Characterization of chlorhexidine-releasing, fast-setting, brushite bone cements

Available online at www.sciencedirect.com

Acta Biomaterialia 4 (2008) 1081–1088 www.elsevier.com/locate/actabiomat

Characterization of chlorhexidine-releasing, fast-setting, brushite bone cements Anne M. Young a,*, Poon Yun J. Ng a, Uwe Gbureck b, Showan N. Nazhat c, Jake E. Barralet d, Michael P. Hofmann e a

Division of Biomaterials and Tissue Engineering, UCL Eastman Dental Institute, 256 Gray’s Inn Road, London WC1X 8LD, UK b Department of Functional Materials in Medicine and Dentistry, University of Wu¨rzburg, 97070 Wu¨rzburg, Germany c Department of Mining and Materials Engineering, McGill University, Montre´al, Que´bec, Canada H3A 2B2 d Faculty of Dentistry, McGill University, Montre´al, Que´bec, Canada H3A 2B2 e Biomaterials Unit, School of Dentistry, University of Birmingham, Birmingham B4 6NN, UK Received 17 September 2007; received in revised form 14 November 2007; accepted 18 December 2007 Available online 15 January 2008

Abstract The effect of antibacterial chlorhexidine diacetate powder (CHX) on the setting kinetics of a brushite-forming b-tricalcium phosphate/ monocalcium phosphate monohydrate (b-TCP/MCPM) cement was monitored using attenuated total reflection Fourier transform infrared spectroscopy. The final composition of the set cement with up to 12 wt.% CHX content before and after submersion in water for 24 h, the kinetics of chlorhexidine release and the total sample mass change in water over four weeks was monitored using Raman mapping, UV spectroscopy and gravimetry, respectively. Below 9 wt.%, CHX content had no significant effect on brushite formation rate at 37 °C, but at 12 wt.% the half-life of the reaction decreased by one-third. Raman mapping confirmed that brushite was the main inorganic component of the set cements irrespective of CHX content, both before and after submersion in water. The CHX could be detected largely as discrete solid particles but could also be observed partially dispersed throughout the pores of the set cement. The percentage of CHX release was found to follow Fick’s law of diffusion, being independent of its initial concentration, proportional to the square root of time and, with 1 mm thick specimens, 60% was released at 24 h. Total set cement mass loss rate was not significantly affected by CHX content. On average, cements exhibited a loss of 7 wt.% assigned largely to surface phosphate particle loss within the initial 8 h followed by 0.36 wt.% per day. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Brushite; Chlorhexidine; Drug release; Drug cement interaction; Setting kinetics

1. Introduction Calcium phosphate cements, either forming hydroxyapatite (HA) or brushite (dicalcium phosphate dihydrate, DCPD) as final product, are used as artificial bone substitute materials [1,2]. A major problem with HA bone cements, comprising the majority of developed calcium phosphate cement systems, is their inability to degrade in order to allow full replacement by the surrounding tissues *

Corresponding author. Tel.: +44 2079152353. E-mail address: [email protected] (A.M. Young).

[1,2]. For this reason there has been in recent years increased interest in the development of degradable brushite-forming cement formulations [3–8]. Many calcium phosphate cements have been investigated as potential drug delivery systems. A large number of studies have focused on the release of antibiotics due to the common problem of bone infection after surgery and the many difficulties involved in treating this. Most studies have used the more robust non-resorbable HAforming systems [9–11], although controlled release of the antibiotic gentamicin from brushite cements has also been demonstrated [12,13]. For the treatment of common

1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2007.12.009

1082

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

infections in the oral cavity and jaw bone, extensive use of antibiotics, however, is of concern due to the potential for antibiotic resistance. A highly effective non-antibiotic, antibacterial alternative is chlorhexidine (CHX, [(CH2)3NH(C(@NH)NH)2C6H4Cl]2). This is used extensively in dentistry, including within crosslinked gelatine (PeriochipTM) for the treatment of periodontal infections [14,15] and as a mouthwash (e.g. CorsodylTM). It has also been incorporated within various dental restorative materials to reduce recurrent caries [16, 17] and is used in antiseptic creams and sprays (e.g. SavlonTM). One clinical problem with solid, periodontal antibacterial devices such as PeriochipTM has been difficulty in placement. Other concerns include movement from the site of application soon after placement due to lack of fixation and the relatively rapid release rate of the antibacterial agent [18]. An alternative, initially fluid, slower CHX-releasing brushite cement would enable easier filling of the periodontal pocket or infected bone around a perio-implant. Solidification of the cement would then provide fixation. After release of CHX, the cement itself could degrade, providing calcium and phosphate ions to aid surrounding bone repair. In previous studies the period of release of gentamicin from brushite cement was limited due to the high porosity of these materials [12,13]. More recently, through improvements in setting retardant action, it has been possible to increase the powder-to-liquid ratio of brushite cements and thereby reduce the final porosity [19]. The presence of CHX, however, might interfere with the hydrolysis reaction of any brushite cement system, as their chemistry and setting rate can be very sensitive to the addition of additives [4,19,20]. The aim of this investigation was therefore to assess the effect of varying CHX concentration on the chemistry of brushite cements. The effects of drug concentration on the setting kinetics and final product chemistry/ drug stability were assessed using attenuated total reflection Fourier transform infrared (ATR-FTIR) and Raman mapping, respectively. In addition, drug release and cement degradation rates were quantified using UV spectroscopy and gravimetry. It has been hypothesized that there may be a maximum drug level that can be added before any serious detrimental effects on cement setting chemistry occur. In addition, drug release is likely to be governed by a diffusion-controlled process but its rate may be reduced in a higher powder content formulation. 2. Materials and methods 2.1. Materials The reactants for the cement powder mixture were phase pure sintered b-tricalcium phosphate (b-TCP) [21] and monocalcium phosphate monohydrate (MCPM) powder (Rhodia, Birmingham, UK) with median particle sizes of 11 and 62 lm, respectively (as determined by laser diffrac-

tion particle sizing) [22]. b-TCP was prepared by sintering a 2:1 molar mixture of dicalcium phosphate anhydrous (DCPA, Mallinckrodt-Baker, Griesheim, Germany) and calcium carbonate (Merck, Darmstadt, Germany) at 1050 °C for 24 h. The b-TCP sinter cake was crushed in a mortar until it passed through a 355 lm sieve and afterwards dry milled for 1 h in a planetary ball mill (PM400 Retsch, Haan, Germany) unidirectionally at 200 rpm in 500 ml agate jars with a load of 125 g b-TCP and four agate balls (30 mm). Chlorhexidine diacetate salt hydrate (CHX) powder (Sigma, UK) was used as received (ACS grade, Sigma– Aldrich, UK). Chlorhexidine diacetate solubility in aqueous buffer solutions ranges from 3.3 to 10.1 mg ml1 at pH 7 and 4, respectively [17]. Use of this salt in powder form, instead of, for example, the more water-soluble chlorhexidine digluconate (as in Periochip), should reduce interaction between drug and cement components. Spectra of pure potential products were obtained using DCPD, and dicalcium phosphate anhydrous (DCPA or monetite), both from Sigma–Aldrich. 2.2. Cement preparation Equimolar amounts of b-TCP and MCPM were combined with increasing levels of CHX. These powders were then mixed with 800 mM citric acid solution at a powderto-liquid ratio of 3.3 g ml1 to form a cement paste and to start the setting reaction. A CHX content of more than 12 wt.% was found to make the mix too dry to form a homogeneous paste. Formulations with 0, 3, 6, 9 and 12 wt.% (relative to the total cement paste weight) CHX were therefore studied. In the presence of water, b-TCP and MCPM reacted to form a brushite cement structure, according to the following equation [3]: b-Ca3 ðPO4 Þ2 þ CaðH2 PO4 Þ2  H2 O þ 7H2 O ! 4CaHPO4  2H2 O

ð1Þ

Solid discs 8 mm in diameter and 1 mm deep were prepared using metal washer rings as moulds. Excess cement was removed with a razor blade before sealing top and bottom with acetate sheet. Samples were then left to set at 37 °C for 24 h. 2.3. FTIR setting reaction monitoring In order to investigate the kinetics of setting, the unset cements were placed on the diamond of an ATR-FTIR Perkin Elmer Series 2000 spectrometer temperature controlled at 37 °C (Perkin–Elmer Beaconsfield, UK, with Timebase software) within 60 s of mixing the powder and liquid. This method is designed to mimic how the material might set in clinical use: at room temperature during mixing by the clinician but then raised to body temperature soon after placement. Spectra between 500 and 4000 cm1 were then obtained with a resolution of 4 cm1, every 12 s from 60 s

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

until spectral absorbance changes became minimal. Results were compared with the spectra of b-TCP, MCPM, CHX, DCPD and DCPA with the same wavenumber range and resolution. Since the absorbance change profiles varied with time only in absolute intensity and not with wavenumber, the reaction process remained unchanged at all times. The reaction extent, f, could therefore be calculated using the following equation: f¼

A t  A0 DAmax

ð2Þ

where A0, and At are the absorbance initially and at time t, and DAmax is the maximum absorbance change at any wavenumber. In this study the 980 cm1 PO stretch peak of brushite was used as changes were most pronounced around this wavenumber. From this, the half-life of the reaction t1/2 was obtained. This was defined as the time from the beginning of mixing until DAt = DAmax/2 (or f = 0.5). Average values of three samples for each formulation were obtained. In order to correlate reaction times from FTIR with the handling properties of the cement pastes, portions of the samples prepared for the FTIR investigation were kept aside for visual determination of the working and setting times. (It should be noted, however, that these results are at room temperature (22 °C) and not body temperature as is possible with FTIR.) These were taken as the times at which viscosity increases rapidly, leaving the paste unworkable, or the paste hardens, respectively. Immediately following the working or setting time the cement will no longer readily flow when mixed with a spatula or changes to an obviously much drier solid, respectively. This is a simple test but can provide numbers that are as reproducible as the FTIR study. These times and t1/2 could all be determined to an accuracy of ±10 s. 2.4. Raman mapping studies For Raman mapping studies sample discs were allowed to set sealed for 24 h at room temperature. Spectra were then obtained over a wavenumber range of 700– 1650 cm1 before and after placement in 30 ml water for a further 24 h. A LabRam spectrometer (Horiba Jobin Yvon, Stanmore, UK) equipped with a 633 nm laser, 1800 grating and 50 magnification objective was used. Four scans, each of 4 s duration, were obtained every 2 lm across an area 100  60 lm2 for each specimen providing 1500 spectra in 8 h. Spectra of b-TCP, MCPM, CHX, DCPD and DCPA were also obtained from the average of two 30 s scans each. LabRam software was used to generate average spectra of the cement samples and compositional maps of the cement surfaces. 2.5. Mass loss and chlorhexidine release Three discs of each composition, of initial mass m0, were placed in tubes containing 10 ml of distilled water such that

1083

both sides of each specimen were fully exposed to the liquid. After 1, 2, 4, 8, 24, 72, 168, 336 and 672 h of storage at 37 °C the specimens were removed and mass mt determined before placement in fresh distilled water. Percentage mass change was calculated using m0  mt : ð3Þ Dm ð%Þ ¼ 100 m0 In order to confirm that CHX was the only component having a noticeable effect on UV absorbance, spectra of each storage solution were obtained using a Unicam UV 500 spectrometer (Thermo Spectronic, UK) over the range 200–400 nm, and the full spectra compared with those of standard solutions. These results indicated that quantification of CHX could readily be determined with sufficient accuracy provided its level in the cements was 6 wt.% or more. Upon lowering drug concentrations below this level, scattering due to particles released from the cements at early times could cause some error in CHX concentration determination. The concentration of CHX in storage solutions was determined using the Beer–Lambert law with a mass extinction coefficient of 46.5 cm2 mg1 at 255 nm. Data were converted to cumulative CHX release and then the percentage of total CHX content released from the sample calculated. For diffusion-controlled release this would be expected to be given by Fick’s equation for diffusion from a thin disc of thickness 2l [16,23] 1 8 X 1 M=M 1 ¼ 1  2 p n¼0 ð2n þ 1Þ2 ! ð2n þ 1Þ2 p2 Dt  exp ; ð4Þ 4l2 where t is the time after placement in water, D is the diffusion coefficient, M is the cumulative mass of drug released at time t and M1 is the mass of drug in the sample. Eq. (4) was calculated using Microsoft Excel with integer n values from 0 to 50 as a function of Dt/l2. This curve could then be fitted to experimental values of M/M1 vs. time with the diffusion coefficient of CHX as the only variable parameter. 3. Results 3.1. FTIR spectra of cements Representative FTIR spectra for cement pastes containing 9 wt.% CHX after different setting times are given in Fig. 1a. Comparison of this figure with that of the starting materials and expected products (Fig. 1b) indicated that initially the cement spectra were largely comparable with the combined spectra of the reactants b-TCP (broad PO stretch at 1000 cm1), MCPM (sharp PO stretch peaks at 1075, 955 and 850 cm1) and water (1640 cm1). Peaks due to CHX were only just visible (several overlapping C@O and C@C stretch and NAH deformations between 1400 and 1700 cm1) when its level exceeded 6 wt.%. With

1084

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

0.8

1.0

0.7

0.9 0.8

0.6

Reaction Extent

Absorbance

0.7 0.5 1200s 0.4

600s

0.3

120s

0.2 0.1

0.6 0.5 0.4

0wt%

0.3

9wt%

0.2

12wt%

0.1 0.0 1800

1600

1400

1200

1000

800

0.0

600

0

200

Wavenumber (/cm)

DCPD

1000

1200

0 500

TCP

400

1.0 MCPM 0.5

800

600

Time (s)

Absorbance

1.5

600 Time (s)

Fig. 2. Reaction extent as a function of time for cements containing 0, 9 and 12 wt.% CHX determined using FTIR.

2.5 2.0

400

CHX

0.0 1800

3 6 9

300

12

200 1600

1400

1200

1000

800

600

Wavenumber (/cm) Fig. 1. FTIR spectra of (a) cement containing 9 wt.% CHX at 120, 600 and 1200 s, and (b) of CHX, b-TCP, MCPM and DCPD (brushite).

cement setting, sharpened peaks appeared at 3530, 3465, 3255, 3160 (region not shown) and 1640 cm1 (Fig. 1a) due to OH stretching of bonded water in the brushite structure. In addition, PO peaks intensified at 1110, 1050, 980 and 860 cm1 (Fig. 1a). 3.2. Setting characteristics of cements Examples of reaction extent and reaction half-lives at 37 °C are given in Figs. 2 and 3, respectively. In all cases there was a delay before any reaction was observed. Addition of up to 9 wt.% CHX had no significant effect on the delay or reaction half-lives, but higher levels caused both to decrease. Conversely at 22 °C, the time of rapid viscosity increase (working time) and setting time both decreased more gradually when the concentration of CHX was raised (see Fig. 3). 3.3. Raman spectra of set cements Average Raman spectra of the set cements with 0, 6 and 12 wt.% CHX before placement in water are shown in Fig. 4a. Comparison of these spectra with those in

100 0 Viscosity increase

Hardening

FTIR t1/2

Fig. 3. Comparison of reaction half-life calculated from FTIR at 37 °C with the working (viscosity increase) and setting (hardening) times at 22 °C for cements containing 3–12 wt.% CHX.

Fig. 4b demonstrates that they matched with the sum of the brushite and CHX spectra except for the presence of a single sharp additional peak at 1043 cm1. The heights of the CHX relative to brushite peaks were proportional to the CHX content in the cement paste. After placement in water for 24 h no systematic changes in the average surface spectra or significant reduction in CHX relative to brushite peak heights could be detected. Fig. 5 shows compositional maps of samples containing 3, 6 and 12 wt.% CHX before and after 24 h submersion in water determined using Raman mapping. In all the compositional maps obtained there were regions that could be assigned to pure brushite crystals (blue) or pure CHX (red) and others that had spectra consistent with a mixture of the two components (mauve). Water-filled pores in the cement structure gave regions of only weak scattering (black areas). In some of the images it was possible to detect short, rod-like brushite crystal shapes, but in others these structures were less clear. At 3 wt.%, CHX could be detected in

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

tion as expected from Eq. (4) (see Fig. 6). A fit based on Eq. (4) fitted closely to the release data at all times with a diffusion coefficient of 7.5  109 cm2 s1 and an average squared multiple correlation coefficient R2 of 0.99.

140 120

986

Intensity (a.u.)

100

wt% CHX

1043

80

1286 / 1292 60

876

12 6

20

0

60

1000

1200 1400 Raman shift (/cm)

1600

989

1800

DCPD

Intensity (a.u.)

50 40

943 / 968

TCP

30 901 / 912 MCPM

20 10

1288 / 1294

1600 CHX

0 800

1000

1200 1400 Wavenumber (/cm)

3.5. Sample mass change

1597

40

0 800

1085

1600

1800

Fig. 4. Average Raman spectra of (a) set dry cements containing 0, 6 and 12 wt.% CHX and (b) of CHX, b-TCP, MCPM and DCPD (brushite).

only a low percentage of the surface spectra and may have been partially hidden under a surface layer of brushite. As the drug concentration was raised CHX was detectable in more of the spectra and by 12 wt.% drug most of the spectra of the surface had peaks that could be assigned to drug and therefore appear red or mauve in Fig. 5. At the higher drug concentrations the CHX regions were larger, suggesting possibly some agglomeration of particles and/or dissolution of the drug in water-filled pores. After placement in water the surface images could be quite variable and in some cases indicated an increase in visible CHX, possibly due to some surface loss of brushite. Additionally, in all maps a small percentage of the surface had regions 1–10 lm in diameter with an unidentified spectrum with two intense sharp peaks at 740 and 1043 cm1. Very few regions that might be attributed to unreacted MCPM or b-TCP (see, for example, green spots in Fig. 5b) could be detected in any of the compositional map spectra. 3.4. Chlorhexidine release rates The cumulative percentage of chlorhexidine release (assessed using UV) was independent of CHX concentra-

The mass loss rate of the set cements increased only slightly upon raising CHX content (see Fig. 7). There was on average a reduction in mass of around 7% during the first 8 h followed by a slower linear decline of around 0.36 wt.% per day. 4. Discussion The above FTIR results have shown that CHX content in the paste of a brushite-forming cement can be increased up to 9 wt.% without affecting the setting reaction kinetics at 37 °C. Addition of 12 wt.% CHX, however, accelerated the chemical setting process, making the formulations drier and more difficult to handle. This is contrary to the effect of gentamicin sulphate which caused a significant lengthening of the setting process of a similar brushite cement at low drug concentrations [13]. In the cements of this new study, citric acid is used as a retardant to slow the setting reaction. Previous FTIR studies have shown that in such cements a reactive intermediate citrate dicalcium phosphate complex is formed that slows the precipitation of brushite [19]. It is possible, however, that with high levels of CHX the citrate ions could replace the acetate ions and chelate with positively charged chlorhexidine ions instead of dicalcium phosphate. At 12 wt.% CHX there are comparable numbers of moles of CHX and citric acid in the formulations. Removal of only approximately one-third of the citrate ions via interaction with CHX, however, would be sufficient to account for the level of decrease in setting rate that is observed. Contrary to FTIR studies, both working and setting time of the cement declined when drug concentration exceeded 6 wt.%, possibly because these parameters were obtained at room temperature as opposed to 37 °C with FTIR. Previous FTIR studies showed that reaction halflives at room temperature are approximately double those observed at 37 °C [19]. Therefore, at lower temperatures, CHX may have greater effects on setting because the drug particles have more time to dissolve and thereby interact with other cement components. The maximum 12 wt.% (20 vol.%) CHX content in the pastes investigated is comparable with the amount present in PeriochipTM [14]. It should be noted that the size of devices required for periodontal treatment is sufficiently small that at these drug loadings, even with very rapid drug release rates, the level of CHX would not be toxic after release from the pocket and dilution in the oral cavity. Although PeriochipTM releases most of its drug content into water within a few hours, the interaction of CHX with tissues then enables it to be retained within the periodontal

-30

0.30

-20

0.40

0

0.25

-10

0.20

0.30

10 20

0.15

0

0.20

30

Y (µm)

0.35

-10

Intensity (cnt)

-20

0.10

10

0.05

20

10 µm

40 0 X (µm)

50

-30

0 X (µm)

50

-30

0

0.25

10

0.20

20

Y (µm)

0.30

Intensity (cnt)

-10

0.15

10 µm

0.40

-20

0.35

-10

0.30

0

0.25 0.20

10 20

30

5 µm

Intensity (cnt)

0.35

-20

Y (µm)

0.10

10 µm

-50

Intensity (cnt)

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

Y (µm)

1086

0.15

30

-50

0 X (µm)

50

0 X (µm)

0.6

10 0.4

20 30

0.2

10 µm

-20

0.60

-10

0.50

0

0.40

10

0.30

20

0.20 10 µm

30

40 0 X (µm)

Y (µm)

Y (µm)

0

Intensity (cnt)

0.8

-10

50

-50

Intensity (cnt)

-20

0.10

0 X (µm)

Fig. 5. Surface compositional maps (60 lm  100 lm) of cements determined via Raman before (a, c and e) and after (b, d and f) 24 h in water showing areas containing particles of CHX (red), brushite crystals (blue), porous regions (black) and low levels of b-TCP (green). Cements contain 12 wt.% CHX (a, b), 6 wt.% CHX (c, d) or 3 wt.% CHX (e, f). (In black and white images black areas are pores but lighter regions have spectra consistent with primarily CHX in a, brushite in e and f and both components in b, c and d).

120

25

wt% CHX 20

80

Mass loss (%)

CHX release (%)

100

6

60

9

40

12 model

0 5

10

15

15 10 5

20

0

0 6 9 12

20

sqrt (time / hr) Fig. 6. Percentage of total drug release as a function of the square root of time from cements containing 6, 9 and 12 wt.% CHX. Line is the best fit of Eq. (4) through the average of the results.

0 0

100

200

300 400 Time (hr)

500

600

700

Fig. 7. Mass loss as a function of time for cements containing 0, 6, 9 and 12 wt.% CHX. The straight line is a linear fit through the data after 8 h. Error bars are SD with n = 3.

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

pocket for a more prolonged period. In the presence of PeriochipTM the concentration of CHX within the periodontal pocket can exceed 2000 and 800 lg ml1 at 2 h and 2 days, respectively, but by 6 days it falls below its minimum inhibitory concentration for most oral pathogens within a biofilm (typically 150 lg ml1) [24–27]. Clinical opinion as to the efficacy of PeriochipTM is divided [28,29]. Comparison of CHX release from the above brushite cements with that from PeriochipTM suggests, however, that sufficient is released from the new cements to be effective against bacterial biofilms at early times but that their slower release might provide more prolonged and therefore effective treatment. The rate of CHX release in this study was slower than previously observed with gentamicin release from similar brushite cements. This may be a consequence of the lower powder content in the earlier brushite cement studies [13]. With higher liquid fraction, a greater volume of the final material consists of pores that are filled with ‘‘unreacted” water that is not used in the formation of brushite, making drug diffusion easier. The average Raman spectra of cements confirmed that most of the MCPM and b-TCP reactants in the cement converted into brushite irrespective of CHX concentration. This is consistent with previous X-ray diffraction (XRD) studies of identical but non-CHX-containing cement [19]. To our knowledge this is the first study to use Raman microscopy to assess the chemistry of brushite cements. The advantages over, for example, XRD include that any changes in the chemistry of non-crystalline components such as drugs can be detected, as well as low levels of unreacted MCPM or b-TCP particles. Previous studies have shown that Raman is highly sensitive to changes in the chemistry of CHX incorporated within polymers [30]. Since particles of CHX within the above cement had identical Raman spectra to those of pure drug, this confirms that the drug can be entrapped without any chemical change. In addition, microscale inhomogeneities can be observed with Raman mapping. The observation of regions containing only brushite in the Raman maps was consistent with brushite crystals forming largely separately from the chlorhexidine. At a powder-to-liquid ratio of 3.3 g ml1 with 800 mM citric acid, the porosity of brushite cement is expected to be 27% [19]. At low drug concentrations this was comparable with the area of black regions in Raman maps. Initially the particles of chlorhexidine constitute 5– 20 vol.% of the set cement for the 3–12 wt.% CHX formulations, respectively. At higher drug concentrations, dissolved and particulate CHX could occupy both the porous region and additionally its original volume. This would explain why in Raman maps of cements with high drug concentrations, much more than 20% of the area is darker red (likely to be pores containing aqueous CHX) or bright red (CHX particles). Using UV spectroscopy, it was found that more than 95% CHX release was achieved after two weeks in water. This, in combination with the observation that the full

1087

UV spectra of the storage solutions were comparable with those of pure CHX solutions, provides further confidence that the cement components do not affect the drug chemistry. At 24 h, 60% of the drug was released from the specimens. Raman mapping studies indicated, however, no obvious reduction in CHX level relative to brushite on the material surface with 24 h immersion. This might be explained if surface loss of calcium phosphate occurs as well as the chlorhexidine release from the bulk. The CHX release profile was consistent with a single diffusion-controlled process. The data also showed that doubling the concentration of the drug will indeed double the amount released in a given time period. Conversely, the percentage gentamicin release rate from brushite cements previously decreased on raising drug concentration, presumably due to much greater interaction between the drug and cement [13]. From Eq. (4) it should additionally be expected that by increasing the diameter and thickness of the samples, the amount of CHX and time of drug release respectively could be further increased. The thickness of the samples is restricted for periodontal applications but may be less so in other bone-filling applications. The mass loss studies (when compared to CHX release data) suggested that after release, the drug was replaced by water of similar density. Additionally, it suggested that the dispersed drug had a negligible effect on the early or the long-term dissolution rates of the cement matrix in water. The initially rapid mass loss in all cases is likely to be due to some particles less securely bound to the material surface rather than dissolution of, for example, unreacted MCPM as the level of this reactant in the set but non-water-exposed material surface is negligible. 5. Conclusions The study showed that up to 9 wt.% CHX content can be incorporated into brushite cements without significant effects on the material setting reactions at 37 °C or damage to the drug. This may be due to the drug largely remaining as a separate solid phase in the cement during the rapid set period. All of the CHX entrapped was released into water at a rate determined by Fick’s law of diffusion. Therefore, by increasing the sample dimensions and the drug content the rate and time of release can be controlled. The levels and time of drug release are suitable for an antibacterial periodontal device. References [1] Bohner M. Calcium orthophosphates in medicine: from ceramics to calcium phosphate cements. Injury 2000;31:37–47. [2] Bohner M, Gbureck U, Barralet JE. Technological issues for the development of more efficient calcium phosphate bone cements: a critical assessment. Biomaterials 2005;26:6423–9. [3] Mirtchi AA, Lemaitre J, Terao N. Calcium phosphate cements: study of the beta-tricalcium phosphate-monocalcium phosphate system. Biomaterials 1989;10:475–80.

1088

A.M. Young et al. / Acta Biomaterialia 4 (2008) 1081–1088

[4] Bohner M, Merkle HP, van Landuyt P, Trophardy G, Lemaitre J. Effect of several additives and their admixtures on the physicochemical properties of calcium phosphate cement. J Mater Sci: Mater Med 2000;11:111–6. [5] Grover LM, Gbureck U, Wright AJ, Barralet JE. Cement formulations in the calcium phosphate H2O–H3PO4–H4P2O7 system. J Am Ceram Soc 2005;88:3096–103. [6] Grover LM, Gbureck U, Young AM, Wright AJ, Barralet JE. Temperature dependent setting kinetics and mechanical properties of beta-TCP-pyrophosphoric acid bone cement. J Mater Chem 2005;15:4955–62. [7] Barralet JE, Grover LM, Gbureck U. Ionic modification of calcium phosphate cement viscosity Part II: hypodermic injection and strength improvement of brushite cement. Biomaterials 2004;25:2197–203. [8] Lilley KJ, Gbureck U, Wright AJ, Knowles JC, Farrar DF, Barralet JE. Brushite cements from polyphosphoric acid, calcium phosphate systems. J Am Ceram Soc 2007;90:1892–8. [9] Ginebra MP, Traykova T, Planell JA. Calcium phosphate cements: Competitive drug carriers for the musculoskeletal system? Biomaterials 2006;27:2171–7. [10] Ginebra MP, Traykova T, Planell JA. Calcium phosphate cements as bone drug delivery systems: a review. J Control Release 2006;113:102–10. [11] Schnieders J, Gbureck U, Thull R, Kissel T. Controlled release of gentamicin from calcium phosphate–poly(lactic acid-co-glycolic acid) composite bone cement. Biomaterials 2006;27:4239–49. [12] Bohner M, Lemaitre J, Merkle HP, Gander B. Control of gentamicin release from a calcium phosphate cement by admixed poly(acrylic acid). J Pharm Sci 2000;89:1262–70. [13] Bohner M, Lemaitre J, van Landuyt P, Zambelli PY, Merkle HP, Gander B. Gentamicin-loaded calcium phosphate bone cement as antibiotic delivery system. J Pharm Sci 1997;86:565–72. [14] Tzafriri AR. Mathematical modeling of diffusion-mediated release from bulk degrading matrices. J Control Release 2000;63: 69–79. [15] Heasman PA, Heasman L, Stacey F, McCracken GI. Local delivery of chlorhexidine gluconate (PerioChipTM) in periodontal maintenance patients. J Clin Periodontol 2001;28:90–5. [16] Leung D, Spratt DA, Pratten J, Gulabivala K, Mordan NJ, Young AM. Chlorhexidine-releasing methacrylate dental composite materials. Biomaterials 2005;26:7145–53. [17] Anusavice KJ, Zhang NZ, Shen C. Controlled release of chlorhexidine from UDMA-TEGDMA resin. J Dent Res 2006;85: 950–4.

[18] Young AM, Leung D, Ho SM, Sheung YT, Pratten J, Spratt DA. New injectable antibacterial composites for bone and tooth repair. In: Medical polymers 2004. Shawbury, Shropshire: Rapra Technology Ltd; 2004. p. 147–58. [19] Hofmann MP, Young AM, Gbureck U, Nazhat SN, Barralet JE. FTIR-monitoring of a fast setting brushite bone cement: effect of intermediate phases. J Mater Chem 2006;16:3199–206. [20] Bohner M, Lemaitre J, Ring TA. Effects of sulfate, pyrophosphate, and citrate ions on the physicochemical properties of cements made of beta-tricalcium phosphate–phosphoric acid–water mixtures. J Am Ceram Soc 1996;79:1427–34. [21] Gbureck U, Grolms O, Barralet JE, Grover LM, Thull R. Mechanical activation of b-TCP. Biomaterials 2003;24:4123–31. [22] Hofmann MP, Nazhat SN, Gbureck U, Barralet JE. Real-time monitoring of the setting reaction of brushite-forming cement using isothermal differential scanning calorimetry. J Biomed Mater Res Part B: Appl Biomater 2006;79:360–4. [23] Crank J. Mathematics of diffusion. 2nd ed. New York: Oxford University Press; 1957. [24] Noiri Y, Okami Y, Narimatsu M, Takahashi Y, Kawahara T, Ebisu S. Effects of chlorhexidine, minocycline, and metronidazole on porphyromonas gingivalis strain 381 in biofilms. J Periodontol 2003;74:1647–51. [25] Soskolne WA, Chajek T, Flashner M, Landau I, Stabholtz A, Kolatch B, et al. An in vivo study of the chlorhexidine release profile of the PerioChip(TM) in the gingival crevicular fluid, plasma and urine. J Clin Periodontol 1998;25:1017–21. [26] Schwach-Abdellaoui K, Vivien-Castioni N, Gurny R. Local delivery of antimicrobial agents for the treatment of periodontal diseases. Eur J Pharm Biopharm 2000;50:83–99. [27] Wilson M, Patel H, Fletcher J. Susceptibility of biofilms of Streptococcus sanguis to chlorhexidine gluconate and cetylpyridinium chloride. Oral Microbiol Immun 1996;11:188–92. [28] Daneshmand N, Jorgensen MG, Nowzari H, Morrison JL, Slots J. Initial effect of controlled release chlorhexidine on subgingival microorganisms. J Periodont Res 2002;37:375–9. [29] Salvi GE, Mombelli A, Mayfield L, Rutar A, Suvan J, Garrett S, et al. Local antimicrobial therapy after initial periodontal treatment—a randomized clinical trial comparing three biodegradable sustained release polymers. J Clin Periodont 2002;29:540–50. [30] Jones DS, Brown AF, Woolfson D, Dennis AC, Matchett LJ, Bell SEJ. Examination of the physical state of chlorhexidine within viscoelastic, bioadhesive semisolids using Raman spectroscopy. J Pharm Sci 2000;89:563–71.