| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;76:S2216-S2219
© 2003 The Society of Thoracic Surgeons
a Cachet Medical Limited, Columbia Heights, Minnesota, USA
* Address reprint requests to Dr Haworth, 4022 7th St NE, Columbia Heights, MN 55421, USA.
e-mail: billhaworth{at}aol.com
Presented at the symposium, "Gibbon & His Heart-Lung Machine: 50 Years & Beyond," Philadelphia, PA, May 2, 2003.
Abstract
From 1953 when Gibbon first successfully supported a patient with extracorporeal circulation to about 1980 many different types of oxygenators were developed. Since their introduction in the early 1980s, microporous hollow fiber oxygenators with blood flow outside the fiber have become the dominant type of oxygenator in use. Their success has been due to both the ability to specify the required properties for a good oxygenator and the application of modern design tools, especially computational fluid dynamics, to the design process. The result has been the availability of many oxygenators from different manufacturers that differ to some extent in their performance but all of which provide adequate performance for successful and safe clinical use.
The oxygenator used by Gibbon in 1953 for the first successful support of a patient [1] used vertical metal screens down which the blood flowed. These were contained in a chamber through which oxygen was flowing and this enabled carbon dioxide to diffuse out of the blood while oxygen diffused in. Gibbon reviewed his development of the oxygenator, circuit, and technique in a 1978 publication [2].
This ground-breaking achievement stimulated other researchers in the field and encouraged many more to enter it. They built their own equipment to support patients during cardiac arrest and developed new surgical techniques to repair cardiac defects. It further prompted the development of replacement heart valves, which could now be placed in the anatomic position. It led in addition to the invention and development of pacemakers, used initially to treat heart block resulting from the surgical procedure but soon applied to pace hearts with irregular beats from causes other than surgery. It really was the birth of revolutions in both heart surgery and cardiology.
During the 2 decades after Gibbon’s first successful case the equipment used, the heart-lung machine, developed rapidly and in many different directions. The technology current in 1976 was summarized in a contemporary review by Bartlett and Harken [3]. The history of oxygenator development has also been reviewed recently by Leonard [4].
There were essentially two different approaches being pursued: those involving direct contact between blood and oxygen, and membrane oxygenators in which a membrane was placed between blood and oxygen.
Direct contact devices included bubble oxygenators in which, as the name implies, oxygenation was achieved by bubbling oxygen directly through the blood. For such oxygenators oxygenation of the blood was less an issue than was the separation of the bubbles and foam generated to ensure that no gas emboli were delivered to the patient. Other direct contact devices used screens as in Gibbon's oxygenator, and in another approach the blood was filmed by being drawn over a series of discs rotating in the blood.
Membrane oxygenators were produced as coiled membrane envelopes or as stacks of flat sheets of membrane. A variety of different membrane materials were used, with silicone rubber becoming an early favorite because of its high gas permeability. The carbon dioxide permeability of silicone is about five times greater than the oxygen permeability and this partially compensates for the smaller available driving force for carbon dioxide transfer. Microporous nonwettable membranes were also introduced and because their gas transfer rates are significantly higher than silicone, they allowed the use of lower membrane area. These microporous membranes were initially made from polytetrafloroethylene. Polypropylene microporous membranes were introduced soon after and displaced the polytetrafloroethylene membranes because they were lower cost and were stiffer, which offered better control of the blood channel geometry, a critical factor affecting oxygenator transfer performance.
By 1980 for routine surgery a bubble oxygenator would be used and there were several different brands available on the market. It was known that the damage to blood caused by the bubble oxygenator was progressive; as a result a patient could be successfully supported only for the limited time of a few hours. If therefore the case was anticipated to be difficult or protracted, a membrane oxygenator would be used. Long-term support of a patient became known as ECMO (extracorporeal membrane oxygenation) and for these cases only a silicone membrane oxygenator was suitable. That is because plasma is eventually able to wet a microporous membrane, a phenomenon called plasma breakthrough, and when that happens gas transfer rates through the membrane are substantially reduced. A solid silicone membrane does not suffer from this effect and so may be used for support of a patient for several days without significant decline in performance.
The perceptions of the relative performance of the bubble and membrane oxygenator at that time (1980) were that the membrane devices had better blood handling (because they could be used for longer perfusions) but that they suffered from a number of poor characteristics. They were considered to have higher prime volumes and to be somewhat compliant so that the hold-up volume in the oxygenator would change with the back pressure on the device. They were regarded as more difficult to use. The silicone membrane products required special priming techniques to ensure that all air was displaced from the system before perfusion of the patient was started. Some early microporous oxygenators used two pumps to protect the oxygenator from high pressures because of concerns about the strength and compliance of the membrane. They had lower gas transfer rates than did bubble oxygenators and so it was very important not to "get behind" the patient's requirements for oxygen because then it would be difficult to increase oxygen delivery rates enough to compensate. Membrane oxygenators were more expensive and were more likely to fail during a case, most often by leaking.
Almost certainly as the direct result of these perceptions, several things happened early in the 1980s. The integrated reservoir was introduced. This combined the venous reservoir and the oxygenator into a single unit, which looked much more like a bubble oxygenator and was run in a similar way. Manufacturing methods borrowed from hemodialyzers helped reduce the manufacturing costs and make an oxygenator less liable to leak. These changes were introduced first into a flat plate membrane oxygenator. But probably the most far reaching change was the introduction of microporous hollow fibers and the manufacture of oxygenators from these membranes. Hollow fibers had been used previously with blood flow inside the fiber and gas flow outside. This new design reversed this flow arrangement with the blood flowing over the outside of the fibers and the gas flowing down the fiber lumen. This is a significantly more efficient arrangement for gas transfer, so that smaller membrane area will provide adequate gas transfer capability, and it is also substantially noncompliant so prime volume does not change with back pressure on the oxygenator.
The combination of all these factors addressed the concerns with the early membrane oxygenators at the same time as reducing the manufacturing costs to a level much closer to those for bubble oxygenators. Over the next decade a profusion of hollow fiber designs was introduced that used blood flow outside the fiber. In general they offered consistent high levels of gas transfer and other performance features at prices comparable to bubble oxygenators. Bubble oxygenators were displaced from the market.
During the past 20 years the almost universal design choice for blood oxygenators has become the microporous hollow fiber oxygenator with blood flow outside the fibers. That is the modern oxygenator of the title. The history of the device however is just the most obvious part of the story. Less visible but central to this evolution is the history of the design methodology, which enabled the development of the modern oxygenator.
All early designs were almost entirely empirical, that is to say prototypes were constructed and then laboratory testing was conducted to measure performance. There would be an iterative process in which adjustments were made to the initial prototype based on the performance tests and using experience, intuition, and hypothesis to construct a subsequent version that was tested. The process was likely repeated many times over in order to arrive at a design meeting clinical requirements.
Because the clinical procedures were also developing in the early days, this design process would undoubtedly have been a moving target. It was only later during the development of cardiac surgery and as procedures became more standardized that it became possible to identify a design specification for an acceptable oxygenator. The technical development process could then be directed toward a reasonably well defined target. Such a process has certainly been possible for the last 15 years at least and this is the period that has seen the burgeoning of hollow fiber oxygenator designs.
A design specification would include quantitative targets for the basic performance measures of oxygen and carbon dioxide transfer capability, blood and gas side pressure drop, prime volume, membrane area, heat exchanger efficiency, and manufacturing costs. Other variables might also be included in the list and although these numbers almost certainly differed in detail between manufacturers they would not be very different. A second list of qualitative targets might also be identified. These might include requirements for consistency, ease of prime, clarity, compactness, good blood handling, and so on.
It is not possible to address each of these issues in detail and in any case the development approach is likely to have differed for each manufacturer of oxygenators. Each manufacturer almost certainly differed in the tools available to the development group. But by the start of the 1990s computer-aided design and computer-aided machining were widely available. Computer-aided design enabled the design of components with complex curved shapes that often could not have been captured with precision by a traditional draftsman and computer-aided machining enabled the production of prototypes faithful to the design and also allowed the manufacture of tools for molding the production components.
Another technology that became available was that of rapid prototyping. There is a variety of different approaches to the objective of producing functional prototype plastic components directly from the plastic in the form of monomer, beads or thin rod. They all avoid the costly and time-consuming process of machining a component from solid plastic and allow more designs to be evaluated more readily.
A third technology that also became much more widely available at about the same time was that of computational fluid dynamics (CFD) [5, 6]. Computational fluid dynamics is used to obtain a numerical solution for the partial differential equations that govern fluid flow, the continuity equation, and the Navier-Stokes equations. This process can be used to generate a full map of the velocity and pressure in a conduit of complex geometry such as a blood heat exchanger and oxygenator. Heat and gas transfer could then be calculated from the CFD velocity field by using mass transfer correlations such as those determined by Mockros [7] or Yang and Cussler [8].
Blood is rheologically complex and that complicates somewhat the process of solving the fluid flow equations. A more significant concern is that the species being transferred, oxygen and carbon dioxide, are present both as dissolved species in plasma and chemically combined with hemoglobin in red cells. Carbon dioxide is further present in plasma in equilibrium with bicarbonate ions. Despite this level of complexity, with the refinement of the CFD programs and the increasing memory and speed of computers it has become possible to solve the equations for heat and for mass transfer numerically. That enables both the heat exchanger performance and gas transfer performance of a candidate device to be determined computationally.
Of course the advent of additional design tools did not preclude the use of the traditional empirical approach. Indeed it would be a foolhardy development group that relied exclusively on computational methodology. An approach that combines both computational studies and empirical testing will probably continue to be most effective for the foreseeable future. However as experience and confidence is gained with the computational results, the amount of empirical testing necessary to prove a design conforms to the target specification is likely to continue to decline.
During the last couple of decades therefore both the specification for the oxygenator has become much more consistent and the design tools available to achieve these targets have become much more powerful. Despite this it is not possible to construct a specification and then use a computer to solve a set of complex equations in order to identify one or more designs capable of meeting the specification. That is because there are many different flow configurations through the hollow fiber bundle of an oxygenator and each of these configurations impose different boundary conditions on the equations to be solved. However once a flow configuration is selected, the equations can be cast in terms of the design variables available and solutions found identifying the effect of changes in these variables.
Suppose that a configuration has been chosen for the oxygenator in which for example gas flows from top to bottom and blood flows from a hollow center radially through the fiber bundle. Now CFD can be used to calculate the pressure and flow field throughout the oxygenator. This result gives the pressure drop across the oxygenator directly. The velocity data can be combined with mass transfer correlations to calculate oxygen and carbon dioxide transfer rates. Alternatively a CFD program that incorporates mass transfer capability can be used to calculate gas transfer rates directly. The local shear stress for any position in the blood flow path can be readily determined from the velocity field and this can be used to determine the extent of shear induced platelet activation, a key feature of the blood handling of an extracorporeal circuit. These calculations can be performed for different values of each design variable, such as fiber diameter, total surface area, and packing fraction of the fibers.
Flow shunting is the result of preferential pathways for flow through the oxygenator. It can result from poor fit of the fiber bundle to the case or from regions of lower and higher packing fraction within the bundle. Because it inevitably degrades performance it differs in nature from the design variables identified above. Shunting must nevertheless be considered by the designer and all possible steps taken to avoid it. That is probably the principle reason for the increasing popularity of fiber bundles in which the fiber spacing is controlled by knitting the fibers into a mat.
The information that results from such calculations as those described above is specific and quantitative. It can conveniently be presented qualitatively in tabular form (Table 1), which identifies the choices available to the designer for meeting the performance specification. It is clear from such a representation that the process is one of optimization because a change in any of the design variables produces both desirable and undesirable performance changes. The qualitative table relating performance to design variables will be similar for any flow configuration but the quantitative values will differ depending on the flow arrangement selected. Despite this the final performance measures for a variety of current oxygenators fall close to each other [9]. There are differences but they are relatively small and are only distinguished with diligence or when the performance of the oxygenator is challenged. Modern oxygenators are designed so they are not challenged under normal conditions of perfusion.
|
References
This article has been cited by other articles:
|
|
 |
|
  D. M. Hoganson, J. L. Anderson, E. F. Weinberg, E. J. Swart, B. K. Orrick, J. T. Borenstein, and J. P. Vacanti Branched vascular network architecture: A new approach to lung assist device technology J. Thorac. Cardiovasc. Surg., November 1, 2010; 140(5): 990 - 995. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Murphy, E. A. Hessel II, and R. C. Groom Optimal Perfusion During Cardiopulmonary Bypass: An Evidence-Based Approach Anesth. Analg., May 1, 2009; 108(5): 1394 - 1417. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
