ARTICLE IN PRESS

Biomaterials 25 (2004) 3743–3750

A novel polymeric chlorhexidine delivery device for the treatment

of periodontal disease
Isaac C. Yuea,b, Jason Poffb, Mar!ıa E. Corte! sb,d, Ruben D. Sinisterrab,c, Caroline B. Farisa,
Patrice Hildgenb, Robert Langerb, V. Prasad Shastrib,*,1
a
Harvard School of Dental Medicine, Boston MA, USA
b
Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge MA, USA
c
Department of Chemistry, Federal University of Minas Gerais, Belo Horizonte, Brazil
d
Restorative Dentistry Department, Federal University of Minas Gerais, Belo Horizonte, Brazil
Received 12 May 2003; accepted 21 September 2003

Abstract

An implantable, anti-microbial delivery device for the treatment of periodontal disease has been developed. In this polymer-based
delivery system, the encapsulation efficiency, release characteristics, and bioactivity of anti-microbial agent were controlled by the
complexation of the drug with cyclodextrins of differing lipophilicity. Microparticles of poly(dl-lactic-co-glycolic acid) (PLGA)
containing chlorhexidine (Chx) free base, chlorhexidine digluconate (Chx-Dg) and their association or inclusion complex with
methylated-b-cyclodextrin (MBCD) and hydroxypropyl-b-cyclodextrin (HPBCD) were prepared by single emulsion, solvent
evaporation technique. It was observed that encapsulation efficiency and release of the chlorhexidine derivatives from the
microparticles was a function of the lipophilicity of the cyclodextrin. Complexation of the poorly water soluble Chx with the more
hydrophilic HPBCD resulted in 62% higher encapsulation efficiency and longer duration of sustained release over a 2-week period
than complexation with the more lipophilic MBCD. In contrast, the complexation of the more water-soluble derivative of
chlorhexidine, Chx-Dg, with the more lipophilic MBCD improved encapsulation efficiency by 12% and prolonged its release in
comparison to both the free Chx-Dg and its complex with HPBCD. Furthermore, it was observed that the initial burst effect could
be diminished by complexation with CD. Preliminary studies have shown that the chlorhexidine released from PLGA chips is
biologically active against bacterial population that is relevant in periodontitis (P. gingivalis and B. forsythus) and a healthy
inhibition zone is maintained in agar plate assay over a period of at least a 1-week. The PLGA/CD delivery system described in this
paper may prove useful for the localized delivery of chlorhexidine salts and other anti-microbial agents in the treatment of
periodontal disease where prolonged-controlled delivery is desired.
r 2003 Elsevier Ltd. All rights reserved.

Keywords: Drug delivery device; Microencapsulation chlorhexidine; Chlorhexidine digluconate; Cyclodextrin; Periodontal disease; PLGA;
Biodegradable polymers

1. Introduction disease is Gram negative, anaerobic bacteria [4]. These

Dental caries and periodontal disease are generally
considered to be the major oral health problems around
the world [1]. Approximately 30% of the American
population (70 million people) suffers from adult
periodontitis, which is defined as periodontal pockets
greater than 4 mm [2,3]. The etiology of periodontal
*Corresponding author. Department of Materials Science and
Engineering, University of Pennsylvania, 3231 Walnut Street, Phila-
delphia, PA 19104, Tel: +1-215-590-7524; fax: +1-215-590-6633.
E-mail address: shastriv@seas.upenn.edu (V.P. Shastri).
1
Current address: School of Medicine and Engineering and Applied
Science, University of Pennsylvania, PA.
bacteria form biofilms called plaques, which contain
more that 1 ! 1011 bacteria/cm3. Over 400 species of
bacteria have been isolated from plaques, but only a few
species of bacteria release toxins and induce a host
inflammatory response resulting in the destruction of
alveolar bone and connective tissues that supports the
dentition [5]. With the destruction of the periodontal
apparatus, patients often lose their teeth bringing about
a decrease in function and esthetics [6].
Currently, the most commonly used procedure for the
treatment of severe periodontitis is the use of mechanical
disruption of the bacterial flora by a procedure called
scaling and root planing. However, several studies have

0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biomaterials.2003.09.113
 ARTICLE IN PRESS
3744 I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750
shown such mechanical disruption is insufficient in
altering the make-up of flora so as to prevent a
recurrence of infection at the sites [7–9]. It has been
suggested that the use of an adjunctive localized anti-
microbial therapy may eliminate the need for secondary
surgical intervention and improve clinical outcomes [10–
12]. Such an adjunctive therapy should involve sustained
delivery of an anti-microbial agent or agents at
efficacious levels for a duration of at least 7 days in
order to alter the flora associated with various clinical
forms of periodontitis [13,14].
In the past 30 years, chlorhexidine (Chx), a bacterial
membrane permeabilizing agent and its water-soluble
derivative chlorhexidine digluconate (Chx-Dg) has been
used extensively in mouth rinses to control bacteria
biofilms on teeth [15–17]. Over the past decade, a cross-
linked collagen based delivery system for Chx-Dg called
PerioChips has been developed [18–22]. In this system,
varying the cross-linking of the collagen matrix and the
concentration of Chx-Dg in the matrix alters the release
of Chx-Dg. The cross-linked collagen delivery system
has certain shortcomings and they include poor hand-
ling characteristics and sub-optimal release profiles
[23,24]. Therefore, the goal of this study is to develop
an implantable system for the delivery of Chx, based on
degradable polymers, that offers improved release
kinetics and handling characteristics. Cyclodextrins
(CD), which are cyclic polysugars, have been used
extensively to alter physico-chemical characteristics of
drug molecules in formulation of oral dosage forms [25–
27]. Based on this observation it was hypothesized that
complexation of a drug with CD prior to encapsulation
in a biodegradable polymer matrix would offer an
additional means of controlling drug encapsulation and
release that is based on the solubility and lipophilicity of
the drug-CD complex. It has been demonstrated
recently that encapsulation and release of inorganic
complexes from biodegradable polymeric microspheres
could be controlled by complexation with cyclodextrins
[28]. In this study, this general idea has been extended
and applied toward the encapsulation and release of
large organic molecules such as Chx and its derivatives
from biodegradable poly(dl-lactic acid-co-glycolic acid)
(PLGA) microspheres by complexation with CD of
varying lipophilicity namely, hydroxypropyl (HPBCD)
and methylated beta-cyclodextrin (MBCD).
2. Experimental section
2.1. Materials
Hydroxypropyl (HPBCD) and methylated beta-cy-
clodextrin (MBCD) was a generous gift from Cerestar
Company (formerly known as Amaizo, Corp., Ham-
mond, Indiana, USA). Chx-Dg was a generous gift from
Degussa Chemicals. Poly(vinyl alcohol) (PVA) was
purchased from Aldrich (MW 10–12 kDa, 80% hydro-
lyzed). Poly(dl-lactic acid -co-glycolic acid) (PLGA)
(RG502, 20 kDa) was purchased from Boehringer
Ingelheim Chemicals (Indianapolis, USA). All other
chemicals were purchased from Aldrich Chemical Co.,
and used as received.
2.2. Delivery device preparation and analytical methods
The development of this drug delivery device consists
of three production steps: (1) complexation of the drug
with the CD, (2) encapsulation of drug/CD complex
into polymer microspheres, and (3) compression of
polymeric microspheres into rectangular chips.
2.3. Preparation of drug complex
The association (inclusion) complex between the CD
and Chx/Chx-Dg was prepared by freeze-drying of Chx
and CD aqueous solution. In brief, CD and Chx/Chx-
Dg in a 1:1 molar ratio was dissolved in distilled water,
and mixed with a magnetic stirrer. The solution was
filtered to remove impurities, and the resulting clear
solution was frozen in liquid nitrogen and lyophilized
into a powder. The formation of an inclusion complex
was confirmed by FT-infrared spectroscopic analysis
samples dispersed in KBr pellets (Nicolet Magna IR-550
FTIR) and differential scanning calorimetry (DSC)
(Perking Elmer Series-7, sample size 5–10 mg) under
dynamic nitrogen atmosphere.
2.4. Preparation of microspheres
The polymer microspheres were prepared by an
emulsion-solvent evaporation technique. In brief,
100 mg of polymer was dissolved in 0.5 ml of methylene
chloride, mixed with the drug or complex of the drug in
a powder form, and sonicated for 30 s using a Vibra Cell
ultra-sonicator (Sonic Materials, Norwalk, CT) to
prepare polymer-drug dispersion. This was then dis-
persed in 100 ml of 1% PVA using a Silverson
homogenizer (Silverson, Nottingham, UK) at 7000 rpm
for 2 min and subsequently stirred at room temperature
for 1 h to harden the microspheres. The microspheres
were isolated by centrifugation and subsequently
washed three times with distilled/de-ionized water to
remove surface adsorbed PVA and resuspended in
minimal volume of water, frozen in liquid nitrogen
and then freeze-dried for 1 week to a free flowing
powder. The mean diameter of the microspheres was
determined using a Coulter Multisizer II. The theore-
tical loading was 20% of polymer mass unless stated
otherwise.
 ARTICLE IN PRESS
I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750
2.5. Fabrication of chip delivery device
Microspheres (10 mg) were compressed into rectan-
gular chips (4 ! 5 ! 0.5 mm) in a hardened steel mold
using a laboratory Carver Press (P=10,000 lbs/cm2,
5 min). The chips were sterilized by exposure to
ultraviolet irradiation in a laminar flow hood for 24 h
before use.
2.6. Release studies
The chips (10.070.2 mg) were placed in 1.5 ml
Eppendorf tubes containing 1 ml of distilled water
release medium (pH 6–6.5 at 37" C) and then mounted
on a Lab-line orbital shaker to ensure thorough mixing.
To ensure sink conditions, the release medium was
completely removed and replaced with fresh medium at
pre-determined time intervals. Release studies were
carried out for two weeks. The total drug encapsulation
was assessed after ninety days at which point the matrix
had completely disintegrated.
Although a serum-rich medium would mimic the
gingival crevicular fluid of the periodontal pocket with
respect to its protein composition, distilled water was
chosen as the release medium as its pH of 6–6.5 is
similar to what is found in the periodontal pocket at
least in the early stages of periodontal disease. Wata-
nabe et al. have shown that patients with gingivitis have
a mean pH of 6–7 [34]. It is also worth noting that Sela
et al. also used distilled water as the release medium in
their earlier studies with PerioChip [23]. Furthermore,
the presence of proteins in the release medium would
make it difficult to establish the baseline release
characteristics of the delivery system. Yet another
consideration was the observation that Chx-Dg pre-
cipitates upon prolonged exposure to PBS due the
formation of the insoluble phosphate salt of Chx, which
is formed upon exchange of the gluconic acid anion with
phosphate ions. Finally, over the course of the release
studies no appreciable changes in release medium pH
from initial conditions were observed suggesting that the
released drug was not altering the release medium.
2.7. Quantification of Chx and Chx-Dg
The Chx and Chx-Dg released was quantified by
reverse phase High Performance Liquid Chromatogra-
phy using a C18 Nova-Pak column (3.9 ! 150 mm,
Waters, Milford, MA) on a Waters HPLC system
equipped with two pumps (Model 510), an autosampler
(WISP 712) and a multiwavelength UV-Visible detector
(Waters 490) interfaced to a Digital DEC PC. Elution
was carried out under isocratic conditions, using a
mobile phase of methanol:water (pH 3.0, 0.1 m phos-
phoric acid, 0.01 m 1-pentanesulfonic acid (36:64, v/v) at
a flow rate of 1.25 ml/min and peak detection was
3745
carried out at 280 nm. Linear calibration curves were
established for both Chx and Chx-Dg. The retention
times for Chx and Chx-Dg were approximately 3.5 and
2.9 min, respectively. The chromatograms were analyzed
using a Waters Millennium software package.
2.8. Bacteriological studies
Studies were undertaken to determine the effective-
ness of the complexation with respect to the drug alone
and to ascertain whether the observed in vitro release
profiles would translate into effective inhibition of
pathogenic bacteria. Bacteriological plates were pre-
pared as follows: NHK bacteriological plates inoculated
with P. gingivalis bacteria were prepared and placed for
two days in an anaerobic chamber at 37" C in an
atmosphere consisting of 80% nitrogen, 10% hydrogen,
and 10% carbon dioxide. Chips (Chx, Chx-Dg, Chx-
HPBCD and Chx-Dg-MBCD) with 20% theoretical
drug loading were placed in wells and 100 ml of Milli-Q
filtered water was added on top of the chips to hydrate
them and aid in drug diffusion. The choice of Chx-
HPBCD and Chx-Dg-MBCD was based on their
superior release profile. Plates were then placed in the
anaerobic chamber and the zones of inhibition were
measured daily using electronic calipers. The formation
of zone of inhibition was followed over a 10-day period.
Preliminary inhibition studies were also carried out with
B. forsythus.
2.9. Statistical analysis
Data was collected from at least three batches and
summarized using appropriate descriptive statistics
through STATA statistical package. Multiple regression
was used to calculate p values, and in p-value of o0.05
was considered statistically significant.
3. Results
3.1. Characterization of Chx-CD complex
The formation of an inclusion complex between Chx
and CD was demonstrated by infrared spectroscopy.
The IR spectrum of HPBCD revealed a strong band
around 3400 cm#1 and 1100 cm#1 associated with v #OH
and v C#O#C, respectively. These bands were signifi-
cantly narrow in comparison to free CD, which is
suggestive of the disruption of hydrogen bonds within
the CD cavity upon inclusion [29]. The formation of an
inclusion complex was further confirmed by the absence
of the melting transition around 133" C, associated with
uncomplexed Chx (Fig. 1). This result is in agreement
with the literature where similar observations were made
in the Chx:beta-CD system [29].
 ARTICLE IN PRESS
3746 I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750
3.2. Microsphere size
The mean diameter of microspheres was measured
using a Coulter Multisizer II. The size of the micro-
spheres ranged from 27 to 45 mm (Table 1). In general,
microspheres containing of Chx-Dg were larger than
microspheres containing Chx. The complexation of Chx
and its derivatives with cyclodextrins did not signifi-
cantly alter the size of the microspheres within each
group with the exception of Chx-HPBCD, which were
about 20–30% larger.
3.3. Encapsulation efficiency
The encapsulation efficiency of Chx, Chx-Dg, and
complexes with CD into PLGA microspheres is shown
in Table 2. It was observed that the complexation of the
water insoluble, lipophilic Chx with the more hydro-
philic HPBCD resulted in an almost 62% increase in
Chx encapsulation in comparison to uncomplexed and
MBCD complexed Chx. However, in the case of the
water-soluble derivative Chx-Dg, complexation with the
more hydrophilic HPBCD diminished the encapsulation
by almost 50% while no statistical differences were
observed between uncomplexed and Chx-Dg-MBCD
Fig. 1. Differential scanning calorimetric curves of (A) Chx-HPBCD
(inclusion compound) and (B) Chx (free).
Table 1
Mean diameters (mm) of PLGA microspheres containing Chx, Chx-Dg
and their complexes with HPBCD and MBCD
Chx
No CD 26.770.6
MBCD 27.372.6
HPBCD 40.472.2
Table 2
Encapsulation efficiency of Chx, Chx-Dg and their complex with CD’s
in PLGA microspheres prepared by simple emulsion-solvent evapora-
tion technique (loadings are reported as mass of drug in 10 mg of
microspheres)
Chx! Chx-Dg!
No CD 9.3% (20) 9.2% (10)
— 10.2% (20)
— 9.8% (40)
MBCD 9.0% (20) 16.8% (10)
— 11.5% (20)
— 8.1% (40)
HPBCD 15.1% (20) 6.1% (20)
SEM o72%.
! Numbers in parentheses represent theoretical loading values.
Chx multiple regression: No CD and HPBCD, p ¼ 0:02; MBCD
and HPBCD, p ¼ 0:017:
Chx-Dg multiple regression: No CD and MBCD, p ¼ 0:05; No CD
and HPBCD, p ¼ 0:001:
complex. In the systems with increasing loading (Chx-
Dg and Chx-Dg-MBCD) the encapsulation efficiency of
uncomplexed Chx-Dg remained relatively unchanged
around 9% (sd o71%) (Table 2, Fig. 5). This
translates into a linear proportionality to Chx-Dg
concentration in the microspheres and loading. How-
ever, in the case of the Chx-Dg-MBCD increasing
loading resulted in a diminution of encapsulation
efficiency (Table 2, Fig. 6). Although the drug concen-
tration increased in the microspheres, it did not scale
proportionally as a function of loading. The diminution
may be attributed to two factors, namely, (1) significant
increase in mass: the mass of the drug:CD complex is
almost 5 times greater than the free drug and (2)
increased osmolality resulting in increased leaching of
the drug-complex during the hardening of the micro-
sphere.
3.4. Characteristics of PLGA chips
The optical micrographs of PLGA microspheres
containing Chx-Dg-MBCD compressed into rectangular
chips are shown in Fig. 2. The microspheres prepared by
the emulsion-solvent evaporation technique yielded
particles that had a narrow and reproducible size
distribution, reproducible loading efficiency, and good
packing characteristics. These characteristics are ideal
for compression molding as dense structures can be
obtained using minimal pressures. The chips exhibited
good handling characteristics even after 3–5 days
incubation in distilled water with no appreciable
Chx-Dg
swelling. However, over the course of the studies the
45.174.2 chips adhered to the wall of the Eppendorf tube making
38.871.5 retrieval and any further analysis with SEM or an
45.174.6
optical microscope extremely difficult. Prior to using the
 ARTICLE IN PRESS
I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750 3747

120
Chx

Cumulative Chx Release (ug)

100 Chx-MBCD
Chx-HPBCD
80

60

40

20

0
0 5 10 15
Time (Days)
Fig. 3. Release profiles of free and CD complexed Chx from PLGA
chips. Release studies were carried out in distilled water at 37" C under
perfect sink conditions.

Fig. 2. Optical micrographs of rectangular chips compressed from

Chx-Dg containing PLGA microspheres. 100
ChxDg20
ChxDg-MBCD20

Cumulative Chx-Dg Release (ug)

ChxDg-HPBCD20
chips in our release studies we compared the release 75
profile of Chx-Dg from free microspheres and chips and
observed similar release patterns (data not shown).
Fracture analysis of the chips using optical microscopy 50
revealed that the majority of microspheres remained
intact under the fabrication conditions employed
although some compression was observed (data not 25
shown). In fact Gliadels, which is the only FDA
approved delivery system for local administration of
chemotherapeutic agent for the treatment of Glioblasto- 0
ma Multiforme, a fatal form of brain tumor, is 0 5 10 15
fabricated by compression molding of poly(anhydride) Time (Days)
microspheres containing BCNU [30]. Fig. 4. Release profiles of free and CD complexed Chx-Dg from
PLGA chips. Release studies were carried out in distilled water at 37" C
3.5. Effect of cyclodextrin on Chx and Chx-Dg release under perfect sink conditions.
from PLGA microspheres

The release curves of Chx and Chx-Dg from PLGA
chips over a 14-day period are shown in Figs. 3 and 4,
respectively. In the first 24 h the following trends were
observed. In general, complexation of Chx and Chx-Dg
with HPBCD and MBCD resulted in larger drug
concentration over the first 24 h. Within the CD’s,
complexation with the more hydrophilic derivative
HPBCD resulted in a 1.5–2 fold increase in solution
drug concentration over the initial 24 h period when
compared to complexation with MBCD. The similarity
in release trends between CD complexes of Chx and
Chx-Dg suggests that one of the dominant factors
governing drug release from this system may be the
solubility of the drug-CD complex.
Studies were carried out to ascertain the effect of
increased loading of the drug-CD complex on release
behavior. For these studies the Chx-Dg system was
chosen as a model system as the excellent water
solubility of Chx-Dg would result in substantial release
of the drug during the initial 24-hour period. The
comparison of Chx-Dg was made to the Chx-Dg-
MBCD complex as it exhibited encapsulation efficiency
that was comparable to the uncomplexed drug. The
release profiles of Chx-Dg and Chx-Dg-MBCD from
PLGA microspheres over a theoretical loading range of
10–40% are shown in Figs. 5 and 6, respectively. We
observed that in the uncomplexed Chx-Dg system, as
expected, an increase in theoretical loading resulted in
an increase in drug release over the first 24-h period.
However, in the Chx-Dg-MBCD system the initial
release appeared to be statistically similar to one another
and quite independent of drug loading. For the
remaining duration of the release (2 weeks), the
concentration of Chx-Dg from the uncomplexed system
showed little change while that from complexed system
appeared to be a function of the drug-complex loading.
 ARTICLE IN PRESS
3748 I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750

25 60
Chx-HPBCD
ChxDg-M BCD
Cumulative Chx-Dg Release (ug)

Diameter of Zone of Inhibition (mm)

ChxDg10 50
20 Chx
ChxDg20 Chxdg
40
ChxDg40
15
30

10
20

5 10

0 0
0 2 4 6 8 10
0 5 10 15
Time (Day) Time (days)

Fig. 5. Release profile of Chx-Dg from PLGA chips, as a function of Fig. 7. Evolution of zone of inhibition in presence of Chx, Chx-Dg,
loading (theoretical loading: 10%, 20% and 40%, Encapsulation Chx-HPBCD and Chx-Dg-MBCD, as a function of time in agar plates
efficiency: 9.2%, 10.2%, and 9.8%, respectively). Numbers associated inoculated with P. gingivalis. (N ¼ 3). Error bars represent sem.
with the curve legends correspond to theoretical loading. Release
studies were carried out in distilled water at 37" C under perfect sink
conditions.

125
Cumulative Chx-Dg Release (ug)

100

75

50
ChxDg -MBCD10
25 ChxDg -MBCD20
ChxDg -MBCD40
0
0 5 10 15
Time (Day)
Fig. 6. Release profile of Chx-Dg-MBCD from PLGA chips, as a
function of loading (theoretical loading 10%, 20%, and 40%,
Encapsulation efficiency: 16.8%, 11.5%, and 8.1%, respectively).
Numbers associated with the curve legends correspond to theoretical Fig. 8. Optical micrograph of an agar plate inoculated with P.
loading. Release studies were carried in distilled water at 37" C under gingivalis, containing PLGA chip composed of Chx-Dg-MBCD
perfect sink conditions. microspheres after 1 week. The arrows point to the zone of lysis (1)
and the zone of inhibition (2). Note that after 1 week, the zone of lysis
(lysed red blood cells (RBC)) (1) is about 20 mm in diameter and the
zone of inhibition (2) is about 40 mm in diameter. We observed that the
zone of inhibition is established quite rapidly within a 24-h period and
3.6. Bacterial inhibition studies
the zone of lysis is prominent after 2–3 days.

Preliminary studies to ascertain the biological activity

of the Chx and derivatives released from these chips
have been carried out. In these studies the formation
and maintenance of a zone of inhibition on agar plates
inoculated with P. gingivalis was followed over a 10-day
period (Figs. 7 and 8). PLGA polymer by itself and
cyclodextrins demonstrated no inhibition of the bacteria
species and demonstrated no lysis of the red blood cells
in the agar plate (data not shown). However, all chips
containing Chx or its derivatives showed a clear zone of
inhibition and lysis beginning on the first day. These
zones continued to rapidly increase in size throughout
the duration of this study except in the case of the
uncomplexed Chx and Chx-Dg where after day-3 the
zone of inhibition appeared to reach a plateau. It is
interesting to note that the evolution of zones of
inhibition as a function of time for both Chx and
ChxDg were quite similar. The zone of inhibition and
lysis obtained with chips containing Chx/HPBCD and
 ARTICLE IN PRESS
I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750
ChxDg/MBCD were larger (>30 mm at day 10) and
showed an upward trend (Fig. 7). An optical micro-
graph of a bacterial agar plate inoculated with P.
gingivalis, containing a chip composed of microspheres
of Chx-Dg-MBCD, that is representative of this study is
shown in Fig. 8. In general the zone of lysis was
approximately half the size of the zone of inhibition for
all samples. Two factors could potentially contribute to
the lysis of RBC (1) the well-established surfactant
nature of Chx-Dg and (2) in the case of the complexed
drug the increased osmolality of the local environment
due to a high concentration of water soluble CD.
4. Discussion
In general the mean particle diameters of the micro-
spheres containing the poorly water-soluble Chx were
smaller than those containing the more water-soluble
Chx-Dg. However, the complexation with CD did not
affect the size of the microspheres within each group.
The larger size of Chx-Dg containing microspheres may
be a result of increased water uptake during the
microsphere hardening phase, as the more water soluble
Chx-Dg and its complex with CD is more capable of
increasing the osmolality within the polymer micro-
spheres in comparison to the more lipophilic Chx and its
complex with CD.
It was observed that the complexation of Chx and
Chx-Dg with HPBCD and MBCD respectively did have
an effect on encapsulation of the drugs (Table 2). While
complexation of the poorly water-soluble Chx with the
more lipophilic MBCD did not alter encapsulation
efficiency (B9%) in comparison to uncomplexed Chx
(9.3%), we observed almost 2-fold improvement in
encapsulation efficiency when it was complexed with the
more hydrophilic HPBCD. This trend was almost
opposite in the case of Chx-Dg, where complexation
with MBCD resulted in no change in encapsulation
efficiency and complexation with HPBCD diminished
encapsulation by almost 50% in comparison to un-
complexed Chx-Dg. It was also observed that the
complexation of both Chx and Chx-Dg with HPBCD
resulted in a greater drug concentration during the first
24-h period, which was followed by a relatively steady
release over a period of 2 weeks (Figs. 3 and 4). In
contrast complexation with MBCD resulted in almost a
3-fold decrease in the initial release of Chx, even though
the concentration in the microsphere was almost 2-fold
greater. In the case of Chx-Dg, complexation with
MBCD resulted in a more linear release.
The above observations may be explained on the basis
of local saturation of drug [31–33] and diffusivity of the
free drug/drug:CD complex. Once the microsphere
hydrates, a concentration gradient that involves the
drug in its free or complexed form will be created. The
3749
ability of the drug to diffuse along this gradient will be a
function of both the solubility of the drug-CD complex
and of the free drug and the diffusivity of the drug. Both
parameters are a function of lipophilicity of the moiety,
which in turn will dictate partitioning dynamics. In the
case of the Chx-CD complex, the dissociation of the
drug-CD is favored by thermodynamics (the dissociated
system has greater entropy), however the diffusion of
the drug along the gradient is limited by poor solubility
of the free Chx. In contrast in the case of Chx-Dg due to
its high solubility in water, the main factor governing
drug diffusion would be the solubility of the complex
with CD. This conclusion is consistent with the
observation that the complex of Chx-Dg, with the more
lipophilic MBCD results in a more sustained release in
comparison to its complex with HPBCD. Furthermore,
one cannot rule out the role of the mobility of the
complex within the polymer matrix and the affinity of
the complex to the polymer matrix in the drug release.
Preliminary studies have been carried out using Agar
plates to evaluate the efficacy of the Chx delivery system
against some of the periodontal flora responsible for
periodontal disease. These studies have shown that Chx
and its derivatives are released in an active form and
that the chip configuration is suitable for maintaining a
healthy zone of inhibition for at least a 1-week period
(Figs. 7 and 8).
The polymeric chlorhexidine delivery system appears
to have all the pre-requisites for clinical evaluation.
Since the device is fabricated by the compression of
drug-loaded microspheres tuning the delivery profile to
match clinical needs should be achievable. The poly-
meric nature of the device also enables customization
with respect pocket size. The efficacy and safety of the
delivery device needs to be established in canine and
primate models before proceeding into the clinical
phase. Several important issues should be addressed in
this phase namely: (1) Inflammation at delivery site due
to localized changes in pH due to accumulation of
polymer degradation products. This can be addressed by
incorporating buffer salts in the chip. (2) Although the
device is expected to plasticize upon water uptake and
become pliable, its initial rigid configuration can have an
abrasive effect with potentially adverse impact on the
already inflamed periodontal tissue. This may be
addressed by the addition of lubricious polymers such
as Polyoxs (PEO MW B106 Da) in the Chip during
fabrication. (3) Finally, the shape of the device needs to
be optimized to ensure retention of device in the
periodontal pocket for at least a period of 2 weeks.
In conclusion, in this study it has been shown that the
encapsulation and release of large organic moieties such
as Chx and its derivatives from degradable polymeric
matrices can be modulated by complexation with
cyclodextrins. The chlorhexidine delivery system devel-
oped exhibits good release characteristics for at least
 ARTICLE IN PRESS
3750 I.C. Yue et al. / Biomaterials 25 (2004) 3743–3750
a 2-week period and may be suitable for local delivery of
anti-microbial and anti-inflammatory agents in the
treatment of periodontitis.
Acknowledgements
The authors acknowledge support of this project
through grants from the National Institutes of Health
(GM 26698 & R24-AI47739-03), the MIT-Taiwan
research program, and support to RDS and MEC
respectively from CAPES and CNPq initiative of the
Brazilian Government. ICY would like to thank the
Harvard Office of Enrichment for their support earlier
on. The authors also wish to acknowledge Savanda Sou
and Prof. Max Goodson for their help with the
bacteriological studies.
References
[1] http://www.whoafr.org/textonly/orahealth/epidemiology.html
[2] Oliver RC, Brown LJ, Loe H. Periodontal diseases in the United
States population. J Periodontol 1998;69(2):269–78.
[3] Anonymous. Periodontal diseases: epidemiology and diagnosis.
J Am Dent Assoc 1998;129(Suppl):9S–14S.
[4] Gibbons RJ, van Houte J. Bacterial adherence and the formation
of dental plaques. In: Beachey EH, editor. Bacterial adherence.
London: Chapman & Hall; 1980. p. 62–104.
[5] Havemose-Poulsen A, Holmstrup P. Factors affecting IL-1-
mediated collagen metabolism by fibroblasts and the pathogenesis
of periodontal disease: a review of the literature. Crit Rev Oral
Biol Med 1997;8(2):217–36.
[6] Anonymous. Periodontal diseases: pathogenesis and microbial
factors. J Am Dent Assoc 1998;129(Suppl):58S–62S.
[7] Cugini MA, Haffajee AD, Smith Jr C, Kent RL, Socransky SS.
The effect of scaling and root planing on the clinical and
microbiological parameters of periodontal diseases: 12-month
results. J Clin Periodontol 2000;27(1):30–6.
[8] Hirschfield I, Wasserman B. A longer-term survey of tooth loss in
600 treated periodontal patients. J Periodontol 1978;49:225–37.
[9] Goldman MJ, Ross IF, Goteiner D. Effect of periodontal therapy
on patients maintained for 15 years or longer, a retrospective
study. J Periodontol 1986;57:347–53.
[10] Greenstein G, Polson A. The role of local drug delivery in the
management of periodontal diseases: a comprehensive review.
J Periodontol 1998;69(5):507–20.
[11] Kornman KS. Controlled-release local delivery antimicrobials in
periodontics: prospects for the future. J Periodontol 1993;64:783.
[12] Langer R. Drug delivery and targeting. Nature 1998;392:5–10.
[13] Persson GR, Alves ME, Chambers DA, Clark WB, Cohen R,
Crawford JM, DeRouen TA, Magnusson I, Schindler T, Page
RC. A multicenter clinical trial of PerioGard in distinguishing
between diseased and healthy periodontal sites. (I). Study design,
methodology and therapeutic outcome. J Clin Periodontol
1995;22(10):794–803.
[14] Drisko CL, Cobb CM, Killoy WJ, et al. Evaluation of period-
ontal treatments using controlled-release tetracycline fibers:
clinical response. J Periodontol 1995;66(8):692–9.
[15] Mandel ID. Antimicrobial mouth rinses: overview and update.
J Am Dent Assoc 1994;125(Suppl 2):2S–10S.
[16] Magnusson I. Local delivery of antimicrobial agents for the
treatment of periodontitis. Compend Contin Educ Dent
1998;19(10):953–6.
[17] Ciancio SG. Local delivery of chlorhexidine. Compend Contin
Educ Dent 1999;20(5):427–32.
[18] Soskolne W, Heasman P, Newman H. Sustained local delivery of
chlorhexidine in the treatment of periodontitis: a multi-center
study. J Periodontol 1997;68:32–8.
[19] Heasman PA, Heasman L, Stacey F, McCracken GI. Local
delivery of chlorhexidine gluconate in periodontal maintenance
patients. J Clin Periodontol 2001;28(1):90–5.
[20] Soskolne WA, Chajek T, Flashner M, Landau I, Stabholtz A,
Kolatch B, Lerner EI. An in vivo study of the chlorhexidine
release profile of the PerioChip in the gingival crevicular fluid,
plasma and urine. J Clin Periodontol 1998;25(12):1017–21.
[21] Killoy WJ. The use of locally delivered chlorhexidine in the
treatment of periodontitis. Clinical results. J Clin Periodontol
1998;25(11, Part 2):953–8; discussion 978–9.
[22] Jeffcoat MK, Bray KS, Ciancio SG, Dentino AR, Fine DH,
Gordon JM, Gunsolley JC, Killoy WJ, Lowenguth RA,
Magnusson NI, Offenbacher S, Palcanis KG, Proskin HM,
Finkelman RD, Flashner M. Adjunctive use of a sub-gingival
controlled-release chlorhexidine chip reduces probing depth and
improves attachment level compared with scaling and root
planning alone. J Periodontol 1998;69(9):989–97.
[23] 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(7):393–8.
[24] MacNeill SR, Johnson VB, Killoy WJ, Yonke M, Ridenhour L.
The time and ease of placement of the chlorhexidine chip local
delivery system. Compend Contin Educ Dent 1998;19(11):
1158–62, 1164–7.
[25] Szente L, Szejtli J. Drug/cyclodextrin/hydroxy acid multi-compo-
nent systems. Properties and pharmaceutical applications.
J Pharm Sci 2000;89(1):1–8.
[26] Ammar HO, Ghorab M, El-nahhas SA, Omar SM, Ghorab MM.
Improvement of some pharmaceutical properties of drugs by
cyclodextrin complexation. 5. Theophylline. Pharmazie 1996;
51(1):42–6.
[27] Uekama K, Otagiri M. Cyclodextrins in drug carrier systems. Crit
Rev Therap Drug Carrier Systems 1987;3(1):1–40.
[28] Sinisterra RD, Shastri VP, Najjar R, Langer R. Encapsulation
and release of rhodium(II) citrate and its association complex
with hydroxypropyl-b-cyclodextrin from biodegradable polymer.
J Pharm Sci 1999;88(5):574–6.
[29] Cort!es ME, Sinisterra RD, Avila-Campos MJ, Tortamano N,
Rocha RG. The chlorhexidine:b-cyclodextrin inclusion com-
pound: preparation, characterization and microbiological evalua-
tion. J Inclus Phenom Macrocyclic Chem 2001;40:297–302.
[30] Dang W, Daviau T, Brem H. Morphological characterization of
polyanhydride biodegradable implant gliadel during in vitro and
in vivo erosion using scanning electron microscopy. Pharm Res
1996;13(5):683–91.
[31] Wimmer R, Aachman K, Larsem KL, Petersen SB. NMR
diffusion as a novel for measuring the association constants
between cyclodextrins and guest molecules. Carbohydrate Res
2002;337:841–9.
[32] Pariot N, Edwards-Levy F, Andry MC, Levy MC. Cross-linked -
b-cyclodextrins microcapsules II. Retarding effect on drug release
through semi-permeable membranes. Int J Pharm 2002;232:
175–81.
[33] Loftsson T, Masson M. Cyclodextrins in topical drug formula-
tions: theory and practice. Int J Pharm 2001;225:15–30.
[34] Kobayashi K, Kobayashi K, Soeda W, Watanabe T. Gingival
crevicular pH in experimental gingivitis and occlusal trauma in
man. J Periodontol 1998;69:1036–43.