Biomaterials 23 (2002) 37873797

Release of gentamicin sulphate from a modied commercial bone cement. Effect of (2-hydroxyethyl methacrylate) comonomer and poly(N-vinyl-2-pyrrolidone) additive on release mechanism and kinetics
* P. Frutosb, E. Diez-Penaa, G. Frutosc, J.M. Barrales-Riendaa,*
a Departamento de Qu!mica-F!sica de Pol!meros, Instituto de Ciencia y Tecnolog!a de Pol!meros, Consejo Superior de Investigaciones Cient!cas (C.S.I.C.), Calle Juan de la Cierva 3, E-28006 Madrid, Spain b ! Departamento de Farmacia y Tecnolog!a Farmaceutica, Facultad de Farmacia, Universidad Complutense, Ciudad Universitaria, E-28040 Madrid, Spain c ! Departamento de Estad!stica e Investigacion Operativa, Facultad de Farmacia, Universidad Complutense, Ciudad Universitaria, E-28040 Madrid, Spain

Received 10 May 2001; accepted 14 January 2002

Abstract The inuence of the (2-hydroxyethyl methacrylate) (HEMA) monomer addition as a comonomer to the cement liquid component and of a polymer, poly(N-vinyl-2-pyrrolidone) (PVP) to the solid component of a standard CMW-1s bone cement on gentamicin sulphate (GS) on its drug release properties have been studied. The addition of HEMA modies the habit of the delivery curves. The incorporation of PVP into the cement matrix, apparently, did not very much modify the shape of the HEMA modied cement release curves, but led to a remarkable increase of the maximum amount of GS released. This effect was proportional to the PVP concentration incorporated. From the matrix composition and SEM data, a model based on the morphology of the matrix has been proposed. The cumulative amount of GS released by each slab Mt is most adequately tted and represented by the equation Mt c at1=2 b1 expnt; which corroborates that the release occurs according to the model proposed, by means of three discrete mechanisms, namely: (i) a short-term initial elution due to the imperfections in the poly(methyl methacrylate) covering of the most external GS beads, burst effect by the buffer solution; (ii) followed by a fracture by stress cracking in an active media of the coated GS beads located on the external surface of the matrix where water molecules enter to dissolve GS molecules releasing them into the buffer solution by a diffusion-controlled process; and (iii) a third process in which the buffer solution penetrates into the internal voids and cracks creating a series of channels in a labyrinthic structure, which may facilitate the access of water molecules to the plastic-coated GS beads within the bulk matrix. This third process is enhanced by the incorporation of PVP beads as dissolved molecules within the matrix. This water-soluble additive is leachable, generating a highly porous structure in the cement. This HEMA and PVP modied cement may be used as a drug delivery system to modulate the GS release rate. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Gentamicin sulphateFGS; Acrylic bone cementFCMW-1s; Drug release; 2-hydroxyethyl methacrylateFHEMA; Poly(N-vinyl-2pyrrolidone)FPVP; Modied bone cement; Voids; Stress cracking; Channels; Capillars; Control release model

1. Introduction Gentamicin sulphate (GS), an aminoglycoside antibiotic with a wide antibacterial spectrum, has been incorporated into acrylic bone cements to prevent
*Corresponding author. Tel.: +349-1-562-2900; fax: +349-1-5644853. E-mail address: jmbarrales@ictp.csic.es (J.M. Barrales-Rienda).

infections. In the past, studies on its release from cement have produced variable and inconsistent results. In vivo and in vitro investigations carried out by Baker and Greenham [1] showed that methyl methacrylate (MMA) bone cement that has no defects is impervious to gentamicin. The release of gentamicin probably reects the presence of defects in the cement. For this reason, in the acrylic bone cement most of the antibiotic remains in the polymeric matrix.

0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 2 ) 0 0 0 2 8 - 5

3788

P. Frutos et al. / Biomaterials 23 (2002) 37873797

The release kinetic of gentamicin from poly(methyl methacrylate) (PMMA) bone cements is controlled by a combination of surface roughness and porosity [2]. It is largely inuenced by the penetration of dissolution uids into the polymer matrix, which requires a certain porosity of the cement [1,3]. Keeping all this in mind, the present authors [4] have put forward that the GS delivery from a commercial acrylic bone cement is a function of the GS concentration in the matrix. Basically, it occurs through a series of defects by a dissolutiondiffusion mechanism. In the past, some modications on drug release from bone cements have been reported [1,57]. Modications are necessary because in PMMA, much of the drug may be retained within the PMMA matrix, sometimes as much as 90% of the load [811]. Zhang et al. [12] have explored the use of antibiotic-impregnated biodegradable beads with a potential use in the local delivery of gentamicin to debrided osteomyelitic bone. They proved that the devices provided a small initial burst effect, followed by a gradual and sustained release of gentamicin. More recently, modications of the bone cements by introducing biodegradable components into the formulation have also been carried out [1315]. In our work, poly(N-vinyl-2-pyrrolidone) (PVP) was used as a biocompatible water-soluble additive. Due to the fact that PMMA presents very augmented hydrophobic characteristics, another approach to modify the polymeric matrix was to add a biocompatible hydrophilic monomer (2-hydroxyethyl methacrylateFHEMA) to the liquid component, giving place to a copolymer of MMA and HEMA in the matrix using a unique monomer feed. Therefore, the main task of the present work is to obtain a deeper insight into the release mechanism of GS from a commercial acrylic bone cement. This can lead to a precise control on the sustained GS delivery using an acrylic bone cement as source.

benzoyl peroxide, which acts as initiator for the radical polymerization of the monomer from the liquid component and 9.10 wt% barium sulphate used as radiopacier. (ii) The liquid component is composed of 98.215 wt% of MMA monomer, 0.816wt % of N,N-dimethyl-ptoluidine (DMPT) which is an activator or promotor, often called accelerator, employed to break down the initiator from the powder component and to initiate the polymerization by means of free radicals, 0.945 wt% ethanol, and trace amounts of ascorbic acid (0.022 wt%) and hydroquinone (0.002 wt%), which act as inhibitors or stabilizers to prevent premature polymerization of the monomer (MMA). Pyrogen-free poly(N-vinyl-2-pyrrolidone) Kollidon 12 % PF (PVP) of number-average molecular weight Mn 4 3:0 10 was supplied free of charge by BASF Hispania, S.A. (Barcelona, Spain). It was used as received. HEMA monomer was purchased from Fluka-Chemie A.G. (Buchs, Switzerland). It was puried before polymerization by extracting with n-hexane followed by vacuum distillation to eliminate the inhibitor (methyl ethyl hydroquinone) [16]. The following materials were also used as received without further purication: anhydrous di-sodium hydrogen phosphate, anhydrous potassium hydrogen phosphate, sodium chloride and sodium borate (Panreac Qu!mica SA Madrid, Spain), o-phthaldialdehyde and 2-mercaptoethanol (Sigma Aldrich Chemie GmbH, Steinheim, Germany), 2-propanol and methanol (Scharlau-FEROSA, Barcelona, Spain). 2.2. Solubility tests In order to know how the morphology of the solid components (PMMA, GS, PVP beads, etc.) in the mixture is affected or altered by the composition of the modied liquid (MMA plus HEMA) component during the time of preparation of the slabs, solubility tests have been carried out. The time of preparation of the slabs is approximately represented by the mixing time of the two components plus the moulding times necessary for the slabs to acquire a solid dimensional stability. 2.2.1. Poly(methylmethacrylate)(solid beads) Methylmethacrylate (liquid monomer) system During the time of mixing plus the time of moulding (23 plus 56 min, i.e. around 79 min), it does not seem possible that the solid plastic material constituted by the PMMA beads, from the solid component be dissolved by its liquid monomer from the liquid component. The time of mixing would not be enough, due to the

2. Experimental procedures 2.1. Materials Commercial CMW-1s acrylic surgical radiopaque bone cement with GS was purchased from De Puy ! Iberica, SA (Madrid, Spain) as a kit of two components: one, a sterile glass ampoule (18.37 g) containing the liquid component, and the other one, a sterile package (40 g) of the powder formulation. The composition (% w/w) of a typical packet of surgical CMW-1 gentamicin bone cement according to the manufacturer is as follows: (i) The solid component is composed of 4.22 wt% of GS, 84.73 wt% of PMMA beads, 1.95 wt% of

P. Frutos et al. / Biomaterials 23 (2002) 37873797

3789

molecular weight of PMMA beads to dissolve within its own monomer, perhaps PMMA beads may swell only very supercially by the MMA monomer. Therefore, we must consider that the beads remain almost intact in the polymeric matrix as a disperse phase (ller). 2.2.2. Methylmethacrylate (MMA):Hydroxyethyl methacrylate(HEMA)-Liquid 50:50 (v=v) mixtureGentamicin sulphate (GS) (solid beads) system GS solid beads are insoluble in MMA and very soluble in HEMA at moderate concentrations. However, our GS is insoluble in a 50:50 (v=v) MMA and HEMA mixture. This means that at the concentrations of MMA/HEMA mixtures employed in the present work, GS must remain insoluble during the process of mixing. On long standing periods, let us say 45 h, GS beads are probably solvated by HEMA, due to a preferential adsorption of HEMA molecules on GS beads. 2.2.3. Methylmethacrylate (MMA)Hydroxyethylmethacrylate (HEMA)-Poly(N-vinyl2-pyrrolidone) (PVP) (solid beads) system The PVP, used as modier in the present work, is highly soluble for a short period of time in both MMA and HEMA and their mixtures. 2.3. Reagents 2.3.1. Buffered saline solution Phosphate buffered saline solution, pH 7.4, was prepared as described in the British Pharmacopoeia [17]: anhydrous di-sodium hydrogen orthophosphate 2.38 g; anhydrous potassium dihydrogen phosphate 0.19 g and sodium chloride 8 g to 1000 ml with deionized water (Milli Q). 2.3.2. o-Phthaldialdehyde reagent o-Phthaldialdehyde reagent was prepared following a procedure given by Zhang et al. [12]. It was formulated by adding 2.5 g of o-phthaldialdehyde, 62.5 ml methanol and 3 ml of 2-mercaptoethanol to 560 ml 0.04 m sodium borate in deionized water solution. The reagent was stored in a brown bottle in darkness and settled for at least 24 h prior to use. 2.4. Methods 2.4.1. Matrix plate preparation (release devices) As mentioned in the introduction, from the commercial CMW-1 GS acrylic bone cement two different types of devices were prepared. One approach was to modify the commercial cement by incorporating different percentages of HEMA into the liquid component (4% and 25%). In the other types of devices, different

percentages of PVP polymer were added to the powder component. At the same time, 50% of HEMA monomer was added to the liquid component. The liquid and powder components were prepared separately by mixing the desired substances in the desired proportions, namely, weighted amounts of HEMA or PVP were added directly into the CMW-1 gentamicin bone cement liquid or into the powder components, respectively. Bone cement slabs were prepared following a protocol described by Bayston and Milner [18] and Wasserlauf et al. [19] using a home-made teon mould in a hydraulic press. Release devices were prepared at room temperature (25721C) and humidity (30%710%). Open mixing of powdered and liquid components was performed in glass beakers. After 23 min of soft careful hand mixing with a metal spatula, the cement paste was hand moulded and introduced into the teon mould to obtain slabs by mould compression. The reaction is exothermic. The characteristics of these specimens are presented in Table 1. 2.4.2. Drug release measurements in vitro Release proles of GS from the bone cement were measured as follows: one release device was introduced into a 300 ml thermosthatized glass reactor [20] containing 250 ml of the buffered solution (pH=7.4). The test solution was maintained at 371C and mechanically stirred at 150 rpm. Aliquot amounts (3 ml) of the solution were withdrawn at suitable time intervals, passed through a 0.45 mm millipore membrane lter, and diluted with the buffered reagent, when necessary, for measurements in the spectrophotometer. All analysis were carried out by analysing two slabs of each specimen in separate release systems. The cumulative percent drug release plotted in the gures represents the mean value. The high solubility of GS in the buffered solution allows us to assure that these assays are kept within sink conditions for long periods. 2.4.3. Gentamicin sulphate (GS) determination A wide variety of methods can be used to quantify aminoglycoside antibiotics, including uorescence polarization immunoassay (FPIA) [21], enzyme-linked immunosorbent assay (ELISA) [22,23], enzyme-immunoassay (EMIT) [24], microbiologic [25] and also chromatographic and spectrophotometric methods. Chromatographic methods are the ones used in most pharmacopoeias [26,27] for the quantitative analysis of gentamicin. In spite of its reliability, the main drawback of chromatographic methods is that they are timeconsuming. Radioenzymatic methods and radioimmunoassays, also quite reliable, are very expensive. Spectrophotometic methods are usually rapid, sensitive and economical [28]. Since gentamicin does not absorb ultraviolet or visible light, an indirect method is required for its

3790 Table 1 Compositions by weight (wt%) of the liquid and powder components used to prepare the modied cements, percentages by weight (wt%) of the drug and modifying components in the polymerized cements (wt%) and some characteristics of the slabs, prepared with these modied cements, employed in the release experiments Sample Composition modied liquid component (wt%) Liquid component UNMODIFIED HEMA 1 HEMA 2 (HEMA+PVP) 1 (HEMA+PVP) 2 (HEMA+PVP) 3 1A 1B 1C 2A 2B 3A 3B 4A 4B 5A 5B 6A 6B 100 96 75 50 50 50 4 25 50 50 50 HEMA Composition modied powder component (wt%) Powder component 100 100 100 80 90 97.5 F F 2.5 10 20 PVP Polymerized cement (wt%) Slabs characteristics P. Frutos et al. / Biomaterials 23 (2002) 37873797

GS

HEMA

PVP F F F 1.92 7.69 15.38

Weight (g) 3.14 2.92 3.29 2.89 2.94 2.82 3.03 3.42 3.47 3.09 3.41 3.11 2.97

Area (cm2) 28.05 27.95 28.38 27.74 27.81 27.82 27.89 28.50 29.09 28.05 28.65 28.79 28.06

Gentamicin content (mg) 91.70 85.28 96.08 83.66 85.02 81.54 87.73 108.28 109.84 90.15 99.71 80.76 77.13

Thickness (cm) 0.20 0.29 0.32 0.20 0.20 0.20 0.21 0.23 0.24 0.21 0.24 0.23 0.22

2.89 2.89 2.89 2.59 2.92 3.16

F 1.26 7.86 11.54 11.54 11.54

P. Frutos et al. / Biomaterials 23 (2002) 37873797

3791

spectrophotometric assay. New spectrophotometric procedures were developed for the analysis of tobramycin and other aminoglycosides [29], using the derivatizing agents o-phtaldialdehyde, uorescamine and dansyl chloride. o-phthaldialdehyde reacts with aminoacids in alkaline medium in the presence of a reducing agent such as 2-mercaptoethanol, yielding strongly uorecing compounds [30]. This allows uorometric assay of amino acids down to the nanomole range. For our measurement the o-phthaldialdehyde reagent modied by Zhang et al. [12] was used as a derivatizing agent [29]. It reacts with the amino groups of GS to yield chromophoric products. The reaction was carried out making 2 ml of our GS problem in solution to react with 2 ml of isopropanol (to avoid the precipitation of the products formed) and 2 ml of o-phthaldialdehyde reagent. The GS concentration was estimated by measuring the absorbance at 332 nm in a Beckman DU-70 Spectrophotometer. Validation of this analytical procedure is required to assure a minimum level of quality in the experimental measurements. The validation of the spectrophotometric method was done following recommendations given by ICH [31]. Characteristic parameter value obtained from the validation method have been given elsewhere [20]. 2.4.4. Scanning electron microscopy (SEM) The surface morphological characteristics of the untreated and treated specimens were studied using a Jeol JSM-840 Scanning Electron Microscope. Liquid nitrogen fractured surfaces were coated with gold palladium alloy in a sputter coating apparatus before the SEM analysis.

Fig. 1. Scanning electromicrograph showing the fractured surface of unmodied commercial CWM-1 acrylic bone cement: (a) before the drug release assay, (b) the same sample after being immersed in the buffered solution at 371C for 52 days.

3. Discussion of results 3.1. Matrix composition and morphology Microstructural examination of polymer/drug matrices was carried out by SEM. In Figs. 1 and 2, some SEM microphotographs obtained from the fractured surface of some of the bone cement samples used in this study are shown. With the aim of establishing differences between the morphology of the six specimens analysed, the unmodied commercial cement and the most modied one (HEMA+PVP) 3 have been selected. Fig. 1 shows the fractured surface of the commercial bone cement, before (Fig. 1(a)) and after (Fig. 1(b)) being immersed in the buffered solution at 371C for almost 2 months. Fig. 2 shows the modied cement (HEMA+PVP) 3, sample before (Fig. 2(a)) and after (Fig. 2(b)) being immersed in the buffered solution at 371C for almost 15 days. As seen in Fig. 1(a), the topography after the fracture of the unsoaked commercial gentamicin-loaded cement

revealed a considerable number of craters and cracks in the surface. These voids are typical and are produced because of the incompatibility shown between PMMA matrix and GS beads. They resemble the replica of GS beads. However, as shown in Fig. 1(b), the microphotographs of the same specimen after GS release also show bubble-like voids and cracks in the cement which have probably been created by stress cracking in an active medium (the buffered saline solution). They may have served as conduits for antibiotic elution. Fig. 2(a) of the unsoaked (HEMA+PVP) 3 cement shows, in the rst place, a heterogeneous mixture in which GS and radiopacier may be heterogeneously dispersed in the polymeric matrix. However, they are covered by the material forming the matrix. Replicas of any of the beads or particles (GS, PMMA, radiopacier) are not observed. This behaviour can be explained by the fact that the matrix, now constituted by a copolymer, poly(MMA-co-HEMA), with some hydrophilic character, exerts some type of moistening increasing the adhesive properties between the beads and the matrix.

3792

P. Frutos et al. / Biomaterials 23 (2002) 37873797

3.2. The model It may be interesting and appropriate to give some thought to the possible mechanisms of GS release and the manner in which it might be modelled. For complex systems such as the (HEMA+PVP) modied bone cement series the best approach is to establish a semiempirical model based on the morphology and on the bone cement components. The release mechanism must necessarily be complex, due to the fact that the cement, as well as the modied cement, have a heterogeneous and complex structure. In the case of a modied cement, the situation is still more complex. We have replaced in the cement liquid component certain amounts of MMA monomer by HEMA. This means that in the HEMA modied cements, the matrix will be composed of a poly(MMA-co-HEMA) statistical copolymer. Varma and Patnaik [34] have studied the copolymerization in bulk of MMA with HEMA at 601C using benzoyl peroxide as initiator. They have observed an ideal copolymerization reaction between MMA and HEMA. This means that even for low conversions both monomers will be incorporated into the copolymers. For high conversions the mole fraction of HEMA in copolymer will be almost the same as that in the monomer feed. Glass transition temperatures of syndiotactic and isotactic PMMA and PHEMA have almost the same values [35,36]. From this point of view, the Tg of our poly(MMA-co-HEMA) copolymer may be not quite different from that of both individual copolymers. However, due to the strong solvating effect of water molecules on the HEMA comonomeric units, water may act as an excellent plasticizer, having a strong relaxing effect on the polymer chain as a whole. Verhoeven et al. [37] have carried out a detailed study on the inuence of water and pH on the Tg of PHEMA homopolymer. They found that water exerts a strong plasticization effect of PHEMA. Thus, for instance for a 15% water content (w/w) the Tg of PHEMA is decreased at around 68 K, i.e. from 363.5 to 295.2 K. Sung et al. [35] have shown that the Tg of PHEMA is markedly affected by the water content and by the stereochemistry of the polymer. The lowering of Tg is probably due to the removal of barriers to the rotational and translational motions of the PHEMA molecule chain segments because of their interactions with water molecules. Therefore, we have passed from a hydrophobic system to a hydrophilic one which, in addition, can also be plasticized by the buffer solution. In the case of PVP modied cements, this polymer is incorporated as a liquid into the matrix, because it is soluble in the MMA+HEMA mixture, until it is removed after the immersion of the slabs into the buffer solution, because of its solubility in water. In this way, it creates a labyrinthic structure in the matrix, which may facilitate the release of GS molecules. During the

Fig. 2. Scanning electromicrograph showing the fractured surface of a modied CWM-1 bone cement containing 11.54% HEMA and 15.38% PVP (HEMA+PVP)3: (a) before the drug release assay, (b) the same sample after being in the buffer solution at 371C for 13 days.

In Fig. 2(b), an SEM microphotograph of a fractured surface from the modied cement after soaking shows a texture composed of voids and a cavernous structure. Initially, PVP beads are dissolved in the liquid MMA+HEMA and after polymerization, they are dispersed as macromolecules within the P(MMA-coHEMA) matrix; these macromolecules are randomly distributed. Upon dissolution, these macromolecules leave behind an interconnected network of pores and channels through which GS from new sites in the matrix may dissolve and diffuse. Thus, the random pore structure contributes with certain tortuosity or labyrinthic structure to the release picture. According to Kuu et al. [32] these types of interconnected pores would form solvent-lled channels. Obviously, due to the solubility of PVP in the matrix, the pore size and length were much larger than that of the GS molecules, the restriction of the pores on diffusion have no signicance. Thus, following Lambov et al. [33], we may put forward the hypothesis that the interconnected water channels in the present matrices may be the main route for GS release for long time periods.

P. Frutos et al. / Biomaterials 23 (2002) 37873797

3793

modied cement preparation step, PVP molecules could probably exert a second effect, such as an amalgamation of the GS molecules, namely, creating a heterogeneous disperse phase in the matrix composed of a solid blend of GS beads in touch with PVP molecules. Some of these hypotheses may be sustained not only by the SEM microphotographs but also by the release experimental results. To analyse the present drug release results we propose the following semiempirical equation: Mt c at1=2 b1 expnt; 1

where the three processes implicated in the model are assumed, namely: (i) a short-term initial elution occurring rapidly and due to the defects in the PMMA covering most of the external GS beads. This initial surface wash-off or burst effect by the buffer solution is associated with the rst term of the right-hand side of Eq. (1); (ii) followed by a fracture by stress cracking in an active media of the polymeric coating of GS beads located on the external surface, cracks, etc. of the matrix. This produces delivery orices of different topologies, where water molecules enter as a penetrant to dissolve GS molecules releasing them into the buffered solution through these orices by a diffusioncontrolled process. This is represented by the second term of the right-hand side kt1=2 ; associated to a Fickian diffusion process, and (iii) a third process in which internal GS molecules within the matrix are dissolved and released through a series of interconnected capillar tubes or channels. This labyrinthic structure is created by internal voids and cracks which may facilitate the access of water molecules to the plastic-coated GS beads within the bulk matrix producing new fractures by stress cracking. This effect is enhanced by the solution of PVP molecules from the matrix. It is represented by the third term b1 expnt being associated with a Noyes and Whitney dissolution process [38,39]. All these steps can be summarized in the model shown in Fig. 3 for the GS release from (HEMA+PVP) modied commercial acrylic surgical radiopaque bone cements. It has some similarities with that proposed by Kuhn and Wilson [40,41] and Verbeeck et al. [42] for the uorine complete release process from dental cements. 3.3. Discussion of the drug delivery proles: kinetics Experimental cumulative amount of GS released by a series including the commercial unmodied and of HEMA and (HEMA+PVP) modied bone cement samples as a function of time are given in Fig. 4. As discussed before, HEMA and MMA upon copolymerization will yield a statistical copolymer; therefore, the nature of the bone cement polymeric matrix with respect to the commercial cement has changed its polarity. Fig. 4(a) shows the effect of

Fig. 3. Scheme of the model proposed showing the three release processes occurring in a HEMA+PVP modied bone cement and the various mechanisms by which these occur: (a) dissolution of the supercial GS by the buffer solution (burst effect), (b) diffusion dissolution from easily accessible (cracks, bubble-like voids, etc.) GS on most external layers of the matrix to the buffered solution, and (c) three stages mechanism consisting of: (i) diffusion from GS easily accessible or neighbouring the internal surface created by PVP removal (channels, capillars, tubes), (ii) diffusion of the GS dissolved in the buffer solution within channels, capillars or tubes and (iii) diffusion to the buffer solution from external ends of the channels.

HEMA on the drug-released proles of the acrylic bone cement systems containing 4% and 25% HEMA, compared with the unmodied bone cement. The shape of the release curve (percentage of drug released from the cements) changed appreciably after the modication

3794

P. Frutos et al. / Biomaterials 23 (2002) 37873797

0.8 0.6

0.6 0.4

(a)

HEMA 1
Dissolution

10 . M t / A, (mg . cm-2)

0.4 0.2 0.0

30 20 10 0 0 2 4
Commercial PVP 1 PVP 2 PVP 3 Commercial HEMA 1 HEMA 2

103 . (Mt /Mi) / A, (cm-2)

0.2 0.0 1.0

Diffusion Burst effect

(b)

HEMA 2

0.5

Burst effect Diffusion Dissolution

10

12 14

0.0

10-2 . t, (h)
Fig. 4. Experimental cumulative amount of gentamicin sulphate released by: (a) the commercial and two HEMA and (b) the commercial and three (HEMA+PVP) modied acrylic bone cement samples as a function of time.

10

12

10-2 . t , (h)
Fig. 5. Experimental (white circles) and predicted (straight line) cumulative amounts of gentamicin sulphate released by the three delivery processes (burst effect (dashdot line), dissolution (dashed line), and diffusion (dot line)) occurring in: (a) 4% and (b) 25% HEMA modied bone cements as a function of time.

of the polymeric matrix with HEMA. The resulting copolymer poly(MMA-co-HEMA) possesses a more hydrophilic nature and, therefore, a better compatibility with hydrophilic substances such as the GS. As a result of this, GS will increase its solubility in the copolymer, as it was proposed above. The hydrophilic nature of the copolymer can also help in the penetration of water into the polymeric matrix. All these features will lead to an optimized release of GS from HEMA modied cements. This can explain the change in the shape of the release curves, and the slightly higher GS release. PVP is dissolved within the poly(MMA-co-HEMA) copolymer matrix, so the PVP molecules are randomly distributed. PVP molecules undergo their dissolution by the buffer solution leaving a randomly connected network of micropores, microtubes or channels, as a pastry structure, as seen in Fig. 2(b). Through this network of pores GS molecules after being dissolved must diffuse out of the polymeric matrix. GS molecules will not probably nd a straight pore or tube path to the outside of the matrix because of their labyrinthic structure. In other words, GS molecules are released out of the matrix through water-lled pores where they are dissolved. Fig. 4(b) shows the effect of PVP on the drug-released proles of the PMMA bone cement systems containing 1.92, 7.69 and 15.38 wt% of PVP. The percentage of drug released from the cements changed dramatically after the addition of PVP. This

proves, as previously stated, that the PVP addition enhances the third mechanism of the model proposed, i.e. the interconnected network of pores produced by its dissolution allows a higher amount of the GS to be released. Even a 100% release can be achieved with the third system in a very short period. All these facts indicate that an approach which had more physical reality than a single model would be more accurate. This is based on the assumption that the mechanism of elution must necessarily be complex, since these modied cements have a heterogeneous and complex structure. Different simple models, mixed mechanisms of diffusion were applied to evaluate their suitability for describing GS release from the modied acrylic bone cements. Several relationships were found to hold for the data under analysis in the present study. However, from a statistical point of view, the best tting was obtained when experimental data were tted to the model described by Eq. (1). Parameters in the models were estimated by minimizing the sum of squared residuals using a tting function for non-linear models. For analysing release curves by means of mathematical models, several criteria were considered: (i) a graphical plot of predicted values against observed data, as seen in Figs. 5 and 6, (ii) estimated parameters and their standard errors, (iii) goodness of the t and (iv) physical relevance of estimated parameters. Parameters obtained for this model and their correlation coefcients are shown in Table 2. The goodness of the t is quantied

P. Frutos et al. / Biomaterials 23 (2002) 37873797

3795

by their correlation coefcients R2 which were the highest when Eq. (1) was chosen (>0.9900). The agreement between experimental values and values predicted by Eq. (1) was very good as seen in Figs. 5 and 6.

4. Conclusions
*

0.6 0.4 0.2

(a)

PVP 1
Dissolution Diffusion Burst Effect
*

0.0 2

(b)

PVP 2
Dissolution Diffusion Burst Effect

1 0 3
2 1 0 0 2

(c)

PVP 3
Dissolution Diffusion Burst Effect

In order to understand the inuence of the polymeric matrix composition on the diffusion properties of GS, an unmodied and ve HEMA modied acrylic bone cement slabs with different compositions and bearing different quantities of a hydrophilic additive PVP have been prepared and studied. It is shown that variations in the polymer matrix composition by copolymerization of MMA with HEMA only slightly modify the amount of GS released, though the shape of the release curve is greatly affected, probably due to the increased matrix hydrophilicity. However, when PVP is added as a modier of the matrix hydrophobicity, the release properties are tremendously increased. From a mechanistic point of view, the release of GS molecules from the polymeric matrix involves three consecutive and/or simultaneous processes: (i) surface wash-off of GS from the external surface of the slabs, burst effect; (ii) release by an appropriate mechanism of the GS beads located in the layers near the external surface and (iii) development of channels or capillars in the matrix by PVP dissolution, together with a release by an appropriate mechanism of the GS beads located near the channels followed by a diffusion of the GS molecules to the outside through these solvent-lled channels. From the proles of the release studies and the data from the tted model, it was determined that the release rate changes with respect to the unmodied bone cement as the HEMA fraction in the copolymer forming the polymeric matrix is increased. The diffusion of GS in the matrix increases very markedly as the PVP additive amount is increased.

102 . (Mt / Mi) / A, (cm-2)

10

12

10-2 . t, (h)
Fig. 6. Experimental (white circles) and predicted (straight lines) cumulative amounts of gentamicin sulphate released by the three delivery processes (burst effect (dashdot line), dissolution (dashed line), and diffusion (dot line)) occurring in: (a) 50% HEMA+1.92% PVP; (b) 50% HEMA+7.69% PVP and (c) 50% HEMA+15.38% PVP modied bone cements as a function of time.

The two major practical conclusions and benets of the present work may be: (i) the extension of the knowledge of the mechanism and the kinetics of GS release from acrylic bone cements and (ii) the possibility of calculating the desired released amount and composition of devices to achieve very dened drug release proles. The kinetic equation derived from the model that considers three mechanisms can be very useful analysing drug release from modied bone cements.

Table 2 Statistical parameters and correlation coefcient R2 according to Eq. (1) for the GS release data from the different systems studied in the present work Sample Polymerized cement HEMA HEMA 1 HEMA 2 (HEMA+PVP) 1 (HEMA+PVP) 2 (HEMA+PVP) 3 1.26 7.86 11.54 11.54 11.54 PVP F F 1.92 7.69 15.38 Parameters of Eq. (1) c 0.0078 0.0331 0.073 0.2551 0.2808 a 0.0006 0.0012 0.0085 0.0253 0.0760 b 0.0244 0.0137 0.3361 1.3806 1.4365 n 0.190 0.208 0.006 0.010 0.143 0.9911 0.9926 0.9989 0.9992 0.9994 Correlation coefcient R2

3796

P. Frutos et al. / Biomaterials 23 (2002) 37873797 [11] Levin PD. The effectiveness of various antibiotics in methyl methacrylate. J Bone Jt Surg 1975;57:2347. [12] Zhang X, Wyss UP, Pichora D, Goosen MFA. Biodegradable controlled antibiotic release devices for osteomyelitis: optimization of release properties. J Pharm Pharmacol 1994;46:71824. [13] Shinto Y, Uchida A, Korkusaz F, Araki N, Ono K. Calcium hydroxyapatite ceramic used as a drug delivery system for antibiotics. J Bone Jt Surg 1992;74B:6004. [14] Garvin KL, Miyano JA, Robinson D, Giger D, Novak J, Radio S. Polylactide/polyglycolide antibiotic implants in the treatment of osteomyelitis: a canine model. J Bone Jt Surg 1994;76A: 15006. [15] Mader JT, Calhoun J, Cobos J. In vitro evaluation of antibiotic diffusion from antibiotic-impregnated biodegradable beads and polymethylmethacrylate beads. Antimicrob Agents Chem 1997;41:4158. [16] Brinkman UA, Van Schalk TAM, De Vries G. American Chemical Society Symposium Series. 1976;31:10518. [17] British Pharmacopoeia.Vol. II, Appendix ID. HMSO. London. 1993. p. A79. [18] Bayston R, Milner RDG. The sustained release of antimicrobial drugs from bone cement. J Bone Jt Surg 1982;64B:4604. [19] Wasserlauf S, Warshawsky A, Arad-Yelin R, Mazur Y, Salama R, Dekel R. The release of cytotoxic drugs from acrylic bone cement. Bull Hosp Joint Dis 1993;53:6874. * [20] Frutos Cabanillas P, D!ez Pena E, Barrales-Rienda JM, Frutos G. Validation and in vitro characterization of antibiotic-loaded bone cement release. Int J Pharma 2000;209:1526. [21] Gurtler F, Kaltsatov V, Boisrame B, Deleforge J, Gex-Fabry M, Balant LP, Gurney R. Ocular availability of gentamicin in small animals after topical administration of a conventional eye drop solution and a novel long acting bioadhesive ophthalmic durg insert. Pharm Res 1995;12:17915. [22] Ara J, Gans Z, Sweedy R. Dot-ELISA for the rapid detection of gentamicin in milk. J Clin Lab Anal 1995;9:3204. [23] Kolosova AY, Blintsov AN, Samsonova JV, Egorov AM. Development of an enzyme-linked immunosorbent assay for gentamicin in human blood serum. Fresenius J Anal Chem 1998;361:32930. [24] Miglioli PA, Pea F, Mazzo M, Lanzafame P. Possible inuence of assay methods in studies of the pharmacokinetics of antibiotics. J Chemother 1993;5:2731. [25] Reamer RH, Dey BP, White CA, Mageau RP. Comparison of monolayer and bilayer plates used in antibiotic assay. J AOAC Int 1998;81:398402. [26] British Pharmacopoeia, 1998, vol 1 HMSO, London. p. 302. * [27] Real Farmacopea Espanola, 1997. Ministerio de Sanidad y Consumo, Madrid. [28] Ayad MM, Shalaby AA, Abdellatef HE, Elsaid HM. Spectrophotometric and atomic absorption spectrometric determination of certain cephalosporins. J Pharm Biomed Anal 1999;18: 97583. [29] Sampath SS, Rabinson DH. Comparison of new and existing spectrophotometric methods for the analysis of tobramycin and other aminoglucosidos. J Pharm Sci 1990;79:42831. [30] Roth M. Fluorescence reaction for amino acids. Anal Chem 1971;43:8802. [31] International Conference of Harmonization. The European Agency for the Evaluation of Medical Products. Note for guidance on Validation of Analytical Procedures: Methodology. ICH Topic Q 2B, 1996. [32] Kuu WY, Wood RW, Roseman TJ. Factors inuencing the kinetics of solute release. In: Kydonieus A, editor. Tratise on controlled drug delivery, fundamentals, optimization, applications. New York: Marcel Dekker, 1992. p. 37154.

Mechanical properties of these materials are being currently investigated in order to establish as to how these modications may affect their characteristic parameters.

Acknowledgements This work was supported in part by a grant from the ! ! ! Comision Asesora de Investigacion Cient!ca y Tecnica ! ! de la Subdireccion General de Promocion de la ! ! Investigacion from the Ministerio de Educacion y Ciencia PB95-0134-C02-01 and MAT99-1127-CO4-02 ! and by a grant from the Direccion General de ! ! Investigacion de la Consejer!a de Educacion y Cultura ! de la Comunidad Autonoma de Madrid 07M/0069/ 1998. The authors would like to extend their gratitude to Dr. M.I. Estrella Pedrola of the Instituto de Fermentaciones Industriales for providing facilities for using the U-VI spectrophotometer. No benets in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References
[1] Baker AS, Greenham LW. Release of gentamicin from acrylic bone cement: elution and diffusion studies. J Bone Jt Surg Am 1988;70A:15517. [2] van de Belt H, Neut D, Uges DRA, Schenk W, van Horn JR, van der Mei HC, Busscher HJ. Surface roughness, porosity and wettability of gentamicin-loaded bone cements and their antibiotic release. Biomaterials 2000;21:19817. [3] Kuechle DK, Landon GC, Musher DM, Noble PC. Elution of vancomycin, daptomycin and amikacin from acrylic bone cement. Clin Orthop 1991;264:3028. * [4] Diez Pena E, Frutos P, Frutos G, Barrales-Rienda JM. Gentamicin sulphate (GS) release from a commercial bone cement. I. Inuence of the GS concentration and some other factors on the release process mechanism and kinetics, to be submitted for publication. [5] Meyer JD, Falk RF, Kelly RM, Shively JE, Withrow SJ, Dernell WS, Kroll DJ, Randolph TW, Manning MC. Preparation and in vitro characterization of gentamycin-impregnated biodegradable beads suitable for treatment of osteomyelitis. J Pharm Sci 1998;87:114954. [6] Stabile DE, Jacobs AM. Development and application of antibiotic-loaded bone cement beads. J Am Pediat Med Ass 1990;80:3549. [7] Buchholz HW, Elson RA, Heinert K. Antibiotic-loaded acrylic cement: current concepts. Clin Orthop 1984;190:96108. [8] Goodell JA, Flick AB, Hebert JC, Howe JG. Preparation and release characteristics of tobramycin impregnated poly (methyl methacrylate) beads. Am J Hosp Pharm 1986;43: 145461. [9] Calhoun JH, Mader JT. Antibiotic beads in the management of surgical infections. Am J Surg 1989;157:4439. [10] Miclau T, Dahners LE, Lindsey RW. In vitro pharmaco kinetics of antibiotic release from locally implantable materials. J Orthop Res 1993;11:62732.

P. Frutos et al. / Biomaterials 23 (2002) 37873797 [33] Lambov N, Peikov P, Belcheva N, Gesheva K, Zlatkov A, Danchebv N. Biodegradable cross-linked produrg of bronchial dilator vephylline. I. Mechanism of the release. J Controll Rel 1997;45:1937. [34] Varma IK, Patnaik S. Copolymerization of 2-hydroxyethyl methacrylate with alkyl acrylates. Eur Polym J 1976;12:25961. [35] Sung YK, Gregonis DE, Russell GA, Andrade JD. Effect of water and tacticity on the glass transition temperature of poly (2-hydroxyethly methacrylate). Polymer 1978;19:13623. [36] Fox TG, Goode WE, Gratch S, Huggett CM, Kincaid JF, Spell A, Stroupe JD. Temperature dependence of the stereospecicity in the free radical polymerization of methyl methacrylate. J Polym Sci 1958;31:1737. [37] Verhoeven J, Schaeffer R, Bouwstra JA, Junginger HE. The physico-chemical characterization of poly (2-hydroxyethyl) methacrylate-co-methacrylic acid): 2. Effect of water PEG 400

3797

[38] [39]

[40] [41]

[42]

and PEG 6000 on the glass transition temperature. Polymer 1989;30:194650. Noyes AA, Whitney WR. The rate of solution of solid substances in their own solutions. J Am Chem Soc 1897;19:9304. . . Noyes AA, Whitney AD. Uber die auosungsgeschwindgkeit von . festen stoffen in ihren eigenen losungen. Z Physik Chem 1897;23:68992. Kuhn AT, Wilson AD. The dissolution mechanism of silicate and glass-ionomer dental cements. Biomaterials 1985;6:37882. Wilson AD, Groffman DM, Kuhn AT. The release of uoride and other chemical spacies from a glass-ionomer cement. Biomaterials 1985;6:4313. Verbeeck RMH, De Moor RJG, Van Even Martens LC. The short-term uoride release of a hand-mixed vs capsulated system of a restorative glass-ionomer cement. J. Dent Res 1993;72: 57781.