Translating the ‘shunt fraction’ method to derive native cardiac output during VA ECMO support

I am grateful to Bachmann and colleagues for taking the time to read and comment on my work ¹ , ² Indeed, it was remiss of me not to have cited their contributions to this subject. ³ I will address three points that Bachmann and colleagues have raised in their Letter.

Firstly, the ‘generalized solution’ described by Bachmann and colleagues is a logical mathematical derivation of the Fick principle, which is reassuringly consistent with the equations used in my report. Based on the Equation (10) in their Letter, Bachmann and colleagues proposed that there is no need for a ‘no sweep gas flow’ approach as described in my report.

$$\frac{\text{Qecmo}}{\text{Qlung}}=\frac{\mathrm{sO}2\text{aorta}-\mathrm{sO}2\mathrm{pa}-\left(\mathrm{sO}2\text{left}\text{atrium}-\mathrm{sO}2\mathrm{pa}\right)}{\mathrm{sO}2\text{postmem}-\mathrm{sO}2\mathrm{pa}-\left(\mathrm{sO}2\text{aorta}-\mathrm{sO}2\mathrm{pa}\right)}$$

However, they also agreed that the ‘no sweep gas flow’ method could ‘improve the accuracy’ because of the greater difference in oxygen content. With the potential for greater accuracy without risks to the patient under close monitoring of oxygenation, Bachmann and colleagues are in effect, providing the premise for my simplified approach with the following equations:

$$\frac{\text{Qecmo}}{\text{Qlung}}=\frac{\mathrm{sO}2\text{left}\text{atrium}-\mathrm{sO}2\mathrm{pa}}{\mathrm{sO}2\text{aorta}-\mathrm{sO}2\mathrm{pa}}-1$$

or

$$\text{Qlung}=\left(\frac{\mathrm{sO}2\text{aorta}-\mathrm{sO}2\mathrm{pa}}{\mathrm{sO}2\text{left}\text{atrium}-\mathrm{sO}2\mathrm{pa}}-1\right)\times \text{Qecmo}$$

Secondly, that left atrial oxygen saturation of 100% is not physiological is not disputed, but that was not the central point of my report. As described in the report, this assumed left atrial oxygen saturation of 100% will underestimate true native cardiac output. As such, if native cardiac output is adequate with an assumed left atrial oxygen saturation of 100%, sampling of the left atrial blood is not required. Sampling of left atrial blood is challenging in clinical practice. The ‘wedged’ blood sample from pulmonary artery catheterization may approximate pulmonary venous and by extension, left atrial blood oxygen, but risks underestimating left atrial blood oxygen saturation if inadequately ‘wedged’. Such underestimation of left atrial oxygen saturation risks overestimating native cardiac output, arguably a less desirable scenario in the context of weaning.

Thirdly, Bachmann and colleagues raised the ‘less often recognized’ issue of venous differential hypoxemia. It should be noted that venous differential hypoxemia is a well‐recognized physiological phenomenon due to the differences in oxygen delivery and oxygen consumption in different tissues/vascular beds. Veno‐arterial extracorporeal membrane oxygenation (VA ECMO), depending on the position of the cannula, changes/exacerbates the oxygen delivery: oxygen consumption in different vascular beds. Thus, VA ECMO alters/exaggerates the physiological difference in venous oxygen content between different vascular beds. Bachmann and colleagues rightly highlighted the absence of arterial and venous differential hypoxemia as pre‐requisites for this ‘shunt’ method. This point is implicit in the calculation of mixed venous oxygen saturation and shunt fraction. ⁴

In conclusion, I applaud Bachmann and colleagues' contributions from their animal studies, and I think we can agree that the concept has indeed ‘found its successful way to the bedside’, albeit with my ‘modification’ of their ‘existing approach’. But that is the essence of translation. What is translational medicine if not the incorporation of the pragmatism of clinical practice onto idealized physiological principles?

Lim, H. S. (2024) Translating the ‘shunt fraction’ method to derive native cardiac output during VA ECMO support. ESC Heart Failure, 11: 2473–2474. 10.1002/ehf2.14747.