

Monitor showing ventricular fibrillation in a patient supported by VA ECMO.


Side note: Is a PEA arrest on ECMO a “thing”?

  • These are patients with loss of native cardiac output or various degrees of cardiogenic shock. The term PEA whilst technically correct is generally not used in these circumstances ,instead we talk about loss of pulsatility secondary to poor native cardiac output or poor aortic valve opening that results in an undetectable pulse.
  • If the patient has severe cardiogenic shock they may not have a palpable pulse or even visible pulsatility on their arterial line trace. Whilst this would technically fit the definition of PEA, an echo may actually show some cardiac movement. In this setting loss of pulsatility reflects severe heart failure rather than an arrest. Increasingly the term “pseudo-PEA” is being used for cases such as this, where true “electro-mechanical dissociation” is absent.

Q6. What are the treatment priorities in a cardiac arrest in a VA ECMO patient?

There is no universally standardised approach for these patients with mechanically supported circulation. In particular adrenaline boluses are NOT indicated. Ensure that support with VA ECMO is adequate.

The following principles apply whilst trying to identify and correct the underlying cause:

Shockable VT/VF

  • Defibrillation — 150J
    • some experts use an initial 100J for VT on VA ECMO
    • consider delaying subsequent shocks until after correction of reversible factors and loading with anti-arrhythmics if refractory
  • Amiodarone 300mg IV after the third shock
  • Avoid 1mg adrenaline boluses, consider titrating oppressors to target MAP

Non-shockable Asystole/PEA

  • Avoid 1mg adrenaline boluses, consider titrating inopressors to target MAP

Some other things to remember when managing cardiac arrest on VA ECMO:

  • Get help! Ensure the ICU Consultant and ECLS Coordinator are notified ASAP.
  • Decrease lung ventilation (pulmonary blood flow is usually minimal). Required respiratory rate may typically be 2-4 breaths per minute during cardiac arrest on VA ECMO.
  • After reversion, ventilator, ECMO blood flow settings, and inotrope and vasopressor support will need to be adjusted

Note that it is possible for a patient receiving VA ECMO to be awake, at least to some degree, during a cardiac arrest… which can have important management implications! Always consider the need for explanation and the need for analgesia and sedation before administering uncomfortable therapeutic interventions.

Q7. What can cause an arrest on VA ECMO?

Important causes grouped according to the traditional ALS ‘Hs and Ts’:

  • Hypoxia
    • turn both the ventilator and ECMO oxygen blender to FiO2 to 1.0
    • ensure O2 is flowing to the oxygenator
    • check for colour change across the oxygenator (bright red blood in the return line)
    • check a pre-oxygenator and post-oxygenator blood gas if required
  • Hyperkalaemia
    • check K+ on arterial blood gas and consider haemolysis
  • Hypovolaemia
    • consider giving fluid bolus
    • check Hb
    • consider retroperitoneal bleeding from cannula insertion
  • Hypothermia
    • check body temperature and ECMO heater unit function
    • cooling occurs rapidly if the ECMO heater unit is not functioning (e.g. ECMO initiation, during transfer)
  • Tamponade
    • focused transthoracic echocardiography (TTE) initially, transoesophageal echocardiography (TOE) TOE if indeterminate and patient is sufficiently stable.
    • If present, or other causes excluded, contact Cardiothoracics and prepare for sternotomy
    • tamponade physiology is complex in VA ECMO, but can cause access insufficiency
  • Tension pneumothorax
    • uncommon in VA ECMO
    • consider especially if difficult to ventilate; focused thoracic ultrasound may help confirm pleural sliding
    • may occur if an intercostal catheter is blocked or displaced
  • Thrombus
    • consider ongoing or new myocardial ischaemia, or pulmonary embolism
  • Toxins
    • check history to see if toxins were the cause of going on VA ECMO (e.g. calcium channel blocker overdose) as specific antidotal management may be required
    • loss of pulsatility can occur with even relatively low doses of beta blockers or sedatives in patients with cardiogenic shock on VA ECMO

Also consider:

  • progression of the underlying disease state (e.g. fulminant myocarditis causing progressive cardiogenic shock or arrhythmias)
  • a problem with the arterial line mimicking PEA (always check the pulse the old-fashioned way!)

Q8. What complication may occur with loss of native cardiac output, even if the patient is fully supported on VA ECMO?

A lack of forward blood flow through the left ventricle has resulted in stasis of blood in the left ventricle.

Thrombosis occurs when forward blood flow and aortic opening is not restored in a timely fashion. This complication is usually catastrophic.


Prolonged impaired pulsatility can also lead to progressive left ventricular distention and potential for aortic regurgitation, which can cause acute pulmonary oedema and render VA ECMO support ineffective.

Key References

  • Nickson CP. Pulseless Electrical Activity (PEA). Critical Care Compendium, Lifeinthefastlane.com. 6 September 2015. [Accessed: 1 November 2017]. Available at URL: https://lifeinthefastlane.com/ccc/pulseless-electrical-activity-pea/
  • Pellegrino V, Sheldrake J, Murphy D, Hockings L, Roberts L. Extracorporeal Membrane Oxygenation (ECMO). Alfred ICU Guideline, 2012.

The post Cardiac arrest on ECMO… Does it still count? appeared first on INTENSIVE.