Ex vivo and in vivo evaluation of the blood compatibilityof surface-modified polyurethane catheters

Hiroyuki Inoue, Keiji Fujimoto, Yoshikimi Uyama, and Yoshito Ikada*Research Center for Biomedical Engineering, Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606, Japan

Catheter model tubes were prepared from a medical-gradepolyetherurethane and their outer surfaces modified by sur-face-graft polymerization of acrylamide and dimethyl acryl-amide (DMAA). The surface-graft layer was characterizedby means of dry staining, scanning electron microscopy(SEM), X-ray photoelectron spectroscopy, and protein ad-sorption. Ex vivo evaluation for the blood compatibility ofthe surface-modified polyurethane was carried out using thepolyurethane tube as an arterio–venous shunt between thecarotid artery and the jugular vein of rabbits. When the sur-face density of grafted polymer was in the range of 10–30mg/cm2, the in vitro adsorption of IgG exhibited a minimumvalue and platelet adhesion to the grafted polyurethane sur-face was insignificant, in marked contrast with that to thevirgin (nonmodified) surface. The in vivo blood compatibil-ity of polyurethane was evaluated by implanting the cath-eter tube in the inferior vena cava of rabbits from the femoralvein after ligation of a distal site of the exposed femoral vein.After remaining there for predetermined periods of time, the

implanted catheters were taken out together with the veinsof the rabbits that had been heparinized and sacrificed justprior to excision of the veins. After exchange of the blood inthe veins for saline, the excised veins were opened by cut-ting longitudinally to inspect for clot formations on the sur-faces of the implanted catheters. Occlusion of the inferiorvena cava was not observed for any of the catheters, nor wasthere any apparent damage or microembolizations in thelungs and kidneys. Many small-sized clots were observed onthe surfaces of the nonmodified polyurethane tubes after a2-week implantation whereas the catheter surfaces graftedwith DMAA polymer chains had a much smaller number ofclots. When the blood compatibility of polyurethane sur-faces was graded for relative evaluation from one (markedclotting) to five (no clotting) based on the size and numberof the clots, the evaluation results were as follows: 3.1 (vir-gin, 2 weeks), 4.0 (grafted, 1 week), 4.1 (grafted, 2 weeks),and 3.5 (grafted, 1 month). © 1997 John Wiley & Sons, Inc.

INTRODUCTION

Various polymeric biomaterials have been appliedto medical devices that come in direct contact withfresh blood. Their medical use is either temporary, asin extracorporeal circulations, or permanent, as in vas-cular grafts. Since the blood-contacting surface shouldbe blood-compatible unless anticoagulants are used,many attempts have been made over the past twodecades to produce blood-compatible surfaces. Theyinclude controlled release of heparin (hepariniza-tion),1 immobilization of fibrolytic proteins,2 poly(eth-ylene glycol) fixation,3 hydrogel-like thin layer for-mation,4 plasma polymerization,5 and so on. It is ofprime importance to establish an assessment methodrelevant to the expected clinical application of bioma-terials, when they are studied for improvement ofblood–surface compatibility. Still a problem in the de-

velopment of clinically applicable blood-compatiblesurfaces is the difficulty of evaluating blood compat-ibility. Any blood-contacting material should be sub-jected to either short-term or long-term evaluation, de-pending on the duration of its clinical use. The presentstudy is not concerned with blood-contacting materi-als for permanent clinical use, but only with those fortemporary use. Even for blood-contacting biomaterialsused only temporarily, it seems reasonable to subdi-vide the duration of blood-contact into short-term andlong-term for the assessment. Thus for the purposes ofthis study short-term blood contact means contactwith blood for no more than one day and long-termfor about one month.

Segmented polyetherurethane (PU) was selected forthis study as a polymer substrate for surface modifi-cation. The PU modification was carried out by sur-face graft polymerization of the water-soluble mono-mers acrylamide (AAm) and dimethyl acrylamide(DMAA). This surface modification technology maybe classified as hydrogel-like thin-layer formation. We*To whom correspondence should be addressed.

Journal of Biomedical Materials Research, Vol. 35, 255–264 (1997)© 1997 John Wiley & Sons, Inc. CCC 0021-9304/97/020255-10

previously reported that this surface modification isvery promising at least for short-term blood compat-ibility.6 The present work includes both short-termand long-term (given the above parameters) bloodcompatibility evaluation. A well-known application ofsuch long-term blood contact is the indwelling cath-eter for intravenous hyperalimentation (IVH) anddrainage.

MATERIALS AND METHODS

Materials

Pellethanet (2363-90AE), manufactured by DowChemicals Co. (Midland, MI), was the selected PUsubstrate. It was molded into films and tubes by ex-trusion, followed by purification with Soxhlet extrac-tion with methanol for 8 h. The thickness of the PUfilm was 100 mm. AAm, DMAA, and acrylic acid(AAc) were purchased from Nacalai Tesque Inc., Kyo-to and used after conventional purification. Ferrousammonium sulfate (Mohr’s salt) was obtained fromNacalai Tesque and used as received.

Surface graft polymerization

The details of the graft polymerization were re-ported elsewhere.7 Briefly, the PU film, or tube, wassubjected to oxidation by ozone generated from anozone generator (Type O-1-2, Nippon Ozone Co., Ltd.,Tokyo). After ozonization at an ozone concentration of15 g/m3 of the dry oxygen in the reactor for 1 min,peroxides were introduced onto the surface region ofPU. It was immersed in aqueous solution of monomercontaining Mohr’s salt as a chain transfer agent. Fol-lowing a degassing procedure, graft polymerizationwas allowed to proceed at 35°C for predeterminedperiods of time. The grafted PU was washed with dis-tilled water to remove as thoroughly as possible thehomopolymer that had formed, and then the graftdensity was determined with a dyeing method usingSky Blue 6B for DMAA8 and with the ninhydrinmethod for AAm graft polymerization.9 The AAc graftdensity was determined by staining with toluidineblue.10

X-ray photoelectron spectroscopy (XPS) was carriedout using Shimadzu ESCA 750 manufactured by Shi-madzu Corp., Kyoto. A cross-section of the grafted PUfilm was observed with an optical microscope afterstaining with Sky Blue 6B. Scanning electron micros-copy (SEM) was conducted with the S-450 type, Hita-chi Co., Ltd, Tokyo, after coating with platinum. Afterlabeling with I125 by the Chloramine T method,11 ad-

sorption of immunoglobulin G (IgG) onto PU filmswas performed at 37°C in phosphate-buffered salinesolution (PBS) containing 1 mg/mL of IgG.

A–V shunting

An A–V shunt circuit was established between thecarotid artery and the jugular vein of rabbits accordingto the method reported elsewhere.6 A PU catheter tubeof 50 cm in length, 0.25 cm in inner diameter, and 0.30cm in outer diameter was inserted into the artery andvein with the use of vessel tips made from polytetra-fluoroethylene.

Implantation in femoral vein

A PU tube of 17 cm in length and 18 G in cathetersize (0.65 mm inner diameter and 1.15 mm outer di-ameter) was implanted into the inferior vena cava of arabbit, as illustrated in Figure 1. Rabbits were anes-thetized by intravenous injection of Nembutalt (25ml/kg). After being shaved and disinfected with Po-vidone-Iodine, the skin was incised with a scalpel to

Figure 1. Schematic representation of catheter implanta-tion in the rabbit inferior vena cava.

256 INOUE ET AL.

expose the femoral vein. After ligation of the vein andproximal incision, the PU catheter was inserted intothe vein and then the proximal portion of the vein wasligated together with the inserted catheter, as depictedin Figure 1. Following removal of the excess tubing theskin was sutured. After a predetermined period oftime, the rabbits were sacrificed, the catheters weretaken out together with the vena cava, and the bloodwas flushed out with saline solution. The outer surfaceof the PU catheter was inspected macroscopically andwith SEM to evaluate the clot formation on the sur-face.

RESULTS

Surface modification

Figure 2 shows representative results of graft poly-merization carried out in the presence of 1.0 × 10−2Mof Mohr’s salt for 1 h. As can be seen, the graft densityincreases with the monomer concentration both forAAm and DMAA graft polymerization. The C1s XPSspectra of AAm-grafted PU surface with differentgraft densities are given without smoothing in Figure3 together with those of the virgin PU and the AAmhomopolymer. The peaks at 285.0, 286.5, 288, and289.4 eV are assigned to the C1s of −CH2−, −C−O−C−,−N−C=O, and −O−CO−NH−, respectively.12 XPS

Figure 2. Dependence of the graft density on the monomerconcentration (1 × 10−2M Mohr’s salt, 35°C, 1 h). s = AAm,d = DMAA.

Figure 3. C1s spectra of the virgin (a), AAm-grafted PUfilms (b–d), and AAm homopolymer (e). Graft density (mg/cm2), (b) 2.0, (c) 8.0, (d) 26.

Figure 4. Cross-section of the virgin (a) and DMAA-grafted PU film (b) with a graft density of 200 mg/cm2.

257EVALUATION OF BLOOD COMPATIBILITY

spectra of the grafted surface are clearly different fromthose of the virgin surface even when the graft densityof AAm polymer chains is as low as 2.0 mg/cm2.

The localization of the graft polymerization to thesurface region of PU is evidenced by the optical ob-servation of the cross-section of the graft PU. Figure 4shows an example of the cross-sectional microphoto-graph of DMAA-grafted PU with a graft density of 200mg/cm2 after staining with Sky Blue 6B. It can be seenclearly that only a thin outer surface region of thecross-section is stained with Sky Blue 6B. A SEM mi-crophotograph taken horizontally of the same graftedPU surface as shown in Figure 4 is shown in Figure 5.A textured structure is clearly observed in Figure 5(c),as already reported.13 After simple air drying of thehydrated surface (instead of freeze drying), such a tex-tured structure could not be observed on the graftedsurface when subjected to SEM observation. When thegraft density of AAm polymer was as low as a fewmg/cm2, the stained layer was much thinner than thatshown in Figure 4, and we could not observe any dis-tinct textured structure such as that shown in Figure5(c).

Protein adsorption from single protein solutionsprovides information on the surface graft polymeriza-tion. Figure 6 represents a result of IgG adsorption tografted PU surfaces as a function of the graft density.The protein adsorption to the virgin PU surface was0.17 ± 0.02 mg/cm2. It can be seen that IgG adsorptionremarkably decreases upon graft polymerization ofAAm and DMAA even if the extent is as low as 10mg/cm2. After that protein adsorption is slightly en-hanced, with an increase in the graft density. This maybe ascribed to the protein sorption into the interior of

the thin hydrogel-like graft layer, as reported in a pre-vious paper.14

Evaluation of the catheter lumen surface by exvivo shunting

Short-term evaluation of the blood compatibility ofgrafted PU surfaces was conducted using fresh blood

Figure 6. Dependence of IgG adsorption on the graft den-sity of grafted PU films. s = AAm, d = DMAA.

Figure 5. SEM microphotographs of the virgin (a) and DMAA-grafted PU film with a graft density of 100 mg/cm2 after airdrying (b) and freeze drying (c).

258 INOUE ET AL.

Figure 7. SEM microphotographs of the virgin and the DMAA-grafted PU tube with a graft density of 30 mg/cm2 after 1h (a) and 12 h (b) exposure to blood in the A–V shunt in a rabbit.

259EVALUATION OF BLOOD COMPATIBILITY

with an A-V shunting technique. A result for theDMAA-grafted catheter with a graft density of 30 mg/cm2 is shown in Figure 7. The inner surface of the PUcatheter was subjected to graft polymerization ofDMAA and brought into contact with the flowingblood. It is obvious from Figure 7 that the graftedsurface exposed to the blood for 1 and 12 h is free ofadhered platelets whereas the virgin PU surface un-dergoes remarkable platelet adhesion even after 1 h ofexposure; and at 12 h there is a significant clot forma-tion on the surface. The variety of platelet adhesionper area was very small. When the PU surface wasgrafted with AAc polymer chains, a bulky clot wasformed, as demonstrated in Figure 8, where a surfaceexposed to the blood for 3 h is shown, suggesting thatthe AAc-grafted surface is inferior in blood compat-ibility to the virgin PU surface.

Evaluation of the external catheter surface by invivo implantation

PU catheters, the outer surfaces of which were sub-jected to graft polymerization of DMAA, were evalu-ated for long-term blood compatibility by placing the

grafted catheters in the inferior vena cava of rabbits.The catheter were taken out of the veins for grossmacroscopic observation of the catheter surfaces. En-tire occlusion of the veins by the presence of the im-planted catheters was not observed. Gross inspectionof the lung, spleen, and kidney of the rabbits receivingthe implanted PU catheters revealed that there wasneither embolized nor necrotic regions in those or-gans. Representative optical photographs of the outerappearance of the virgin and the DMAA-grafted cath-eters that had been in contact with blood for 2 weeksand 1 month are shown in Figure 9. It is seen thatdifferent-sized clots are formed discontinuously onthe surface of implanted catheters. Large clots werenoticed at the distal end of the catheter, and those maybe due to the stagnant flow of blood resulting from theligation of the femoral vein. Large-sized clots wereclearly red, indicating the entrapment of red bloodcells in the clots. Indeed, SEM observation of such redclots revealed the presence of red blood cells in theclots. On the other hand, apparently clot-free regionshad no red blood cells attaching to the catheter surfacewhen observed by SEM. However, closer inspectionrevealed the presence of proteinaceous deposits evenon such clot-free surfaces. Seemingly, there is a dis-

Figure 8. SEM microphotographs of the virgin (a) and the AAc-grafted PU tube with a graft density of 50 mg/cm2 (b) after3 h exposure to blood in the A–V shunt in a rabbit.

260 INOUE ET AL.

Figure 9. Clot formation on the outer surface of the virgin and the DMAA-grafted PU catheter with a graft density of 30mg/cm2 after implantation in the rabbit inferior vena cava for different periods of time: (a) Virgin, 2 weeks, grade 3; (b)DMAA-grafted, 2 weeks, grade 5; (c) DMAA-grafted, 1 month, grade 4.

261EVALUATION OF BLOOD COMPATIBILITY

tinct difference in clot formation between the virginand the grafted surfaces. In addition, prolonged im-plantation of grafted catheters apparently results inenhanced clot formation.

Although some problems are associated with thequantitative expression of the extent of clot formation,we attempted to give numerical grading on the clotformation for the relative blood compatibility evalua-tion. The semiquantitative grading employed in thisstudy as a measure of clot formation is described inTable I. Although the number of clots and the clotdiameter varied in a complex manner, for the pur-poses of comparison, we roughly graded the extent ofclot formation as presented in Table I. The level ofclotting evaluated on the basis of this grading isshown for each of the optical photographs in Figure 9.The assessment of the blood compatibility of PU sur-faces, based on this 5-graded level, is given in Table IIas a function of the implantation time. The graft den-sity of the DMAA-grafted surfaces used for this long-term evaluation is around 30 mg/cm2, which provedto be optimal for the minimum adsorption of IgG. Ascan be seen from Table II, the grading is widely scat-tered even on the surface brought into contact withblood for the same period, but it is apparent that thelowest blood compatibility is noticed on the virginsurface. Clot formation on the grafted surface is likelyto increase with the time of implantation.

DISCUSSION

As shown in Figures 2 to 5, surface graft polymer-ization of acrylic monomers readily took place ontoPU surfaces if they had been oxidized with ozoneprior to the graft polymerization. The oxidation had tohave been localized only on the surface region of theozone-treated PU since no significant deterioration ofthe bulk mechanical properties of PU was observed.As pointed out before,15 the exact structure of the graftlayer, such as the mixing extent with the substratepolymer, the graft chain length, and the water content,is unclear because analytical means required for thesecharacterizations are not available at present. Never-

theless, the results of protein adsorption and plateletadhesion given in Figure 6 and 7 suggest a significantimprovement of the blood compatibility of PU sur-faces by the graft polymerization of AAm and DMAA.Predominant clot formation on the AAc-grafted sur-face in contact with the following blood for only 3 h, asdemonstrated in Figure 8, probably are due to the car-boxyl groups on the modified PU surface, which mayinteract with solutes in the blood, such as calcium ion,in contrast with the nonionic surfaces, such as DMAAand AAm-grafted.

When a new surface is to be created for its betterblood compatibility through any method, includingsurface modification, new material synthesis, or con-trolled release of anticoagulants, evaluation shouldbe undertaken to assess its blood compatibility,ideally under conditions identical to the clinical use ofthe material. However, for long-term blood compat-ibility evaluation there is no widely accepted methodthat is simple, reproducible, and reliable.16 For short-term evaluation, there are several well-known meth-ods, including the Lee–White test,17 the Imai–Nosemethod,18 the Chandlerloop method,19 and the exvivo A–V shunting.20–25 Very few methods have beenproposed for long-term evaluation. Among them arethe vena cava ring method26 and the anastomoticmethod.27 All of them have advantages and disadvan-tages. For instance, the anastomotic method, whichhas been used mostly for studies of vascular grafts, isassociated with clot formation at the anastomotic site,often independent of the surface property of the ma-terial to be tested.

The present study did not employ such an anasto-motic technique, but rather simply fixed the samplecatheter by ligation, as illustrated in Figure 1. To avoidstrip-off of clots from the catheter surface during thepulling-out of the catheter from the rabbit vein, wesacrificed the animals, excised the inferior vena cavatogether with the implanted catheter, and cut open theveins along their length to examine with the naked eyethe clot formations on the surfaces of the implantedcatheters. This method is similar to actual catheteriza-tion into blood vessels for hyperalimentation or drain-age. A disadvantage of this method is the unstable

TABLE IGrading for Blood Compatibility Evaluation

Grade

Clots Formed

Diameter (mm)Number on One Catheter

Surface

1 > 1.0 > 52 1.0–1.5 3–43 1.0–1.5 1–24 0.5–1.0 2–55 < 0.5 0

TABLE IIIn Vivo Evaluation of Blood Compatibility of Grafted

Surfaces Implanted for Different Periods of Time

SurfaceImplantation

Time Evaluation Grade

DMAA-grafted PU 1 week 4, 4 (average = 4.0)DMAA-grafted PU 2 weeks 5, 4, 4, 5, 4, 4, 3, 4

(average = 4.1)DMAA-grafted PU 1 month 4, 3 (average = 3.5)Virgin PU 2 weeks 3, 2, 4, 4, 3, 3, 3

(average = 3.1)

262 INOUE ET AL.

blood flow along the catheter surface. If there is abifurcation in the vein where the catheter was in-serted, the blood will flow quickly around the bifur-cation site; on the other hand, the blood will flowslowly at the place where the catheter surface is inclose contact with the inner surface of the vein. Thedifference in the blood flow should result in differentextents of clot formation. This might be a major reasonfor the scattering of evaluation results seen in Table II.

Nevertheless, the present assessment method,which involves implanting a catheter in the rabbit in-ferior vena cava, seems to be effective in evaluatingthe blood compatibility of a polymer surface for 1month or less. The finding that the DMAA-grafted PUsurface exhibits less clot formation than the non-grafted surface is in agreement with the short-termevaluation result obtained by the A–V shunt method.The better blood compatibility of the grafted PU sur-face than that of the virgin surface may be due toweaker interaction of the grafted surface with theblood components, as demonstrated elsewhere.3,4,6,8,14

The lessening with time of the difference in bloodcompatibility between the virgin and the grafted sur-faces may be ascribed to the gradual attachment ofblood components onto the DMAA-grafted surface. Itis probable that an equilibrium state of attachmentand detachment of blood components on the surfacewill be attained within a limited period of time.

In this work we did not attempt to study the de-tachment of adhered platelets and clots. This detach-ment and the change of the platelet count in the bloodby the presence of PU are the targets of our futurework.

References

1. A. Albanese, R. Barbucoi, J. Belleville, S. Bowry, R. Iloy,H. D. Lemke, and L. Sabatini, ‘‘In vitro biocompatibilityevaluation of a heparinizable material (PUPA), basedon polyurethane and poly(amido-amine) components,’’Biomaterials, 15, 129–136 (1994).

2. A. Sugitachi, M. Tanaka, T. Kawahara, and K. Takagi,‘‘Antithrombogenicity of UK-immobilized polymersurfaces,’’ Trans. Am. Soc. Artif. Intern. Org., XXVI, 274–278 (1980).

3. T. Akizawa, K. Kino, S. Koshikawa, Y. Ikada, M. Ya-mashita, and K. Imamura, ‘‘Efficiency and biocompat-ibility of a polyethylene glycol grafted cellulosic mem-brane during hemodialysis,’’ Trans. Am. Soc. Artif. In-tern. Org., 35, 333–335 (1989).

4. K. Fujimoto, M. Minato, and Y. Ikada, ‘‘Poly(vinyl al-cohol) hydrogels prepared under different annealingconditions and their interactions with blood compo-nents,’’ ACS Symp. Ser., 540, 228–242 (1994).

5. A. S. Hoffman, ‘‘Biomedical applications of plasma gasdischarge processes,’’ J. Appl. Polym. Sci. Appl. Polym.Symp., 42, 251–267 (1988).

6. K. Fujimoto, H. Tadokoro, Y. Ueda, and Y. Ikada,‘‘Polyurethane surface modification by graft polymer-

ization of acrylamide for reduced protein adsorptionand platelet adhesion,’’ Biomaterials, 14, 442–448 (1993).

7. K. Fujimoto, Y. Takebayashi, H. Inoue, and Y. Ikada,‘‘Ozone-induced graft polymerization onto polymersurface,’’ J. Polym. Sci., Polym. Chem., 31, 1035–1043(1993).

8. K. Fujimoto, H. Inoue, and Y. Ikada, ‘‘Protein adsorp-tion and platelet adhesion onto polyurethane graftedwith methoxy-poly(ethylene glycol) methacrylate byplasma technique,’’ J. Biomed. Mater. Res., 27, 1559–1567(1993).

9. M. Suzuki, A. Kishida, H. Iwata, and Y. Ikada, ‘‘Graftcopolymerization of acrylamide onto a polyethylenesurface pretreated with a glow discharge,’’ Macromol-ecules, 19, 1804–1808 (1986).

10. E. Uchida, Y. Uyama, and Y. Ikada, ‘‘Surface graft po-lymerization of ionic monomers onto poly(ethylene ter-ephthalate) by UV-irradiation without degassing,’’ J.Appl. Polym. Sci., 47, 417–424 (1993).

11. J. S. Garvey, N. E. Cremer, and D. H. Sussdorf, ‘‘125I- or131I-labeled proteins,’’ in Methods in Immunology: a labo-ratory text for instruction and research, 3rd ed., J. S.Garvey, N. E. Cremer, and D. H. Sussdorf (eds.), Mas-sachusetts, 1977, pp. 171–182.

12. A. Dilks, ‘‘The identification of peroxy-features at poly-mer surfaces by ESCA,’’ J. Polym. Sci. Polym. Chem. Ed.,19, 1319–1327 (1981).

13. H. Inoue, Y. Uyama, E. Uchida, and Y. Ikada, ‘‘Scan-ning electron microscope observation of lubricious sur-face for medical use,’’ Cells and Mater., 2, 21–28 (1992).

14. E. Kulik and Y. Ikada, ‘‘In vitro platelet adhesion tononionic and ionic hydrogels with different water con-tents,’’ J. Biomed. Mater. Res., 30, 295–304 (1996).

15. E. Uchida, Y. Uyama, and Y. Ikada, ‘‘XPS analysis ofthe poly(ethylene terephthalate) film grafted withacrylamide,’’ J. Polym. Sci. Polym. Chem., 28, 2837–2844(1990).

16. J. M. Courtney and C. D. Forbes, ‘‘Thrombosis on for-eign surfaces,’’ Br. Med. Bull., 50, 966–981 (1994).

17. R. I. Lee and P. D. White, ‘‘A clinical study of the co-agulation time of blood,’’ Am. J. Med. Sci., 145, 495–503(1913).

18. Y. Imai and Y. Nose, ‘‘A new method for evaluation ofantithrombogenicity of materials,’’ J. Biomed. Mater.Res., 6, 165–172 (1972).

19. C. H. Cholakis and M. V. Sefton, ‘‘In vitro platelet in-teractions with a heparin-polyvinyl alcohol hydrogel,’’J. Biomed. Mater. Res., 23, 399–415 (1989).

20. H. G. Downie, E. A. Murphy, H. C. Rowsel, and J. F.Mustard, ‘‘Extracorporeal circulation: A device for thequantitative study of thrombus formation in flowingblood,’’ Circ. Res., 12, 441–448 (1963).

21. M. D. Lelah, L. K. Lambrecht, and S. L. Cooper, ‘‘A ca-nine ex vivo series shunt for evaluating thrombus de-position on polymer surfaces,’’ J. Biomed. Mater. Res., 18,475–496 (1984).

22. E. L. Chaikof, J. E. Coleman, K. Ramberg, R. J. Con-nolly, E. W. Merrill, and A. D. Callow, ‘‘Developmentand evaluation of a new polymeric material for smallcaliber vascular prostheses,’’ J. Surg. Res., 47, 193–199(1989).

23. D. K. Han, S. Y. Jeong, Y. H. Kim, B. G. Min, and H. I.Cho, ‘‘Negative cilia concept for thromboresistance:Synergistic effect of PEO and sulfonate groups graftedonto polyurethanes,’’ J. Biomed. Mater. Res., 25, 561–575(1991).

24. G. Ryu, D. Han, Y. Kim, and B. Min, ‘‘Albumin immo-

263EVALUATION OF BLOOD COMPATIBILITY

bilized polyurethane and its blood compatibility,’’ Am.Soc. Artif. Intern. Org., 38, 644–648 (1992).

25. J. Caix, G. Janvier, B. Legault, L. Bordenave, F. Rouais,B. Basse–Cathalinat, and C. Baquey, ‘‘A canine ex vivoshunt for isotopic hemocompatibility evaluation of aNHLBI DTB primary reference material and of aIUPAC reference material,’’ J. Biomat. Sci. Polym. Edn.,5, 279–291 (1994).

26. V. L. Gott, D. E. Koepke, R. L. Doggett, W. Zarnstoff,and W. L. Young, ‘‘The coating of intravascular plastic

prostheses with colloidal graphite,’’ Surgery, 50, 382–389 (1961).

27. B. K. Kusserow, R. W. Lanow, and J. E. Nichols, ‘‘Ob-servations concerning prosthesis-induced thromboem-bolic phenomena made with an in vivo embolus testsystem,’’ Am. Soc. Artif. Intern. Org., 16, 58–62 (1970).

Received April 7, 1995Accepted May 28, 1996

264 INOUE ET AL.