membranes

Article
Newly Developed Pediatric Membrane Oxygenator
that Suppresses Excessive Pressure Drop in
Cardiopulmonary Bypass and Extracorporeal
Membrane Oxygenation (ECMO)
Makoto Fukuda 1, * , Asako Tokumine 1 , Kyohei Noda 2 and Kiyotaka Sakai 3
1 Department of Biomedical Engineering, Kindai University, 930 Nishimitani, Kinokawa-city,
Wakayama 649-6493, Japan; tokumine@[Link]
2 Department of Medical Engineering, Siga University of Medical Science Hospital, Seta, Tsukinowa-cho,
Otsu-City, Siga 520-2192, Japan; noda0611@[Link]
3 Department of Chemical Engineering, Waseda University, 3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan;
kisakai@[Link]
* Correspondence: fukuda@[Link]; Tel.: +81-736-77-0345 (ext. 4308); Fax: +81-736-77-4754




Received: 8 October 2020; Accepted: 19 November 2020; Published: 21 November 2020


Abstract: This article developes a pediatric membrane oxygenator that is compact, high performance,
and highly safe. This novel experimental approach, which imaging the inside of a membrane
oxygenator during fluid perfusion using high-power X-ray CT, identifies air and blood retention in the
local part of a membrane oxygenator. The cause of excessive pressure drop in a membrane oxygenator,
which has been the most serious dysfunction in cardiovascular surgery and extracorporeal membrane
oxygenation (ECMO), is the local retention of blood and air inside the oxygenator. Our designed
blood flow channel for a membrane oxygenator has a circular channel and minimizes the boundary
between laminated parts. The pressure drop in the blood flow channel is reduced, and the maximum
gas transfer rates are increased by using this pediatric membrane oxygenator, as compared with the
conventional oxygenator. Furthermore, it would be possible to reduce the incidents, which have
occurred clinically, due to excessive pressure drop in the blood flow channel of the membrane
oxygenator. The membrane oxygenator is said to be the “last stronghold” for patients with COVID-19
receiving ECMO treatment. Accordingly, the specification of our prototype is promising for low
weight and pediatric patients.

Keywords: membrane oxygenator; hollow fiber membrane; cardiopulmonary bypass (CPB);

extracorporeal membrane oxygenation (ECMO); COVID-19

1. Introduction
Membrane oxygenator is the artificial organ (artificial lung) used in cardiopulmonary bypass
(CPB), such as coronary artery bypass surgery, valve replacement surgery, implantable artificial heart
surgery. Additionally, it is used in extracorporeal membrane oxygenation (ECMO) for respiratory
assistance therapy to acute respiratory distress syndrome, such as that caused by new coronavirus
infection (2019 novel coronavirus disease, COVID-19) [1–3]. Membrane oxygenator is said to be the
“last stronghold” for patients with COVID-19 [1].
The most serious dysfunction (adverse event) of a membrane oxygenator is abnormally
increased pressure drop across oxygenating and heat exchanging sections of a membrane oxygenator.
Platelet-fibrin thrombus and coagulation in membrane oxygenator block blood flow channel and
increase pressure drops, making it difficult to perfuse blood into the membrane oxygenator with a

Membranes 2020, 10, 362; doi:10.3390/membranes10110362 [Link]/journal/membranes

Membranes 2020, 10, 362 2 of 26

systemic centrifugal pump despite adequate anticoagulation [4–8]. In this case, it is necessary to
perform an emergency oxygenator changeover that is originally used with another unused membrane
oxygenator. During this replacement, CPB and ECMO treatment are discontinued. It is a serious
incident that puts the patient’s life endanger. To reduce such incidents and accidents, it is necessary to
detect the occurrence of excessive pressure drops across membrane oxygenator in advance, and the
pressure drops and the blood flow channel pressures at the inlet and outlet of the membrane oxygenators
(1st pressure, 2nd pressure) are recommended to be monitored [5–8]. Fisher et al. reported that the
incidence of excessive pressure drop was 0.4–2.5% [6]. On the other hand, the Working group of the
Japanese Society of Cardiovascular Surgery in Japan found that the incidence of membrane oxygenator
replacement, due to excessive pressure drop was 1/3578 (0.03%) in adults and 1/576 (0.17%) in children
(2016), and reported that there were six times more incidents of membrane oxygenator replacement in
children than those in adults [9,10].
Although many membrane oxygenators for adults with completely different designs have been
developed for practical use in Japan, there are very few pediatric membrane oxygenators. Since the
frequency of membrane oxygenator replacement was six times higher in children than in adults,
it was a problem from the viewpoint of medical safety that there were not many options for pediatric
membrane oxygenators. When a conventional membrane oxygenator for adults is used for a low
weight patient, the blood velocity is slow because of its large priming volume, so that the gas transfer
rate is low. Additionally, the actual retention time is long so that blood coagulation occurs and blocks
the blood flow channel. Therefore, a properly designed small oxygenator is urgently needed for low
weight or pediatric patients.
Numerous studies have been conducted on the design of membrane oxygenators [11–18]. Matsuda
et al. clarified the relationship between the appropriate shape of hollow fiber bundle and gas transfer
rate of an outside blood flow membrane oxygenator by an experimental approach from a transport
phenomenon perspective [12,13]. On the other hand, a lot of knowledge has been obtained by
Computational Fluid Dynamics (CFD) [14,16–18], Zhang et al. developed the design methodology
to optimize a design of membrane oxygenator by calculating blood velocity and pressure drop
distributions in local areas inside a blood flow path using the CFD simulation [16]. However, incidents
are still occurring, due to excessive pressure drop. As a matter of fact, it has not been clear what kind
of phenomenon or membrane oxygenator design cause an excessive pressure drop, and it has not been
solved yet [10,11]. CFD simulation (mathematical model) is useful, but it has limitations in grasping
the actual phenomena. Therefore, the problem is solved by understanding the actual phenomena
through an experimental approach. Here, we focused on the local retention of blood and air inside
of a membrane oxygenator as one of the causes of the excessive pressure drop, and developed an
original experimental method using high-power X-ray computed tomography (CT) for verification.
Needless to say, platelet-fibrin thrombus and coagulation in membrane oxygenator are not only caused
by hydrodynamic reasons, but also materials and surface roughness.
The objective of the present study is to perform internal imaging of the membrane oxygenator
during fluid perfusion using a high-power X-ray CT device, and to locate the local blood and air
retention. To reduce the incidents, due to excessive pressure drop in cardiovascular surgery and
ECMO, we propose the design concept of a membrane oxygenator that suppress such blood retention
and excessive pressure drop. In addition, we aim to make a prototype and evaluate the pediatric
membrane oxygenator.

2. Material and Methods

2.1. Hollow Fiber Membrane Oxygenators

The membrane oxygenators used in this study were commercial membrane oxygenators; oxia®
ACF (JMS Co., Ltd., Tokyo, Japan, Sample A), CAPIOX® FX05 (Terumo Co., Ltd., Tokyo, Japan, Sample
C), and the newly developed pediatric membrane oxygenator (prototype), and commercial available
Membranes 2020, 10, 362 3 of 26

pediatric membrane oxygenator, which is currently the smallest in the world (LivaNova Co., Ltd.,
Sorin Group, Mirandola, Italy, Sample B) as a control. Table ?? shows the specifications of the hollow
fiber membrane oxygenators. These are the outside blood flow membrane oxygenators.

Table 1. Specification of newly developed pediatric membrane oxygenators and other membrane
oxygenators.

Commercially Commercially
Commercially Newly Developed
Available Membrane Available Membrane
Sample Available Membrane Pediatric Membrane
Oxygenator 1) Oxygenator
(Product) Oxygenator Oxygenator
Kids D100 CAPIOX® FX05
Sample A Prototype
Sample B Sample C
LivaNova Co., Ltd.,
Terumo Co., Ltd.,
Manufacturer JMS Co., Ltd., Japan — Sorin Group,
Japan
Italy
Membrane area
1.7 0.39 0.22 0.5
[m2 ] 2)
Polypropylene Polypropylene
Material of hollow
(PP), (PP), Polypropylene Polypropylene
fiber membrane
asymmetric pore asymmetric pore (PP) (PP)
Pore structure 3)
structure structure
Inner diameter of
lumen 238 ± 5 238 ± 5 233 ± 14 195 ± 8
[µm] (n = 30)
Membrane
thickness 35 ± 1 35 ± 1 34 ± 1 57 ± 6
[µm] (n = 30)
Number of hollow
fibers 19,500 9000 - -
[-]
Length of hollow
fiber 90 45 - -
[mm]
Air permeability 4)
35.1 ± 1.4 33.8 ± 0.4 - -
[sec]
Poly(2- Poly(2- poly(2-
Antithromboge-nic poly(hydoroxyethylm
methacryloyloxyethyl methacryloyloxyethyl methacryloyloxyethyl
material coating for etharylate)
phosphoryl choline) phosphoryl choline) phosphoryl choline)
blood flow channel (PHEMA)
(PMPC) (PMPC) (PMPC)
Priming volume
245 37 31 43
[mL]
Range of blood
flow rate 1.0–7.0 0.5–2.0 0.1–0.7 −1.5
(Max QB ) (7.0) (2.0) (0.7) (1.5)
[L/min]
Calculated
2.1 1.1 2.6 1.7
retention time [s] 5)

1) The commercially available pediatric membrane oxygenator, which is currently the smallest in the world.
2) Membrane area = outer diameter of hollow fiber lumen × π × length of hollow fiber × number of hollow fibers.
3) Tortuous pore diameter is not measured. 4) The gas permeability of the membranes was measured by the

Gurley method using an air resistance tester defined in ISO5636-5 [19]. 5) Calculated retention time = Priming
volume/Maximum blood flow. The smaller the value, the larger the gas permeability.

The membrane specifications of the new pediatric oxygenator and Sample A are exactly the same.
The gas permeabilities of the prototype and Sample B membranes are almost the same.
Calculated or ideal retention time is the time that blood is ideally or uniformly perfused once
inside a device. The actual retention time is different from the calculated retention time.
Membranes 2020, 10, 362 4 of 26

Membranes 2020, 10, 362 4 of 28

2.2. Imaging a Calculated

MembraneorOxygenator Inside
ideal retention During
time is Fluid
the time Perfusion
that blood Using
is ideally High-Power
or uniformly X-Ray
perfused CT
once
inside a device. The actual retention time is different from the calculated retention time.
We have developed the non-destructive flow analysis experiment and evaluated the retention
of air in the
2.2. Imaging channel
blood a MembraneofOxygenator
membrane Insideoxygenators [20]. Using
During Fluid Perfusion When fluid flows
High‐Power X‐ray inside
CT a membrane
oxygenator composed of the metal heat exchanger, gas exchanger, and others made from various
We have developed the non‐destructive flow analysis experiment and evaluated the retention
materials isofinvestigated using
air in the blood a medical
channel X-rayoxygenators
of membrane CT (tube voltage output,
[20]. When 120 kV),
fluid flows artifacts
inside are generated,
a membrane
oxygenator
false images composed of occur
and obstructions the metal heat exchanger,
during imaging,gassoexchanger,
accurateand others made
imaging from variousTherefore,
is impossible.
materials is investigated using a medical X‐ray CT (tube voltage output, 120 kV), artifacts are
high-power (430 kV) X-ray CT was used for high-resolution imaging of membrane oxygenators
generated, false images and obstructions occur during imaging, so accurate imaging is impossible.
composedTherefore,
of various materials, such as stainless-steel and polypropylene hollow fiber membranes
high‐power (430 kV) X‐ray CT was used for high‐resolution imaging of membrane
during fluid perfusion.
oxygenators composedThe ofresolution is significantly
various materials, higher than
such as stainless‐steel that of medical
and polypropylene X-ray
hollow fiber (120 kV).
membranes
Figure 1 shows the during fluid perfusion.
photograph of the The resolution is significantly
experimental [Link] than 2that
Figure of medical
shows X‐ray
the prototype of a
(120 kV). Figure 1 shows the photograph of the experimental apparatus. Figure 2 shows the prototype
closed extracorporeal circuit.
of a closed extracorporeal circuit.

Figure 1. Experimental apparatus for analyzing fluid flow in a device; the non‐destructive
Figure 1. Experimental apparatus for analyzing fluid flow in a device; the non-destructive experimental
experimental study using high‐power X‐ray computed tomography (X‐ray CT) [20]. TOSCANER‐
study using high-power
24500twin X-ray
(Toshiba Co.,computed tomography
Ltd., Tokyo, (X-ray
Japan), installed at CT) [20]. TOSCANER-24500twin
the Industrial Technology Center of (Toshiba
Co., Ltd., Tokyo,
Wakayama
Membranes
Japan), installed
2020, 10,Prefecture.
362
at the Industrial Technology Center of Wakayama Prefecture.
5 of 28

Figure 2. Schematic of an experimental apparatus for analyzing fluid flow in a device; a closed‐circuit
Figure 2. Schematic of an experimental apparatus for analyzing fluid flow in a device; a closed-circuit
connected to a soft reservoir.
connected to a soft reservoir.
We briefly show the experimental method based on the patent already reported [20]. A closed‐
circuit system including membrane oxygenator, centrifugal pump, etc., were installed in the 2‐stage
set of the frame, and only a membrane oxygenator was imaged.
An extracorporeal circuit with a small priming volume was created by hand so that a required
flow rate (0.5–8 L/min) could be perfused with a small fluid volume, and it was connected to a soft
reservoir for sealing the circuit (closed‐circuit, Figure 2). The centrifugal pump was connected to the
Membranes 2020, 10, 362 5 of 26

We briefly show the experimental method based on the patent already reported [20].
A closed-circuit system including membrane oxygenator, centrifugal pump, etc., were installed
in the 2-stage set of the frame, and only a membrane oxygenator was imaged.
An extracorporeal circuit with a small priming volume was created by hand so that a required flow
rate (0.5–8 L/min) could be perfused with a small fluid volume, and it was connected to a soft reservoir
for sealing the circuit (closed-circuit, Figure 2). The centrifugal pump was connected to the controller.
Next, reverse osmosis (RO) water was injected into a soft reservoir, membrane oxygenator, and a
circuit to perform priming (priming volume ≈ 700 mL). After that, RO water and an X-ray contrast
agent (RO aqueous solution of 10 wt% BaSO4 , BaSO4 particle size 5 µm, specific gravity 4.3 g/cm3 )
were perfused at an average flow rate of 5 L/min (0.7 L/min for a pediatric membrane oxygenator) for
60 min. The flow was in a steady-state, during which X-ray imaging was performed. RO aqueous
solution of 10 wt% BaSO4 (BaSO4 particle size 5 µm, specific gravity 4.3 g/cm3 ) was used as an X-ray
contrast agent for the following reasons.

(1) A test solution that can be used for X-ray CT imaging (high specific gravity of particles).
(2) It can localize in the region of the blood flow channel where it is expected to stay easily in between
the hollow fibers (the red blood cell diameter is 8 µm as a guide, and since the BaSO4 particles do
not deform during fluid perfusion, 5 µm particle was used as the BaSO4 particle size which is
slightly smaller than 8 µm).
(3) On the other hand, during the X-ray CT imaging, there should be no flow channel blockage that
would prevent the test solution from flowing.

The tests were carried out on several copies of the prototype. They were not reused.

2.3. High-Power X-Ray Computed Tomography (CT)

TOSCANER-24500twin (Toshiba Co., Ltd., Tokyo Japan, installed at the Industrial Technology
Center of Wakayama Prefecture) was used as the High-power (430 kV) X-ray computed tomography
(CT) equipment. The high-power X-ray computed tomography (CT) equipment was operated in
compliance with the Regulation on Prevention of Ionizing Radiation Hazards [21]. The area around
the high-power X-ray CT system is outside the controlled area with an effective dose of 2 µSv/h or
less (≈1.3 mSv/every three months or less). The operation of the high-power X-ray CT system was
performed by a certified Operation Chief of Work with X-rays.
The imaging area obtained by the detector was decomposed with a pixel size of 2048 × 2048 to
create voxels. Voxels change the color of CT images depending on the proportion of substances in a
unit volume. When the substance density is high, it is displayed in white, and when there is a lot of
air, it is displayed in black. The plane cut view image from the top was taken every 2 mm of height
(thickness) of the devices.

2.4. Oxygen and Carbon Dioxide Transfer Rates, Pressure Drop of Blood Flow Channel
Fresh bovine blood containing heparin Na was adjusted to standard venous blood characteristics
(Hb = 12.0 g/dL, Base Excess (BE) = 0 mmEq/dL, 37 °C, PV CO2 = 47 ± 1 mmHg) [21]. This was perfused
once through a membrane oxygenator to evaluate gas transfer rate. The ratio of blood flow rate to gas
flow rate (V/Q) was V/Q = 0.5, 1, 2, and oxygen fractional concentration (FiO2 ) of inspired gas was
100%. Bovine blood was sampled at the blood inlet and outlet of the membrane oxygenator, and pH,
PO2 , PCO2 , and oxygen saturation were measured with the blood gas analyzer (Rapid Lab 348EX,
Siemens Healthcare Diagnostics Co. Ltd., Vienna, Austria). The oxygen contents of arterial and venous
blood were calculated, and the oxygen and carbon dioxide transfer rates were calculated based on
Fick’s equation (law of conservation of mass) (n = 3) [22,23],

Sa O2 − Sv O2 QB
 
KO2 = × 1.34 × Hb + 0.00314 × (Pa O2 − Pv O2 ) × (1)
100 100
Membranes 2020, 10, 362 6 of 26

Cv CO2 − Ca CO2 QB 2.226 × Tv CO2 − Ta CO2 QB

   
KCO2 = × = × (2)
100 100 100 100
where KO2 is the oxygen transfer rate (mL/min), SaO2 the arterial saturation (%), SvO2 the venous
saturation (%), PaO2 the arterial partial pressure (mmHg), PvO2 the venous partial pressure, Hb the
hemoglobin concentration (g/dL), QB the flow rate of blood side (mL/min), 1.34 the Hemoglobin oxygen
capacity (mL/g), 0.00314 the oxygen solubility (vol%/mmHg, 37 ◦ C), KCO2 the carbon dioxide transfer
rate (mL/min), CaCO2 the arterial carbon dioxide content (mL/dL), CvCO2 the venous carbon dioxide
content (mL/dL), TaCO2 the arterial carbon dioxide content (mmol/dL), TvCO2 the venous carbon
dioxide content (mmol/dL), 2.226 the conversion factor ((mmol/L)→(mL/dL), 1 mol = 22.26L at CO2 ).
The blood inlet pressure PBi and the outlet pressure PBo of a membrane oxygenator were measured
with the digital pressure sensor PPX-R01N-6M (CKD CORPORATION Ltd., Tokyo, Japan) at the blood
flow rate QB = 0.1–2.0 L/min. The pressure drop ∆PB in the blood flow channel was calculated.

∆PB = PBi − PBo (3)

Data on the membrane oxygenators were treated using a two-way analysis of variance (two-way
ANOVA, p < 0.05).

3. Results

3.1. Non-Destructive Visualization of a Hollow Fiber Membrane Oxygenator Using High-Power X-Ray
Computed Tomography
Figure 3 shows the high-power X-ray CT images of Sample A. Figure 3a is the external view
of the device, and Figure 3b is the vertical cross-sectional view. Figure 3c shows the photographic
image of the vertical cross-sectional view. The heat exchangers appear white, and the packed hollow
fiber membrane (gas exchanger) containing air is clearly observed, due to the difference in CT value
(Hounsfield Unit) of the material and structure. The packed hollow fiber membrane layer containing
air is arranged in the circular blood channel. The outer periphery of the blood channel is in a state
where the adhesive has infiltrated into the hollow fibers and solidified, and it is distinguished from the
circular blood channel.
Figure 4a shows the high-power X-ray CT image of the vertical cross-sectional view of Sample C,
and Figure 4b shows the photographic image of the external view.
Figure 5 shows the plane cut views of the yellow arrow in the vertical view image of Figure 3b.
Figure 5(b)-1 shows the interface between the heat exchanger and blood flow channel. A large number
of stainless-steel tubes forming the heat exchanger are arranged in the lateral direction and appear to
overlap the blood flow channel. Figure 5(b)-2 shows the plane cut view about 2 mm above Figure 5(b)-1.
Figure 5(b)-2 is the interface between the heat exchanger and spacers, the air layer in between the
spacers was observed to be black. Figure 5(b)-4 is the interface between the heat exchanger and the
packed hollow fiber membrane. Figure 5(b)-5 is the packed hollow fiber membrane layer, and the area
inside the yellow dotted circle is the blood flow channel. Figure 5(b)-6 is the interface between the
packed hollow fiber membrane layer and the arterial filter. The wavy arterial filter was observed to be
white. As described above, non-destructive visualization by high-power X-ray computed tomography
(CT) enables detailed observation of the internal structure of the membrane oxygenator composed of
various materials.
 of stainless‐steel tubes forming the heat exchanger are arranged in the lateral direction and appear to
overlap the blood flow channel. Figure 5(b)‐2 shows the plane cut view about 2 mm above Figure
5(b)‐1. Figure 5(b)‐2 is the interface between the heat exchanger and spacers, the air layer in between
the spacers was observed to be black. Figure 5(b)‐4 is the interface between the heat exchanger and
the packed hollow fiber membrane. Figures 5(b)‐5 is the packed hollow fiber membrane layer, and
the area inside the yellow dotted circle is the blood flow channel. Figure 5(b)‐6 is the interface between
the packed hollow fiber membrane layer and the arterial filter. The wavy arterial filter was observed
to be white. As described above, non‐destructive visualization by high‐power X‐ray computed
Membranes 2020, 10, 362 tomography (CT) enables detailed observation of the internal structure of the membrane oxygenator
7 of 26
composed of various materials.

Membranes 2020, 10, 362 8 of 28

(a)

(b)

(c)
Figure 3. Images of the hollow fiber membrane oxygenator (sample A) by high‐power X‐ray
Figure 3. Images of the hollow fiber membrane oxygenator (sample A) by high-power X-ray computed
computed tomography; (a) external view, (b) vertical cross‐sectional view. (c) Photographic image of
the vertical cross‐sectional view. Colored (red, pink, yellow) arrows represent images of blood
tomography; (a) external view, (b) vertical cross-sectional view. (c) Photographic image of the vertical
streamlines. (b) (b)‐1 Interface between the heat exchanger and blood flow channel. (b)‐2 Spacer, 1st
cross-sectional view. Colored (red, pink, yellow) arrows represent images of blood streamlines. (b) (b)-1
layer, the interface between the heat exchanger and spacer. (b)‐3 Spacer, 2nd layer. (b)‐4 Interface
between the heat exchanger and packed hollow fiber membrane (gas exchanger). (b)‐5 Packed hollow
Interface between the heat exchanger and blood flow channel. (b)-2 Spacer, 1st layer, the interface
fiber membrane (gas exchanger). Although the color is dark, the same circular channel as in (b)‐4 is
confirmed. (b)‐6 Interface between the packed hollow fiber membrane layer and the arterial filter. (c)
between the heat exchanger and spacer. (b)-3 Spacer, 2nd layer. (b)-4 Interface between the heat
(c)‐1 Heat exchanger. (c)‐2 Interface between the heat exchanger and blood flow channel. (c)‐3 Spacer,
exchanger and packed hollow fiber membrane (gas exchanger). (b)-5 Packed hollow fiber membrane
(gas exchanger). Although the color is dark, the same circular channel as in (b)-4 is confirmed.
(b)-6 Interface between the packed hollow fiber membrane layer and the arterial filter. (c) (c)-1 Heat
exchanger. (c)-2 Interface between the heat exchanger and blood flow channel. (c)-3 Spacer, 1st layer,
the interface between the heat exchanger and spacer. (c)-4 Spacer, 2nd layer. (c)-5 Interface between
the heat exchanger and packed hollow fiber membrane (gas exchanger). (c)-6 Packed hollow fiber
membrane (gas exchanger). (c)-7 Interface between the packed hollow fiber membrane layer and the
arterial filter.
metal □ □ 400~2000□□
water□□□□□□ 0□□□□□□
hollow fiber membrane (water is included) −200~−300
hollow fiber membrane (air is included) □ −700~−900
Membranes 2020, 10, 362 8 of 26
air□□□□□ −1,000□

(a)

(b)
Figure 4. Images of the hollow fiber membrane oxygenator (sample C) by high-power X-ray computed
tomography; (a) vertical cross-sectional view. (b) Photographic image of external view.
 Membranes 2020, 10, 362 10 of 28

Membranes 10, Images

2020, 4.
Figure 362 9 of 26
of the hollow fiber membrane oxygenator (sample C) by high‐power X‐ray
computed tomography; (a) vertical cross‐sectional view. (b) Photographic image of external view.

[Link]
Figure Planecut
cutview
view(from
(fromtop)
top)images
imagesofofthe
thehollow
hollowfiber
fibermembrane
membraneoxygenator
oxygenator(sample
(sampleA)A)byby
high-power X-ray computed tomography.
high‐power X‐ray computed tomography.
3.2. Non-Destructive Visualization in a Hollow Fiber Membrane Oxygenator during RO Water and X-Ray
3.2. Non‐Destructive
Contrast Agent Perfusion Visualization in a Hollow Fiber Membrane Oxygenator during RO Water and X‐Ray
Contrast Agent Perfusion
Figure 6 shows the plane cut views of Sample A during RO water perfusion. There were air bubbles
Figure
displayed 6 shows
in black theinterface
at the plane cut viewsthe
between of spacers
Sample and
A during ROperipheral
the outer water perfusion. There
part (Figure were air
6(b)-2~4).
bubbles displayed in black at the interface between the spacers and the outer peripheral
The blood inlet is located in the bottom center of the device, the blood flow velocity is higher (Figurepart in the
6(b)‐2~4).
center The blood
and slower inlet
in the is located
outer in theTherefore,
periphery. bottom center of the
uneven flowdevice, the blood
(channeling) andflow
air/velocity is higher
blood retention
in to
are theoccur
center
at and slower
the outer in the outer
periphery periphery.
of the Therefore,
device. From Figureuneven flow
6, air and (channeling)
blood retention and air/atblood
occurs the
retention are to occur at the outer periphery of the device. From Figure 6, air and
boundary surface and outer periphery of spacers constituting the laminated structure, which affects blood retention
occurs
blood at the boundary
coagulation reactionssurface and outer periphery of spacers constituting the laminated structure,
and thrombosis.
which affects blood coagulation reactions and thrombosis.
Figure 7 shows the plane cut views of Sample A during RO aqueous solution of 10 wt% BaSO4
perfusion. X‐ray contrast agent was displayed in white and air bubbles were displayed in black.
Membranes 2020, 10, 362 10 of 26
Membranes 2020, 10, 362 12 of 28

Figure
Figure 6. [Link]
Theplane
planecut
cut view
view images
images (from
(from top)
top) ofof the
the membrane
membrane oxygenator
oxygenator (sample
(sample A,A, ® ®ACF,
oxia
oxia ACF,
JMS Co., Ltd., Japan) by the X‐ray computed tomography (during RO water perfusion).
JMS Co., Ltd., Japan) by the X-ray computed tomography (during RO water perfusion).

Figure 7 shows the plane cut views of Sample A during RO aqueous solution of 10 wt% BaSO4
perfusion. X-ray contrast agent was displayed in white and air bubbles were displayed in black.
Compared with Figure 6, retention of bubbles and X-ray contrast agent particles (BaSO4 ) were more
noticeable in the heat exchanger, 1st, and 2nd spacers (Figure 7(b)-1~4). There were many air and
BaSO4 particles at the interface between the spacers and the outer peripheral parts. From the results
shown in Figure 7, there is a risk of excessive pressure drop, due to retention at the peripheral parts and
the interfaces of the blood flow channel. This is consistent with the results of local blood coagulation
inside a membrane oxygenator after clinical use. The local retention area of the fluid is grasped in
more detail when the X-ray contrast agent (BaSO4 particles, specific gravity 4.3 g/cm3 ) is perfused
Membranes 2020, 10, 362 11 of 26

than when RO is perfused. Therefore, the methodology in which the X-ray contrast agent is perfused
is more useful as a critical test that reproduces a clinically serious situation, such as blood retention
and coagulation.
Membranes 2020, 10, 362 13 of 28

Figure
Figure7. [Link]
Planecut view
cut view (from
(fromtop) images
top) imagesofof
the membrane
the membrane oxygenator
oxygenator (sample
(sample A)A)
bybyhigh-power
high‐power
X-ray computed tomography (during RO aqueous solution
X‐ray computed tomography (during RO aqueous solution of 10wt% BaSO of 10wt% BaSO perfusion).
4 4 perfusion).Compared
Compared
with
withFigure
Figure 6, retention
6, retentionof bubbles andand
of bubbles X-ray contrast
X‐ray agent
contrast particles
agent (BaSO
particles 4 ) were
(BaSO more
4) were noticeable
more in
noticeable
the
inheat exchanger,
the heat 1st, and
exchanger, 2nd 2nd
1st, and spacers (Figure
spacers 7(b)-1~4).
(Figure There
7(b)‐1~4). werewere
There many air and
many BaSO
air and 4 particles
BaSO at
4 particles
the interface between the spacers and the outer peripheral parts (Figure 7(b)-1~4).
at the interface between the spacers and the outer peripheral parts (Figure 7(b)‐1~4).

[Link] the other

Design Concepthand,
of thefor comparison,
Newly Developed there wereMembrane
Pediatric no X-ray contrast agent particles (BaSO4 ) in the
Oxygenator
blood flow channel of the gas exchanger (in between hollow fibers) (Figure 7(b)-5). It was feared
Figure
that the hollow 8 fibers
showswould
the vertical cross‐sectional
easily contact each other,photographic images
and an effective gasof the newly
transfer developed
area would be
pediatric membrane 2 oxygenator (prototype) and commercially available
smaller than 1.7 m , but there was no such concern from Figure 7(b)-5. It was considered that thepediatric membrane
oxygenator
X-ray contrastdevice
agent was(Sample
trappedB). in
Figure 9 shows
the arterial filterthe three‐dimensional
of Figure diagram
7(b)-6 after passing to explain
through the gasthe
geometry of the prototype and the flow of gas and bloodstreams.
exchanger (in between hollow fiber membranes) of Figure 7(b)-5. Based on these findings, we designed
the newThe pressure
pediatric drop of
membrane a membrane
oxygenator oxygenator
and made is affected
a prototype, as shownby three factors,
in Section the blood
3.3. Although the
characteristics
quantitative of the
analysis patient
is not (hematocrit
sufficient value, viscosity,
in these experiments, temperature)
it is possible and
to clarify the blood flow
local air and rate,
bloodin
additioninside
retention to thethecomplicated flow channel
membrane oxygenator, of ahas
which membrane
not been oxygenator.
understood by During cardiopulmonary
the conventional CFD.
bypass (CPB), perfusionists change the blood flow rate and monitor the inlet pressure and outlet
pressures of a membrane oxygenator, but cannot control the blood properties (hematocrit, viscosity,
temperature). Therefore, even if the risk of patient factors cannot be controlled, it is desired to design
a membrane oxygenator that allows perfusionists to flexibly handle the operation conditions. For this
purpose, it is necessary to design a compact, high performance and highly safe membrane oxygenator
as compared with the conventional oxygenators.
Membranes 2020, 10, 362 12 of 26

We have previously conducted similar experiments and found that it is difficult to detect the
tracer in bovine blood with X-rays. Therefore, we tried to visualize the internal fluid flow inside a
membrane oxygenator using only pure water and barium aqueous solution.

3.3. Design Concept of the Newly Developed Pediatric Membrane Oxygenator

Figure 8 shows the vertical cross-sectional photographic images of the newly developed pediatric
membrane oxygenator (prototype) and commercially available pediatric membrane oxygenator device
(Sample B). Figure 9 shows the three-dimensional diagram to explain the geometry of the prototype
and the flow of gas and bloodstreams.
The pressure drop of a membrane oxygenator is affected by three factors, the blood characteristics of
the patient (hematocrit value, viscosity, temperature) and blood flow rate, in addition to the complicated
flow channel of a membrane oxygenator. During cardiopulmonary bypass (CPB), perfusionists change
the blood flow rate and monitor the inlet pressure and outlet pressures of a membrane oxygenator,
but cannot control the blood properties (hematocrit, viscosity, temperature). Therefore, even if the
risk of patient factors cannot be controlled, it is desired to design a membrane oxygenator that
allows perfusionists to flexibly handle the operation conditions. For this purpose, it is necessary to
design a compact, high performance and highly safe membrane oxygenator as compared with the
conventional oxygenators.
Membranes 2020, 10, 362 15 of 28

Figure 8. Vertical cross-sectional photographic images of hollow fiber membrane oxygenators, (a) newly
Figure 8. Vertical cross‐sectional photographic images of hollow fiber membrane oxygenators, (a)
developed pediatric membrane oxygenator (prototype) and (b) commercially available membrane
newly developed pediatric membrane oxygenator (prototype) and (b) commercially available
oxygenator (Sample B). Colored (red, pink, yellow) arrows represent images of blood streamlines.
membrane oxygenator (Sample B). Colored (red, pink, yellow) arrows represent images of blood
streamlines.
 Figure 8. Vertical cross‐sectional photographic images of hollow fiber membrane oxygenators, (a)
newly developed pediatric membrane oxygenator (prototype) and (b) commercially available
membrane oxygenator (Sample B). Colored (red, pink, yellow) arrows represent images of blood
Membranes 2020, 10, 362 13 of 26
streamlines.

Figure
Figure 9.
9. Diagram
Diagram toto explain
explainthethegeometry
geometryofofthethe newly
newly developed
developed membrane
membrane oxygenator
oxygenator and and the
the flow
flow of gas and bloodstreams. The reed screen sheets composed of hollow fiber membranes
of gas and bloodstreams. The reed screen sheets composed of hollow fiber membranes are alternately are
alternately
laminated laminated to attain
to attain the the multilayer
multilayer structure structure of thefiber
of the hollow hollow fiber bundle.
bundle. ColoredColored (red,
(red, pink, pink,
yellow)
yellow) arrows represent
arrows represent images ofimages of blood streamlines.
blood streamlines. The heat
The heat carrier carrier inside
is perfused is perfused
of the inside of the
stainless-steel
stainless‐steel tube (heat
tube (heat exchanger) exchanger)
from the back from
of thethe backtoward
figure of the figure toward the front.
the front.

The basic design of the prototype is the structure in which heat exchanger (stainless-steel tubes),
gas exchanger (packed hollow fiber membrane), filters, and spacers that partition them are laminated.
The reed screen sheets composed of hollow fiber membranes are alternately laminated to attain the
multilayer structure of the hollow fiber bundle. Blood flows from the bottom center of the device
and is perfused in between the stainless-steel tubes and then crossflows through packed hollow fiber
membrane, which is defined as the outside blood flow membrane oxygenator (Figure 8a, Figure 9).
The contacting pattern between the blood and the exchange fluid (gas, heat carrier) is the crossflow.
In the crossflow, the fluid streams flow perpendicular to each other to achieve the turbulent flow
of blood, reducing the boundary layer of blood side and increasing gas transfer rate. The outside
blood flow membrane oxygenator also has the advantages of lower blood priming volume and blood
pressure drop, which allow the wide range of blood flow rate at clinical operations, compared with the
inside blood flow membrane oxygenator [12,13,24]. The design of the membrane device determines
the performance and safety of the membrane oxygenator. The characteristics of the currently used
membrane oxygenators have been shown in the previous study [24]. Compared with these membrane
oxygenators, the membrane oxygenator produced in our study has key parameters, such as membrane
area, blood flow rate, and priming volume, that are an order of magnitude smaller [24], so our
membrane oxygenator has completely different characteristics compared with previous membrane
oxygenators. Therefore, our study is extremely useful from a practical point of view.
On the other hand, since the blood flow channel of a membrane oxygenator is much more
complicated than that of human blood vessels, uneven flow (channeling) and retention of blood are
likely to occur in the blood flow channel of a membrane oxygenator [11–18]. Uneven flow and retention
generate blood coagulation reaction, block parts of the blood flow channel, and reduce the blood flow
channel area, thus increasing the pressure drop. Therefore, the blood flow channel of the membrane
oxygenators has a simple circular shape to suppress uneven flow and retention of blood. This design
concept suppresses an uneven flow and retention and makes the blood flow velocity uniform.
Membranes 2020, 10, 362 14 of 26

From the results shown in Section 3.2, we focused on the fact that there were many local air
retentions at the boundaries of the laminated parts in the membrane oxygenator. The concrete design
concept is indicated as follows, (1) Assuming that air and blood retention is one of the causes of
excessive pressure drop across membrane oxygenator, we designed the blood flow channel that
minimizes the boundary parts of the laminated parts to suppress this. To obtain a smooth blood flow
(to prevent blood retentions), a circular channel having a continuous heat exchanger and gas exchanger
is designed. (2) The size of the prototype is aimed to be smaller than that of the commercially available
pediatric membrane oxygenator, which is currently the smallest in the world (Figure 8). On the other
hand, (3) To minimize pressure drop, the membrane area is set to 0.39 m2 , and the priming volume
is set to 37 mL, which is larger than the membrane area (0.22 m2 ) and priming volume (31 mL) of
conventional pediatric membrane oxygenator. Moreover, (4) The maximum blood flow rate is set to
2.0 L/min so that the flow rate operation range becomes wider and increase the perfusionists’ degree of
freedom, and the blood retention time is shortened to suppress the blood coagulation reaction.
On the other hand, the concentric circle type (Figure 4) is the mainstream in the design of
membrane oxygenators worldwide, and the existence of such a completely different design concept is
intriguing. In terms of biofluid mechanics, this design can maintain the pressure drop suppression
and high gas transfer rate. This is because the blood flow is temporarily stored in the blood reservoir,
and dynamic pressure of blood is converted into static pressure so that the blood flows uniformly into
the packed capillary membranes.

3.4. Non-Destructive Visualization of the Newly-Developed Membrane Oxygenator during Fluid Perfusion
Figure 10 shows the vertical cross-sectional images of the newly-developed membrane oxygenator
during RO water and RO aqueous solution of 10 wt% BaSO4 perfusion. Figure 11 shows the plane cut
views during RO water perfusion, which is indicated by the red arrow in Figure 10a.
There were no air bubbles compared with Figure 10, and it was found that RO water was uniformly
perfused in local areas. Figure 11(a)-4 and 5 are hollow fiber membrane bundle, and the area inside the
yellow dotted circle is the blood flow channel. RO water is uniformly perfused in between the hollow
fibers in the circular channel. The hollow fiber membranes and the housing adhered to each other with
an adhesive around the area outside the yellow dotted circle.
Membranes 2020, 10, 362 15 of 26
Membranes 2020, 10, 362 17 of 28

Figure
Figure 10.10. Vertical
Vertical cross‐sectional
cross-sectional images
images ofofthethe newly
newly developed
developed hollow
hollow fiber
fiber membrane
membrane oxygenator
oxygenator
(prototype)
(prototype) byby high‐power
high-power X‐ray
X-ray computed
computed tomography,
tomography, (a) during
(a) during RO water
RO water perfusion,
perfusion, (b) during
(b) during RO
RO aqueous
aqueous solutionsolution
of 10 ofwt%10 wt%
BaSOBaSO4 4 perfusion.
perfusion. (a) (a)
(a)-1 (a)‐1
Heat Heat exchanger.
exchanger. (a)‐2
(a)-2 Interface
Interface between
between thethe
heat
heat exchanger
exchanger andand blood
blood flow
flow channel.(a)-3
channel. (a)‐3 Interface
Interface betweenthe
between the heat
heat exchanger
exchanger andand blood
blood flow
flow
channel.(a)-4
channel. (a)‐4 Packed
Packed hollow
hollow fiber
fiber membrane
membrane (gas
(gas exchanger).
exchanger). (a)‐5
(a)-5 Packed
Packed hollow
hollow fiber
fiber membrane
membrane
(gas
(gas exchanger)2mm
exchanger) 2mmaway awayfromfromthe theimage
imagein in (a)-4.
(a)‐4. Although
Although the thecolor
colorisisdarker
darkerthan
thanthatthatinin
(a)‐4,
(a)-4,the
same
the samecircular
circularchannel
channelasasin in (a)‐4
(a)-4 and (a)‐5 isisconfirmed.
and in (a)-5 confirmed.(a)-6 (a)‐6Interface
Interfacebetween
betweenthe thespacer
spacer andand
arterial
arterial side
side filter.
filter. (b)(b) (b)‐1
(b)-1 BaSO
BaSO 4 particles
4 particles (white)are
(white) aretrapped
trappedaround
aroundouter outerparts
partsofofthethe blood
blood flow
flow
channel.
channel. (b)‐2
(b)-2 Interface
Interface between
between thethe heat
heat exchanger
exchanger and and blood
blood flowflow channel.
channel. (b)‐3
(b)-3 Interface
Interface between
between
the
the heatexchanger
heat exchangerand andblood
bloodflowflow channel.
channel. (b)‐4
(b)-4Packed
Packedhollowhollowfiber fibermembrane
membrane (gas
(gas exchanger).
exchanger). (b)‐
5 Packed
(b)-5 Packedhollowhollowfiberfibermembrane
membrane(gas (gasexchanger)
exchanger)2 2mm mmawayawayfrom fromthe theimage
imageinin(b)‐4.
(b)-[Link]
Although the
the color
color is is darker
darker than
than that
that in in (b)-4,
(b)‐4, thethe same
same circular
circular channel
channel as inas(b)‐4
in (b)-4
andand in (b)-5
in (b)‐5 is confirmed.
is confirmed. (b)‐6
(b)-6 Interface
Interface between
between thethe spacer
spacer andand arterial
arterial sideside .
filter.
filter
Membranes 2020, 10, 362 16 of 26
Membranes 2020, 10, 362 18 of 28

Figure 11. The plane cut view (from top) images of the newly developed membrane oxygenator
(prototype) by high-power X-ray computed tomography (during RO water perfusion).
Figure 11. The plane cut view (from top) images of the newly developed membrane oxygenator
Figure 12 shows
(prototype) by the plane cut
high‐power views
X‐ray of Figure
computed 10b during
tomography an aqueous
(during solution
RO water of 10 wt% BaSO4
perfusion).
perfusion, which is indicated by the red arrow in Figure 10b. Compared with Figure 7, there were no
air bubbles, and few X-ray contrast agent particles (BaSO4 ) were observed. This is because the new
design minimizes the circular flow path and boundary parts. Thus, the risk of excessive pressure drop
in the prototype is less than that in sample A.
Membranes 2020, 10, 362 17 of 26
Membranes 2020, 10, 362 19 of 28

Figure 12. The plane cut view (from top) images of the newly developed membrane oxygenator
(prototype) by high-power X-ray computed tomography (during RO aqueous solution of 10 wt% BaSO4
perfusion). There
Figure 12. Theis aplane
slightcut
amount
view of X-raytop)
(from contrast
imagesagent retention
of the newlyatdeveloped
the outermost periphery
membrane of the
oxygenator
blood flow path by
(prototype) in Figure 12(b)-1.
high‐power X‐ray computed tomography (during RO aqueous solution of 10 wt%
BaSO4 perfusion). There is a slight amount of X‐ray contrast agent retention at the outermost
However, theofproblem
periphery the bloodthat
flowapath
slight amount
in Figure of X-ray contrast agent retention is observed at the
12(b)‐1.
outermost periphery of the blood flow channel in Figure 12(b)-1 is for future work.
3.5. Pressure Drop of Blood Flow Channel
3.5. Pressure Drop of Blood Flow Channel
Figure 13(a) shows the relationship between the blood flow rate and the pressure drop of both
Figure 13a shows the relationship between the blood flow rate and the pressure drop of both
pediatric membrane oxygenators within the recommended range of blood flow rate at clinical
pediatric membrane oxygenators within the recommended range of blood flow rate at clinical practices.
practices. The pressure drop of the newly developed pediatric membrane oxygenator (prototype)
The pressure drop of the newly developed pediatric membrane oxygenator (prototype) was much
was much smaller than that of the commercially available membrane oxygenator (Sample B). This is
smaller than that of the commercially available membrane oxygenator (Sample B). This is mainly
mainly because the priming volume of the prototype (37 mL) is larger than that of Sample B (31 mL)
because the priming volume of the prototype (37 mL) is larger than that of Sample B (31 mL) and the
and the length of the blood flow channel is shorter. The pressure drop at the maximum blood flow
length of the blood flow channel is shorter. The pressure drop at the maximum blood flow rate was
rate was also smaller in the new membrane oxygenator (Table 2). Thus, a risk of excessive pressure
also smaller in the new membrane oxygenator (Table 2). Thus, a risk of excessive pressure drop in the
drop in the prototype is less than that in Sample B.
prototype is less than that in Sample B.
and blood retention is longer, and/or due to uneven flow (channeling). So it is essential to
appropriately use the membrane oxygenator depending on the set blood flow rate at clinical
practices. In the future, we will confirm the incidence, due to the excessive pressure drop at clinical
practices using this pediatric membrane oxygenator.
Since the pressure drop with respect to the blood flow rate increases quadratically, the blood
Membranes 2020, 10, 362 18 of 26
flow is turbulent flow and contributes to the increase in gas transfer rate [12–16].

Membranes 2020, 10, 362 21 of 28

(a)

(b)
Figure 13.
Figure 13. The relationship between pressure drop of the blood flow channel and blood flow rate.
**Two‐way
Two-way analysis
analysis of
of variance
variance(two‐way ANOVA,p p< <0.05).
(two-wayANOVA, 0.05).Blue andand
Blue yellow areas
yellow represent
areas the
represent
recommended
the recommended ranges of of
ranges blood flow
blood flowrate.
rate.(a)
(a)newly
newly developed
developed pediatric oxygenator
pediatric membrane oxygenator
(prototype)vs.
(prototype) [Link]
commercially available
available membrane
membrane oxygenator
oxygenator (Sample(Sample B). (b)developed
B). (b) newly newly developed
pediatric
pediatric membrane
membrane oxygenator oxygenator (prototype)
(prototype) vs. commerciallyvs. commercially available
available membrane membrane
oxygenator oxygenator
(Sample A).
(Sample A).

3.6. Determination of Oxygen and Carbon Dioxide Transfer Rates

Figure 14 (a) shows the relationship between the blood flow rate and oxygen transfer rate of both
pediatric membrane oxygenators. The oxygen transfer rates of both oxygenators were almost the
same, and there was no significant difference. This is because that the new membrane oxygenator has
a larger priming volume (37 mL > 31 mL) and the blood flow velocity is low, and the resistance in the
blood‐side diffusion boundary layer is larger during oxygen transfer, thereby decreasing oxygen
transfer rate per membrane area [11–13]. Figure 14 (b) shows the relationship between V/Q and
oxygen transfer rate at QB = 0.7 L/min. There is no significant difference in the oxygen transfer rate
between the two membrane oxygenators at V/Q = 0.5, 1, and 2. Additionally, the oxygen transfer rate
Membranes 2020, 10, 362 19 of 26

Table 2. Maximum oxygen and carbon dioxide transfer rates and pressure drop of membrane
oxygenators.

Commercially
Commercially
Newly Developed Available
Available
Sample Pediatric Membrane MEMBRANE
Membrane
(Product) Oxygenator Oxygenator
Oxygenator
Prototype Kids D100
Sample A
Sample B
Maximum oxygen
433 116 43
transfer rate
(Max QB = 7 L/min) (Max QB = 2 L/min) (Max QB = 0.7 L/min)
[mL/min, STP]
Maximum carbon
dioxide transfer rate
355 98 38
[mL/min, STP]
(V/Q = 1.0)
Pressure drop of the
106 89 155
blood flow channel
(Max QB = 7 L/min) (Max QB = 2 L/min) (Max QB = 0.7 L/min)
[mmHg]
Pressure drop of the blood flow channel (packed capillary membranes and heat exchanger).

On the other hand, Figure 13b shows the relationship between the blood flow rate and the pressure
drop of the prototype and Sample A. Within the recommended range of blood flow rate, the pressure
drop of the prototype was about the same as that of Sample A.
However, at the same blood flow rate, the pressure drop of the prototype is higher than that
of Sample A. One of the reasons that pediatric oxygenators are replaced more frequently than
adult membrane oxygenators in CPB [6,9,10] is that the pressure drop (design value) of a pediatric
oxygenator is higher than that of an adult oxygenator at the same blood flow rate, which can be seen in
Figure 13b. Moreover, excessive pressure drop leading to incidents is likely to occur when air and blood
retention is longer, and/or due to uneven flow (channeling). So it is essential to appropriately use the
membrane oxygenator depending on the set blood flow rate at clinical practices. In the future, we will
confirm the incidence, due to the excessive pressure drop at clinical practices using this pediatric
membrane oxygenator.
Since the pressure drop with respect to the blood flow rate increases quadratically, the blood flow
is turbulent flow and contributes to the increase in gas transfer rate [12–16].

3.6. Determination of Oxygen and Carbon Dioxide Transfer Rates

Figure 14a shows the relationship between the blood flow rate and oxygen transfer rate of both
pediatric membrane oxygenators. The oxygen transfer rates of both oxygenators were almost the
same, and there was no significant difference. This is because that the new membrane oxygenator has
a larger priming volume (37 mL > 31 mL) and the blood flow velocity is low, and the resistance in
the blood-side diffusion boundary layer is larger during oxygen transfer, thereby decreasing oxygen
transfer rate per membrane area [11–13]. Figure 14b shows the relationship between V/Q and oxygen
transfer rate at QB = 0.7 L/min. There is no significant difference in the oxygen transfer rate between
the two membrane oxygenators at V/Q = 0.5, 1, and 2. Additionally, the oxygen transfer rate does not
increase even if the gas flow rate is increased, under which the blood flow rate is limited.
 Membranes
Membranes 2020,
2020, 10, 362
10, 362 22 of2028of 26

(a) V/Q = 1.0

(b) QB = 0.7 L/min

Figure 14. 14.
Figure The The
relationship
relationshipbetween oxygen
between transfer
oxygen raterate
transfer andand
blood flowflow
blood raterate
of both pediatric
of both pediatric
membrane oxygenators, (a) V/Q and (b) Q B=0–2 L/min, V/Q=0.5–2. *Two‐way analysis of variance
membrane oxygenators, (a) V/Q and (b) QB = 0–2 L/min, V/Q = 0.5–2. Two-way analysis of variance
(two‐way ANOVA,
(two-way ANOVA, p < p0.05).
< 0.05).

Figure 15a shows the relationship between the blood flow rate and the carbon dioxide transfer
rate of both pediatric membrane oxygenators. The carbon dioxide transfer rate of the prototype was
higher at QB > 0.7 L/min. This is because the prototype has a larger membrane area. Figure 15b shows
the relationship between V/Q and carbon dioxide transfer rate at QB = 0.7 L/min. Unlike the oxygen
transfer rate results, the higher the V/Q, the higher the carbon dioxide transfer rate. This is because the
 Membranes 2020, 10, 362 21 of 26

gas flow rate (flow velocity) is higher and the carbon dioxide partial pressure at the gas side outlet is
lower, so the overall difference of carbon dioxide partial pressure between the blood side and the gas
side (driving
Membranes force)
2020, 10, 362 is larger, thereby the carbon dioxide transfer rate is increased. 23 of 28

(a) V/Q = 1.0

(b) QB = 0.7 L/min

Figure
Figure15.
[Link] relationship
The relationshipbetween
betweencarbon
carbondioxide
dioxidetransfer
transferrate
rateand
andblood
bloodflow
flowrate
rateofofboth
bothpediatric
pediatric
membrane oxygenators, (a) V/Q and (b) Q = 0–2 L/min, V/Q = 0.5–2. *Two‐way analysis
membrane oxygenators, (a) V/Q and (b) QB = 0–2 L/min, V/Q = 0.5–2. * Two-way analysis of variance
B of variance
(two‐way ANOVA,pp<<0.05).
(two-wayANOVA, 0.05).

4. Discussion
As mentioned above in Section 3.3, the pressure drop of a membrane oxygenator is affected by
three factors, the blood characteristics of the patient (hematocrit value, viscosity, temperature) and
blood flow rate, in addition to the complicated flow channel of the membrane oxygenator. During
cardiopulmonary bypass (CPB), perfusionists change the blood flow rate and monitor the inlet
pressure and outlet pressures of a membrane oxygenator, but cannot control the blood properties
Membranes 2020, 10, 362 22 of 26

Table 2 shows the numerical values extracted from Figures 13–15. The gas transfer rates at each
maximum blood flow rate were higher in the new membrane oxygenator. It was feared that the hollow
fibers of the new membrane oxygenator would be easily contacted with each other, and the effective
gas exchange area would be smaller, but there was no such concern.

4. Discussion
As mentioned above in Section 3.3, the pressure drop of a membrane oxygenator is affected
by three factors, the blood characteristics of the patient (hematocrit value, viscosity, temperature)
and blood flow rate, in addition to the complicated flow channel of the membrane oxygenator.
During cardiopulmonary bypass (CPB), perfusionists change the blood flow rate and monitor the inlet
pressure and outlet pressures of a membrane oxygenator, but cannot control the blood properties
(hematocrit, viscosity, temperature). Therefore, even if the risk of patient factors cannot be controlled,
it is desired to design a membrane oxygenator that allows perfusionists to flexibly handle the
operation conditions.
Since the blood flow channel of a membrane oxygenator is much more complicated than that of
human blood vessels, uneven flow (channeling) and retention of blood flow are likely to occur [11–14].
Uneven flow and retention generate blood coagulation reaction, and reduce the blood flow channel
area, thus increasing the pressure drop. Therefore, the blood flow channel of the newly developed
membrane oxygenator is a simple circular shape to suppress uneven flow and retention. This design
concept suppresses an uneven flow and retention and makes the blood flow velocity uniform [12–16].
On the other hand, a blood flow channel that increases gas transfer rate by making turbulent blood
flow, such as the crossflow, is desired [12,15,16].
Although the quantitative analysis was not sufficient in these experiments, it was possible to clarify
the local air and blood retention inside the membrane oxygenator, which has not been understood by
the conventional CFD. This is consistent with the results of local blood coagulation inside a membrane
oxygenator after clinical use. In addition, the amount of air retention inside the membrane oxygenator
is affected by the priming skill of each perfusionist. In clinical practice, it is ideal to completely remove
the air inside the membrane oxygenator by the priming operation. However, it is difficult to tell
whether the air was actually completely removed. In these experiments as well, since air may not have
been removed completely after the priming operation, it was possible that bubble retentions were
observed when fluids were perfused (Figures 6 and 7).
Membrane oxygenator dysfunctions in ECMO are caused by excessive pressure drop, due to blood
coagulation and thrombus formation, and reduction of gas transfer rate, due to the reduction of effective
membrane area, and plasma leakage [1,24,25]. The same emergency supports are performed in ECMO
as for the pulmonary dysfunction in CPB. However, it has been reported that, blood coagulation and
thrombosis are likely to occur, due to vascular inflammation in patients with COVID-19, and hemofilter
is easily clogged in continuous hemodiafiltration (CHDF) treatment [1]. There is concern that the ECMO
treatment of patients with COVID-19 will cause a more serious excessive pressure drop. Therefore,
our pediatric membrane oxygenator that suppresses high-pressure drop is very useful for ECMO,
which is used for longer periods than CPB in cardiovascular surgery.
On the other hand, plasma leakage is likely to occur in ECMO. It has been reported that when hollow
fiber membranes of an oxygenator deteriorated, plasma components in the blood leaked into the lumen of
hollow fiber (plasma leakage), a liquid of yellow foam leaks from the gas outlet port of an oxygenator
during an ECMO treatment for critically ill patients with COVID-19 [1]. At that time, there is concern
that COV-19 in plasma may permeate through the pore and diffuse as an aerosol from the gas outlet
port, which is still one of the issues in operating ECMO. If the pore diameter of a typical hollow fiber
membrane used for membrane oxygenators is on the order of 0.1 µm and the porosity is 40–60% [26],
COV-19 with the diameter of 100 nm may permeate through a pore from the outside to inside lumen of
hollow fiber membrane. Therefore, first of all, it is necessary to verify the permeability of COV-19, and the
appropriate pore size on the outside of the membrane, and anisotropic pore structure design [27–30] to
Membranes 2020,
Membranes 10, 362
2020, 10, 362 23of
25 of28
26

gradually replaced with plasma (the pores are gradually filled with plasma) over time, resulting in
prevent COV-19 from permeating the anisotropic pore. It would be effective to use a non-porous structure
the permeation of plasma into a hollow fiber membrane lumen. Therefore, it is useful to design a pore
membrane, such as a silicone membrane, to suppress plasma leakage [23–25], but a silicone membrane
size and pore structure on the outer side of a membrane such that plasma does not reach the
has low gas permeability [26]. Figure 16 shows the observations of the inner lumen and outer surfaces
membrane thickness. In the future, it is desired to study the pore structure and materials of the hollow
of the capillary membrane of Sample A for reference. The method reported in our previous study [31]
fiber membrane to fulfill the various functions of a membrane oxygenator.
was employed to observe the inner lumen and outer surfaces of the capillary membranes using a field
The membrane oxygenator is said to be the “last stronghold” for patients with COVID‐19. A
emission scanning electron microscope (FE-SEM) (JSM-7610F, Jeol Ltd., Tokyo, Japan) at an accelerating
membrane oxygenator for artificial lung must be further evolved [24] by taking into consideration of
voltage of 1.5 kV and an emission current of 47.2 µA
the problems in patients with COVID‐19 that we have not experienced before [1].

Figure 16. FE-SEM images of the surfaces of the capillary membrane (Sample A), (a) inner lumen
Figure 16. FE‐SEM images of the surfaces of the capillary membrane (Sample A), (a) inner lumen
surface, and (b) outer surface.
surface, and (b) outer surface.
Polypropylene membrane of the membrane oxygenators is hydrophobic and has very small
5. Conclusions
pores [26], there is sufficient surface tension to prevent plasma infiltration. Hence, these pores are
This experimental
gas-filled, approach identified
resulting in significantly air and blood
higher transport ratesretention
of oxygeninand
the local
carbonpart of the than
dioxide oxygenator.
if pores
Our design
are filled concept
with [Link],
a membrane duringoxygenator suppressesit air
ECMO treatment, and blood retention
is considered that the air andin excessive
the pores
pressure drop.
is gradually By using
replaced with this pediatric
plasma (the membrane oxygenator,
pores are gradually the
filled abnormally
with plasma) overincreased pressure
time, resulting
drop
in theinpermeation
the blood flow channel
of plasma intoisareduced, and membrane
hollow fiber the maximum [Link] transferit rate
Therefore, is raised.
is useful It is
to design
possible to reduce
a pore size and pore thestructure
incidentson andtheaccidents,
outer sidedueof atomembrane
excessive pressure
such thatdrop thatdoes
plasma havenot occurred
reach thein
clinical
membrane practices. Therefore,
thickness. even initthe
In the future, world’sto
is desired unprecedented
study the poreCOVID‐19 pandemic,
structure and materials thisofmembrane
the hollow
oxygenator
fiber membranefor low weight
to fulfill theand pediatric
various patients
functions of aismembrane
promising oxygenator.
for ECMO treatment of patients with
COVID‐19.
Membranes 2020, 10, 362 24 of 26

The membrane oxygenator is said to be the “last stronghold” for patients with COVID-19.
A membrane oxygenator for artificial lung must be further evolved [24] by taking into consideration
of the problems in patients with COVID-19 that we have not experienced before [1].

5. Conclusions
This experimental approach identified air and blood retention in the local part of the oxygenator.
Our design concept for a membrane oxygenator suppresses air and blood retention and excessive
pressure drop. By using this pediatric membrane oxygenator, the abnormally increased pressure
drop in the blood flow channel is reduced, and the maximum oxygen transfer rate is raised. It is
possible to reduce the incidents and accidents, due to excessive pressure drop that have occurred in
clinical practices. Therefore, even in the world’s unprecedented COVID-19 pandemic, this membrane
oxygenator for low weight and pediatric patients is promising for ECMO treatment of patients
with COVID-19.

Author Contributions: M.F.: concept, design of the experiment, performed the experiment, data
analysis/interpretation, wrote the paper. A.T.: design of the experiment, data analysis/interpretation, wrote the
paper. K.N.: performed the experiment, data analysis/interpretation. K.S.: concept, design of the experiment,
data analysis/interpretation, wrote the paper. All authors have read and agreed to the published version of
the manuscript.
Funding: This research was funded by JMS Co., Ltd. as part of its future research program of engineering
development.
Acknowledgments: The authors gratefully thank Naomi Backes Kamimura (Kindai University) for helpful
suggestions during preparation of the manuscript. The authors gratefully thank Shinichi Tokumoto (Industrial
Technology Center of Wakayama Prefecture) for helpful operations of the high-power X-ray CT system.
Conflicts of Interest: Makoto Fukuda: This work was funded by JMS Co., Ltd. as part of its future research
program of engineering development. The other authors have no conflicts of interest to declare.

Abbreviations
CaCO2 arterial carbon dioxide content (mL/dL)
CvCO2 venous carbon dioxide content (mL/dL)
FiO2 inspired oxygen fractional concentration (%)
KCO2 carbon dioxide transfer rate (mL/min)
KO2 oxygen transfer rate (mL/min)
Hb hemoglobin concentration (g/dL)
P pressure (mmHg)
∆P pressure drop (mmHg)
PaO2 arterial partial pressure (mmHg)
PvO2 venous partial pressure (mmHg)
QB flow rate of blood side (mL/min)
SaO2 arterial oxygen saturation (%)
SvO2 venous oxygen saturation (%)
TaCO2 arterial carbon dioxide content (mmol/dL)
TvCO2 venous carbon dioxide content (mmol/dL)
V flow rate of gas side (mL/min)
Greek letters
a artery
B blood
i inlet
o outlet
v vein
Membranes 2020, 10, 362 25 of 26

References
1. Japanese ECMOnet for COVID-19. Available online: [Link] (accessed on
1 August 2020).
2. Ohshimo, S.; Shime, N.; Nakagawa, S.; Nishida, O.; Takeda, S.; Committee of the Japan ECMO Project.
Comparison of extracorporeal membrane oxygenation outcome for influenza-associated acute respiratory
failure in Japan between 2009 and 2016. J. Intensive Care 2018, 6, 38. [CrossRef] [PubMed]
3. Peek, G.J.; Mugford, M.; Tiruvoipati, R.; Wilson, A.; Allen, E.; Thalanany, M.M.; Hibbert, C.L.; Truesdale, A.;
Clemens, F.; Cooper, N.; et al. Efficacy and economic assessment of conventional ventilatory support
versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): A multicentre
randomised controlled trial. Lancet 2009, 374, 1351–1363. [CrossRef]
4. Blomback, M.; Kronlund, P.; Aberg, B.; Fatah, K.; Hansson, L.O.; Egberg, N.; Moor, E.; Carlsson, K.
Channelologic fibrin formation and cold-induced clotting of membrane oxygenators during cardiopulmonary
bypass. J. Cardiothorac. Vasc. Anesth. 1995, 9, 34–43. [CrossRef]
5. Eugene, A.H., II. Chapter 5: Circuitry and Cannulation Techniques. In Cardiopulmonary Bypass Principles and
Practice, 3rd ed.; Gravlee, G.P., Davis, R.F., Stammers, A.H., Ungerleider, R.M., Eds.; Lippincott Williams &
Wilkins: Philadelphia, PA, USA, 2008; pp. 63–113.
6. Fisher, A.R.; Baker, M.; Buffin, M.; Campbell, P.; Hansbro, S.; Kennington, S.; Lilley, A.; Whitehorne, M.
Normal and abnormal trans-oxygenator pressure gradients during cardiopulmonary bypass. Perfusion 2003,
18, 25–30. [CrossRef]
7. Myers, G.J.; Weighell, P.R.; McCloskey, B.J.; Holt, A.M.; McTeer, S.; Maxwell, S.L. A multicenter investigation
into the occurrence of high-pressure excursions. J. Extra Corpor. Technol. 2003, 35, 127–132.
8. Ündar, A.; Owens, W.R.; McGarry, M.C.; Surprise, D.L.; Kilpack, V.D.; Mueller, M.W.; McKenzie, E.D.;
Fraser, J.C.D. Comparison of hollow-fiber membrane oxygenators in terms of pressure drop of the membranes
during normothermic and hypothermic cardiopulmonary bypass in neonates. Perfusion 2005, 20, 135–138.
[CrossRef]
9. Izumi, R. To reduce complications in pediatric extracorporeal circulation. Jpn. J. Extra Corpor. Technol. 2019,
46, 274. (In Japanese)
10. Iwaki, S.; Tani, S.; Yoshida, Y.; Taka, H. Why do the excessive pressure drops in pediatric membrane
oxygenator not decrease? Jpn. J. Extra Corpor. Technol. 2019, 46, 283. (In Japanese)
11. Sakai, K.; Yanagisawa, M.; Hosoya, N.; Ohmura, T.; Sakagami, M.; Kuwano, K.; Nakanishi, H. Comparison
of oxygenation and flow characteristics of inside and outside blood flow membrane oxygenators.
Artif Organs Today 1993, 3, 57–80.
12. Matsuda, N.; Sakai, K. Blood flow and oxygen transfer rate of an outside blood flow membrane oxygenator.
J. Membr. Sci. 2000, 170, 153–158. [CrossRef]
13. Matsuda, N.; Sakai, K.; Yamamoto, K.-I.; Iwasaki, H. Effects of hollow fiber packing fraction on blood flow
pattern and gas transfer rate of an intravascular oxygenator (IVOX). J. Membr. Sci. 2000, 179, 231–241.
[CrossRef]
14. Funakubo, A.; Taga, I.; McGillicuddy, J.W.; Fukui, Y.; Hirschl, R.B.; Bartlett, R.H. Flow vectorial analysis in an
artificial implantable lung. ASAIO J. 2003, 49, 383–387. [PubMed]
15. Catapano, G. Turbulent flow technique for the estimation of oxygen diffusive permeability of membranes for
the oxygenation of blood and other cell suspensions. J. Membr. Sci. 2004, 230, 131–139. [CrossRef]
16. Zhang, J.; Nolan, T.D.; Zhang, T.; Griffith, B.P.; Wu, Z.J. Characterization of membrane blood oxygenation
devices using computational fluid dynamics. J. Membr. Sci. 2007, 288, 268–279. [CrossRef]
17. Tabesh, H.; Amoabediny, G.; Poorkhalil, A.; Khachab, A.; Kashefi, A.; Mottaghy, K. A theoretical model
for evaluation of the design of a hollow-fiber membrane oxygenator. J. Artif. Organs 2012, 15, 347–356.
[CrossRef]
18. Hirano, A.; Yamamoto, K.-I.; Matsuda, M.; Inoue, M.; Nagao, S.; Kuwana, K.; Kamiya, M.; Sakai, K.
Flow Uniformity in Oxygenators with Different Outlet Port Design. ASAIO J. 2009, 55, 209–212. [CrossRef]
[PubMed]
19. ISO5636-5(2013), Paper and Board – Determination of Air Permeance (Medium Range)—Part 5: Gurley Method;
ISO Technical Committee: Geneva, Switzerlnd, 2013.
Membranes 2020, 10, 362 26 of 26

20. Fukuda, M.; Izumi, R. Analysis of fluid flow in a membrane oxygenator device non-destructively.
Patent 045272, 31 August 2017.
21. Ministry of Health, Labour and Welfare. Chapter II Controlled Areas; Exposure Dose Limits and Measurements
(Indications of Controlled Areas) Article 3, Regulation on Prevention of Ionizing Radiation Hazards, Ministry of
Labour Order No. 41; Ministry of Health, Labour and Welfare: Tokyo, Japan, 1972.
22. ISO 7199. Third Edition: 2016 Cardiovascular Implants and Artificial Organs—Blood-Gas Exchangers (Oxygenators);
International Organization for Standardization: London, UK, 2016; pp. 11–15.
23. Galletti, P.M.; Richardson, P.D.; Snider, M.T.; Friedman, L.I. A standardized method for defining the overall
gas transfer performance of membrane oxygenators. ASAIO J. 1972, 18, 359–368. [CrossRef]
24. Evseev, A.K.; Zhuravel, S.V.; Alentiev, A.Y.; Goroncharovskaya, I.V.; Petrikov, S.S. Membranes in
Extracorporeal Blood Oxygenation Technology. Membr. Technol. 2019, 1, 201–211. [CrossRef]
25. Meyns, B.; Vercaemst, L.; Vandezande, E.; Bollen, H.; Vlasselaers, D. Plasma Leakage of Oxygenators in
ECMO Depends on the Type of Oxygenator and on Patient Variables. Int. J. Artif. Organs 2005, 28, 30–34.
[CrossRef]
26. Fournier, R.L. Chapter 9: Extracorporeal devices. In Basic Transport Phenomena in Biomedical Engineering,
4th ed.; Fournier, R.L., Ed.; CRC Press: Boca Raton, FL, USA, 2017; pp. 451–514.
27. Hagiwara, K.; Innami, K.; Yokoyama, K.; Kitoh, H.; Nogawa, A.; Muramoto, T.; Tatebe, K.; Seita, Y.;
Fukasawa, H. An approach to the microporous hollow fiber for the ECMO oxygenator—Micopore
characterization of the gas exchange performance, plasma leakage, and hydrophilization of the inner
surface of fibers. Jpn. J. Artif. Organs 1992, 21, 720–726.
28. Yamamoto, K.-I.; Hayama, M.; Matsuda, M.; Yakushiji, T.; Fukuda, M.; Miyasaka, T.; Sakai, K. Evaluation of
asymmetrical structure dialysis membrane by tortuous capillary pore diffusion model. J. Membr. Sci. 2007,
287, 88–93. [CrossRef]
29. Yamazaki, K.; Matsuda, M.; Yamamoto, K.-I.; Yakushiji, T.; Sakai, K. Internal and surface structure
characterization of cellulose triacetate hollow-fiber dialysis membranes. J. Membr. Sci. 2011, 368, 34–40.
[CrossRef]
30. Fukuda, M.; Saomoto, H.; Mori, T.; Yoshimoto, H.; Kusumi, R.; Sakai, K. Impact of three-dimensional
tortuous pore structure on polyethersulfone membrane morphology and mass transfer properties from a
manufacturing perspective. J. Artif. Organs 2020, 23, 171–179. [CrossRef] [PubMed]
31. Fukuda, M.; Yoshimoto, H.; Saomoto, H.; Sakai, K. Validity of Three-Dimensional Tortuous Pore Structure
and Fouling of Hemoconcentration Capillary Membrane Using the Tortuous Pore Diffusion Model and
Scanning Probe Microscopy. Membranes 2020, 10, 315. [CrossRef] [PubMed]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional
affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license ([Link]