Definition of Biocompatibility

The selection of the materials used in the construction of prostheses and implants is basically focused on their ability to maintain mechanical, chemical and structural integrity and on various characteristics which allow this function to substitute any organ or tissue properly and exhibit safe, effective performance within the body. For a number of years, biocompatibility has been defined as the ability of a material to perform with an appropriate host response in a specific application. Under this definition, which appears relatively ambiguous and vague, any material used satisfactorily in orthopedic surgery may be, for instance, inappropriate for cardiovascular applications because of its thrombogenic properties. Another widely used material may have deleterious effects if used under stress-strain conditions because of wear particles generation. Biocompatibility is by no means a measurable entity. One may simulate observing the biocompatibility of a material by comparing its behaviour to reference materials in standardized experimental conditions.

In a clinical context, the behaviour of autologous grafts or tissues becomes the gold standard for any given material to be considered biocompatible. In fact, biocompatibility is a complex notion which has to be interpreted as a series of events or interactions happening at the tissue/material interface whose outcome must be satisfactory or optimal. These interactions are influenced by intrinsic characteristics of the materials but also by the confrontational circumstances namely, the biological site destined for implantation, and most of all, the inflammatory setting induced by the surgical act and maintained by the presence of the material. In the last few decades, considerable progress has been made regarding the tissue/material interactions upon implantation which allowed the identification of materials and surface characteristics which are more biocompatible. Scientists are now elaborating new strategies to facilitate the integration of various prostheses and implants to their respective organ or tissue sites by proposing new bioactive and biocooperative materials.

Biocompatibility Tests

A few decades ago, it was generally thought that biomaterials destined for implantation had to be chemically inert and consequently were believed to play no major role in any physiological process. However, it has recently become very clear that materials which stimulate prosthetic incorporation in the tissue offer considerable advantages and are likely to be more successful than those that do not heal satisfactorily. With the development of sophisticated analytical techniques, progress has been made to elucidate which molecular or cellular responses are critical in host/material reactions and what surface characteristics and/or polymer chemistry are important in mediating these reactions toward host tolerance. It is only through local tissue response that a dynamic interaction will occur between activated inflammatory cells and the secretion of cytokines to stimulate angiogenesis, smooth muscle cells and fibroblast proliferation and collagen production in a cascade of events leading to the encapsulation and satisfactory healing of prostheses or implants.

With the increasing number of synthetic materials being introduced into the field of medicine, there is an increasing demand for more discriminating tools to evaluate their safety and efficacy. Any severe reactions by the host toward the biomaterial will probably result in failure. Consequently, the need for standardized methods and protocols for assessing the biological response of materials has never been greater. A list of tests being used to assess the biological response of materials is shown in Table 1.

Table 1

Test methods used to evaluate biological response.

At present, a variety of different types of cell culture methods are frequently used. They involve in vitro studies which may assess the morphology, cytotoxicity, or secretory functions of different cell types (usually those that will be in contact with the material during implantation). This may be achieved by either a direct or indirect contact assay, or by adding a diluted extract from the biomaterial to the culture medium. Other types of tests determine various cell functions such as cell membrane integrity, replication, phagocytosis, the production of reactive oxygen species, secretion, activation, chemotaxis, and chemokinesis.

Blood contact assays have been developed and include tests investigating the adhesion or activation of blood cells, proteins, and macromolecules such as those found in the complement or coagulation cascades. Other biocompatibility tests have been tentatively proposed and involve analytical testing or observations of physiological phenomena, reactions or surface properties attributable to a specific application such as anti-bacterial surface testing, protein adsorption characteristics, calcification or mineralization processes.

In spite of this plethora of possible test methods, the protocols currently required by North American and European regulatory agencies for the evaluation of biocompatibility do not incorporate such quantitative test procedures aimed at determining the nature and intensity of cellular reactions (Table 1). Only qualitative assessments are recommended, and these involve the estimation of acute toxicity through the gross microscopic examination of tissue, which relies on the subjective expertise of investigators in identifying cell morphology and the severity of the reaction. While such toxicity testing is essential, this method alone should not be used when evaluating the biocompatibility of a material. Before claiming the biocompatibility of one given biomaterial, one must perform a series of tests requiring the evaluation of cell compatibility, toxicity, mutagenicity as well as additional in vitro tests for a specific application which will guarantee the function and inocuousness of the device. It is generally accepted that in vitro tests are essential prior to conducting in vivo trials. Such preliminary screening tests are quick to perform, reproducible and inexpensive. While recognizing that their main limitation is their relative simplicity compared to the complex interactions that occur in vivo, their sensitivity in discriminating between biocompatible and nonbiocompatible materials (with proper control materials) can be used to reject unsatisfactory materials before performing in vivo implantation studies. Thus, there is a need to find the most appropriate and sensitive in vitro tests which correlate well with the in vivo response.

In order to determine the biocompatibility of materials used in the construction of synthetic vascular prostheses, we recently attempted to establish a direct correlation between such in vitro results and the in vivo healing of arterial grafts. The usefulness of six different in vitro tests in predicting the in vivo healing performance of vascular grafts was evaluated in those grafts known to exhibit a variety of different healing responses as observed during previous in vivo trials in a canine thoraco-abdominal bypass model. The results were able to demonstrate that some in vitro tests, namely, polymorphonuclear cell (CD₁₈) and lymphocyte (IL-2) receptor expression assays, direct contact assay using endothelial cells and an extract dilution assay on mouse fibroblasts were useful in predicting the in vivo behaviour of vascular prostheses. Of particular interest were the uncleaned arterial grafts which had shown poor healing in previous in vivo studies and induced elevated expression of CD₁₈and IL-1 receptors. These grafts caused poor endothelial cell migration and viability and generated a cytotoxic response from fibroblasts in an extract dilution assay.¹

Decade 1970–1985

With the pioneer work of Boretos in the late 1960s² and of Lyman early 1970s,³ who both claimed the blood compatibility of polyurethane-urea polymers, these elastomer materials have since been widely used for biomedical applications such as the artificial heart,³intra-aortic balloons,⁴pacemaker leads,⁵heart valves,⁶and hemodialysis membranes.⁷ It was soon realized that blood compatibility was intimately related to their microphase separated structure composed of hard and soft segment domains.^8,9The first significant studies on the blood response characteristics of polyurethanes (PUs) were reported in the early 1980s.¹⁰The first emphasis was put on the effect of chemical composition with the studies of da Costas et al on a series of segmented polyurethanes of various surface composition using three different polyethers as soft segments, namely, polytetramethylene oxide (PTMO), polypropylene oxide (PPO), and polyethylene oxide (PEO), and two diisocyanates, 2,4-toluene diisocyanate (2,4-TDI) and 4,4'-diphenylmethane diisocyanate (MDI). Using a simple homemade in vitro test system, the platelet retention index of various PTMO, PPO and PEO synthetized PUs showed lower indexes for PEO-PUs as compared to PTMO-PUs, PPO-PUs or hard segment analogs.¹¹The relationship between blood response and hard/soft segment concentrations was further confirmed with Biomer¹² and other studies which showed that the hard segments of PUs were highly thrombogenic in platelet retention experiments.¹³ Two reasons were proposed to explain the high thrombogenicity of hard segments: first, the high crystallinity of the polymers and second, they exhibit surfaces with strong hydrogen bonding. Merrill et al suggested that surface mobility may play a role in the interaction of polymers with biological systems.¹³Using scanning electron microscopic techniques, Takahara et al also demonstrated that the platelet adherence and morphology on PU films was significantly modified by the size and characteristics of the microphase separation in the polymer.¹⁴

In an attempt to reduce the thrombogenic potential of polyether-urethanes (PEUs), Lelah et al investigated the effects of ionic domains on platelet and fibrinogen deposition at the surface of PEUs.¹⁵ On the hypothesis that negatively-charged surfaces show better blood compatibility, they chemically modified their chain extender N-methyldiethanolamine (MDEA) into zwitterionomer-, anionomer-, and cationomer-type polyurethanes as compared to neutral PEUs. As expected, cationization of PEUs increased their acute thrombogenic potential in terms of platelet and fibrinogen adhesion, whereas zwitterionic surfaces (positive and negative charges) were shown to be more thromboresistant than anionomeric surfaces. They explained their results by the embedding process of positive ions in the bulk and the preferential presence of negative ions exposed at the end of the short side chain on the surface. Another explanation was also given for this phenomenon. Lelah et al suggested that a synergistic effect of both charges (zwitterionomer-PEUs) on protein adsorption created a favorable surface for improved blood compatibility.¹⁵ Although not frequently discussed, one may have to take into consideration the effect of the isoelectric point (pI) of the protein and pH of the environment (to determine the net charge of the protein), as well as the zeta potential of the PU surface (to determine the net charge bearing on the polymer's surface) when designing these experiments.

It was also during the 1980s that the concept of hydrophilicity of polyurethanes was first discussed. Hydrophilicity describes a surface characteristic promoting water absorption in the polymer which has been associated with blood response. An optimization of the hydrophobic-hydrophilic ratio as surface modification has been shown to improve blood compatibility by reducing platelet adhesion.^14,16,17

Lelah et al have also examined the hydrophilic status of their ionized polyurethanes and found contrasting results. In contrast to neutral more hydrophilic PEUs, the slightly hydrophilic cationomer PEU exhibited greater platelet and fibrinogen uptake than its neutral parent PEU. They finally concluded that the simple concept of hydrophilicity alone should not be used to correlate with blood responses.¹² In fact, microporous vascular grafts made of hydrophilic polyether-urethane-urea (PEUUs) were shown to induce a low level of hemolysis in vitro but exhibited a high rate of occlusive thrombosis in vivo.^18–20

Chemical, Morphological, and Structural Modifications (Decade 1988–1998)

In the late 1980s, a number of studies investigating various chemical, morphological and structural modifications to PUs have shed some light on the blood response induced by these polyurethanes. Newly developed and innovative methods of synthesis and surface modifications such as chemical incorporation, grafting, and coating techniques were attempted with the aim of increasing blood compatibility of polyurethanes.

Extraction Methods

Extraction methods on some polyurethane block copolymers have been found to improve the blood compatibility of an experimental polyurethane-urea material containing higher proportions of soft segment than Biomer. This high soft segment polymer was precipitated and extracted with methanol prior to recast from solution. The ex vivo platelet deposition levels on the extracted surface were significantly lower than that of the nonextracted material.²¹ Biomer submitted to the same extraction process gave similar results. However, Lelah et al were unable to determine if the extraction process modified the platelet response by removing low molecular weight oligomers from the block copolymer or if extraction removed low molecular weight additives.²¹

The results of another study examining the solvent cast and acetone-extracted Biomer indicated that the presence of an extractable fraction, when removed, also improved the compatibility of the material.²² This study was followed by that of Grasel et al who investigated the effects of extraction on the blood response of a standard polyurethane block copolymer containing no additives, stabilizers, or processing aids.²³ Three well-known extraction media were retained, namely, methanol, toluene and acetone. The blood response of the nonextracted and extracted polyurethanes coated on polyethylene tubings was tested in a canine ex vivo shunt experiment by measuring¹¹¹ In-labelled platelets and ¹²⁵I-labelled fibrinogen deposited on the surface. Results showed no significant difference in both platelet and fibrinogen deposition following extraction in either media. As extracted Biomer was previously shown to have improved blood compatibility,²¹ these results were attributed to a difference in chemistry (Biomer having urea functionality) or to the presence of an unknown agent in Biomer.

Chemical Composition

In recent years, the chemical composition of PUs was modified by several investigators with the intention of improving blood compatibility and discovering which factors influence most blood contacting responses. One method was to synthetize PUs or segmented polyether-urethane-ureas with highly hydrophilic soft segments. Producing PEUUs with MDI and PTMO, Takahara et al introduced long hydrophilic side chains with sodium sulfonate of various concentrations or methoxy end groups into the PEUU soft segment. Results demonstrated that platelet and fibrinogen uptake was lower with the least concentrated sulfonated PEUU whereas long side chain diol containing methoxy end group PEUU was more thrombogenic than a PEUU having a comparable concentration of sodium sulfonate groups.²⁴Further investigations on the influence of the chemistry of polyol on blood interactions were carried out by the group of Cooper. They investigated different PUs containing various polyol constituents PEO, PTMO, polybutadiene (PBD), hydrogenated-PBD (HPBD) and polydimethylsiloxane (PDMS), as well as all interesting materials with favourable blood-contacting properties. These PUs were prepared by conventional two-step solution polymerization. Using a canine ex vivo shunt experiment, PU-coated polyethylene tubings were tested for short term platelet and fibrinogen deposition. PEO-based PU was found to be more thrombogenic than nonpolar HPBD-based PU with PDMS-based PU being the least thrombogenic.²⁴The results with PEO were found to be similar to those published by Okkema et al,²⁵ but were in contrast to those reported by da Costas et al¹¹ whose platelet counting method might be arguable in terms of reproducibility and accuracy. The relatively good results observed with PDMS-based PU were imparted to the hydrophobic nature of the polymer surface. Here again, the authors concluded that the initial rate of platelet adhesion increased with an increase in hydrophilicity of the various polyol used in the synthesis of the PUs, thus, confirming that the blood compatibility of polyurethanes may depend on a combination of factors including microphase separation, surface heterogenicity, and surface hydrophilicity.²⁴

In a poorly reported study, Affrossman et al studied the effect of molecular weight (Mw) variation of different soft segments on the blood response of two series of PEUs synthesized with PEO or PTMO while the hard segment (MDI) and the chain extender (not stated) were kept constant. In a questionable in vitro blood chamber experiment, they reported data on platelet behaviour for PTMO polyurethanes while omitting those of the PEO polyurethanes. They found no significant difference between PTMO-based polyurethanes of molecular weight ranging between 650 and 2000 Mw and provided no substantial conclusion.²⁶

Other investigators developed innovative sulfonated polyurethanes on the basis that functional groups incorporated in the PU may impart some degree of anticoagulant activity or as others reported, heparin-like character or activity. These polyurethane materials have been shown to display varying degrees of antithrombotic activity.^27–30 Santerre et al synthesized sulfonated PUs having various sulfur contents (1.4–3.1%) and containing lysine or aspartic acid. Thrombin time experiments showed no significant heparin-like properties to sulfonated PUs with lysine or aspartic acid although an increase in thrombin time was shown for those sulfonated PU with increased sulfonate content.³¹Their findings were in contrast to those reported for sulfonated polystyrene resins.³²

More recently, a novel blood-compatible polymeric material, 2-methacryloyloxyethyl phosphorylcholine (MCP), was investigated for its surface blood compatibility properties by blending in PU. Ishihara et al have incorporated various MCP polymers copolymerized with cyclohexyl methacrylate or 2-ethylhexyl methacrylate (EHMA) into Tecoflex60 using the same solvent. Following incubation in whole blood or platelet-rich plasma (PRP), platelet adhesion was assessed by scanning electron microscopy (SEM). Examination of photomicrographs revealed a reduced adhesion of platelets at the surface of PU-MCP membranes with MCP composition as low as 5 wt.%. The authors suggested that a reduction of plasma protein adsorption on PU-MCP membranes might explain their results.^33,34

With the aim of developing polyurethane biomaterials which are more stable in vivo, PUs were prepared without polyester soft segments. Li et al introduced new PUs based on cholesterol and phosphatidylcholine analogous moieties. Polybutadiene (PBD), polyisoprene (PIP), hydrogenated-PIP (HPIP) glycols were used as polydiol soft segments in a conventional two-step solution polymerization procedure using MDI and two additional chain extenders as hard segments.³⁵ Qualitative results of platelet deposition after 60 min of incubation in PRP and viewing under SEM revealed that the apparent number of platelets as well as the morphological changes were the lowest for HPIP-PU while maintaining excellent mechanical properties. The hydrophobic nature of the hydrocarbon-based polydiols was the only argument used by the authors to explain their results.

In a second study, Li et al were interested in blending phospholipid diols with long chain alkyl groups (C₁₆ to C₂₀) because of their claimed blood compatibility, excellent mechanical properties, and nonether soft segment composition. PUs were synthesized using the previously investigated HPIP and four different phospholipid diols as polyols with MDI and BD as hard segments. Again, the apparent number of platelets and their morphological changes after incubation in PRP for 60 min were assessed by SEM. The results showed a better trend of low adherent platelets on phospholipid PUs as compared to PU alone or BioSpan® as controls. The authors suggested better blood compatibility properties for HPIP-based phospholipid PUs with saturated long chain alkyl groups (C₁₆, C₁₈, and C₂₀) and claimed their potential for wide biomedical applications.³⁶

Finally, in a recent study, a solution polymerization was used to prepare PU and α-hydroxyethyl methacrylate (HEMA)-terminated PU (HPU) interpenetrating polymer networks (IPNs), with PU based on two different polyether-type polyols of various Mw.³⁷Different wt.% of HPU were used in the preparation of IPNs and tested for their blood response. A relative index platelet adhesion (RIPA) was used as a hemocompatible parameter. The amount of whole blood platelets was calculated on PU membranes and compared to that of glass plate in a ratio of adhered platelet on PU/adhered platelet on glass. Results showed lower platelet adhesion with IPNs containing 25 wt.% of HPU and interpreted as an optimal hydrophilic/hydrophobic domain ratio induced by changes in surface soft segments to hard segment ratio. XPS analysis confirmed these results by showing that RIPA correlated well with the surface O/N and C-O-C ether group ratios.

Surface Modifications of Polyurethanes

In the last decade, a number of surface modifications have been proposed to improve the antithrombotic properties of polyurethanes, bearing in mind that the primary factor in determining the blood compatibility of any materials is the material surface chemistry. The methods include coating or impregnation processes, photo-chemical immobilization reactions and various grafting techniques using surface derivatization or oxidation, ionization, or polymerization.

Grafting Techniques

In a chronic ex vivo arterio-venous shunt experiment, McCoy et al studied six different materials simultaneously in terms of radiolabelled platelet deposition and SEM viewing.³⁸ A sulphonated PEU and a C₁₈ alkyl grafted derivative of PEU were compared to the same PEU, Biomer, low density polyethylene (LDPE), and PDMS as controls. The results demonstrated that the alkyl grafted C₁₈ PEU was the most thrombogenic of all materials in terms of platelet deposition. Their findings were in disagreement with those of Grasel et al³⁹ and Li and Nakaya³⁶although the platelet adhesion test differed slightly. Again, sulfonated PEU materials appear to exhibit nonthrombogenic behaviour as previously reported.^15,27,28

The blood response toward sulphonated grafted polyurethane surfaces obtained by surface derivatization was further assessed in vitro and in vivo. The oxidizing properties of Cerium (IV) ions were used to graft 2-acrylamido-2-methylpropane sulphonic acid (AMPS) on Pellethane. In a series of blood incubation tests, AMPS-PU was shown to reduce the generation of fibrinopeptide A, b-thromboglobulin and C_3a as well as decreasing the adherence of platelets and neutrophils in vitro. However, conflicting results were observed in vivo as a greater amount of adherent thrombus formation and increased attraction of macrophages to the material were observed with AMPS-PU. Although confering heparin-like characteristics to AMPS derivatization grafting on biomaterial surfaces, the authors expressed caution in extrapolating in vitro results to the in vivo situation.⁴⁰

In another oxidation reaction, this time by ozone and peroxides adsorption, a surface-grafted polymerization of acrylamide and dimethylacrylamide (DMAA) on polyether-urethane was attempted by Inoue et al and further evaluated by ex vivo arterio-venous shunt and in vivo by catheter tube implantation in the inferior vena cava of rabbits.⁴¹ Their findings revealed that the DMAA-grafted PU surface exhibited less clot formation than the nongrafted surface. They suggested that the better blood compatibility of the DMAA-grafted PU may be due to weaker interaction of the grafted material with the blood components as frequently reported for other materials.^42,43

Phosphorylcholine groups were also attached to polyether-urethane surfaces by a photo-chemical method using UV irradiation. As previously reported by Ishihara et al,³³phosphorylcholine-adsorbed PEU surfaces exhibited prolonged clotting times and reduced platelet adhesion with increased concentrations of phosphorylcholine at the surface compared to unmodified PEU surfaces.⁴⁴ Photo-oxidation should be carefully assessed before using UV irradiation of PU materials.

Finally, Saito et al⁴⁵ have recently reported the grafting of photo-reactive α-propylsulphate-poly(ethylene oxide) (PEO-SO₃) on PU surfaces via a photo-chemical technique. Using a flow-controlled chamber method, they showed anti-factor X_a activity on the grafted surface as well as reduced platelet adhesion compared to unmodified PU surface. Their study confirms previously reported studies showing the benefits of PEO and of SO₃ in improving the blood compatibility of PU surfaces.⁴⁵

Heparin Immobilization

For the past three decades, many investigators have focused on the development of antithrombogenic polyurethanes by immobilizing heparin either by chemical reaction of functional groups contained in a spacer and introduced in the PU backbone or by a grafting method on graft polymerization of functional groups. Reviewing all the research with heparin would be fastidious and only recent innovative approaches will be discussed here. Through the years, surfaces bearing ionically bound heparin have encountered major difficulties, namely, temporary anticoagulant activity and elution from the surface which may jeopardize long-term applications. Heparinizable PUs may be obtain by different methods. First, a chain extension reaction may be performed with chain extenders containing amine on hydrolizable ester groups in their backbone or side chain. Heparin may then be covalently bound by coupling reactions between the free hydroxyl or amine groups on heparin and the free isocyanate group on an hydrophilic spacer such as PEO.^46,47 A second approach consists in the introduction of spacers such as diamine diisocyanate⁴⁸ or poly (amino-amine)⁴⁹ onto the urethane linkage and subsequent heparinization of the PU by the dipping technique. Among the strategies that increase the immobilizing site, graft polymerization may be an effective method using functional group grafting by oxygen plasma glow discharge followed by graft polymerization. Heparin-immobilized PUs have been prepared by coupling reactions of NH₂ and COOH functional groups with heparin,⁵⁰ and more recently with acrylic acid,⁵¹ and acryloylbenzothiazole (AB).⁵² Results of the blood compatibility of heparinized-PUs by functional group grafting demonstrated lower activation of platelets and plasma proteins which leads to reduced thrombus formation as compared to nonheparinized functional group grafted PUs. Peripheral blood mononuclear cells were also shown to adhere less and secrete lower tumor necrosis factor after contact with heparinized PUs. Authors acknowledged that their results with AB grafted on PU surfaces were not promising with regard to blood contact applications. Moreover, the residual bioactivity of heparin was found to be approximately 25%, slightly higher than those reported elsewhere.^47,49,53

Immobilization of Antithrombotic Molecules

Other novel approaches to achieve blood-compatible PUs besides albumin and heparin are the coating or chemical immobilization of various antithrombotic drugs⁵⁴ or molecules such as urokinase derivatives,⁵⁵ prostacyclin,⁵⁶ ADPase,⁵⁷ dipyridamole,⁵⁸ glucose,⁵⁹ and more recently hirudin⁶⁰ or silver atoms.^61,62 However, some of these studies have been anecdotal without any in vivo or clinical follow-ups while others are still under development or currently under investigation.

In Vitro Biocompatibility Testing of Polyurethanes

The biocompatibility of polyurethanes has been assessed in vitro using various cell culture techniques. For the last two decades, the group of Anderson at Case Western Reserve University has been involved in elucidating the cellular interactions with biomaterials, especially those induced by a segmented polyether-urethane, Biomer. Their studies also included various polymers for comparison purposes. Results demonstrated that Biomer induced low monocyte reactivity in terms of interleukin-1 (IL-1) secretion which was found to be similar to PDMS.⁷⁶ However, Biomer was unable to stimulate fibroblast proliferation.^77,78 These studies were able to confirm the biocompatibility of Biomer, but also indicated that the intensity of the fibrotic or healing response was low. These findings may have a serious impact on the biofunctionality of prostheses or implants made of Biomer upon implantation. In other words, a biocompatible polymer does not necessarily mean that the healing of the device will be optimal and satisfactory.

When reviewing the literature on the in vitro biocompatibility testing of polyurethanes, we found only a limited number of papers describing the biocompatibility of well-characterized polyurethanes.^79–82 Other studies reported the compatibility testing of polyurethane devices including catheters,^61,83 stents,^74,84 tympanic membranes,⁸⁵ and artificial heart⁸⁶ without any detailed description of the chemistry involved in the fabrication of the biomaterial. Most of these studies report good biocompatibility of polyurethanes in general.

When conducting biocompatibility studies on PUs, one must include appropriate control materials or materials with clinical relevance for a specific application. In a vascular application, various polyurethane vascular grafts were compared for cytotoxicity and endothelial cell behaviour using organotypic culture assays. A polyester-urethane arterial graft, Vascugraft, was shown to promote the growth of a continuous monolayer of endothelial cells equivalent to that observed with polytetrafluoroethylene vascular grafts, Impra and Goretex^®,87whereas a Tecoflex-based polyurethane vascular graft exhibited no cell proliferation.⁸⁸ Furthermore, a hydrophobic polyether-urethane-urea graft was found to exhibit superior cell migration characteristics than those observed with a hydrophilic graft of similar chemistry, but having different surface texture.^87,88 None of the biomaterials tested were shown to release cytotoxic contaminants. The authors recognized that the surface characteristics of a biomaterial may have a potential effect on cell behaviour.

In another study, Ertel et al reported on the intrinsic toxicity of various materials including a polyether-urethane.⁸⁰ Following evidence that no toxic leachables were released from any of the materials investigated, they suggested that the toxicity generated by most of their materials could be attributable to surface morphology or chemistry.

Investigating a number of vascular grafts having different chemistry and surfaces, we have recently been able to identify to some extent which type of chemistry or structure may generate optimal cell growth using an organotypic culture technique.⁸⁹It was found that woven polyester structures showed greater cell growth than polyurethane or PTFE structures having porous structures. As previously reported, textured porous surfaces as opposed to smooth surfaces have been shown to stimulate both cell growth and metabolism. Moreover, cell adhesion is said to be regulated by substrate wettability, surface charge, and roughness.^90,91

The effect of the surface of prosthetic devices on cell behaviour was further confirmed by Sank et al who found that polyurethane foam such as those found in breast implants are a poor substratum for fibroblast attachment and proliferation when compared to smooth or textured silicone surfaces.⁹² In contrast, endothelial cells showed slightly better proliferation properties on the PU surface. Lee et al also suggested that there are two crucial factors which should be considered when determining cell attachment and proliferation properties at the surface of a polyurethane. The first factor is the surface morphology which may be regulated by dispersion of hard segment phase in the polymer and second, the hydrophilic property which is induced by high chain mobility at 37°C for instance.⁸¹

At this time, there is no real consensus as to the best surface characteristics to promote optimal cell behaviour on polyurethanes. These elastomers possess attractive chemical and mechanical properties and exhibit relatively good biocompatible characteristics. However, extrapolating in vitro results to the in vivo situation should be done with proper caution.

In Vivo Biocompatibility Studies of Polyurethanes

In the last three decades, a substantial amount of in vivo studies has been published on various polyurethanes used in the construction of the artificial heart, heart valves, pacemakers, vascular grafts, stents, endoprostheses and catheters. Although many studies confirmed the excellent mechanical properties and favourable biofunctionality of these devices, only a few reports have been concerned with the chemical stability (oxidation) and the degradation (mineralization, environmental stress-cracking) of segmented PUs. Consequently, a number of modifications have been attempted to reduce mineralization and oxidation, stabilize the hydrophobic-hydrophilic domains, increase the resistance to degradation and enhance the mechanical properties through innovative polymer synthesis. Recently published, more significant papers which compare PUs of various chemical composition or modification will be reviewed to assess those exhibiting good compatibility after implantation.

As reviewed by Coury et al,⁹³ it was only during the 1980s that calcification, environmental stress-cracking, and oxidation phenomena in vivo were brought to the attention of scientists involved in the synthesis of PU elastomers and devices. The first interest was in the segmented polyether-urethane Biomer frequently used in several biomedical applications. Biomer was found to be relatively stable with implantation time ^94,95 although some microscopic defects at the surface were reported.⁹⁶It was also demonstrated that Biomer induced low inflammatory reactions in vivo.^94,97,98

A few years later, a study by Zhao et al brought new evidence that adherent macrophages and foreign body giant cells (FBGC) may be involved in the chemical degradation and stress-cracking phenomena.^99,100 The authors demonstrated that the presence of localized surface cracking on a PEU cast film was produced under adherent FBGC (see Chapter 5).¹⁰⁰

In a more detailed study which included three different Pellethane PEUs with varying weight percent in hard segment as well as one PEUU and other materials as controls, Anderson's group indicated, using a theoretical analysis,¹⁰¹ that an increase in Pellethane weight percent of hard segment, surface hardness, and hydrophobicity resulted in an increase in total protein adsorption, density of adherent macrophages participating in FBGC formation and subsequent FBGC density. They suggested that implant surface properties influence the inflammatory response shortly after implantation and modulated further cell activation and proliferation. On the other hand, they recognized that their model was not able to generate any correlations between the materials studied and cell fusion kinetics.¹⁰² In a study on aliphatic polyurethane Tecoflex-based membranes, Lindner et al demonstrated that an increase of plasticizer weight percent in Tecoflex membranes corresponded to an increase in the inflammatory reaction shortly after implantation. When compared to PVC, Tecoflex-based membranes were shown to elicit a greater chronic inflammatory response. The authors suggested that the plasticizer released during the course of implantation might have triggered the inflammatory response.¹⁰³

More recently, the effect of surface charge on the inflammatory response of polyether-urethane was investigated by William's group. In their experiment, the net charge of PEU was modified by introducing different concentrations of sulphonate ionic groups in the PEU backbone giving a range of negative charge. After intramuscular implantation in rats, they reported that the response to PEU was largely inflammatory and that the net surface negative charge had a significant effect on the acute inflammatory response (>1 <2 weeks) by reducing neutrophil invasion and macrophage activation.¹⁰⁴

With the view to stabilized PEUs, antioxidant additives such as synthetic compounds Santowhite and Irganox have been conventionally incorporated in the polymer (see Chapter 3). However, it has been recognized that the use of these compounds may compromise the biological response if too concentrated or if their oxidation degradation products have undesirable toxicity. The use of a natural antioxidant vitamin E was recently investigated in a PEUU elastomer for surface degradation characteristics, chemical stability and inflammatory response in a cage implant exudate study. It was shown that vitamin E prevented surface cracking and chemical degradation up to five weeks in vivo. However, after 10 weeks, an estimated 18% chemical degradation was observed for the PEUU with vitamin E whereas 82% chemical degradation was reported for the PEUU without antioxidant. Leukocyte count and adhesion studies revealed that the inflammatory response was lower on the PEUU with vitamin E than on the PEUU without antioxidant.¹⁰⁵

In recent years, a number of modifications or substitutions of the soft segment component have been attempted in order to enhance the biostability and biocompatibility of segmented PUs. Changes to the soft segment chemistry may include substitution of polyether segments with polybutadiene, polydimethylsiloxane, polycarbonate, and other aliphatic hydrocarbon segments. Incorporation of PDMS in PUs was shown to exhibit good blood compatibility, low toxicity, good thermal and oxidative stability, low modulus, and anti-adhesive properties (see Chapter 6). Studies have also confirmed that polycarbonate soft segment was more stable than polyether soft segment. With studies lacking findings on the biostability and biological response of these modified PUs, Mathur et al recently reported the effects of chemical composition variation of the soft segment on the chemical stability, the rate of degradation and the inflammatory response to modified PUs. Two unmodified polyether-urethanes, one PDMS modified PEU, and two polycarbonate-urethanes were investigated. Results demonstrated lower cell densities and adhesion on PDMS-PEU than on unmodified PEU. These findings were attributed to the hydrophobic nature of PDMS end groups. Polycarbonates showed slightly less macrophage attack as compared to PEUs. In a similar fashion, the biodegradation rate of PDMS-PEU was also less extensive than those of unmodified PEUs. Again, the authors suggested that PDMS might provide a shield against oxygen radicals secreted by inflammatory cells and consequently reduced the rate of biodegradation. In addition, it was found that only a minor amount of biodegradation was seen on polycarbonate-urethanes as compared to unmodified PEUs and PDMS-PEU, thus confirming the oxidative stability of the carbonate linkage.¹⁰⁶

Nair et al¹⁰⁷ have also studied the effect of hydrophilicity on tissue response to polyurethane interpenetrating polymer networks (IPNs). They found that highly hydrophilic polyurethane-polyvinyl pyrrolidone or, in contrast, highly hydrophobic polyurethane-poly(methyl methacrylate) IPNs elicited inert responses in vivo. The authors concluded that the relative hydrophobicity or hydrophilicity of polymer surfaces is an important factor determining tissue compatibility. They also found that interfacial energy had no correlation with tissue responses whereas an interfacial energy near zero has been shown to be a requirement for blood compatibility. More recently, Hunt et al further investigated the effect of changing the hydrophilicity of polyurethanes on the in vivo biological response.¹⁰⁴ They obtained a relatively narrow range of hydrophilicity (42–67 contact angle degrees) from various polyurethanes prepared with PEO, PTMO, PHMO, POMO, and PDMO as macrodiols and MDI as hard segment constituents. After intramuscular implantation, the acute and chronic inflammatory responses were evaluated by counting specific cell types and quantifying cytokines released from these cells. Although they state that the decrease of hydrophilicity increases macrophage population, the data reported by Hunt et al on macrophage ED-2 and MHC III cellularities clearly showed that the more hydrophobic PDMO-based PU was less reactive than the other materials particularly during the acute phase of inflammation (>1 week, <1 month).

In conclusion, the effect of hydrophilicity, net charge, antioxidant incorporation, and chemical composition on the hemocompatibility, biocompatibility or biostability of PU surfaces is still under debate and requires clear and well-designed studies to bring quantifiable and reproducible parameters that will discriminate between polyurethanes of various chemical composition, structure and morphology.