- Poster presentation
Artificial Lung Gas Exchanges Depend On Ecmo Settings
Intensive Care Medicine Experimentalvolume 3, Article number: A514 (2015)
Artificial membrane lung (AL) gas exchanges are usually evaluated according to PaO2/FiO2 ratio. in addition, dead space ventilation and shunt fraction can be measured by the same equations used for native lungs .
To study the effect of AL settings - gas flow (GF), blood flow (BF) and FiO2 - on AL PaO2/FiO2 ratio, dead space and shunt, to suggest how to properly monitor these parameters.
We performed three different tests:
a) GF changes (from 1 to 10 L/min) in 8 AL at clinically set BF and FiO2;
b) BF changes in 6 AL at constant FiO2 (1) and GF (10 L/min);
c) FiO2 = 1 vs. clinically set FiO2 in 10 AL at clinically set BF and GF.
We performed pre- and post-oxygenator blood gas analysis and measured CO2 at AL exhaust port by sidestream capnography, in order to evaluate PaO2/FiO2 ratio, dead space and shunt.
a) At clinically set BF (4.0 ± 0.9 L/min) and FiO2 (0.87 ± 0.15), PaO2/FiO2 ratio did not correlate with GF, whereas dead space progressively increased with GF (R = 0.7904, p < 0.0001) (Figure 1).
b) Data were collected at basal (3.3 ± 0.7 L/min), increased (4.1 ± 0.8 L/min) and decreased (2.5 ± 0.6 L/min) BF. With the progressive increase of BF, dead space did not change, whereas PaO2/FiO2 ratio decreased and shunt increased (p < 0.001) (Figure 2).
c) At clinically set BF (3.5 ± 1.1 L/min) and GF (5.1 ± 2.4 L/min), the mean difference ± standard deviation of PaO2/FiO2 ratio and shunt obtained at clinically set (0.80 ± 0.20) vs. FiO2 = 1 was -76 ± 109 mmHg and 2.1 ± 11.6%, respectively.
To properly monitor AL CO2 removal and oxygen transfer, evaluations should be performed at constant GF and at FiO2 = 1 and constant BF, respectively.
Castagna L, Zanella A, Scaravilli V, Magni F, Deab SA, Introna M, et al: Effects on membrane lung gas exchange of an intermittent high gas flow recruitment maneuver: preliminary data in veno-venous ECMO patients. J Artif Organs. 2015 Mar 26, [Epub ahead of print]