The Irenabyss Gallery - Equipment - Matt Brain

Return to The Irenabyss Gallery Home Page

VV-ECMO Oxygen Saturation Calculation Tool

An Education Tool to Calculate Arterial Saturations in Venovenous Extracorporeal Membrane Oxygenation

3D Chart of Arterial Hb O2 Saturations at ECMO Flow 4L/min and VO2 200ml/min for a range of cardiac outputs and [Hb]

Some features and demonstrations have been removed pending further upgrades.

Venovenous ECMO is an increasingly used support for severe pulmonary disease where systemic oxygenation is impaired and cannot be corrected despite breathing 100% inspired oxygen. The inflamed lungs become progressively less compliant resulting in harmful ventilator pressures to achieve lung tidal volumes that adequately ventilate the patient with secondary lung injury and barotrauma resulting.

In the Alfred Intensive Care Unit, VV-ECMO is commonly performed with a multistage (side fenestrations as well as an end lumen) access cannula in the inferior vena cava (IVC) and a return cannula ejecting into the right atrium. The blood passes via a rotary pump through an oxygenator before returning to the patient. The cannula are typically placed via the common femoral veins - the right atrial return cannula is usually positioned under transthoracic echo guidance visualising the tip in the right atrium, the access cannula is then positioned around 10cm more distal in the IVC by comparison of the distance at the skin and abdominal ultrasound of the IVC. If higher ECMO circuit flow is required (in a patient with high oxygen consumption) then a second access cannula may be placed in the internal jugular vein as the IVC will collapse at the higher flow rates immediately and suddenly leaving the patient without any support.

A classic physiology question to help staff understand the oxygenation effects of VV-ECMO proposed by Dr. David Tuxen is as follows:

A young patient with severe ARDS secondary to influenza is at steady state fully supported by VV-ECMO. His chest x-ray demonstrates bilateral white out and for the purposes of this question it can be assumed that his lungs are not contributing to his systemic oxygenation at all.

His cardiac output is 4L/min, the ECMO flow rate is 5L/min. Hemoglobin is 10g/L, mixed central venous haemoglobin oxygen saturation is 65% and arterial saturation is 95%. What will be the effect on his saturations if his cardiac output rises from 4L/min to 6L/min.

Solving this is a useful exercise and doesn’t require any further information – all the equations can be found in the chapter on oxygen in Nunn’s Applied Respiratory Physiology and probably most other physiology texts. Apart from the non-contribution of the lungs, other assumptions include that the shape of the oxygen haemoglobin dissociation curve is normal and that recirculation between the access and return lumens of the ECMO cannula is negligible.

I developed this concept further in a vbamodule that returns the arterial and venous haemoglobin oxygen saturations for any cardiac output and VV-ECMO flow rate. This also requires that the hypothetical patient VO2 be entered and hence results in 4 independent variables : the cardiac output, the ECMO flow rate, the haemoglobin concentration, and the VO2.

VV ECMO 2D Chart Output of Arterial and Venous Saturations vs Cardiac Output and Hemoglobin


The VO2 is entered as a target value – ie what the patient is trying to achieve. It should be noted that VO2 does not reach this target if the amount of oxygen available per minute is less than the amount consumed. In these circumstances, VO2 becomes supply limited and can be seen when the green crosses in the 2D right graph fall below the VO2 target, it could be expected lactic acidosis and signs of tissue hypoperfusion may occur under these conditions. I have factored a relatively modest rise in VO2 per rise in cardiac output, this factor is adjustable and may in reality be quite steep.

A fifth independent variable is recirculation between the access and return cannula. In reality this will change depending on cardiac output : if the heart stops, there will be no venous return and recirculation might approach 100% - blood will enter the ECMO circuit from the IVC and return via the right atrium, flowing in a retrograde fashion back down the IVC resulting in no systemic oxygenation (amongst other adverse effects!) At high cardiac outputs and hence high venous return, recirculation will be minimal as the oxygenated blood will be 'washed through' the right atrium into the right ventricle. This graph below demonstrates the fall off in the central mixed venous oxygen saturation at lower cardiac outputs.

The model allows any amount of recirculation to be factored in (such as might occur if the access and return cannula were connected the wrong way or too close to each other) however in the 3D graphs displayed below only a minimal amount of recirulcation (~2%) is assumed when the cardiac output is greater than the ECMO flow rate.

VV ECMO 3D Chart Output of Venous Oxygen Saturations vs Cardiac Output and Hemoglobin at target VO2 200

After sitting on this project for a while it became clear that a 3D graph would be an ideal teaching tool to display how these variables interact. The model only runs in Excel 2010 with macros enabled however to allow near universal viewing, I have let it run through multiple incremental values with an automatic screen shot at each cycle. This can then be displayed as a movie – however it should be noted that despite the smooth changes in flow rates seen in animation, the model assumes a steady state at each point ie that after any change in ECMO flow rate a short period of time to allow oxygen consumption has occurred.

The youtube link is below and a better resolution version is at the bottom of the page

Note - this page is predominantly a teaching tool. All the numbers are generated within the limits of the assumptions described and have not been validated. As such the slopes on the graphs represent what the direction of change that might be expected with changing conditions however other factors particularly temperature, acid base state and ECMO performance will impact significantly on actual numbers.

Matt Brain: May 2012

Return to The Equipment Home Page

Return to The Irenabyss Gallery