56: Pressors, Fluid, or Flow – Optimizing ECMO Physiology

A post arrest patient just got initiated on ECMO.  Do you give fluids, add pressors, or increase flow?  Marc Dickstein, an anesthesiologist from Columbia University and an expert in the physiology of ECMO, talks with Zack about how to manage these patients, what diagnostics we need and how to optimize your use of the machine.  This talk is a must for everyone starting ECPR in their departments.

Photo: Marc Dickstein

Marc’s ECMO physiology website Harvi

Marc’s ASAIO article on ECMO physiology –

Dickstein ML. The Starling Relationship and Veno-Arterial ECMO: Ventricular Distension Explained. ASAIO J. 2018 Jul/Aug;64(4):497-501. doi: 10.1097/MAT.0000000000000660. PubMed PMID: 29076945.

Zack’s recent Resus Editorial on Impella

Shinar Z. Is the "Unprotected Heart" a clinical myth? Use of IABP, Impella,
and ECMO in the acute cardiac patient. Resuscitation. 2019 May 21. pii:
S0300-9572(19)30173-X. doi: 10.1016/j.resuscitation.2019.05.005. [Epub ahead of
print] PubMed PMID: 31125528

Keep up with the EDECMO Project

  • Comment below if it is about this post
  • Contact us with thoughts, general comments, and suggestions
  • Follow us on Twitter
  • Join our Google+ Page
  • Like us on Facebook
  • Subscribe, rate, and review in iTunes
  • Call us at (470) EDECMO1, that's (470) 333-2661

This post was brought to you by .


  1. John Stokes says

    Great episode and love the discussion. Wanted to add a simple point that I didn’t hear mentioned. It may have not been mentioned because it is obvious to everyone.

    Lots of discussion about whether the aortic valve is opening and adjusting flow, pre-load, after-load to get the valve opening. Remember the simplest, real-time way to assess whether the valve is opening is your right radial or axillary arterial line. The patient needs it anyway. If you have a pulsatile waveform on that arterial line, the valve is opening. If you don’t have a pulsatile waveform, the valve isn’t opening. Additionally, trending the pulse pressure (difference between systolic and diastolic) on your a-line can give you an imperfect estimation regarding how much the LV is ejecting. Importantly, LV may be distended even if you have a pulsatile waveform so still need to be alert to this. But as you get someone on pump and start adjusting pressures, pump flow, etc, get a right upper extremity arterial line in and keep an eye on that arterial line waveform; it gives you a lot of beat by beat information that can guide your resuscitation.

    • Marc Dickstein says

      Great comments. Spot on! What you mention is so important, it’s worth repeating: No pulsatility on the a-line means there’s no ejection. But if there is pulsatility (ie ejection), you still need to determine the filling pressure required for the LV to eject; could be reasonable (pcwp in the teens) or it could be unacceptably high. No way of knowing that from the a-line waveform. And a distended ventricle won’t recover, and acute pulmonary edema really complicates management!

  2. David Hersey says


    Excellent podcast. It is going to take me awhile to work through this one.

    Keep up the hard work because the Reanimate Conference and this podcast is making a difference (not sure how you find the time!).

    With luck, Dr. James Gould, will be getting our own ‘Code ECMO’ program up and running here in Nova Scotia.

  3. Waleed Sneij says

    Love the podcast. Thank you for taking the time to do it. I’m still struggling with CVP measurements once on VA ECMO. Would you elaborate on this some more? In my mind adding a parallel circuit (with ECMO) would lower RV filling pressures but it does not seem to be the case? I’m thinking about RV failure from PE for example.

    • There are a number of factors that will determine CVP on ECMO. Venous drainage would lower CVP. And lowering RV SV would allow for a lower CVP. But, a higher pulmonary venous pressure would raise RV afterload and, for a given target SV, would increase CVP. Consider a situation whereby total systemic flow and MAP do not change after initiation of ECMO (ie native CO reduces by the exact same amount as the ECMO flow): in this case pulmonary flow is reduced by exactly the ECMO flow rate, RV SV is reduced, RV afterload is unchanged, and therefore CVP is reduced (think basic Starling curve). In the case of an RV failure from an acute PE, the big benefit of VA ECMO comes from the reduction (diversion) of pulmonary flow and reduction in PAP and reducing RV SV and afterload. However, ECMO is most often initiated with the express purpose of increasing total systemic flow. In a situation whereby initial CO is very low (and MAP is very low) and ECMO flow is simply additive to the total systemic flow (in a parallel circuit), then the venous return to the RV is unchanged and the RV SV is unchanged, but it is ejecting against a higher afterload (as pulmonary venous pressures increase in response to the higher LV afterload); in this case the CVP would have to be higher to maintain the same SV when ejecting against a higher afterload. There are also some other more complicated factors related to ventricular interactions (ie septal position and motion) that will influence CVP. Check out the Harvi simulator to get a better feel for how all of these parameters interact and ultimately determine the CVP!

Speak Your Mind