Volume 80 Number 2

Metronidazole-Loaded Bioabsorbable Films as Local Antibacterial Treatment of Infected Periodontal Pockets

Yael Shifrovitch,* Itzhak Binderman,* Hila Bahar,* Israela Berdicevsky, and Meital Zilberman*
Background: Periodontal disease is infectious in nature and leads to an inammatory response. It arises from the accumulation of subgingival bacterial plaque and leads to the loss of attachment, increased probing depth, and bone loss. It is one of the worlds most prevalent chronic diseases. In this study we developed and studied metronidazole-loaded 50/50 poly(DL-lactide-co-glycolide) (PDLGA), 75/25 PDLGA, and poly(DL-lactic acid) (PDLLA) lms. These lms are designed to be inserted into the periodontal pocket and treat infections with controlled-release metronidazole for 1 month. Methods: The structured lms were prepared using the solution-casting technique. Concentrated solutions and high solvent-evaporation rates were used to get most of the drug located in the bulk, i.e., in whole lms volume. The effects of copolymer composition and drug content on the release prole, cell growth, and bacterial inhibition were investigated. Results: The PDLLA and 75/25 PDLGA lms generally exhibited a low- or medium-burst release followed by a moderate release at an approximately constant rate, whereas the 50/ 50 PDLGA lms exhibited a biphasic release prole. The drug released from lms loaded with 10% weight/weight metronidazole resulted in a signicant decrease in bacterial viability within several days. When exposed to human gingival broblasts in cell culture conditions, these lms maintained their normal broblastic features. Conclusions: This study enabled the understanding of metronidazole-release kinetics from bioabsorbable polymeric lms. The developed systems demonstrated good biocompatibility and the ability to inhibit Bacteroides fragilis growth; therefore, they may be useful in the treatment of periodontal diseases. J Periodontol 2009;80:330-337. KEY WORDS Biocompatibility; drug delivery; metronidazole.

* Department of Biomedical Engineering, Tel Aviv University, Tel Aviv, Israel. Department of Microbiology, Technion Israel Institute of Technology, Haifa, Israel.

eriodontal disease, a localized inammatory response due to infection of a periodontal pocket arising from the accumulation of subgingival plaque, is one of the worlds most prevalent chronic diseases. With 36.8% of American adults estimated to have the disease, the prevalence of periodontal disease is greater than cancer, heart disease, arthritis, obesity, acquired immunodeciency syndrome, and many other diseases.1 The topical administration of antibacterial agents in the form of mouthwashes is ineffective in controlling disease progression because a limited amount of the drugs actually reach the periodontal pocket. Moreover, the drugs are constantly ushed because of a very high uid-clearance rate (an estimated 40 replacements of the uid per hour within a 5-mm pocket).2 There are numerous benets for local drug delivery to the pocket in the form of subgingivally placed systems. Fortunately, periodontal diseases are localized in the immediate environment of the pocket, which is easily accessible for the insertion of a delivery device by syringe or tweezers, depending on the physical form of the delivery system. The critical period of exposure of the pocket to the antibacterial drug is between 7 and 10 days.3 Intrapocket delivery systems can be divided into bioabsorbable and non-bioabsorbable systems. Nonbioabsorbable systems must be removed or discharged from the pocket after their
doi: 10.1902/jop.2009.080216

330

J Periodontol February 2009

Shifrovitch, Binderman, Bahar, Berdicevsky, Zilberman

drug-release function has been accomplished. Examples of non-bioabsorbable systems include cellulose acetate bers loaded with tetracycline,4-6 chlorhexidine,7 or metronidazole8 and lm- or slab-based devices made of poly(methyl methactylate) and ethylcellulose loaded with the three aforementioned drugs.9-13 The most studied systems showed promising clinical results in the maintenance of periodontal pockets over a 2-year period.12,13 Bioabsorbable systems are usually polymeric or protein in nature and undergo natural degradation in response to gingival uid components. The rst bioabsorbable systems to be developed were based on hydroxypropylcellulose loaded with various agents: tetracycline, chlorhexidine, and ooxacin.14 The researchers reported on the fast drug release from the lm within 2 hours with tetracycline remaining within the pocket for 24 hours after insertion. Several modications have been made to overcome the rapid degradation and short duration of drug release, e.g., incorporating methacrylic acid copolymer particles into the lm to get ooxacin release for 7 days.14 A bioabsorbable device based on hydrolyzed gelatin cross-linked by formaldehyde, as reported by Steinberg et al.,15 has evolved to the commercial chlorhexidine in a gelatin matrix. This United States Food and Drug Administrationapproved device releases chlorhexidine. A different system is based on a water-free mixture of melted glycerol mono-oleate and metronidazole to which sesame oil was added to improve its ow properties in the syringe. The gel ows deeply into the periodontal pockets and readily adapts to root morphology. It sets in a liquid crystalline state when it comes in contact with water. The matrix is degraded as a result of neutrophil and bacterial activity within the pocket.16 Effective doses of metronidazole within the pocket are maintained for 24 to 36 hours. Metronidazole-loaded poly(vinyl alcohol) lms demonstrated a biphasic release prole, but the drug was released within several hours because of the high hydrophilic nature of the host polymer.17 Metronidazole and amoxicillin were loaded in poly(DLlactide-co-glycolide) (PDLGA) and poly(DL-lactic acid) (PDLLA) lms.18 The drug-release study18 showed that during the rst 16 days, the released quantities of drugs were higher than the minimum inhibitory concentration (MIC) needed against various microbes causing periodontal diseases. Metronidazole is a partially hydrophilic drug, which effectively inhibits anaerobic microorganisms and protozoan infections.19,20 This drug is used for the treatment of many infections, including periodontal and vaginal infections.12,21,22 This drug was incorporated in drug-delivery systems for various applications, such as tablets for the treatment of peptic ulcers,23 microspheres for the treatment of diseases

associated with the colon and the gastric mucosa,24 alginate gel beads for gastric applications,25 and various systems for the treatment of periodontal diseases, as described above. Solution casting of polymers is a well-known method for preparing polymer lms. To incorporate a drug using this method, the polymer is dissolved in a solvent and mixed with the drug prior to casting. Then the solvent is evaporated, and the polymer/drug lm is created. We reported a method for controlling drug location/dispersion within the lm.26,27 In this process, the solvent evaporation rate determines the kinetics of drug and polymer solidication and, thus, the drug dispersion/location within the lm. Solubility effects in the starting solution also contribute to the postcasting diffusion processes and occur concomitantly with the drying step. In general, two types of polymer/drug lm structures were created and studied: a polymer lm with large drug crystals located on the lm surface (A type) and a polymer lm with small drug particles and crystals distributed within the bulk (B type). This structure enables control of the drug-release prole and results in desired release behavior. The two types of lms, A and B, were developed and studied for two types of drugs: highly hydrophobic drugs, such as dexamethasone,26 and highly hydrophilic drugs, such as gentamicin.27 Metronidazole has moderate hydrophilicity (10 mg/ml). Therefore, it is of interest to study the controlled-release characteristics of polymer/metronidazole-loaded lms. The aim of the present study was to develop and evaluate metronidazole-loaded bioabsorbable lms, designed to be inserted into the periodontal pocket and treat infections during metronidazole controlled-release phase. This study focused on the effects of drug content and the type of host polymer on the drug-release prole. Selected lms were studied for cell growth and bacterial inhibition. We believe that a relatively long release time (4 weeks) would be benecial for the healing process. MATERIALS AND METHODS All experiments and analyses were conducted in the Biomaterials Laboratory at Tel-Aviv University. Materials Bioabsorbable polymers. The following three bioabsorbable polymers were used: PDLLAi with an inherent viscosity (IV) of 0.62 dl/g in ChCl3 at 30C, which corresponds to a molecular weight (MW) of 90,177 Da; 75/25 PDLGA with an IV of 0.65 dl/g in ChCl3 at 30C, which corresponds to a MW of 100,050 Da;
i PerioChip, Perio Products, Jerusalem, Israel. Elyzol, Dumex, Copenhagen, Denmark. Absorbable Polymers International, Pelham, AL. Absorbable Polymers International.

331

Metronidazole-Loaded Bioabsorbable Films

Volume 80 Number 2

50/50 PDLGA# with an IV of 0.57 dl/g in ChCl3 at 30C, which corresponds to a MW of 71,500 Da. The drug used in the study was metronidazole.** Microorganisms. A clinical isolate of Bacteroides fragilis was used in this study. Centers for Disease Control and Prevention (CDC) anaerobic blood agar was used to grow the microorganisms. This medium is highly recommended for the anaerobic growth of bacteria. Film Preparation Polymer lms (0.12 to 0.15 mm thick) consisting of PDLLA, 75/25 PDLGA, or 50/50 PDLGA and metronidazole were prepared by a three-step solutionprocessing method. The polymer (1 g) was mixed with a relatively small volume of methylene chloride (20 ml) at room temperature until it was totally dissolved, giving a clear solution; metronidazole was added to the polymer solution to give a relatively concentrated solution. Two constant drug loadings were used: 20 mg metronidazole (2% weight/weight [w/w]) and 100 mg metronidazole (10% w/w). The drug was fully dissolved in both solutions. The following steps were solution casting into a petri dish and solvent drying under atmospheric pressure at room temperature. The petri dish was not covered, so as to enable a relatively fast evaporation rate of 10 to 20 ml/hour. Afterwards, the solution was cast into a petri dish, and the solvent was dried under atmospheric pressure at room temperature. Following this step, the lm underwent isothermal heat treatment at 33C for 23 hours in a vacuum oven. This heat treatment enabled disposal of residual solvent. Morphologic Characterization Polarized light microscopy was performed using a microscope (transmission mode). In Vitro Weight-Loss Studies The polymer lms (1 1 cm) were weighed and then immersed in phosphate buffered saline (PBS; 50 ml in a petri dish) at 37C for 20 weeks to determine their weight-loss proles. Samples (also those broken into small pieces) were removed every week, dried in a vacuum oven, and weighed. The weight loss was calculated as follows: weight loss % 100X w0 wf ; w0

dazole release from the lms. The lms were weighed before the experiment to permit determination of 100% release. Sodium azide (0.05% weight/volume) was added to prevent contamination by various microorganisms. At the following times, 1.5-ml samples were collected: 1, 6, and 24 hours; 2, 3, and 8 days; and 2, 3, 4, 5, and 7 weeks. The maximum metronidazole concentration in the release medium was 10 times less than its water solubility so as not to affect the release kinetics. The amount removed was replaced with fresh PBS, and the correction factor was applied as follows:  n1 40 ; correction factor 40 1:5 where n is the sequential sample number. The metronidazole content in each sample was determined with an ultraviolet/visible scanning spectrophotometer at 320 nm. The metronidazole working range was 10 to 140 mg/ml; therefore, the samples were diluted. A calibration curve was prepared for each set of measurements, with a correlation coefcient >0.99. Three samples were examined for each lm type. The means and SD values are presented in the relevant gures. Preparation of Bacteria B. fragilis was grown overnight on CDC anaerobic blood agar at 37C under anaerobic conditions. The bacterial cells were collected and resuspended in saline and adjusted to 1 108/ml by visual comparison to a 0.5 McFarland standard. Ten-fold dilutions were performed in Eppendorf tubes. Aliquots of 100 ml were spread on CDC agar, and three repetitions were conducted. Evaluation of Residual Bacteria Samples of 10 or 100 ml were collected or diluted to a nal concentration, at the appropriate time, and spread on CDC anaerobic agar plates. Colony forming units (CFU)/ml were counted after 24 hours of incubation at 37C. MIC The MIC of metronidazole against B. fragilis was determined using the E-test method, and it was found to be 1 mg/ml. Special stripsii for detecting MIC values (using the E-test method) were used. Microbiologic Experiment Design The microbiologic study, i.e., the kinetics of residual bacteria, was performed as follows. The effect of metronidazole released from the lms (in PBS) on bacterial inhibition was studied. The release medium samples
# ** ii Absorbable Polymers International. Sigma, St. Louis, MO. Difco Microbiology, Detroit, MI. Leica, Wetzlar, Germany. Zenyth 200rt, Anthos, Eugendorf, Australia. Biodisk, Salna, Sweden.

where w0 and wf are the weights of the dried lms before and after exposure to water, respectively. Four samples were tested at each point. The means and SD values are presented in the relevant gures. In Vitro Metronidazole Release Study Polymer/metronidazole lms with a diameter of 10 cm were immersed in 40 ml PBS at 37C for 7 weeks (50/ 50 PDLGA lms were immersed for 5 weeks) in semistatic conditions to determine the kinetics of metroni332

J Periodontol February 2009

Shifrovitch, Binderman, Bahar, Berdicevsky, Zilberman

were collected after 0, 5, and 48 hours and 1, 2, and 3 weeks. The organisms were added to a nal concentration of 1 107 CFU/ml. The samples were collected (after 0 and 24 hours and 2 and 7 days in the case of 50/50 PDLGA lms and after 0 and 24 hours and 3 and 8 days in the case of PDLLA lms) for viable counting and were expressed as CFU/ml. Bacteria in PBS served as the control. CDC anaerobic blood agar plates were used for counting the residual bacterial. The calculation established the residual number of bacteria in the presence of metronidazole. Because B. fragilis is an obligatory anaerobe, all incubations were performed in anaerobic jars in the presence of anaerobic bags. Biocompatibility Experiment Films of 5 5 mm, which were loaded with 2% w/w and 10% w/w of metronidazole, were placed in 35-mm cell culture dishes and covered with 70% alcohol solution for 30 minutes to achieve sterilization of the lms. The alcohol was decanted and replaced by three washes of sterile PBS and left to dry in the laminar ow hood, blowing sterile air. Human gingival broblasts (HGFs; three to four subcultures of human gingival explants) were seeded at 1 105 cells/dish, with 2 ml growth medium covering the lms. The cells were cultured in minimum essential medium supplemented with 10% newborn bovine serum. The medium was replaced every 4 days. After 4 to 6 days, subconuent cultures were observed under phase-contrast microscopy. The medium was decanted, and the cells were washed with PBS and xed with 5% formalin in PBS. Two hours later, the formalin was removed, and the cells were washed in PBS and stained with Coomassie blue to expose the cytoskeleton F-actin. Cells on the lms and the culture dish exhibited normal broblastic morphologic features.

RESULTS Microstructure of Polymer/Metronidazole Films Bioabsorbable polymeric lms containing metronidazole were prepared using solution processing, accompanied by a postpreparation isothermal heat treatment. Our structuring technique enabled us to control the drug location/dispersion in the lm. Using this technique, the lms were prepared from concentrated solutions using a fast evaporation rate to get lms with the most drug molecules located in the bulk, i.e., in whole lms volume, rather than on the surface. The microstructure of the metronidazole-loaded 50/50 PDLGA, 75/25 PDLGA, and PDLLA lms is presented in Figure 1. In the 50/50 and 75/25 PDLGA-based lms, drug crystals, particles, and aggregates (2 to 100 mm) are dispersed in the bulk. Only a small amount of the drug is located on the surface of the lm.

Figure 1.
Light photomicrographs of polymer/metronidazole 10% w/w lms. Host polymer: 50/50 PDLGA (A), 75/25 PDLGA (B), and PDLLA (C).

Conversely, in the PDLLA-based lms, more drug crystals and particles are located on the surface of the lm. The effect of this difference in microstructure on the drug-release prole is described in the next section.
Oxoid, Cambridge, U.K.

333

Metronidazole-Loaded Bioabsorbable Films

Volume 80 Number 2

Figure 3.
Weight loss of polymer lms as a function of degradation time. Blue diamonds = 50/50 PDLGA; green triangles = PDLLA.

The weight-loss (erosion) prole of PDLLA and 50/ 50 PDLGA is presented in Figure 3. Although the PDLLA lms lost only 2% of their initial weight during the rst 7 weeks of degradation, the 50/50 PDLGA lms lost 18% of their initial weight. Microbiologic Evaluation of the Effect of Metronidazole Release on Bacterial Viability The purpose of these experiments was to monitor the effectiveness of various concentrations of the antibiotic released from the lms in terms of the residual bacteria compared to the initial number of bacteria. Bacteria present in PBS only served as the control. We chose the following two types of lms with different release proles: 50/50 PDLGA lms loaded with 2% w/w metronidazole, which demonstrated a biphasic release prole with a relatively highburst release of 20% (compared to the other samples containing 2% w/w drug), and PDLLA lms loaded with 10% w/w metronidazole, which demonstrated the highest burst release of ;38% and the highest release rate during the rst 2 weeks of release. Figure 4 presents the two proles for comparison. The proles are expressed as the percentage of the total drug encapsulated (Fig. 4A) and as the drug amount (milligrams; Fig. 4B). The samples were collected after 5 and 48 hours and after 7 and 21 days for 50/50 PDLGA lm containing 2% w/w metronidazole and after 5 and 48 hours and 7 and 14 days for PDLLA loaded with 10% w/w metronidazole. B. fragilis was added to the tubes containing previously released drug, and the number of bacterial cells was monitored. The initial bacterial concentrations used in our experiment were 1 107 CFU/ml, which is relatively high. Most infections involve lower concentrations (usually not more than 1 105 CFU/ml). Therefore, we assume that if our metronidazole-eluting lms were effective against such concentrations, they will be effective against serious infections. The results (bacterial inhibition kinetics) for 50/50 PDLGA lm containing 2%

Figure 2.
In vitro cumulative metronidazole release from polymer/metronidazole lms: A) lms containing 2% w/w metronidazole, B) lms containing 10% w/w metronidazole. Blue diamonds = 50/50 PDLGA; red squares = 75/25 PDLGA; green triangles = PDLLA.

Metronidazole Release From Bioabsorbable Films and Their Weight-Loss Prole The effects of polymer type and drug loading on the metronidazole-release prole were examined. The cumulative metronidazole-release proles from various polymer/metronidazole lms loaded with 2% w/w metronidazole and 10% w/w metronidazole are presented in Figure 2. A small or moderate burst release (during the rst 48 hour) was observed for all six studied lms. The lms loaded with 2% w/w metronidazole released 10% to 20% of their drug within the rst 48 hours (Fig. 2A), whereas the lms loaded with 10% w/w metronidazole released 20% to 38% of their drug (Fig. 2B). The PDLLA/metronidazole and 75/25 PDLGA/metronidazole lms exhibited a constant rate of release during the rst 2 to 3 weeks; thereafter, only a small amount of drug was released through week 7. Conversely, the 50/50 PDLGA/metronidazole lm exhibited a two-phase release: the rst release occurred during the rst 3 days for the 2% w/w loaded lms and during the rst 7 days for the 10% w/w loaded lms, and the second phase started after 14 days for the 2% w/w loaded lms and after 20 days for the 10% w/w loaded lms. All encapsulated drug was released by 35 days.
334

J Periodontol February 2009

Shifrovitch, Binderman, Bahar, Berdicevsky, Zilberman

Figure 5.
Number of CFU of B. fragilis over time. A) After exposure to metronidazole released from 50/50 PDLGA lms containing 2% w/w drug: white bars = control PBS; black bars = 5 hours; blue bars = 48 hours; green bars = 7 days; purple bars = 21 days. B) After exposure to metronidazole released from PDLLA lms containing 10% w/w drug: white bars = control PBS; black bars = 5 hours; blue bars = 48 hours; green bars = 7 days; purple bars = 14 days.

Figure 4.
Metronidazoles release prole from lm samples that were used for microbiologic evaluation expressed as the percentage of total encapsulated drug (A) and as milligrams (B). Blue diamonds = 50/50 PDLGA lm loaded with 2% w/w drug; green triangles = PDLLA lm loaded with 10% w/w drug.

w/w metronidazole are presented in Figure 5A, and the results for PDLLA loaded with 10% w/w metronidazole are presented in Figure 5B. Our results showed that, when B. fragilis was exposed to 50/50 PDLGA lm loaded with 2% w/w metronidazole, the quantity of drug released during the rst 7 days (<30%) eradicated all bacteria within less than a week (Fig. 5A), whereas the drug quantity released during the rst 5 and 48 hours only reduced the bacterial concentration to <1 103 CFU/ml within a week. Hence, relatively small amounts of drug (such as 2% w/w) start to become effective only after 1 week. When B. fragilis was exposed to the PDLLA lm, which was loaded with 10% w/w metronidazole, the drug quantity released for at least 48 hours eradicated all bacteria within 3 days (Fig. 5B). Biocompatibility No differences in cell shape or number were observed between the treated (Fig. 6A) and non-treated lms (Fig. 6C). Also, the gingival broblasts exhibited an elongated shape on the lms (Figs. 6A and 6C) or

on the bottom of the culture dishes (Figs. 6B and 6D). These results indicated that the lms are biocompatible with HGFs. DISCUSSION Our results showed that although most metronidazole particles were dispersed in the bulk of the 50/50 and 75/25 PDLGA lms, many drug particles and crystals were located on the surface of the PDLGA lms (Fig. 1). These differences in drug location/dispersion in the lm can be explained by the fact that lactic acid is more hydrophobic than glycolic acid. PDLLA contains only lactic acid monomers; therefore, it is more hydrophobic than 50/50 PDLGA and 75/25 PDLGA, which contain lactic and glycolic acid monomers in the polymer chains. Although we used concentrated (viscous) solutions and a relatively fast evaporation rate during the lm-preparation step, it seemed that the hydrophilic metronidazole tended to diffuse to the surface of the hydrophobic PDLLA solution; therefore, part of the drug was located on the surface of the lm and was not entrapped in it. The metronidazole release pattern from the 50/ 50 PDLGA lms was different than from the 75/25
335

Metronidazole-Loaded Bioabsorbable Films

Volume 80 Number 2

in the current study resulted from the lm structuring, i.e., most drug molecules were located in the lm rather than on its surface. Also, specic interactions, such as hydrogen bonding, may exist between the metronidazole molecules and the bioabsorbable polyester chains. These may decrease the release rate. Our metronidazole-release proles showed longer periods of release than those obtained for other metronidazole-eluting bioabsorbable systems (several hours to several days).14,17,18 Our microbiologic experiments showed that metronidazole was effective in the tested samples. The lms collection and preparation method did not affect metronidazoles antibiotic potency. Also, our results indicated that moderate drug contents (such as 10% w/w) were very effective and could eradFigure 6. icate the bacterial growth within HGF growth on PDLLA/metronidazole (2% w/w) lms (A), on plate at the vicinity of PDLLA/ several days, even when a large metronidazole (2% w/w) lms (B), on PDLLA/metronidazole (10% w/w) lms (C), on plate at the inoculum (1 107 CFU/ml) was vicinity of PDLLA/metronidazole (10% w/w) lms (D). (Original magnication 10). created. The release prole of metronidazole from PDLLA lms PDLGA and PDLLA lms (Fig. 2). The former lms exloaded with 10% drug was more suitable than that obhibited a constant rate of release during the rst 2 to 3 tained from 50/50 PDLGA lms loaded with 2% drug. weeks, and then only a very small amount of drug was In addition to their effectiveness against the relevant released until week 7. This rst phase of release is bacterial strain, our new structured bioabsorbable polygoverned mainly by diffusion. We expected a second mer/metronidazole lms were biocompatible with phase of release to occur after 7 weeks because of a HGFs and, therefore, may be used to treat periodontal diseases. combination of diffusion and erosion of the host polymer. In the case of 50/50 PDLGA lms, we observed CONCLUSIONS the two phases of release because that lm exhibited the fastest degradation rate, and massive erosion ocBioabsorbable lms containing metronidazole were curred during the rst 4 weeks of exposure to aqueous prepared by solution processing. These lms are demedium (Fig. 3). The release proles of the lms signed to be inserted into the periodontal pockets and loaded with 10% w/w drug showed that larger drug treat infections with controlled-release metronidazole quantities were released from the PDLLA lms than for 1 month. PDLLA and 75/25 PDLGA lms generally exhibited a low or medium burst release followed from the 75/25 PDLGA lms (Fig. 2B). This probably by a moderate release at an approximately constant resulted from differences in the drugs location in the lm, i.e., in the former, larger drug quantities were rate, whereas the 50/50 PDLGA lms exhibited a bilocated near the lms surface (Fig. 1). Our results phasic release prole due to the relatively high degradaalso showed that the drug loading affected the burst tion rate of the host polymer. This study demonstrated release. The higher burst release from lms loaded that the release prole of metronidazole was determined with 10% w/w drug resulted from higher driving force mainly by the host polymer type. Drug loading had a mifor diffusion of drug molecules from polymer domains nor effect on the release prole. The drug contents in the surrounding medium exclose to the surface. ceeded the required minimum effective concentration. Metronidazole is a relatively hydrophilic drug (10 mg/ml); therefore, we expected higher burst release When relatively high drug loading was used (10% and higher rates of release. The slow-moderate burst-rew/w), the released metronidazole resulted in a signilease values and moderate drug-release rate observed cant decrease in bacterial viability within several days,
336

J Periodontol February 2009

Shifrovitch, Binderman, Bahar, Berdicevsky, Zilberman

even with relatively high bacterial concentrations (1 107/ml CFU). When relatively low drug loading was used (2% w/w), the kinetics of bacterial inhibition was relatively slow, and 1 week was required to eradicate bacterial growth. The lm preparation did not affect metronidazoles potency as an antibacterial agent. Our results indicated that the drug-loaded lms exhibited biocompatible properties with regard to HGFs. Hence, our new metronidazole-eluting lms may be useful in the treatment of periodontal diseases. ACKNOWLEDGMENTS A clinical isolate of B. fragilis was kindly provided by the Microbiology Department, Haddassah Ein Karem Hospital, Jerusalem, Israel. The authors report no conicts of interest related to this study. REFERENCES
1. Southard GL, Godowski KC. Subgingival controlled release of antimicrobial agents in the treatment of periodontal disease. Int J Antimicrob Agents 1998;9:239253. 2. Kinane DF. Local antimicrobial therapies in periodontal disease. Ann R Australas Coll Dent Surg 2000;15: 57-60. 3. Soskolne WA. Subgingival delivery of therapeutic agents in the treatment of periodontal diseases. Crit Rev Oral Biol Med 1997;8:164-174. 4. Goodson JM, Holborow D, Dunn RL, Hogan P, Dunham S. Monolithic tetracycline-containing bers for controlled delivery to periodontal pockets. J Periodontol 1983; 54:575-579. 5. Goodson JM, Offenbacher S, Farr DH, Hogan PE. Periodontal disease treatment by local drug delivery. J Periodontol 1985;56:265-272. 6. Newman MG, Kornman KS, Doherty FM. A 6-month multi-center evaluation of adjunctive tetracycline ber therapy used in conjunction with scaling and root planing in maintenance patients: Clinical results. J Periodontol 1994;65:685-691. 7. Coventry J, Newman HN. Experimental use of a slow release device employing chlorhexidine gluconate in areas of acute periodontal inammation. J Clin Periodontol 1982;9:129-133. 8. Wan Yusof WZ, Newman HN, Strahan JD, Coventry JF. Subgingival metronidazole in dialysis tubing and subgingival chlorhexidine irrigation in the control of chronic inammatory periodontal disease. J Clin Periodontol 1984;11:166-175. 9. Addy M, Rawle L, Handley R, Newman HN, Coventry JF. The development and in vitro evaluation of acrylic strips and dialysis tubing for local drug delivery. J Periodontol 1982;53:693-699. 10. Friedman M, Golomb G. New sustained release dosage form of chlorhexidine for dental use. I. Development and kinetics of release. J Periodontal Res 1982;17:323-328. 11. Stabholz A, Soskolne WA, Friedman M, Sela NM. The use of sustained release delivery of chlorhexidine for the maintenance of periodontal pockets: 2-year clinical trial. J Periodontol 1991;62:429-433. 12. Golomb G, Friedman M, Soskolne A, Stabholz A, Sela MN. Sustained release device containing metronidazole for periodontal use. J Dent Res 1984;63:1149-1153.

13. Elkayam R, Friedman M, Stabholz A, Soskolne AW, Sela MN, Golub L. Sustained release device containing minocycline for local treatment of periodontal disease. J Control Release 1988;7:231-236. 14. Deasy PB, Collins AEM, MacCarthy DJ, Russell RJ. Use of strips containing tetracycline hydrochloride or metronidazole for the treatment of advanced periodontal disease. J Pharm Pharmacol 1989;41:694-699. 15. Steinberg D, Friedman M, Soskolne A, Sela MN. A new degradable controlled release device for treatment of periodontal disease: In vitro release study. J Periodontol 1990;61:393-398. m S, Larsson K, Krog N, 16. Norling T, Lading P, Engstro Nissen SS. Formulation of a drug delivery system based on a mixture of monoglycerides and triglycerides for use in the treatment of periodontal disease. J Clin Periodontol 1992;19:687-692. 17. Mallapragada SK, Peppas NA, Colombo P. Crystal dissolution-controlled release systems. II. Metronidazole release from semicrystalline poly(vinyl alcohol) systems. J Biomed Mater Res 1997;36:125-130. 18. Ahuja A, Ali J, Rachman S. Biodegradable periodontal intrapocket device containing metronidazole and amoxicillin: Formulation and characterization. Pharmazie 2006;61:25-29. hkla ER, Koppel T, Saag M, Pa hkla R. Metronidazole 19. Pa concentrations in plasma, saliva and periodontal pockets in patients with periodontitis. J Clin Periodontol 2005;32:163-166. 20. Lamp KC, Freeman CD, Klutman NE, Lacy MK. Pharmacokinetics and pharmacodynamics of nitroimidazole antimicrobials. Clin Pharmacokinet 1999; 36:353-373. 21. Perioli L, Ambrogi V, Rubini D, et al. Novel mucoadhesive buccal formulation containing metronidazole for the treatment of periodontal disease. J Control Release 2004;95:521-533. 22. Austin MN, Meyn LA, Hillier SL. Susceptibility of vaginal bacteria to metronidazole and tinidazole. Anaerobe 2006;12:227-230. 23. Wu Y, Fassihi R. Stability of metronidazole, tetracycline HCl and famotidine alone and in combination. Int J Pharm 2005;290:1-13. 24. Chourasia MK, Jain SK. Design and development of multiparticulate system for targeted drug delivery to colon. Drug Deliv 2004;11:201-207. 25. Murata Y, Kofuji K, Kawashima S. Preparation of oating alginate gel beads for drug delivery to the gastric mucosa. J Biomater Sci Polym Ed 2003;14: 581-588. 26. Zilberman M. Dexamethasone loaded bioresorbable lms used in medical support devices: Structure, degradation, crystallinity and drug release. Acta Biomater 2005;1:615-625. 27. Aviv M, Berdicevsky I, Zilberman M. Gentamicinloaded bioresorbable lms for prevention of bacterial infections associated with orthopedic implants. J Biomed Mater Res A 2007;83:10-19. Correspondence: Dr. Meital Zilberman, Department of Biomedical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv 69978, Israel. Fax: 972-3-6407939; e-mail: meitalz@eng.tau.ac.il. Submitted April 23, 2008; accepted for publication July 30, 2008. 337