Patients with severe ARDS have significant impairment in their pulmonary gas exchange which impairs the ability to oxygenate and decarboxylate the blood. Lung compliance is very low which forces limits on the minute ventilation used to prevent barotrauma to the lungs (referred to as low tidal volume ventilation or lung protective ventilation). Without adequate minute ventilation the patient develops physiologically significant hypoxia and hypercapnia.

Using the Murray score as a guide we can assess whether a patient with ARDS would potentially benefit from venovenous extracorporeal oxygenation (vv-ECMO).

Once on vv-ECMO however, there are two factors that are most crucial in determining adequate blood gas exchange.

Percent ECMO blood flow / cardiac output

EBF/CO is the ratio of ECMO circuit blood flow rate as percent of cardiac output. Both are measured in L/min, and therefore you end up with a dimensionless percentage. Since the membrane lung is providing the oxygenation and CO2 removal, the larger portion of the total cardiac output that is passed through the oxygenator the more gas exchange that can occur.

If EBF/CO can be consistently maintained >60%, then oxygenation saturation (SaO2) is almost always >90%.

vv-ECMO circuits can usually support blood flows up to about 5 Lpm before cavitation/chatter limit further increases. So for typical patients, EBF/CO is not usually a problem unless the patient has a very high cardiac output as seen in liver failure, septic shock, hyperthyroidism, or severe anemia.

Maintaining a patient’s hemoglobin level above 10 g/dL with red blood cell transfusions can improve O2 delivery and achieve adequate SaO2, while allowing for lower EBF/CO ratios and lower total circuit blood flow rates.

Recirculation Blood Flow

Recirculation in ECMO circuits occur when oxygenated blood is returned to the ECMO circuit without passing through the peripheral vasculature. This represents wasted ECMO blood flow, and in can be thought of as reducing the effective EBF/CO.

Recirculation blood flow (RBF measured in Lpm) can be assessed by comparing pre-oxygenator pO2 with a peripheral pO2 (not ScvO2 from the pulmonary artery catheter which should be the same as post-pump pO2). Measuring a true mixed peripheral pO2 is difficult in vv-ECMO because pulmonary artery catheter has received oxygenated blood from the ECMO circuit. A venous sample from the internal/external jugular vein or other peripheral vein can be used as a substitute, but reflect local circulation that may not be reflective of the overall system. A theoretical, but often not practical method is to turn off the sweep gas and measure the gas sample from the pre-oxygenator.

Recirculation occurs because of the inherent geometry of using vv-ECMO cannulas that are both drawing and returning blood to the sample venous blood pool. Cannulas can also be misplaced or misdirected, which is critically important in the double lumen venovenous (DLVV) cannulas (eg Avalon) where recirculation can range from 20-50%.

Once an adequate effective ECMO blood flow has been established, (EBF-RBF)/CO, by ensuring that there is that ECMO flow is >60% of the cardiac output, and that recirculation is minimal, we can optimize the oxygen and CO2 gas exchange. Fortunately, these can generally be set independently on the vv-ECMO circuit using the air/oxygen mixer (blender) and sweep gas flow regulator.

Sweep gas flow rate

Carbon dioxide elimination depends on sweep gas flow rate (set on the the sweep gas regular). The rate of sweep gas flow through the membrane lung determines blood decarboxylation. Carbon dioxide tension (pCO2) is mostly unaffected by FDO2 or ECMO blood flow once it is adequate.

Fraction Delivered O2 Gas

The fraction of delivered O2 gas (FDO2) is set on the blender, and directly affects oxygenation of the blood. The sweep gas flow rate has little effect on blood pO2/SO2.


1. Intensive Care Med 2013. Blood oxygenation and decarboxylation determinants during venovenous ECMO for respiratory failure in adults.