VV-ECMO for Residents: Candidacy, Runs and Oxygenators

Our first episode covered veno-venous ECMO the way it gets examined: sweep for CO₂, flow for oxygen, who qualifies, and how to read CESAR and EOLIA honestly. This episode is everything we didn't say — not the physiology, but the service. How ECMO got here, who actually says yes to a referral, and the things that genuinely shorten a run.

Please note: this reflects local Wythenshawe practice. Take the principles, and check your own guidelines.

We start with the history, because it explains the present. The first patient was a road traffic accident victim in 1960s America — go and look at the photograph of the machinery. Then years in which ECMO was essentially associated with death, and genuine doubt that it solved anything. Then, around 2008 to 2010, everything arrived at once: swine flu, a patient in Scotland who had to be transferred to Sweden because we couldn't offer ECMO here, the political question that followed, CESAR out of Leicester, and observational data from Australia and New Zealand — who were ahead of us because their lung transplant organs travelled further and arrived with longer ischaemic times. NICE looked at all of it, concluded equipoise had been lost, and Wythenshawe won one of the bids.

Then candidacy, which is far less formulaic than it used to be. It's now an MDT decision with two, three or more consultants, and the ideal patient doesn't really exist anymore. We work through two contrasting referrals that show why: a patient with a BMI of 50–60 and acute asthma may be an easier yes than an older patient with a bad pneumonia — because reversibility and expected run length matter more than any single exclusion criterion.

We cover what COVID changed (bifemoral cannulation, awake patients, less sedation, better steroid timing), the actual rest settings for a newly cannulated patient, and the three levers that shorten runs: early tracheostomy, negative diuresis and sedation weaning — with the caveat that each is harmful at the wrong moment. Plus an honest answer to why tracheostomy on ECMO stays a consultant procedure.

Then the part that changes practice most: a failing oxygenator doesn't just impair gas exchange, it causes coagulopathy. A D-dimer in the tens of thousands, a fibrinogen refractory to daily transfusion, unexplained platelet drops with negative HIT screens. It looks like DIC and isn't — and the treatment is changing a membrane that may be oxygenating perfectly well. We finish with hypoxia troubleshooting (including the classic mistake of turning up the sweep gas), decannulation and the microbiology plan, and exactly how you lay out an ECMO patient to get them through a CT scanner.

Chapters

• (00:00) Cold open — everything episode one didn't say
• (00:50) A road accident in the 1960s, and the years ECMO meant death
• (02:40) Swine flu, a patient sent to Sweden, and the political case
• (04:10) CESAR, Australia, and why NICE decided equipoise was lost
• (05:40) Why there won't be more UK trials
• (06:40) Who gets ECMO now: the MDT, and two contrasting patients
• (09:00) What COVID changed: bifemoral, awake, steroids
• (10:30) Rest settings for a newly cannulated patient
• (11:50) The three levers that shorten a run
• (13:10) Why tracheostomy stays a consultant procedure
• (14:20) The failing oxygenator that isn't failing
• (16:00) Hypoxia on ECMO — and the sweep gas mistake
• (17:30) Decannulation and the microbiology plan
• (18:20) Taking an ECMO patient to CT
• (19:10) Wrap-up

Key takeaways

• ECMO went from a therapy associated with death to a commissioned national service because swine flu, a patient sent abroad and CESAR all arrived together
• Equipoise has been lost in the UK, so don't expect further randomised trials here
• Candidacy is an MDT decision — a very obese asthmatic may be an easier yes than an older patient with pneumonia, because reversibility and run length outrank single exclusions
• Since COVID: bifemoral cannulation, awake patients, less sedation, better steroid timing
• Ventilate gently — peak around 20, PEEP around 10, FiO₂ 0.5, rate 12 — because not thrashing the lung is where the benefit lives
• Early tracheostomy, negative diuresis and sedation weaning shorten runs, but each is harmful at the wrong time
• A failing oxygenator causes coagulopathy, not just poor gas exchange — change it even if gas exchange looks fine
• For hypoxia: check cannula position, match ECMO flow to cardiac output, consider sedation — and don't turn up the sweep
• Accept saturations around 92%; chasing a normal number leads to harm
• Plan decannulation with microbiology, and recognise the post-decannulation SIRS response for what it is

References / further reading

• Peek GJ et al. CESAR trial. Lancet 2009
• Australia and New Zealand ECMO Influenza Investigators. ECMO for 2009 influenza A(H1N1) ARDS. JAMA2009
• National Institute for Health and Care Excellence. Extracorporeal membrane oxygenation for severe acute respiratory failure in adults (IPG391), 2011
• Combes A et al. EOLIA trial. NEJM 2018
• Goligher EC et al. Bayesian re-analysis of EOLIA. JAMA 2018
• ELSO Guidelines: Management of Adult Patients Supported with VV-ECMO, 2021
• Camporota L et al. Outcomes of the NHS England National ECMO Service. BJA 2021
• NHS England Adult Respiratory ECMO Service Specification

This podcast is for medical education for healthcare professionals. It is not clinical advice. Practice described reflects local Wythenshawe protocol at the time of recording — always follow your own centre's guidance and your regional ECMO centre.