Biomaterials 25 (2004) 4887–4900

ARTICLE IN PRESS

*Correspondin

2108.

E-mail addres

0142-9612/$ - see

doi:10.1016/j.bio

Long-term in vivo biostability of poly(dimethylsiloxane)/poly(hexamethylene oxide) mixed macrodiol-based

polyurethane elastomers

Anne Simmonsa,*, Jari Hyvarinena, Ross A. Odella, Darren J. Martinc,Pathiraja A. Gunatillakeb, Kathryn R. Noblea, Laura A. Poole-Warrena

aGraduate School of Biomedical Engineering, University of New South Wales, Sydney NSW 2052, AustraliabCSIRO Molecular Science, Private Bag 10, Clayton South MDC Vic 3169, Australia

cSchool of Chemical Engineering, The University of Queensland, Brisbane, Qld 4072, Australia

Received 28 July 2003; accepted 26 December 2003

Abstract

The long-term biostability of a novel thermoplastic polyurethane elastomer (Elast-EonTM 2 80A) synthesized using

poly(hexamethylene oxide) (PHMO) and poly(dimethylsiloxane) (PDMS) macrodiols has been studied using an in vivo ovine

model. The material’s biostability was compared with that of three commercially available control materials, Pellethanes 2363-80A,

Pellethanes 2363-55D and Bionates 55D, after subcutaneous implantation of strained compression moulded flat sheet dumbbells in

sheep for periods ranging from 3 to 24 months. Scanning electron microscopy, attenuated total reflectance-Fourier transform

infrared spectroscopy, and X-ray photoelectron spectroscopy were used to assess changes in the surface chemical structure and

morphology of the materials. Gel permeation chromatography, differential scanning calorimetry and tensile testing were used to

examine changes in bulk characteristics of the materials.

The results showed that the biostability of the soft flexible PDMS-based test polyurethane was significantly better than the control

material of similar softness, Pellethanes 80A, and as good as or better than both of the harder commercially available negative

control polyurethanes, Pellethanes 55D and Bionates 55D.

Changes observed in the surface of the Pellethanes materials were consistent with oxidation of the aliphatic polyether soft

segment and hydrolysis of the urethane bonds joining hard to soft segment with degradation in Pellethanes 80A significantly more

severe than that observed in Pellethanes 55D. Very minor changes were seen on the surfaces of the Elast-EonTM 2 80A and

Bionates 55D materials.

There was a general trend of molecular weight decreasing with time across all polymers and the molecular weights of all materials

decreased at a similar relative rate. The polydispersity ratio, Mw=Mn; increased with time for all materials. Tensile tests indicated

that UTS increased in Elast-EonTM 2 80A and Bionates 55D following implantation under strained conditions. However, ultimate

strain decreased and elastic modulus increased in the explanted specimens of all three materials when compared with their

unimplanted unstrained counterparts.

The results indicate that a soft, flexible PDMS-based polyurethane synthesized using 20% PHMO and 80% PDMS macrodiols

has excellent long-term biostability compared with commercially available polyurethanes.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Polyurethanes; Poly(dimethylsiloxane); Degradation; Biostability; Environmental stress cracking (ESC)

1. Introduction

Since Boretos and Pierce first suggested using poly-urethane elastomers as biomaterials in 1967 [1], this

g author. Tel.: +61-2-9385-3911; fax: +61-2-9663-

s: a.simmons@unsw.edu.au (A. Simmons).

front matter r 2004 Elsevier Ltd. All rights reserved.

materials.2004.01.004

family of polymers has been used in implantable medicaldevices because of their relatively good biocompatibilityand desirable mechanical properties such as strength,abrasion resistance and flexibility [2]. However, im-planted devices containing these elastomers are knownto suffer from polymer degradation (hydrolytic, oxida-tive and enzymatic) and calcification under certaincircumstances [3].

ARTICLE IN PRESSA. Simmons et al. / Biomaterials 25 (2004) 4887–49004888

Many studies have been undertaken to better under-stand the biodegradation mechanisms of polyurethanesin vivo [4–10]. Experimental findings indicate thatpolyester polyurethanes are not suitable for use inlong-term implantable devices due to poor hydrolyticstability [2]. This disadvantage was largely overcome bythe use of polyether polyurethanes with poly(tetra-methylene oxide) being the most common polyether.These polymers are hydrolytically stable yet are subjectto oxidative degradation in various forms includingmetal ion oxidation, auto-oxidation and environmentalstress cracking (ESC) in the in vivo environment. Softergrades of polyether polyurethanes are subject to stresscracking and auto-oxidative phenomena whereas theharder grades are inherently more stable.

In order to produce more oxidatively stable andESC-resistant polyurethanes, modifications or sub-stitutions to the chemical structure of polyurethaneshave been attempted. Softer polycarbonate-basedpolyurethanes [3,11–15] have proven to be moreoxidatively stable in comparison to polyether-basedmaterials of similar hardness but these have been foundto suffer from hydrolytic instability at the carbonatelinkage [16].

The use of poly(dimethylsiloxane) (PDMS) as acoating, a surface modifying end group [16–18] orincorporated into the soft segment of the polymer [19–21] is an attractive alternative due to this material’s goodbiocompatibility, high flexibility, low toxicity, goodthermal and oxidative stability, low modulus and anti-adhesive nature. However, incorporation of a non-polarmacrodiol such as PDMS into the polyurethane back-bone is generally difficult due to its poor compatibilitywith conventional compounds used in polyurethanesynthesis. Gunatillake [22] demonstrated that incorpora-tion of a small amount of a second macrodiol such aspoly(hexamethylene oxide) (PHMO) is effective in thepreparation of siloxane-rich polyurethanes with goodmechanical properties.

In a subsequent study, Martin [23] investigated thebiostability of a range of polyurethanes synthesized withvarying proportions of PHMO and PDMS. Thebiostability of these polymers was assessed by thesubcutaneous implantation of strained dumbbells insheep for 3 months followed by microscopic examina-tion. In that study, it was found that the testformulation containing a soft segment of 80% PDMSand 20% PHMO produced the most desirable mix ofexcellent biostability and relatively low hardness, tensileand flexural moduli. Similar mechanical properties toPellethanes 80A were achieved with biostability com-parable with Pellethanes 55D.

Since medical devices are often required to performreliably and safely while implanted in the body forprolonged periods of time, long-term biostability is acritical requirement for many biomaterials.

Subcutaneous implantation of pre-stressed polyur-ethanes in animal models has been established as amethod for evaluating susceptibility of these materials toESC [24,25] and a subcutaneous ovine model has beenused within our laboratories for more than 10 years [26].The application of longitudinal strain of approximately150% to polyurethane flat sheet samples in this testprocedure was designed to accelerate the rate andseverity of any degradation. Strains of this magnitudeproduce significant degradation in P80A after 3 monthimplantation periods [23].

Many chemical and mechanical alterations have beenreported to be associated with ESC in polyurethanes.These include embrittlement and loss of tensile strength[27,28], mass loss [29] and loss of ether [27,29–32], amide[27,31], carbonyl [27,29,31,32] and aliphatic hydrogen[27,29,31,32] absorbance in attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR).Additionally, loss of ether absorbance [28] and enhance-ment of carbon–oxygen absorbance has been observedusing X-ray photoelectron spectroscopy (XPS) [28,30].Decreases in molecular weight [27,28,30,32] typicallyoccur following degradation.

This study investigated the long-term biostability of asoft, flexible PDMS-based polyurethane synthesizedusing 20% PHMO and 80% PDMS macrodiols by thesubcutaneous implantation of strained dumbbells insheep for 24 months. Changes in the surface topographyof the explanted samples such as cracking, pitting ortearing were examined and rated using scanning electronmicroscopy (SEM) and the assessment methods de-scribed by our laboratory previously [33]. The mechan-ical properties, surface characteristics and changes inmolecular weights of the explanted materials were alsoassessed. Results are compared with those of Pel-lethanes 80A, Pellethanes 55D and Bionates 55D.

2. Materials and methods

2.1. Polyurethanes

The test material used in this study, Elast-EonTM 280A (E2 80A), incorporates hard segments of 4,40-methylenediphenyl diisocyanate and 1,4-butanediol andmixed soft segments of PDMS and PHMO incorporatedas a compatibilizing macrodiol in the ratio of 80:20. Thetest polyurethane was fabricated as reported previously[23] using a one-step bulk polymerization process. Thepolymer incorporated 60% total macrodiol by weightand the isocyanate to hydroxyl ratio was 1.00. Themolecular weights of PDMS and PHMO macrodiolswere 944 and 714 g/mol, respectively.

The control materials used in this study werePellethanes 2363-55D (P55D, negative control, DowChemical Co., USA), Pellethanes 2363–80A (P80A,

ARTICLE IN PRESSA. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4889

positive control, Dow Chemical Co., USA) andBionates 55D (B55D, negative control, The PolymerTechnology Group Inc., USA).

2.2. Sample preparation and in vivo implantation

Test and control materials were compression mouldedinto 1mm-thick flat sheets. Strained flat sheet dumbbellswere prepared as reported previously [33] with a strainof 150% imposed on the narrow section of eachdumbbell. Strained samples were prepared for implanta-tion as well as for use as unimplanted reference samplesfor each material and for each implantation period asindicators of material stability and for reference tobaseline material condition including processing arte-facts.

All strained samples were washed in 2% detergentsolution (Decon 90) for 3 days, rinsed four times inMilli-Q water, soaked again overnight in Milli-Q waterand then allowed to dry in a laminar flow hood. Sampleswere then sterilised using ethylene oxide gas and allowedto degas for at least 7 days before being implanted.

Biostability was assessed by subcutaneous implanta-tion of strained dumbbell samples in the dorsal region offour sheep for each of the 3, 6, 12 and 24 monthtimepoints of the study, a total of 16 sheep. Addition-ally, one spare sheep was implanted for each timepointand included in the study only if a study animal wascensored for any reason. If a spare sheep was notrequired in the biostability study, the samples from thatsheep were used for mechanical and physical testing.Healthy crossbred wethers weighing between 40 and60 kg and 1–2 years old were used.

Two samples of each material were implanted in eachanimal giving n ¼ 8 for each material at each timepoint.Three additional strained, unimplanted reference sam-ples of each material were prepared using techniquesidentical to those used to prepare the implanted samples.

At the conclusion of each implantation period, allspecimens were explanted and, together with theunimplanted reference specimens, cleaned in a 0.1mNaOH solution for 7 days followed by rinsing, washingin 2% detergent solution for a further 3 days, rinsingthoroughly and drying. This process removes excessproteinaceous and lipid deposits to allow visual andchemical analysis of the surface.

2.3. Scanning electron microscopy (SEM)

Specimens were gold sputter coated in preparation forexamination using SEM. A low-voltage SEM (Cam-bridge Instruments Stereoscan 360) was used to collect astandard set of SEM images of explanted specimens andunimplanted reference samples as previously described[33]. Sixteen images were taken for each specimen at

defined sites in a sequence of magnifications from 10�up to 500� .

The 16 images were rated independently by twoexaminers blinded to the material type and implantationperiod. Each image was assessed as to the presence orabsence of defects and a combined weighted score wascalculated as previously described [33]. With this system,specimens could receive a score between 0 (no defects atany magnification) and 50 (specimen fractured ordegradation evident in all images).

2.4. Material characterization

Several analytical techniques were used to determinethe effects of implantation on the mechanical properties,chemical structure and bonding in explanted specimens.Specimens from spare sheep or sheep excluded from thestudy were used for these procedures. Not all tests wereperformed at all timepoints due to the irregularavailability of these additional animals.

2.4.1. Mechanical tests

The mechanical properties of all materials weremeasured following 12 months implantation using anInstron 4302 universal testing machine with a 100Nload cell at a crosshead velocity of 100mm/min untilfailure. Ultimate tensile strength, stress at 100% strain,elongation and Young’s Modulus were calculated fromthe collected force vs. displacement data.

2.4.2. Differential scanning calorimetry (DSC)

DSC was used to characterize the thermal propertiesof P55D, B55D and E2 80A after implantation for 12months. Samples were dried at 45�C for 48 h under avacuum (0.1 Torr) to remove moisture prior to recordingthermograms over a temperature range of �150�C to250�C on a TA Instruments 2920 modulated thermalanalyser that was calibrated for heat flow and tempera-ture. The experiments were carried out at a heating rateof 10�C/min under a nitrogen purge of 20ml/min.

2.4.3. Gel permeation chromatography (GPC)

GPC was performed on all materials at 0, 3, 6, 9.5 and12 months to investigate changes in molecular weightand polydispersity following implantation. The GPCsystem consisted of a Spectra Physics Iso Chrom LCPump and Datajet integrator, Shodex RI Se-61 differ-ential refractometer, ICI HPLC temperature controllerand three Polymer Laboratories Plgel columns of poresize 105, 103 and 100 (A. Dimethylformamide with 0.05mlithium bromide was used as the mobile phase. Sampleswere filtered using a 0.45 mm syringe filter and theninjected at a flow rate of 1ml/min at 80�C.

The GPC was calibrated using seven (7) standardpolystyrenes which covered a molecular weight range of3500–250 000. A calibration curve was generated and

ARTICLE IN PRESSA. Simmons et al. / Biomaterials 25 (2004) 4887–49004890

average molecular weights of the polyurethane samplescalculated from the calibration curve using Cirrus (orSpectra Physics) software. The number average mole-cular weight (Mn), the weight average molecular weight(Mw) and polydispersity were recorded for eachmaterial.

The molecular weights determined by GPC wereanalysed by linear regression of the logarithm ofmolecular weight against implantation time. The regres-sion slopes of the four polymers were compared asdescribed by Weisberg [34].

2.4.4. Attenuated total reflectance-Fourier transform

infrared spectroscopy (ATR-FTIR)

ATR-FTIR was used to study changes in the surfacechemical structure of E2 80A, P80A, B55D and P55Dfollowing 24 months implantation. Analysis was con-ducted using a Perkin Elmer System 2000 FTIRequipped with attenuated total reflectance. Spectra wererecorded between 600 and 4000 cm�1 and the intensitiesof the peaks were normalized with respect to thearomatic absorbance peak at 1414 cm�1. Twenty scanswere averaged to obtain one representative spectrum foreach material and qualitative analysis was used toinvestigate the differences between spectra for theunimplanted and explanted samples.

For those materials with polyether soft segments(P80A, P55D and E2 80A), attention was focused on thespectral regions corresponding to the carbonyl groups(non-hydrogen-bonded urethane carbonyl stretching at1730 cm�1 and hydrogen-bonded urethane carbonylstretching at 1703 cm�1), ether methylene (1365 cm�1)and the ether groups (aliphatic asymmetric C–O–Cstretch near 1110 cm�1 and aliphatic symmetric C–O–Cstretch at 1081 cm�1) [35,36].

Peaks of interest for the polycarbonate polyurethane,B55D, are at 1738 cm�1 (non-hydrogen-bonded C=Oof the carbonate linkage), 1720 cm�1 (hydrogen-bondedC=O of the carbonate linkage), 1247 cm�1 (O–C–Ostretch in carbonate) and 955 cm�1 (symmetric stretch of(O–C–O) in carbonate) all of which are attributed to thepresence of carbonate groups in the soft segment [37].Other peaks of interest are 1700 cm�1 (non-hydrogen-bonded urethane carbonyl stretching) with the hydro-gen-bonded urethane carbonyl stretching expectedbetween 1700 and 1685 cm�1 [38].

2.4.5. X-ray photoelectron spectroscopy (XPS)

The surface chemistry of E2 80A, B55D P80A andP55D after 24 months implantation was analysed byXPS. Samples were analysed using an ESCALAB 220i-XL machine which employs a monochromate Al Kasource. The X-ray gun was operated at a take-offangle of 90� and at 120W (10KV, 15mA) with passenergy of 100 eV for wide scans and 20 eV for regionscans.

2.5. Statistical analysis

The in vivo biostability study was a split-plot design,with sheep as the plot or block [39]. The between-sheepfactor was the implantation time and the within-sheepfactor was the material, with all materials implanted intoeach sheep. The scores of the two raters for theexplanted specimens were averaged and then rankedand analysis of variance was performed on the ranks[40]. P80A was excluded from the analysis. The effect oftime was tested over the sheep (within time) meansquare. The effect of material and the material by timeinteraction were tested over the material by sheepinteraction mean square [39]. Differences betweenspecific materials or times were tested by linear contrastsas described by Searle [41] using the appropriatedenominator as described above. Scores for unim-planted specimens were analysed in the same way asthe explanted specimens as well as by logisticregression after transformation to a binary variable(flawed/unflawed). All analyses were performed usingStata 6.0 [42].

3. Results and discussion

3.1. In vivo biostability

Representative SEM images at 150� magnificationfor all materials at the 3 month and P55D, B55D and E280A at the 6, 12 and 24 month timepoints are shown inFigs. 1–4, respectively. (All P80A samples had fracturedin vivo at the 6, 12 and 24 month timepoints.) P80Ashowed signs of deep, aggressive degradation at alltimepoints as illustrated by the micrograph of thematerial after 3 months implantation shown in Fig.1(d). E2 80A was devoid of significant surface degrada-tion at 12 months as shown in Fig. 3(b) but somesamples displayed occasional surface voids after 24months as shown in Fig. 4(b). At 12 months, P55Dsamples had a relatively clear surface with a hetero-geneous texture possibly due to processing wax migrat-ing to the surface (Fig. 3(c)). P55D displayed some finesurface cracking at 24 months (Fig. 4(c)). B55D had arelatively clear surface at 24 months with occasionalsurface voids as shown in Fig. 4(a). There was somedegree of variability in the surface appearance of allmaterials at 24 months leading to a large variability inratings.

The scores are summarized in Fig. 5. Substantialdegradation of P80A was apparent at all timepoints. E280A, P55D and B55D showed little or no degradation at3 months but thereafter degradation increased over theduration of the study.

Analysis of variance on the ranks of the scoresindicated that the material by time interaction was not

ARTICLE IN PRESS

(a) (b)

(c) (d)

Fig. 1. SEM (150� ) of explanted B55D (a), E2 80A (b), P55D (c) and P80A (d) after implantation for 3 months under 150% strain.

(a) (b)

(c)

Fig. 2. SEM (150� ) of explanted B55D (a), E2 80A (b) and P55D (c) after implantation for 6 months under 150% strain.

A. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4891

significant and that the effects of time and material werehighly significant. Tests of linear contrasts indicated thatthere was no significant difference between E2 80A andB55D and that both had significantly less degradationthan P55D (po0:005).

Unimplanted specimens were initially flawless butexhibited minor degradation over time. The maximumscore was 6 (out of 50) and the mean score was less than3 for all four materials after 24 months (compare Fig. 5).

Analysis of variance on the ranked scores indicated thatthe effect of time was significant (p ¼ 0:001) and that thedifferences among materials was not significant(p ¼ 0:15). The degradation in these specimens mightbe due to sterilisation prior to storage. Previous work inthis laboratory has demonstrated that ethylene oxidesterilisation causes significant degradation in somematerials [43]. This is in agreement with other literaturein the area [44–46].

ARTICLE IN PRESS

(a) (b)

(c)

Fig. 3. SEM (150� ) of explanted B55D (a), E2 80A (b) and P55D (c) after implantation for 12 months under 150% strain.

(a)

(c)

(b)

Fig. 4. SEM (150� ) of explanted B55D (a), E2 80A (b) and P55D (c) after implantation for 24 months under 150% strain.

A. Simmons et al. / Biomaterials 25 (2004) 4887–49004892

3.2. Molecular weight

The number averaged molecular weights (Mn) of allmaterials determined by GPC are plotted in Fig. 6 aslogarithm of molecular weight against implantationtime. There was a general trend of molecular weightdecreasing with time across all polymers. The differencesbetween the slopes for the four polymers were notsignificant. Thus, the fitted lines in Fig. 6 are drawn witha common slope of �0.0128 per month, which isequivalent to a relative rate of decrease of 2.9 percentper month. The results for the weight-averaged mole-

cular weight (Mw) were similar but with a slower rate ofdecrease, 1.7 percent per month, as shown in Fig. 7. Thepolydispersity ratio, Mw=Mn; consequently increasedwith time for all materials.

It is interesting that, as far as these data show, themolecular weights of all materials decreased at a similarrelative rate even though there was a large disparity inthe biostability ratings, especially between P80A and theother polymers. Data generated using GPC methods,and single detection GPC in particular, should betreated with some caution. GPC methods rely on gooddissolution of the polymers in the mobile phase and

ARTICLE IN PRESS

3 months

Sco

re

0

10

20

30

40

50

6 months

12 months

B55D P55D E2 80A P80A

Sco

re

0

10

20

30

40

50

24 months

B55D P55D E2 80A P80A

Fig. 5. Stress crack ratings for B55D, P55D, E2 80A and P80A after implantation for 3, 6, 12 and 24 months. Median scores are shown for all

materials at all timepoints.

B55D

Mw

/ kD

a

150

200

250

100

E2 80A

P55D

time /month0 6 12

Mw

/ kD

a

150

200

250

100

P80A

time / month3 9 0 6 123 9

Fig. 6. Weight average molecular weight (Mw) of unimplanted samples of P80A, P55D, E2 80A and B55D and samples explanted after 3, 6, 9.5 and

12 months. One sample of each material was tested at the zero timepoint. Two samples of each were tested at the 3, 6, and 12 month timepoints with

the exception of P55D at 3 and 12 months and B55D at 6 months where one sample was tested. Four samples of each material were tested at the 9.5

month timepoint.

A. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4893

polyurethanes are multicomponent copolymers withissues in solvent selectivity depending on their composi-tion. Single detection GPC with refractive indexmeasurements is known to be less sensitive in detectingmolecular differences than other methods such as lightscattering. However, single detection GPC can beappropriate for the comparison of the molecular weightsof samples known to be of similar chemical composition

and structure and allows analysis of molecular weighttrends for each material with time.

It may also be relevant that, on explant, all P80Asamples used for GPC analysis had broken in vivowhereas implants of the other polymers were intact.Furthermore, results of the main biostability studyshowed that most P80A samples were broken afterimplantation for 3 months. Consequently, all the P80A

ARTICLE IN PRESS

B55D

Mn

/ kD

a

50

150

100

E2 80A

P55D

time / month0 3 6 9 12

Mn

/ kD

a

50

150

100

P80A

time / month0 3 6 9 12

Fig. 7. Number average molecular weight (Mn) of unimplanted samples of P80A, P55D, E2 80A and B55D and samples explanted after 3, 6, 9.5 and

12 months. One sample of each material was tested at the zero timepoint. Two samples of each were tested at the 3, 6, and 12 month timepoints with

the exception of P55D at 3 and 12 months and B55D at 6 months where one sample was tested. Four samples of each material were tested at the 9.5

month timepoint.

A. Simmons et al. / Biomaterials 25 (2004) 4887–49004894

specimens that were tested for molecular weight hadbeen unstrained for much or most of the implantationperiod. External strain is reported to have an accelerat-ing effect on in vivo ESC [47] and it is possible that thedegradation rate of P80A was less than it would havebeen had the strain being maintained.

It should also be remembered that the biostabilityscores determined in this study reflect phenomenaoccurring on the surface whereas GPC analyses theaverage molecular weight of the polymer specimen as awhole. Consequently, the effect of surface degradationmay be masked by the mass of the bulk polymer and thepossibly more extreme changes in molecular weightcaused by oxidative degradation at the surface of thematerials may not necessarily be detected by GPC of thebulk.

3.3. Mechanical properties

As there were insufficient reference samples preparedaccording to the techniques outlined in Section 2.2, thetensile properties of the materials explanted after 12months were compared with unimplanted, unsterilised,unstrained samples for indicative trends. Representativestress vs. strain curves are shown in Fig. 8 and Table 1summarises the Young’s modulus, UTS and elongationat failure of the explanted and unimplanted materials.All of the P80A samples had fractured in vivo at the 12month timepoint and could not be tested.

It should be emphasised that the unimplanted speci-mens tested were unstrained and unsterilised and theirmechanical properties are compared with those of the

explanted specimens that had been strained at 150% forthe period of implantation. UTS increased in E2 80Aand B55D following implantation under strained con-ditions. Ultimate strain for the explanted specimens ofall three materials was lower than their unimplantedunstrained counterparts. The elastic modulus for theexplanted specimens of all materials was higher thantheir unimplanted unstrained counterparts. These dataindicate that the explanted materials were less flexibleand less extensible than the unimplanted materialsand this type of effect following implantation understrain has been reported previously by others [8]who attributed the change in properties to acombination of stress relaxation, chain orientation andsome strain-induced crystallisation of the polymers.The dramatic increase in UTS exhibited by B55D mayalso be the result of strain-induced ordering of thismaterial.

The DSC thermograms for the unimplanted andexplanted samples of P55D, B55D and E2 80A areshown in Fig. 9. The endotherms are broad and diffusein all three materials and only small variations can beseen between the explanted and unimplanted variants. Ashift of the weak endotherms below 100�C in E2 80Aand B55D suggests some strain-induced ordering in theinterfacial regions and smaller, less well-organised harddomains. This would suggest that the increase in UTSand Young’s Modulus seen in these materials is due tochain alignment with the applied strain. The strain of150% was sufficient to plastically deform the harddomains and partially align the soft segment chains.Subsequently, stiffness and UTS increased in the

ARTICLE IN PRESS

Elast Eon 2 80A

0102030405060708090

0 200 400 600 800 1000Strain (%)

Str

ess

(MP

a) unimplanted

implanted

Pellethane 55D

0

10

20

30

40

50

60

70

80

0 200 400 600 800Strain (%)

Str

ess

(MP

a)

unimplanted

implanted

Bionate 55D

0255075

100125150175200

0 200 400 600 800

Strain (%)

Str

ess

(MP

a)

unimplanted

implanted

Fig. 8. Tensile properties of representative E2 80A, B55D and P55D

samples implanted for 12 months compared with unstrained unim-

planted controls.

Table 1

Mechanical properties of P80A, P55D, B55D and E2 80A samples

implanted for 12 months compared with unstrained unimplanted

controls

Material Implant status n E

(MPa)

UTS

(MPa)

Extension at

failure (%)

P80A Implanted — — —

Unimplanted 2 11.370.3 49.579.0 994723

P55D Implanted 2 161.6719.1 56.4712.0 3675

Unimplanted 2 42.470.7 69.774.1 724730

B55D Implanted 2 134.876.3 163.979.2 164715

Unimplanted 2 39.273.2 59.775.3 621722

E2 80A Implanted 2 64.971.2 72.1714.6 197754

Unimplanted 2 23.371.0 36.173.1 817742

-150 -100 -50 0 50 100 150 200 250

Temperature (°C)

Hea

t fl

ow

unimplantedimplanted

a

b

c

Fig. 9. DSC thermograms of E2 80A (a), B55D (b) and P55D (c)

samples implanted for 12 months compared with unstrained unim-

planted controls. The shift of the weak endotherms in the implanted

specimens of E2 80A and B55D are arrowed.

A. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4895

implanted samples. Our data suggest that recrystallisa-tion due to strain played a minimal role.

3.4. Surface characteristics

In vivo applications demand polymers with appro-priate chemical properties. These properties in thesurface region of the polymers are even more importantsince the surface chemistry between blood/body tissuesand the implanted medical device plays a significant rolein the efficacy of the device. Additionally, chemicalanalysis of the polymer surface can indicate specificmolecular alterations following implantation [36].

FTIR has been used extensively both to quantify andidentify chemical functionality and structure in poly-mers by relating absorbance of electromagnetic radia-tion in the infrared spectra to changes in chemicalvibrational and rotational energy levels. Figs. 10–13illustrate the ATR-FTIR traces for P80A, P55D, B55Dand E2 80A, respectively, in the 2000–700 cm�1 region.Analysis of methylene absorbances in the range2937–2776 cm�1 was complicated by the presence of

ARTICLE IN PRESS

2000.0 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700

cm-1

-1.00

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

A

unimplanted

Fig. 10. ATR-FTIR spectra for Pellethanes 80A implanted for 24

months and unimplanted control (absorbance vs. frequency, cm–1).

2000.0 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700

cm-1

-1.00

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

A

unimplanted

Fig. 11. ATR-FTIR spectrum for Pellethanes 55D implanted for 24

months and unimplanted control (absorbance vs. frequency, cm�1).

2000.0 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700

cm-1

-1.00

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.5

3.0

A

unimplanted

Fig. 12. ATR-FTIR spectrum for BionateTM 55D implanted for 24

months and unimplanted control (absorbance vs. frequency, cm�1).

2000.0 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700

cm-1

-1.00

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

A

unimplanted

Fig. 13. ATR-FTIR spectrum for Elast-EonTM 2 80A implanted for

24 months and unimplanted control (absorbance vs. frequency, cm�1).

A. Simmons et al. / Biomaterials 25 (2004) 4887–49004896

strong peaks at 2918 and 2849 cm�1 in the unimplantedPellethanes samples which may be attributed to thepresence of bis-ethylene stearamide on the surface ofthese materials [36].

The major differences between the IR spectra for theexplanted and unimplanted reference materials are listedin Table 2 with proposed peak assignments from theliterature [35,36]. Analysis of the spectra for theimplanted P80A samples revealed significant polyur-ethane degradation. The chemical changes caused by

biodegradation in the 24 month subcutaneous ovineimplants were consistent with severe oxidation of thealiphatic polyether soft segment and hydrolysis of theurethane bonds joining hard to soft segment. Oxidativechange in the soft segment is reflected in the significantloss of ether absorbance at 1110 cm�1 and methyleneabsorbance at 1367 cm�1. Hydrolysis of the hardsegment junction interphase region is reflected in theloss of the non-hydrogen-bonded carbonyl absorbanceat 1730 cm�1, the urethane C–N absorbance around1227 cm�1 and the urethane C–O–C absorbance at1081 cm�1.

A new shoulder at approximately 1170 cm�1 was seenin the P80A implanted spectrum and this may beattributed to degradation products in this material.Schubert et al. [29] discuss the appearance of a peak at1174 cm�1 which they assign to a branched ether. Therelatively weak new absorbance in P80A at around1170 cm�1 may therefore signify some degree of cross-linking of the soft segment. Alternatively, Wu et al. [31]suggest that this new band may be the O–C–O stretch ofCOOH and Zhao et al. [32] suggest it may be theasymmetric C–O–C stretch in esters. Another newshoulder was observed at an absorbance of approxi-mately 1560 cm�1. This may be attributed to an amide[35] and further confirm hydrolysis in this material.

The changes in the peak intensities observed in P55Dsamples are also consistent with oxidative degradationof the soft segment although to a lesser extent than thatseen in its 80A counterpart. P55D showed a strongreduction in ether absorbance at 1110 cm�1 with aweaker loss of methylene absorbance at 1367 cm�1

following implantation. Less pronounced hydrolysis ofthe hard segment junction interphase region is reflectedin a weak loss of the non-hydrogen-bonded carbonylabsorbance at 1730 cm�1 and the more pronouncedlosses in urethane C–N absorbance at 1227 cm�1 andurethane C–O–C absorbance at 1081 cm�1.

New peaks were seen at approximately 847, 1475 and1640 cm�1. Products of the oxidative degradation of thesoft segment are believed to be carboxylic acids, alcohols

ARTICLE IN PRESS

Table 2

Summary of decreases in the ATR-FTIR spectra of P80A, P55D, B55D and E2 80A samples implanted for 24 months compared with unimplanted

control samples with proposed peak assignments [35,36]

Wave number (cm�1) Proposed peak assignment P80A P55D B55D E2 80A

1738 Non-hydrogen bonded C=O stretch from carbonate — — W —

1730 Non-hydrogen bonded urethane C=O stretch S M — —

1715 Hydrogen bonded C=O stretch from carbonate — — W —

1703 Hydrogen bonded urethane C=O stretch S M — —

1530 Urethane N–H bend and C–N stretch S M — W

1460 Aliphatic CH2 bend S — — —

1367 C–H wag in CH2 S M — W

1311 Urethane N–H bend and C–N stretch and C–H bend in benzene S M — W

1247 C–O–C stretch from carbonate — — M —

1227 Urethane C–N stretch S S — —

1217 O–C–O stretch from carbonate — — W —

1110 Ether C–O–C stretch S S — W

1081 Urethane C–O–C stretch S S — W

1017 Si–O–Si stretch — — — W

955 O–C–O symmetric stretch from carbonate — — W —

S=strong decrease, M=moderate decrease and W=weak decrease.

A. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4897

and aldehydes [29]. In the explanted P55D spectrum,there is evidence of a possible carboxylic acid absor-bance near 1640 cm�1. There is no evidence of analdehyde absorbance near 1740 cm�1 suggesting thatany aldehydes produced by the oxidation are furtheroxidised to carboxylic acids. Alcohol absorbances areoften difficult to detect [36] and there were no obvioussigns of new peaks in the 1420–1260 cm�1 region whereprevious workers [36] have seen possible signs of alcoholpeaks following oxidative degradation. There is noevidence of the new band at 1170 cm�1 seen in P80A.

Analysis of the spectra for the B55D samples revealeda moderate reduction in absorbances associated with thecarbonate bond at 1247 cm�1 and weaker reductions at1738, 1715, 1217 and 955 cm�1 in the explanted sample.Decreases in peak heights were not observed in the1700–1685 cm�1 range suggesting no significant loss inurethane carbonyl. Analysis of changes in the urethanepeak height at 1227 cm�1 was complicated by thepresence of the carbonate-related absorption bands at1247 and 1217 cm�1.

The chemical changes seen in the B55D samplesimplanted for 24 months were consistent with gradualhydrolysis of the carbonate linkage. These findings arein agreement with Tang et al. [38,48] who studied thedegradation of polycarbonate polyurethanes by expos-ing them to cholesterol esterase (CE), a hydrolyticenzyme produced by monocyte-derived macrophages.Within a week of incubation with CE, the materials werefound to undergo degradation at the molecular level. Ina subsequent study, the same group [48] found that bothurethane and carbonate bonds were susceptible to CEcatalysed hydrolysis. Our findings are not definitive forhydrolysis of the urethane linkage due to the complexpeak distribution around the urethane peak at1227 cm�1.

As was the case with P80A, a new shoulder atapproximately 1170 cm�1 emerged in the IR spectra ofB55D after implantation. This may be attributed to abranched ether [29], the O–C–O stretch of COOH [31] orthe asymmetric C–O–C stretch in esters [32].

Analysis of the spectra for the implanted E2 80Asamples revealed distinct peaks at approximately800 cm�1 (Si–C in Si-CH3), 1017 cm�1(Si–O–Si),1081 cm�1 (C–O–C stretch and Si–O–Si), 1227 cm�1

(C–N stretch), 1260 cm�1 (C–H in Si–CH3), 1535 cm�1

(N–H bend and C–N stretch), 1599 cm�1 (C=C stretchbenzene ring and aromatic N–H bending) and1703 cm�1 (hydrogen-bonded urethane carbonyl). Ana-lysis in the ether region was complicated by the presenceof the absorption band for the Si–O–Si bond at1085 cm�1 which overlaps the ether group absorbancesat 1110 and 1080 cm�1.

The explanted E2 80A specimen showed very minorchanges from the unimplanted specimen with a veryweak reduction of ether absorbance at 1110 cm�1

indicative of weak oxidative degradation in the im-planted sample. There was no loss of methyleneabsorbance at 1367 cm�1. Very weak hydrolysis of thehard segment junction interphase region was reflected inthe weak loss of the non-hydrogen-bonded carbonylabsorbance at 1730 cm�1 and the urethane C–O–Cabsorbance at 1081 cm�1. The peak heights at 800 and1260 cm�1 (Si–CH3) did not change significantly overtime indicating that the PDMS did not degrade. Therewere no indications of the branched ether peak around1170 cm�1 seen in P80A nor any signs of additionalaldehydes, alcohols or carboxylic acids.

In summary, biodegradation in polyether polyur-ethanes can be monitored by analysis of the peaksassociated with ether, methylene and urethane linkages[29]. The ATR-FTIR data for P80A, P55D and E2 80A

ARTICLE IN PRESS

Table 3

Atomic concentration of the surfaces of E2 80A, P55D and P80A

specimens implanted for 24 months and unimplanted controls as

determined by XPS

Atomic

concentration (%)

P80A P55D B55D E2 80A

U I U I U I U I

Oxygen 9.9 15.2 13.8 16.1 19.2 14.0 20.5 20.7

Nitrogen 3.5 4.8 3.4 5.9 1.9 2.7 2.9 5.2

Carbon 85.8 79.4 82.3 76.9 67.4 80.3 64.8 63.0

Silicone 0.8 0.6 0.5 1.1 3.5 3.0 11.7 11.1

Oxygen/carbon ratio 0.12 0.19 0.17 0.21 0.28 0.17 0.32 0.33

U=unimplanted, and I=24 month implanted.

A. Simmons et al. / Biomaterials 25 (2004) 4887–49004898

outlined above suggest that the apparent order ofsusceptibility to oxidation of the polyether soft segmentas indicated by loss of ether and methylene absorbancesis P80A>P55D>E2 80A. Hydrolysis of the hardsegment junction as indicated by loss of the peaksassociated with urethane linkages was greater in P80Athan P55D and E2 80A.

The ATR-FTIR results for B55D and the Pellethanematerials confirm previous work suggesting that thepolycarbonate soft segment is more stable than thepolyether soft segment [3,14,15,17]. Polycarbonatepolyurethanes have shown similar mechanical propertiesto conventional polyurethanes with improved oxidativestability [3]. Mathur et al. [16] compared polyether andpolycarbonate polyurethanes using an in vivo cageimplant system and found similar acute and chronicresponses for both polymers. However, it was believedthat the oxidative stability of the carbonate linkage inpolycarbonate polyurethanes reduced the biodegrada-tion in comparison to the polyether polyurethanes.Tanzi et al. [13] compared a polycarbonate polyurethanesimilar to Bionate with Pellethane by incubating thepolymers in a range of solutions. The two polymersexhibited similar changes in molecular weight, thermalproperties and molecular structures at the polymersurface in all solutions except HNO3. In the presence ofHNO3, the chain scission was less pronounced in thepolycarbonate material than its polyether counterpart.

However, the ATR-FTIR results also indicate thatthe PDMS-based polyether polyurethane, E2 80A,showed improved biostability compared with thePellethane materials and similar to that of B55D. Theseresults generally agree with the biostability ratingsderived from the semi-quantitative rating system andresults outlined in Fig. 5.

Preliminary XPS analysis was conducted on E2 80A,B55D, P55D and P80A specimens implanted for 24months and unimplanted controls to determine possiblemodes of degradation. The atomic concentration ofoxygen, nitrogen, carbon and silicon on the surfaces asdetermined by survey scans are shown in Table 3. Thesurface of the unimplanted B55D specimen was also

found to contain small amounts of fluorine possibly acontaminant from the moulding process.

The data indicate that total oxygen increased andtotal carbon decreased at the surface of P80A and P55Dspecimens following 24 months implantation. Theoxygen:carbon ratio (O/C) of these materials increasedduring implantation suggesting surface oxidation of theexplanted samples with oxidation greater in P80A thanP55D. Carbon and oxygen concentrations were similarfor implanted and unimplanted specimens of E2 80A.Total oxygen decreased and total carbon increased onthe surface of B55D in line with findings by Zhang et al.[49]. These findings are in general agreement with theFTIR data outlined above.

Nitrogen levels increase on the surface of all materialsafter implantation. As nitrogen is only present in thehard segment, this raises the possibility that hardsegment concentration has increased on the surfacedue to segmental rearrangement of the polyurethanechains as reported by Zhang et al. [50]. The intensity ofthe nitrogen peak is weak however and interpretationusing this peak should be performed with caution. It isalso possible that small amounts of residual protein onthe surface of explanted specimens may have contrib-uted to the increased nitrogen content on the surface ofthese materials.

There was a small presence of silicone in P80A, P55Dand B55D possibly related to a moulding process duringmanufacturing.

4. Conclusions

This study has demonstrated that a soft, flexiblePDMS-based polyurethane synthesized using 20%PHMO and 80% PDMS macrodiols has excellentlong-term biostability compared with commerciallyavailable polyurethanes. Elast-EonTM 2 80A showedsimilar properties to Pellethanes 80A with biostabilitycomparable to Pellethanes 55D and Bionates 55D.

ATR-FTIR and XPS results on implanted andunimplanted samples confirmed the biostability ratingsderived from the semi-quantitative system of rating theSEM micrographs although ATR-FTIR was a moresensitive indicator of degradation mechanisms. GPCresults indicate that the molecular weights of Elast-EonTM 2 80A, Pellethanes 55D and Bionates 55Ddecrease with time at the same relative rate.However, the possibly more extreme changes inmolecular weight caused by oxidative degradation atthe surface of the materials would not necessarily bedetected by GPC of the bulk. Tensile tests indicated thatUTS increased in Elast-EonTM 2 80A and Bionates 55Dfollowing implantation under strained conditions. How-ever, ultimate strain decreased and elastic modulusincreased in the explanted specimens of all three

ARTICLE IN PRESSA. Simmons et al. / Biomaterials 25 (2004) 4887–4900 4899

materialswhen compared with their unimplanted un-strained counterparts.

As determined by the ovine in vivo model, theincorporation of PDMS together with the compatibiliz-ing polyether, PHMO, into the polymer backboneproduces substantial improvement in the long-termbiocompatibility of polyether polyurethanes with di-verse physical and mechanical properties.

Acknowledgements

The financial support of AorTech Biomaterials andthe Commonwealth Government of Australia throughthe Cooperative Research Centre and R&D Startprograms is gratefully acknowledged. The authors alsoacknowledge the technical support provided by An-thony Bennie, Veronika Tatarinoff, John Klemes, RajuAdhikari and Bill Bin Gong at different stages of thiswork.

References

[1] Boretos JH, Pierce WS. Segmented polyurethane: a new elastomer

for biomedical applications. Science 1967;158:1481–7.

[2] Lelah MD, Cooper SL. Polyurethanes in medicine. Boca Raton,

FL: CRC Press; 1986.

[3] Stokes K, McVenes R, Anderson JM. Polyurethane elastomers

biostability. J Biomater Appl 1995;9:321–54.

[4] Stokes K. The in vivo auto-oxidation of polyether polyurethane

by metal ions. J Biomater Sci Polym Ed 1990;1:207–30.

[5] Zhao Q, Fitzpatrick AM, Anderson JM, Hiltner A, Stokes K,

Urbanski P. Cellular interactions with biomaterials: in vivo

cracking of pre-stressed pellethanes 2363-80a. J Biomed Mater

Res 1990;24:621–37.

[6] Zhao Q, Topham N, Anderson JM, Hiltner A, Lodoen G, Payet

CR. Foreign-body giant cells and polyurethane biostability:

in vivo correlation of cell adhesion and surface cracking.

J Biomed Mater Res 1991;25:177–83.

[7] Hergenrother RW, Wabers HD, Cooper SL. The effect of chain

extenders and stabilizers on the in vivo biostability of polyur-

ethanes. J Appl Biomater 1992;3:17–22.

[8] Hergenrother RW, Wabers HD, Cooper SL. Effect of hard-

segment chemistry and strain on the stability of polyurethanes:

in vivo biostability. Biomaterials 1993;14:449–58.

[9] Xi T, Sato M, Nakamura A, Kawasaki Y, Umemura T,

Tsuda M, Kurokawa Y. Degradation of polyether urethane

by subcutaneous implantation into rats. I. Molecular weight

change and surface morphology. J Biomed Mater Res 1994;28:

483–90.

[10] Sato M, Xi T, Nakamura A, Kawasaki Y, Umemura T, Tsuda M,

Kurokawa Y. Degradation of polyether urethane by subcuta-

neous implantation into rats. II. Changes of contact angles,

infrared spectra, and nuclear magnetic resonance spectra.

J Biomater Sci Polym Ed 1994;6:225–67.

[11] Zhao Q, Casas-Bejar J, Urbanski P, Stokes K. Glass wool-H2O2/

CoCl2 test system for in vitro evaluation of biodegradative stress

cracking in polyurethane elastomers. J Biomed Mater Res

1995;29:467–75.

[12] Szycher M, Reed AM, Siciliano A. In vivo testing of a biostable

polyurethane. J Biomater Appl 1991;6:110–30.

[13] Tanzi MC, Mantovani D, Petrini P, Guidoin R, Laroche G.

Chemical stability of polyether urethanes versus polycarbonate

urethanes. J Biomed Mater Res 1997;36(4):550–9.

[14] Szycher M, Reed AM, Potter J. A solution grade biostable

polyurethane elastomer: ChronoFlex AR. J Biomater Appl 1994;

8:210–36.

[15] Capone CD. Biostability of a non-ether polyurethane. J Biomater

Appl 1992;15:108–29.

[16] Mathur AB, Collier TO, Kao JW, Wiggins M, Schubert MA,

Hiltner A, Anderson JM. In vivo biocompatibility and biost-

ability of modified polyurethanes. J Biomed Mater Res 1997;36:

246–57.

[17] Ward RS, White KA, Gill RS, Lim F. The effect of phase

separation and end group chemistry on in vivo biostability of

polyurethanes. Washington DC: transactions of the 42nd ASAIO

meeting. ASAIO-J 1996;42(2):17.

[18] Ward RS. Surface modification prior to surface formation:

control of polymer surface properties via bulk composition.

Med Plastics Biomater 1995;2(Spring):34–41.

[19] Shibayama M, Suetsugo M, Sakarai S, Yamamoto T, Nomura S.

Structure characterization of polyurethanes containing

poly(methylsiloxane). Macromolecules 1991;24:6254–62.

[20] Chun Y, Kim K, Shin J, Kim K. Synthesis and characterization of

poly(siloxane-urethane)s. Polymer Int 1992;27(2):177–85.

[21] Hergenrother RW, Yu XH, Cooper SL. Blood-contacting

properties of polydimethylsiloxane polyurea-polyurethanes. Bio-

materials 1994;15(8):635–40.

[22] Gunatillake PA, Meijs GF, McCarthy SJ, Adhikari R.

Poly(dimethylsiloxane)/poly(hexamethylene oxide) mixed macro-

diol based polyurethane elastomers. I. Synthesis and properties.

J Appl Polym Sci 2000;76:2026–40.

[23] Martin DJ, Poole-Warren LA, Gunatillake PA, McCarthy SJ,

Meijs GF, Schindhelm K. Polydimethylsiloxane/polyether-mixed

macrodiol-based polyurethane elastomers: biostability. Biomater-

ials 2000;21:1021–9.

[24] Stokes K, Urbanski P, Cobian K. New test methods for the

evaluation of stress cracking and metal catalyzed oxidation in

implanted polymers. In: Planck HI, Egbers G, Dauner Syre M,

editors. Polyurethanes in biomedical engineering, vol. II, Am-

sterdam: Elsevier Science Publishers, B.V; 1987.

[25] Zhao Q, Agger MP, Fitzpatrick M, Anderson JM, Hiltner A,

Stokes K, Urbanski P. Cellular interactions with biomaterials:

in vivo cracking of pre-stressed pellethane 2363-80a. J. Biomed

Mater Res 1990;24:621–37.

[26] Brandwood A, Meijs GF, Gunatillake PA, Noble KR, Schind-

helm K, Rizzardo E. In vivo evaluation of polyurethanes based on

novel macrodiols and MDI. J Biomater Sci Polymer Ed

1994;6(1):41–54.

[27] Meijs GF, McCarthy SJ, E Rizzardo E, Chen YC, Chatelier RC,

Brandwood, A, Schindhelm K. Degradation of medical grade

polyurethane elastomers: the effect of hydrogen peroxide in vitro.

J Biomed Mater Res 1993;27:345–56.

[28] Takahara A, Coury AJ, Hergenrother RW, Cooper SL. Effect of

soft segment chemistry on the biostability of segmented poly-

urethanes. I. In vitro oxidation. J Biomed Mater Res 1991;25:

341–56.

[29] Schubert MA, Wiggins MJ, Schaefer MP, Hiltner A, Anderson

JM. Oxidative biodegradation mechanisms of biaxially strained

poly(etherurethane urea) elastomers. J Biomed Mater Res

1995;29:337–47.

[30] Sutherland K, Mahoney JRI, Coury AJ, Eaton JW. Degradation

of biomaterials by phagocyte-derived oxidants. J Clin Invest

1993;92:2360–7.

[31] Wu Y, Selletti C, Anderson JM, Hiltner A, Lodoen GA, Payet

CR. An FTIR-ATR investigation of in vivo poly(ether urethane)

degradation. J Appl Polym Sci 1992;46:201–11.

ARTICLE IN PRESSA. Simmons et al. / Biomaterials 25 (2004) 4887–49004900

[32] Zhao Q, McNally AK, Rubin KR, Renier M, Wu Y, Rose-

Caprara V, Anderson JM, Hiltner A, Urbanski P, Stokes K.

Human plasma a2-macrogobulin promotes in vitro oxidative

stress cracking of Pellethane 2363-80A: in vivo and in vitro

correlations. J Biomed Mater Res 1993;27:379–89.

[33] Martin DJ, Poole-Warren LA, Gunatillake PA, McCarthy SJ,

Meijs GF, Schindhelm K. New methods for the assessment of

in vitro and in vivo stress cracking in biomedical polyurethanes.

Biomaterials 2001;22:973–8.

[34] Weisberg S. Applied linear regression, 2nd ed. New York: Wiley;

1985.

[35] Dillon JG. Infrared spectroscopic atlas of polyurethanes. Lan-

caster, PA: Technomic Publishing; 1989.

[36] McCarthy SJ, Meijs GF, Mitchell N, Gunatillake PA, Heath G,

Brandwood A, Schindhelm K. In-vivo degradation of polyur-

ethanes: transmission—FTIR microscopic characterization of

polyurethanes sectioned by cryomicrotomy. Biomaterials

1997;18:1387–409.

[37] Zhang Z, King MW, Guidoin R, Therrien M, Pezolet M, Adnot

A, Ukpabi P, Vantal MH. Morphological, physical and chemical

evaluation of the Vascugraftr arterial prosthesis: comparison of a

novel polyurethane device with other microporous structures.

Biomaterials 1994;15(7):483–501.

[38] Tang YW, Labow RS, Santerre JP. Enzyme-induced

biodegradation of polycarbonate-polyurethanes: dependence on

hard-segment concentration. J Biomed Mater Res 2001;56:

516–28.

[39] Zar JH. Biostatistical analysis, 4th ed. London: Prentice-Hall;

1999.

[40] Conover WJ. Practical nonparametric statistics, 2nd ed. New

York: Wiley; 1980. p. 337.

[41] Searle S. Linear models for unbalanced data. New York: Wiley;

1987. p. 87 (Chapter 4).

[42] StataCorp. Stata statistical software: release 6.0. College Station,

TX: Stata Corporation; 1999 (Anova, test, concord, graph, bs,

simul).

[43] Hyvarinen JP, Belgraver W, Noble K, Gunatillake PA, Simmons

A. Effect of sterilisation on the mechanical properties of

polyurethanes. 27th Annual Meeting of the Society for Biomater-

ials, Saint Paul, USA, April 24–29, 2001. p. 128.

[44] Lerouge S, Guignot C, Tabrizian M, Ferrier D, Yagoubi N,

Yahia L’H. Plasma-based sterilization: effect on surface and bulk

properties and hydrolytic stability of reprocessed polyurethane

electrophysiology catheters. J Biomed Mater Res 2000;52:774–82.

[45] Zhang YZ, Bjursten LM, Freij-Larsson C, Kober M, Wesslen B.

Tissue response to commercial silicone and polyurethane elasto-

mers after different sterilization procedures. Biomaterials

1996;17(23):2265–72.

[46] Abraham GA, Frontini PM, Cuadrado TR. Physical and

mechanical behavior of sterilized biomedical segmented polyur-

ethanes. J Appl Polym Sci 1997;65(6):1193–203.

[47] Szycher M. Biostability of polyurethane elastomers: a critical

review. J Biomater Appl 1988;3:297–402.

[48] Tang YW, Labow RS, Santerre JP. Enzyme-induced biodegrada-

tion of polycarbonate-polyurethanes: dependence on hard-seg-

ment chemistry. J Biomed Mater Res 2001;57:597–611.

[49] Zhang Z, Marois Y, Guidoin RG, Bull P, Marois M, How T,

Laroche G, King MW. Vascugraftr polyurethane arterial

prosthesis as femoro-popliteal and femoro-peroneal bypasses in

humans: pathological, structural and chemical analyses of four

excised grafts. Biomaterials 1997;18(2):113–24.

[50] Zhang Z, King M, Guidoin RG, Therrien M, Doillon C, Diehl-

Jone WL, Huebner E. In vitro exposure of a novel polyesterur-

ethane graft to enzymes: a study of the biostability of the

vascugraftr arterial prosthesis. Biomaterials 1994;15(14):

1129–44.