To the Editor:
Tay et al1 provide an excellent overview on the practical management of venovenous extracorporeal membrane oxygenation (VV-ECMO). The physiologic basis for the proposed algorithms should be discussed to provide clarity to the readers.
Firstly, a preoxygenator O2 saturation threshold of >80% was suggested for cannula repositioning to address recirculation. Based on the Fick Equation (assume negligible dissolved O2),
where Qs = systemic flow; Qp = pulmonary flow; VO2 = oxygen consumption; CaO2, CvO2, CpaO2, and CpvO2 = arterial, venous pulmonary artery, and pulmonary venous oxygen content, respectively; and O2SatA, O2SatCV, O2SatPV, and O2SatPA = arterial, central venous, pulmonary venous, and pulmonary arterial oxygen saturation, respectively.
Recirculation is effectively a left to right shunt with the VV-ECMO as the “lungs” (see figure 1, Supplemental Digital Content, http://links.lww.com/ASAIO/A267). From Equation (1), the recirculation fraction equals (O2SatR − O2SatCV)/(O2SatR − O2SatD), where O2SatR and O2SatD equal return and drainage oxygen saturation, respectively, replacing O2SatPV and O2SatPA. Assuming postmembrane O2 saturation, i.e., O2SatR equals 100% and a normal O2SatCV of 70%, a premembrane O2 saturation, i.e., O2SatD of >80% would translate into a clinically significant shunt (recirculation) fraction of >1.5. This provides support to Tay et al1 general suggestion that O2SatD >80% (and more specifically O2SatR − O2SatD of <20%) indicates significant recirculation.
Secondly, VV-ECMO flow of >60% of cardiac output (CO) was proposed. Total CO can be determined from Equation (1), particularly if the lungs fail to effectively oxygenate blood (and assuming negligible recirculation). In this case, the pulmonary blood flow that has bypassed the VV-ECMO effectively becomes a right-to-left shunt. The VV-ECMO becomes the “lungs” and Qp refers to VV-ECMO flow (see figure 2, Supplemental Digital Content, http://links.lww.com/ASAIO/A268). From Equation (1), O2SatPV and O2SatPA are replaced with O2SatR and O2SatD, respectively.
Hence, Qp/Qs equals (O2SatA − O2SatCV)/(O2SatR − O2SatD), where Qp is the VV-ECMO flow and Qs is the CO.
O2SatCV should be similar to O2SatD, so Qp/Qs equals (O2SatA − O2SatD)/(O2SatR − O2SatD), and Qs equals Qp/(O2SatA − O2SatD)/(O2SatR − O2SatD).
High CO relative to VV-ECMO flow will result in drop in arterial oxygen saturation as larger amount of blood flow bypasses the VV-ECMO (see figure 3, Supplemental Digital Content, http://links.lww.com/ASAIO/A269); this “shunt flow” equals Qs − Qp. The better the lungs are at oxygenating the blood, the greater the Qs will be underestimated, as the “shunt” would be underestimated. Tay et al1 proposed VV-ECMO flow of 60% of CO would maintain systemic oxygen saturation in excess of the proposed minimum of 80%, assuming otherwise normal VV-ECMO function.
In summary, it is important that specific recommendations are made, not based on some arbitrary consensus, but sound physiologic principles. It is hoped that the aforementioned concepts based on shunt fraction calculations (with some necessary assumptions and simplifications) provide readers with the physiologic rationale behind the recommendations.
Hoong Sern Lim
Queen Elizabeth Hospital Birmingham
University Hospital Birmingham NHS Foundation Trust
Birmingham, United Kingdom