The hollow fibre ECMO oxygenator

ECMO is half of "Mechanical supports: ECMO/IABP", an L1 topic of Section 2.1.4 in the second edition of the CICM Syllabus for the Second Part Examination, and logically one would have to agree that the oxygenator is half of ECMO, because it contributes the M and the O. Hilariously, serious textbooks written for perfusion specialists introduce it as an item that is indispensible for oxygenation of the circuit, presenting the case that, if it were not present, all of the blood returned to the arterial circulation of a V-A ECMO patient would remain deoxygenated. If we have trained AI on this material, the generative ECMO answers will seem extremely condescending. 

In brief:

  • The ECMO oxygenator is made of nonporous hollow fibres 
    • The membrane material is polymethylpentene (PMP)
    • The total surface area ranges from 0.8 m2 to 2.5m2
    • The hollow fibres contain the sweep gas; the blood washes around them
    • The resistance is low; transmembrane pressure is usually no more than 60-70 ml at 5-7 L/min flows 
    • The membrane is highly biocompatible; the membrane surface presented to the blood is nonporous, and there is no direct blood-gas contact
    • Water can be passed along similar fibres to warm the blood in the oxygenator
  • The function of the oxygenator deteriorates over time:
    • Gradual protein deposition and platelet aggregation can lead to slowly impaired gas exchange in the oxygenator
    • This can be detected by observing an increase in transmembrane pressure (ΔP)  and measuring the PO2 of the return blood, which should remain at least 200mmHg higher than the venous access PO2. 

 Among all the components of the extracorporeal life support machinery, none has undergone more transformations over its history than the design of this gas and heat exchange interface (remember we started with rotating cylinders with oxygen wafted over them).  The best reference for the technical aspects of the design of these devices is Breiter (2011) whose chapter in Biomaterials for Artificial Organs is an excellent explanation of what these things are made of, and why we have to make them like that. 

Design features of the hollow fibre oxygenator for ECMO

The membrane oxygenator is the integrated hollow fibre assembly that allows blood, gas and water to parallel-play together and politely exchange their desirable properties while otherwise keeping their hands to themselves. These vary in design and properties so much that one cannot make generalisable statements about them, but the general range seems to be between 0.8 m2 and 2.5 m2. Moreover they have slightly different design specifications when they are intended to be used over short periods and in severely heparinised patients. The main differences between oxygenators intended for cardiopulmonary bypass operations and those intended for ECMO are related to the porosity of the membrane (Iwahashi et al, 2004). It is, unfortunately, impossible to explain exactly how each system is fit for purpose without going into extensive detail about the membrane characteristics here, and so:

The available membrane materials for extracorporeal oxygenators

This thing has to be:

  • Highly gas-permeable
  • Water-impermeable 
  • Cheap
  • Strong
  • Offering minimal resistance to flow
  • Durable on sustained exposure to blood and oxygen gas 
  • Biologically inert

That's an "ideal" membrane, and many potential materials are available, but their characteristics disqualify the vast majority. Even among the few we have specifically selected for this purpose, all have caveats. For example silicone polymers (polysiloxanes) are extremely gas-permeable, but neither cheap nor especially durable; polypropylene is cheap strong and highly gas-permeable, but only permeable because it forms pores and is therefore not watertight; and polymethylpentene (PMP) is strong, gas-selective and water-impermeable but not cheap. Those are not the only contenders, but they were the most attractive because they ticked the largest number of  boxes. It is fortunately unnecessary to discuss these at any great length, and the reader with the appetite for that sort of thing is redirected to Sheng et al (2025) for the current state of play  in oxygenator material science and Hagiwara et al (2025) for future directions and experimental solutions.

Cardiopulmonary bypass oxygenators use microporous membranes

Bypass membrane oxygenators prioritise efficient gas exchange at high flows.  "Most perfusionists will use flows of 2.2–2.8 L/m2/min", according to  De Somer (2007), which is probably still true, and which equates to something like 5-7 L/min. At this kind of flow the gas exchange membrane does not have very long to exchange gas, and therefore the rate of gas transfer needs to be extremely efficient.  This is offered by porous membranes, which allow areas of direct blood-air interface to be maintained safely for a period of some hours. By "porous", one seriously means porous: the voids are 30-50% of the total volume of the membrane, according to Yokoi et al (2024). This excellent series of SEM images from Fukuda et al (2021) is a fine example of what a typical polypropylene membrane looks like.

On the other hand, these membranes are much less biocompatible. Consider: the  air-blood interface is intensely thrombogenic and immunogenic, which will gradually produce platelet and protein adhesion, and clog some of the membranes. Additionally adsorbed blood components will reduce the hydrophobicity of the membrane and make it hydrophilic, which means that it will soon begin to issue body water (at first), and then actual plasma. The most significant effect (apart from the grossness of leaking all over your floor) is that these membranes degrade in their performance over time. Six hours is the useful lifespan given by most manufacturers, and this is probably shorter in the presence of acidosis (Usui et al, 1997)

ECMO circuit oxygenators use non-porous membranes

ECMO oxygenators prioritise longevity. These devices should expect to be used for several days or weeks, and so they can afford to be expensive and fiddly to produce. Modern ECMO oxygenators are made with polymethylpentene which has the capacity to be thin, strong, gas-permeable but water-exclusive, and relatively biocompatible (largley because of the lack of blood-air interface). The fibres are also often assymmetric, which is where a hardened layer of something thin and water-impermeable is laid down on the blood surface of a hollow fibre, and the rest of it can remain porous o allow free gas movement. 

The other difference is efficiency. ECMO oxygenators need to be able to exchange gas efficiently at a variety of flow conditions, and under most circumstances they will not be running at their maximum capacity for the entire period of their operation. 

And now finally we are up to actually discussing the design features of the ECMO oxygenator, as for  some reason a deep dive into membrane engineering was required as a preamble.    In short, most oxygenators these days are hollow fibre devices with a shell that contains the blood, penetrated by thin fibres that contain gas, as depicted by the Maquet propaganda below. Some also have thin fibres that contain warm water, to do the heat exchange. At one stage the idea was to perfuse the gas into the shell and send the blood through the hollow fibres, but the resistance to flow and shear stress was too great (Mostafavi et al, 2021), and this is the model we have settled on.

maquet propaganda image of their intgerated oxygenator

Performance characteristics of the ECMO oxygenator

So, how good are these devices at their purpose? Well. They may be as good as they are going to get, in the sense that the hollow fibre technology has permeated so far and wide that research is mostly focused on making small incremental changes to the internal haemodynamics of specific fibre bundle arrangements (eg. Fu et al, 2023). In terms of their performance limits, they tolerate quite a lot of beating, and are rather safe even when placed in unusual situations. For example, the usual setting is to have a gas flow rate approximately the same as the blood flow rate, but Shaw et al (2025) cranked up the gas flow rate to double the background and discovered that this did not lead to either circuit rupture or an increase in gaseous microemboli (still, in case your device measures gas inlet pressure, 50mmHg is the most you should tolerate. The maximum transmembrane pressure is obviously going to vary depending on the prevailing conditions, but in general one should expect a gentle linear increase over the useful range of flows, from up to something like 60-70 mmHg at 5-6L flow (Guan et al, 2010,as well as the company propaganda).

ECMO oxygenator malfunctions

Given the simple design and the lack of moving parts, one might ask (risking an overlap with the troubleshooting chapter),  how could this incredibly basic device disappoint me? The most common way is by refusing to oxygenate the patient, or remove CO2, in the presence of adequate sweep gas flow. This sort of thing can go down in two different ways:

  • Insidiously, where the oxygenator gradually becomes coated with fibrin, dead platelets, etc 
  • Abruptly, where it fails catastrophically because of circuit thrombosis or air embolism

The gradual deterioration of the gas exchange membrane is what we are interested in, because usually this is something predictable and preventable, and usually something you watch over days if not weeks. There are telltale signs:

  • The patient becomes gradually coagulopathic, depleted of platelets, with increasing evidence of haemolysis (i.e. high LDH and free haemoglobin levels)
  • The filter gradually increases its transmembrane pressure (ΔP) and the flows become increasingly difficult to maintain, with higher and higher blood flow rates
  • Visible clots begin to appear on the filter
  • The filter begins to leak fluid, which usually means it has become permeable and lost some of its gas exchange surfaces to waterlogging
  • Blood gases demonstrate a decrease in the PO2 of the postoxygenator blood (Zakhary et al, 2020, draws the line at an "A-a gradient" of 200mmHg ) 

References

Miller, Bernard J., John H. Gibbon Jr, and Charles Fineberg. "An improved mechanical heart and lung apparatus: its use during open cardiotomy in experimental animals." Medical Clinics of North America 37.6 (1953): 1603-1624.

Iwahashi, Hidehiko, Koichi Yuri, and Yukihiko Nosé. "Development of the oxygenator: past, present, and future." Journal of Artificial Organs 7.3 (2004): 111-120.

Breiter, S. "Membranes for oxygenators and plasma filters." Biomaterials for Artificial Organs. Woodhead Publishing, 2011. 3-33.

De Somer, Filip. "What is optimal flow and how to validate this." The Journal of ExtraCorporeal Technology 39.4 (2007): 278-280.

Segers, P. AM, et al. "Clinical evaluation of nine hollow-fibre membrane oxygenators." Perfusion 16.2 (2001): 95-106.

Hagiwara, Junya, Jeffrey D. DellaVolpe, and Yuichi Matsuzaki. "What Are the Best Biocompatible Materials for Extracorporeal Membrane Oxygenation." Journal of Functional Biomaterials 16.6 (2025): 226.

Sheng, Donghai, et al. "Preparation and modification of membranes for extracorporeal membrane oxygenator: A review." Chinese Journal of Chemical Engineering (2025).

Yokoi, Ryo, et al. "Development of a membrane oxygenator for long-term ECMO support using fine silicone hollow fiber." Evolving Therapies and Technologies in Extracorporeal Membrane Oxygenation. IntechOpen, 2024.

Fragomeni, Gionata, et al. "The maximal pore size of hydrophobic microporous membranes does not fully characterize the resistance to plasma breakthrough of membrane devices for extracorporeal blood oxygenation." Frontiers in Bioengineering and Biotechnology 7 (2020): 461.

Fukuda, Makoto, et al. "Electron microscopic confirmation of anisotropic pore characteristics for ECMO membranes theoretically validating the risk of SARS-CoV-2 permeation." Membranes 11.7 (2021): 529.

Usui, Akihiko, et al. "A clinical study for the durability of oxygenators on cardiopulmonary support." Artificial organs 21.7 (1997): 772-778.

Mostafavi, Amir Hossein, et al. "Oxygenation and membrane oxygenators: emergence, evolution and progress in material development and process enhancement for biomedical applications." Journal of Membrane Science and Research 7.4 (2021): 230-259.

Leerson, Jack, et al. "Detecting oxygenator thrombosis in ECMO: a review of current techniques and an exploration of future directions." Seminars in Thrombosis and Hemostasis. Vol. 50. No. 02. Thieme Medical Publishers, Inc., 2024.

Modi, Shan P., et al. "Comparative analysis of oxygenator dysfunction in polymethylpentene oxygenators: A pilot study." Perfusion 40.4 (2025): 933-940.

Mouzakis, Foivos Leonidas, et al. "Has Extracorporeal Gas Exchange Performance Reached Its Peak?." Membranes 14.3 (2024): 68.

Zakhary, Bishoy, et al. "How I approach membrane lung dysfunction in patients receiving ECMO." Critical Care 24.1 (2020): 671.

Shaw, Michael, et al. "The effect of excessive gas to blood ratios in an ECMO oxygenator." Perfusion 40.3 (2025): 678-685.