Conventional hollow fibers are mainly made up of microporous polymers such as polypropylene (PPL). Such fibers are extensively used as membranes for the production of current capillary membrane oxygenators (CMOs).1 Microporous hollow fibers (MHFs) are considered efficient mostly because of the high gas exchange performance and the satisfactory antithrombogenic properties they provide2; however, controversies come forth regarding their endurance and durability especially during long-term applications (1–2 weeks).3,4
The main shortcoming of MHF is the so-called plasma wetting. It is an unpredictable phenomenon in which plasma breaks through the micropores of the capillaries into the gas phase leading to failure of the oxygenator mechanism. Its major cause is still ambiguous. Mottaghy et al.5 have reported that the plasma condensation generated by a temperature gradient leads to loss of surface tension and consequently induces a capillary effect through the micropores. Tamari et al.6 have suggested that albumin might act as a wetting agent leading to plasma leakage effect, whereas Montoya et al.7 have reported that plasma wetting might be initiated by phospholipid adsorption at the surface of the micropores of fibers. Yet, these suggestions have not led to a satisfactory solution to handle this problem. Polymethylpentene (PMP) is another microporous capillary, yet classified as a plasma resistant because of its thin outer-coated surface layer. It has also shown a prominent gas exchange performance as PPL and even better hemocompatibility where platelet consumption and resistance to blood flow seem to be greatly reduced.8 However, a study reported that an unjustified incident of leakage occurred after 48 hours for two PMP-based oxygenators.9 PMP may also leak air instead of plasma during long-term gas exchange, resulting in an increase of the inherent inflammatory activation of the extracorporeal circulation. However, this effect of the gas emboli is highly significant for prompting blood vessel occlusion if blood is returned through the arterial circulation.8–10,15
Therefore, a better solution for long-term gas exchange application is a nonporous material such as a silicone membrane. This explains why the old Kolobow silicone spiral coil membrane oxygenator (Medtronic Inc., Anaheim, CA), introduced in 1972, is still widely used for extracorporeal membrane oxygenation in clinics worldwide despite its gas insufficiency and related handling complications.
An innovative approach would be to combine the advantages of both MHF and silicone membranes. A retrospective overview reveals many attempts for fabricating a silicone capillary membrane. The first attempt was in 1963 by Bodell et al.11 Marx12 also introduced a silicone rubber capillary membrane a short time after. Rais-Bahrami et al.13 had even produced the Mera Silox (Mera Kendo Medical Instruments, Japan) commercial oxygenator in 1992 that showed high pressure drops and less transfer rates compared to the microporous oxygenators. Furthermore, Funakubo et al.14 have initiated in 1996 the development of a new silicone rubber hollow fiber with a reduced membrane thickness compared with that in the earlier attempts. It has shown promising results in gas transfer and hemocompatibility during a 2 week (in vivo) study, yet this concept is not commercially realized. Another recent study by Lafayette et al.15 in 2009 has also shown that a novel approach may be applicable in the near future. “True” membranes, unlike microporous ones, obviously create a higher resistance barrier for gas to diffuse. Therefore, for a silicone capillary to act as an effective gas exchange apparatus, the resistance across the membrane has to be as low as possible according to Fick’s first law of diffusion.
Apparently, it is not difficult to fabricate a fine, ultrathin silicone hollow fiber, which is an intuitive solution that can lead to a considerable reduction in the diffusion barrier exhibited by the thickness (δ) of the membrane wall. In fact, the challenge lies in the ability of fabricating such a true membrane capillary while preserving its intactness, and at the same time maintaining adequate stability, that is, the mechanical strength of the membrane walls, which is a major difficulty of earlier studies.16
In this article, we introduce the performance characteristics of a new silicone rubber capillary. Referring to its configuration properties, it is entitled as silicone hollow sphere (SiHSp) fiber (Raumedic AG, Münchberg, Germany). So far, the purpose of the study was to evaluate the effect of modifying the structural design specification of the SiHSp fiber on gas exchange for both O2 and CO2. This evaluation was performed by means of in vitro testing for a series of constructed, miniaturized SiHSp oxygenators.
Materials and Methods
To avoid the previously mentioned shortcomings, we have depicted a novel approach to reduce the gas transfer resistance without reducing the dimensional membrane thickness. The concept is to fabricate a three-dimensional lattice, which can mimic the structure of the bee comb in its architectural properties; in other words, the least solid material is to be used for constructing a structure that preserves a requisite stability as shown in Figure 1. Thus, the membrane resistance toward gas transfer can be significantly reduced. As a result, gas may now diffuse easily in the free spheres (cavities) introduced inside the membrane walls without the need of a drastic reduction in the membrane thickness geometry (Figure 2).
This model is realized via the following manufacturing processes: A commercially available medical-grade silicone elastomer (polymethylvinylsiloxane) is mixed on a roller mill with a catalyst for cross-linking that will later on create the foam structure (spheres) via a propellant. During the extrusion process, the premixed elastomer is extruded to shape the required tubing in the desired dimension. During the extrusion process, no heat is applied, and there is no chemical reaction in the silicone matrix. The extruded tubing is pulled through a vulcanization oven with temperature set above 300°C, where two chemical reactions are initialized by the exposure of the extruded tubing to such a high temperature. First, the platinum-organic catalyst will start a cross-linking reaction of the polysiloxane polymers. This cross-linking reaction results in a three-dimensional network of polysiloxane chains that are covalently bonded together. Second, initialized by the heat exposed to the tubing, the propellant that had been mixed to the polysiloxane will release H2O and nontoxic gases such as CO2. The released gases form a porous foam structure within the silicone rubber matrix of the tubing. The number and size of the air bubbles inside the wall of the tubing are determined by the concentration and distribution of the propellant inside the polysiloxane mixture. Last, the foamed silicone tubing is placed in a curing oven for several hours at an elevated temperature above 150°C. This postcuring process is necessary to finalize and complete the cross-linking reaction within the silicone matrix. The result is a completely crosslinked silicone rubber tubing with a porous foam structure.17
In total, six different configurations of SiHSp fibers were produced. The fibers have geometrical design alterations as shown in Table 1. This is done to experimentally assess its impact on gas exchange while preserving adequate mechanical stability. The produced fibers have the same concentration as the formed bubbles (spheres) inside the walls. The density of the membrane depends mainly on the proportion of the embedded spheres; in this report, the ratio of spheres to dense material is 0.35. The optimal membrane surface area of an oxygenator is a substantial parameter in determining the gas exchange efficiency of the oxygenator.18 Therefore, different geometrical dimensions of fibers in the SiHSp oxygenator modules and the two reference prototypes were carefully adjusted to produce a uniform effective membrane surface area (Aeffmem = 0.02 m2). In contrast to conventional flowing phases in CMOs, blood in SiHSp oxygenators and in reference modules flows within the capillaries. This, of course, weakens the transfer rate. However, the purpose at this stage is not to test the SiHSp modules as clinical oxygenators, but rather the fiber’s material. This way, a laminar flow is generated, and thereby, the influence of the fluid profile on mass transfer would be excluded or at least subordinated to the membrane mass transfer. This is considered to be a very convenient method for a fair comparison where an identical aligning of flexible silicone nonwoven mats of fibers between all modules is unfeasible and would generate a lot of error. The first reference module is constructed from OXYPHAN fibers (a common type of PPL) (Membrana GmbH, Wuppertal, Germany), whereas the second reference module is made up of the same silicone raw material but possessing no spheres, referred to as solid silicone (Solid.Si). The performance of Solid.Si tangibly manifests the impact of spheres on reducing wall resistance of SiHSp. Both reference modules (Figure 3) are nevertheless constructed with geometrical design dimensions similar to that of SiHSp modules.
Fiber Impermeability Inspection
Before gas exchange inspection, the impermeability of the fibers that were used for constructing the test modules had to be evaluated. The leakage test was performed by applying a hydrostatic pressure using a polycarbonate tube column (h: 1 m, ø: 0.25”) connected to the superior top of the module. The column was also connected to the extracapillary compartment inlet of the module positioned vertically. The column and the module were totally filled with perfluorocarbons (FC-43). The physical properties of such a fluid, like high density (1.88 g/cm3), low intermolecular forces, and low surface tension (16 dyn/cm), make it a much favorable media over water and blood for an oxygenator’s leakage test.19 This circuit was able to generate a mean pressure drop (150 mm Hg) between the inlet and outlet, which was considered high enough to force any possible leakage in the fiber walls. Underneath the modules, a beaker, situated at the intracapillary compartment outlet, quantified the amount of leaked FC-43 into the fiber’s inner compartment. By means of such a procedure, the seal and compactness shortcomings, which might have been imposed through the process of manufacturing, can be detected. After 24 hour application, only leakage-free modules can be qualified for the later gas exchange investigation.
Gas Exchange Inspection
In vitro gas exchange investigations were performed in two stages. At first, testing was performed with a regular saline solution as the fluid phase. It has been reported as a suitable fluid for the first-step oxygenator testing. This is attributable to its easy handling property for evaluating the intactness of fibers on one hand, and measuring the diffusion efficiency with a hemoglobin-free medium on the other.16 Later on, the same prototypes were tested with heparinized porcine blood. Out of each of the fiber modules reported in this article, at least five identical prototypes were produced to make sure that the data are reproducible and to enhance their precision as well. During the investigations, tests were performed according to the recommended practice and guidelines of the Association of Advanced Medical Instrumentation (AAMI) standards.20
Figure 4 shows how the in vitro testing is assembled by the heart or lung machine Stöckert SIII (Sorin Group, Milano, Italy). The experimental circuit contains three membrane oxygenators (one test oxygenator and two deoxygenators). The two deoxygenators are used to assemble the standard venous blood gas parameters according to AAMI standards and in accordance to the norms of DIN EN ISO 12022. The deoxygenators are connected to gas inlets (N2, CO2) with different flow rates (VGas: ml/min), whereas the test oxygenator is plugged into a 100% O2 gas inlet. The appropriate polyvinylchloride (PVC) and silicone rubber tubing, two arterial filters (each connected to the deoxygenators with tubing), and the appropriate connectors with stopcocks for blood sampling are set in the circuit. Two PVC reservoirs with a sampling port are used to recirculate the blood for adjusting the inlet blood parameters. Two HILITE 7000 oxygenators (Medos AG, Stolberg, Germany) are used as deoxygenators. Temperature is maintained at 37°C (±2°C) with a heat exchanger connected to a water bath. The blood flow is maintained and monitored with a Transonic T108 Flow Meter (Transonic Systems Inc., Ithaca, NY). The pressure drop is recorded during the runtime testing with a laboratory-made reliable liquid pressure system. It contains four channel digital control units and two pressure sensors positioned at the inlet and outlet of the test oxygenator. Appropriate concentrations of saline solutions and sodium bicarbonate were administrated to control hematocrit level and base excess, respectively, according to the standard values. The blood gases are measured using a blood gas analyzer ABL720 (Radiometer, Copenhagen, Denmark). To have a defined surface area, blood flows inside the capillaries whereas gas flows in a countercurrent manner.
Experiments were conducted up to a maximum flow rate (Q: ml/min) of 200 which is considered as a demeaned flow for such small-scale modules. Each module had to be tested solely under the same inlet conditions. The procedure was repeated to ensure the reproducibility of the obtained results. At least five probes were taken from the outlet and two probes from the inlet at every flow rate to increase the precision of measurements. VGas/QBlood for all tests was performed at 1:1 ratio.
Impermeability inspection results show leakage neither in the walls of the used fibers nor in the potting material in all of the tested SiHSp modules. Besides, SiHSp modules show easy handling characteristics as a sign of adequate stability of the silicone material. As a result, an extra reduction in di and δ is still feasible in future for a better gas exchange medium while ensuring sufficient intactness and stability.
Figure 5 represents the mean ΔpO2 and ΔpCO2 of SiHSp C3771, F3774, and Solid.Si of an identical inner diameter (di = 600 μm). However, initial δ of F3774 is greater than Solid.Si which is greater than C3771 by 10 μm (see Table 1). With such a uniform di and a very nearly uniform δ, a fair comparison for gas exchange is granted. This slight difference in δ was thought to define a boundary for the Solid.Si to fall between the two other fibers in case no effect of spheres is recorded. Results of SiHSp clearly show a distinct improvement in gas transfer for both O2 and CO2 at all flow rates compared with that of Solid.Si. These results strengthen the belief that by increasing the proportion of embedded spheres, gas exchange would be promoted.
The results also show that even the SiHSp with largest di (twice as big as a PPL fiber) and δ (110 µm to <1 µm in PMP) can still oxygenate the inlet value from 40 to 120 mm Hg at maximum flow rate. The results depicted in Figure 6 compare SiHSp fibers to OXYPHAN reference module. Here, results are also in mean ΔpO2 and ΔpCO2. OXYPHAN shows the highest values at all flow rates, for example, at Q = 100 ml/min, OXYPHAN has ΔpO2 = 239.9 (mm Hg) ± 0.37 (mean SD) and ΔpCO2 = 35.7 (mm Hg) ± 0.34, whereas SiHSp fiber A3769 records ΔpO2 = 194 (mm Hg) ± 0.31 and ΔpCO2 = 30 (mm Hg) ± 0.63. The capacity for CO2 removal is also in accordance with the oxygen uptake characteristic.
Figure 7, A and B represent the oxygen uptake and carbon dioxide as transfer rates. The OTR and CTR for SiHSp fiber A3769 are 12.6 (mlO2/min) and 10.4 (mlCO2/min), respectively, at Q = 200 ml/min, whereas OXYPHAN has relatively higher values for both O2 and CO2 of 14.1 mlO2/min and 13.2 mlO2/min, respectively. Solid.Si records the lowest values for both O2 and CO2 among all tested modules.
So far, the purpose of the study was to get sufficient information about the optimal configuration, such as wall thickness, diameter of capillaries, amount of embedded spheres, and enough mechanical stability for fiber arrangements. Using this information, the SiHSp configuration to be developed in future can be optimized to increase its gas exchange potential while preserving the air or blood leakage-free properties and easy handling. Because the raw material is purely made up of silicone, which is well-known for its high hemocompatibility, then the remaining concern regarding hemocompatibility would be the membrane surface profile. In our opinion, the apparent curvatures (dents) that exist are obviously smooth and do not show any sharp edges, as they have to be considered as if they are “injection moldings” of the embedded microspheres and therefore, cannot create any steep sharp edges. On the contrary, we believe that these dents can lead to a formation of microsecondary flow region that would even affect mixing the physically dissolved oxygen in blood. In addition, the typical plasma layer between the membrane and blood cells would be reduced, and therefore, not only the oxygen uptake in the blood plasma will be enhanced but also the diffusion through the erythrocyte membrane will be facilitated.
Until now, no differences in the (negligible) hemolysis results between SiHSp and the other membranes (Solid.Si and OXYPHAN) have been recorded. These are, however, preliminary evidences that cannot be taken whatsoever as an evidence of high hemocompatibility. To this point, our primary concern was to investigate the gas exchange which was the major aspect for evaluating this material. Therefore, larger modules of SiHSp, at the same time, long-term experiments, are planned for the next study, to specifically demonstrate SiHSp’s advantage and dominancy over other materials in improved hemocompatibility as an oxygenator hollow fiber.
The results demonstrate that SiHSp tested fibers have by all means a gas transfer capacity comparable with that of the current oxygenator hollow fibers, such as PPL and OXYPHAN. Although the gas transfer efficiency of SiHSp is slightly lower compared with that of PPL, it is nevertheless higher than the conventional silicone membrane, thus emphasizing the advantage of this concept.
1. Lim MW. The history of extracorporeal oxygenators. Anaesthesia. 2006;61:984–995
2. Broomé M, Palmér K, Scherstén H, Frenckner B, Nilsson F. Prolonged extracorporeal membrane oxygenation and circulatory support as bridge to lung transplant. Ann Thorac Surg. 2008;86:1357–1360
3. Nose Y, Motomura T, Kawahito S Oxygenator-Artificial-Lung, Past, Present, and Future. 2001 Houston, TX ICAOT/ICMT Press
4. Zwischenberger JB, Alpard SK. Artificial lungs: A new inspiration. Perfusion. 2002;17:253–268
5. Mottaghy K, Oedekoven B, Starmans H, et al. Technical aspects of plasma leakage prevention in microporous capillary membrane oxygenators. ASAIO Trans. 1989;35:640–643
6. Tamari Y, Tortolani AJ, Lee-Sensiba KJ. Bloodless testing for microporous membrane oxygenator failure: A preliminary study. Int J Artif Organs. 1991;14:154–160
7. Montoya JP, Shanley CJ, Merz SI, Bartlett RH. Plasma leakage through microporous membranes. Role of phospholipids. ASAIO J. 1992;38:M399–M405
8. Peek GJ, Killer HM, Reeves R, Sosnowski AW, Firmin RK. Early experience with a polymethyl pentene oxygenator for adult extracorporeal life support. ASAIO J. 2002;48:480–482
9. Sato H, Odeleye ME, Chambers SD, et al. Thoracic artificial lung (TAL) development: Determining the most suitable fiber for TAL. ASAIO J. 2005;51:51A
10. Puis L, Ampe L, Hertleer R. Case report: Plasma leakage in a polymethylpentene oxygenator during extracorporeal life support. Perfusion. 2009;24:51–52
11. Bodell BR, Head JM, Head LR, Formolo AJ. Oxygen flow through capillary silastic tubing as a principle for a membrane oxygenator. Q Bull Northwest Univ Med Sch. 1962;36:94–98
12. Marx G. Verfahrenstechnische Grundvorgaenge in Natuerlichen und Kunstlichen Lungen und deren Anwendung. 1968 Dissertation TU Muenchen University
13. Rais-Bahrami K, Mikesell G, Seale WR, Rivera O, Hearty JP, Short BL. In vitro evaluation of the Mera Silox-S 0.5 and 0.8 m 2 silicone hollow-fibre membrane oxygenator for use in neonatal ECMO. Perfusion. 1992;7:315–320
14. Funakubo A, Higami T, Sakuma I, et al. Development of a membrane oxygenator for ECMO using a novel fine silicone hollow fiber. ASAIO J. 1996;42:M837–M840
15. LaFayette NG, Schewe RE, Montoya JP, Cook KE. Performance of a MedArray silicone hollow fiber oxygenator. ASAIO J. 2009;55:382–387
16. Drummond M, Domingo B, Paula A, et al. Technological evolution of membrane oxygenators. Braz J Cardiovasc Surg. 2005;20:432–437
17. . Ziembinski R. Gas exchange membrane in particular for use in artificial lung and method for the production of a gas exchange membrane of this type. Patent application publication 2010:US2010/0050875 A1
18. Kashefi A. Untersuchungen zu Stofftransport und Fluiddynamic bei extrakorporalen Membranoxygenatoren. Dissertation. 2004 RWTH Aachen University ISBN: 3-86130-764-2
19. Mottaghy K, Driessen G, Zander M, Kreisel C, Mendler N, Schmid-Schönbein H. The floating-droplets oxygenator: developments using fluorocarbons with coaxial rotating cylinders. Trans Am Soc Artif Intern Organs. 1977;23:464–469
20. . Association for the Advanced of Medical Instrumentation: Cardiovascular implants and artificial organs. Blood Gas Exchangers. 1996:7199
plasma leakage; cardiopulmonary bypass; membrane oxygenation; hollow fibers