Journal of Controlled Release 47 (1997) 247–260

Fluidized-bed spray coated porous hydrogel beads for sustained

release of diclofenac sodium
Yi-Ming Sun*, Chih-Cheng Chang, Wei-Fung Huang, Huang-Chien Liang
Department of Chemical Engineering, Yuan-Ze Institute of Technology, Chung-Li, Taoyuan 320, Taiwan, ROC

Received 11 October 1996; revised 10 January 1997; accepted 29 January 1997

Abstract

Swellable porous hydrogel beads were loaded with diclofenac sodium and then coated with pseudo-latex ethylcellulose
(EC) in a fluidized-bed spray coater. The drug release profile in-vitro can be modified by adding 25% or more triethyl citrate
(TEC) or dibutyl sebacate (DBS) as a plasticizer. The plasticized coating film can overcome the swelling stress of the core
and function as a barrier for sustained drug release. Adding TEC resulted in a near-zero order kinetics after an initial burst in
the release profile, and adding DBS showed a first initial burst, a second stage of linear release, a third stage of slow release.
The initial burst is due to the flaws or cracks on the surface of the coated-beads and it can be reduced by increasing the
amount of plasticizer or film thickness and also by adjusting the thermal treatment time. A 4- to 6-h treatment at 608C was
found to be the optimal condition to give the smallest initial burst for coating film containing TEC as a plasticizer. Adding
hydrophilic hydroxylpropyl methylcellulose (HPMC) could smooth the initial burst and change the release profile to a
Fickian-type release, and in this case increasing the thermal treatment time could reduce the release rate. Two kinds of
EC-coated beads were tested in-vivo by using rabbits as animal models and showed good sustained release behaviors as the
drug concentration in plasma could be maintained above 0.4 mg / ml and lower than 1.5 mg / ml for 24 h.  1997 Elsevier
Science Ireland Ltd.

Keywords: Microencapsulation; Hydrogel; Diclofenac sodium; Fluidized-bed; Spray coating; Ethylcellulose

1. Introduction release rate from the microcapsules is determined by

the permeabilities of the medium and drug through
Oral delivery is by far the most widely used route the membrane so that sustained release of drug can
for medication. Although a variety of new methods be achieved.
have been proposed for the preparation of controlled Hydrogel beads of PHEMA have been reported as
release dosage forms for oral delivery, microen- excellent drug carriers for controlled release [3,4],
capsulation has been well accepted by the pharma- and a zero-order release can be obtained with a
ceutical industry for its simplicity [1,2]. Granules of special design of drug concentration profile within
active ingredients are coated with a layer of insoluble the beads [5]. This study explores the feasibility of
membrane film to form microcapsules. The drug using porous hydrogel beads as drug carriers and
further coating them with a layer of ethylcellulose
*Corresponding author. film for sustained drug release. Porous beads can

0168-3659 / 97 / $17.00  1997 Elsevier Science Ireland Ltd. All rights reserved
PII S0168-3659( 97 )01649-0
248 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

provide large pore volume and surface area for drug adjust the permeation rate and balance the swelling
loading, and the porous structure causes minimal stress from the inner core, swellable hydroxylpropyl
mass transfer resistance for medium and drug so that methylcellulose (HPMC) was added in coating
the drug release rate can be solely controlled by the formulation as a modifier [16,18,19] in one case. The
coating membrane. The swelling property of the drug release properties of various formulations in
beads can enhance the drug loading process in terms such a system were evaluated in-vitro by using a
of rapid absorption of drug solution into the beads, dissolution method and in-vivo by using rabbits as
but the coating may become fragile during applica- animal models.
tion if it is not strong enough or properly swollen.
This paper demonstrates that such drawbacks can be
overcome by proper selection of the kinds and
amounts of additives and preparation conditions in 2. Experimental
the coating process.
In this study, 2-hydroxyethyl methacrylate (2-
HEMA), ethylene glycol dimethacrylate (EGDMA), 2.1. Materials
and a porous-structure forming diluent were used to
produce the porous PHEMA by inverse suspension 2-Hydroxyethyl methacrylate (HEMA) monomer
polymerization [6–9]. The beads were loaded with was a gift provided by Chung-Chun Chemical
diclofenac sodium (abbreviated as DS) as a model. A (Taiwan) and was purified by vacuum distillation
self-designed Wurster-type fluidized spray-coating with 0.5 wt% cuprous chloride (Merck, Germany) as
apparatus [10,11] was used to encapsulate the drug- inhibitor. Ethylene glycol dimethacrylate (EGDMA),
loaded beads with ethylcellulose (Aquacoat, FMC). used as the crosslinking agent in the porous micro-
Aquacoat is an aqueous pseudo-latex dispersion of sphere synthesis, was obtained from Aldrich (USA)
ethylcellulose with averaged diameter of 0.2 mm and was also purified by vacuum distillation. Azo-
[12]. The use of organic solvents is avoided so that it bisisobutyronitrile (AIBN, Kanto, Japan), the free-
is advantageous for the purpose of environmental radical polymerization initiator, was used without
protection and industrial hygiene and safety. During any purification. Sodium chloride, magnesium chlo-
the coating process, individual ethylcellulose latex ride hexahydrate, hydrochloric acid, and toluene
particles encounter the surface of a bead and were from Shimakyu (Japan) and were used as
coalesce into a continuous film after evaporation of received. Pharmaceutical grade 95% ethanol was
water [12]. The time for the coalescence step to be supplied by Taiwan Tobacco and Wine Monopoly
completed varies with the film formation temperature Bureau. Aquacoat, used as the major coating
so that the post thermal treatment time becomes an material, was obtained from FMC Co. (USA). Hy-
important operational parameter to be optimized. droxylpropyl methylcellulose (HPMC), used as a
Because the inner core used in this study is swellable hydrophilic additive in the formulation, was obtained
in aqueous medium, the coating material should be from Acros Organics (USA). Triethyl citrate (TEC,
strong and tough enough to keep the coated film Lancaster, UK), dibutyl sebacate (DBS, Tokyo
from falling apart. Adding plasticizer is generally Kasei, Japan), dimethyl sebacate (DMS, Tokyo
necessary in the coating process so that the colloidal Kasei, Japan), and triacetin (TRI, Tokyo Kasei,
polymeric particles can soften and coalesce into a Japan) were used as plasticizers in coating formula-
continuous membrane in a shorter time or at a lower tions. Diclofenac sodium and piroxicam (courtesy of
temperature during the post-drying thermal treatment Standard Chemical and Pharmaceutical Co., Taiwan)
[12,13]. In addition, the permeation properties for the were used as model drug and internal standard,
penetrants in the coating film and the drug release respectively. LC-grade acetonitrile, methanol, acetic
rate are modified [14–17]. Four kinds of commonly acid, dichloromethane, hydrochloric acid, and phos-
used plasticizers: triethyl citrate (TEC), dibutyl phoric acid were used in mobile phase or as the
sebacate (DBS), dimethyl sebacate (DMS), and extraction medium in HPLC analysis for samples
triacetin (TRI) were tested [12–18]. In order to obtained from animal study. Heparin sodium (China
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 249

Chemical and Pharmaceutical Co., Taiwan) was used 2.4. Fluidized-bed spray coating
as anticoagulant in blood sample preparation.
The coating solution was prepared by mixing
2.2. Preparation of porous hydrogel beads suitable amounts of pseudo-latex ethylcellulose
(Aquacoat), plasticizers, HPMC, and water. The
The PHEMA beads were prepared by suspension composition of the coating solution (only that with
polymerization similar to the Mueller’s method [6], TEC or DBS as a plasticizer) in each batch is listed
except that toluene was added as a pore forming in Table 1. A small self-designed fluidized-bed spray
agent in order to produce porous structure [7–9]. The coater was made similar to the Wurster column
continuous phase was aqueous solution of sodium [10,11], and its dimension is shown in Fig. 1. There
chloride and magnesium hydroxide. The dispersion is a draft tube in the center of the bed. An air-
phase consisted of HEMA, EGDMA, toluene, and atomizing spray nozzle (Model: 17310-6-1 / 8jj,
AIBN (9:1:5:0.02 by weight). The polymerization Spraying System Co., USA) is located in the bottom
was carried out in a four-head reaction vessel of the bed. In each run, 50 g of drug-loaded
blanketed with nitrogen. The continuous phase was microspheres were used. The fluidization air from
stirred in 150 rev. / min and heated to 758C first. The the bottom distribution plate was maintained at 450–
dispersion phase (600 g) was dropped slowly into 500 l / min and at 608C. The drug-loaded beads were
the continuous phase (1000 ml) in order to form filled into the bed to form stable fluidization, and
proper suspended droplets. The temperature of the then the coating solution, which was atomized by
reaction was maintained at 758C for 3 h and then 0.75-bar air, was sprayed at a rate of between 1.5
958C for 1 h. The polymerization was stopped by and 2.5 ml / min. The temperature in the outlet air
cooling the suspension solution to room temperature, was between 42–458C. The coating continued until
and then the solution was neutralized by adding all the coating solution was dispatched. The coated
concentrated hydrogen chloride. The beads were beads were dried in the bed for another 10 min
filtered out of the solution and were washed with before removal. The products were stored in an
water and aqueous ethanol (95%) several times to environment with relative humidity at about 40% at
remove unreacted monomers and toluene. The beads room temperature before further thermal treatment.
were then dried in vacuum at 908C for 6 h followed The weight of the beads increased about from 22.5%
by screening with vibratory standard sieves (En- to 35% after coating (Table 1). The drug loading
decotts, Octagon 200). Particles of 500–710 mm (DL) was about 12–13% by weight of the finished
were used in the present study. product.

2.3. Drug loading 2.5. Thermal treatment

The drug solution was prepared by dissolving Some of the coated beads were thermally treated
diclofenac sodium (20% by weight) in aqueous immediately after coating. An adequate amount of
ethanol (80 vol.%). One part of the porous beads was coated beads were evenly distributed in seven sample
added to five parts of the drug solution. The slurry dishes, and then they were placed in an oven at 608C.
solution was stirred at 150 rev. / min at 378C for 6 h, The time was counted afterwards, and one sample
then the liquid portion was removed by filtration. dish was taken out of the oven at 1, 2, 3, 4, 6, 12,
The drug-loaded beads were vacuum dried at 908C and 24 h, respectively. The thermally treated samples
for 6 h to remove the residual solvent. Drug particles were also stored in an environment with relative
were precipitated within the porous structure of the humidity at about 40% at room temperature and were
beads. The drug-loaded beads were tumbled in a tested for dissolution within 1 week.
sieve so that loosely loaded drug on the outer surface The beads for the pH-effect and animal studies
of beads could fall off and was removed. Products were stored without any immediate thermal treatment
from 10 batches were mixed to give a large ensemble after coating. Those beads were stored for about 3
for the later coating process. months and were thermally treated in the same way
250 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

Table 1
The composition of the coating solutions and the results of spray coating for samples coated with Aquacoat and TEC or DBS as a
plasticizer
Run Weight Composition of coating solutions (g) Plasticizer Weight Recovery Weight Core:
of beads (%)b after (%)c increase shell e
(g) Aquacoat a TEC DBS H2O HPMC coating percentage
(g) (%)d
1 50 55.5 3 242 20 61.2 87.9 22.4 5:1
2 50 55.5 3.75 241 25 63.7 90.5 27.4 4:1
3 50 55.5 4.5 240 30 61.4 86.3 22.9 5:1
4 50 55.5 5.25 240 35 61.2 85.1 22.5 5:1
5 50 74.1 6 320 30 67.4 86.2 34.8 3:1
6 50 55.5 3 242 20 62.6 89.9 25.3 4:1
7 50 55.5 3.75 241 25 62.9 89.3 25.8 4:1
8 50 55.5 4.5 240 30 63.1 88.7 26.2 4:1
9 50 55.5 5.25 240 35 63.3 88.0 26.6 4:1
10 50 66.7 6 325 2 30 67.6 88.3 35.2 3:1
a
The total solid content is about 30% and EC is about 27% in Aquacoat.
b
Based on the weight of coating polymers (EC and HPMC).
c
Based on the total weight of uncoated beads, solid content in Aquacoat, HPMC (used in Run 10 only), and plasticizer.
d
Based on the uncoated beads.
e
Weight ratio of the uncoated beads to the coating film (assuming the loss of the beads was negligible during spray-coating).

as we described above. All were taken out of the carried out in 1 liter of medium thermosated at 378C
oven at an optimal time, which was determined from (USP XXIII, paddle method, 50 rev. / min). The
previous experiments. The pH-effect and animal concentration of diclofenac sodium in the medium
studies were completed within one and a half months was determined from the absorbance at 276 nm by
after the thermal treatment. using a UV/ Vis spectrophotometer (Hewlett Packard
8452A) equipped with an autosampling dissolution
2.6. Scanning electron microscopy ( SEM) and package. The fractional release of diclofenac sodium
optical microscopy ( OM) into the dissolution medium was plotted versus time
to give the drug release profile curve. Deionized
A scanning electron microscope was used to water was used as the dissolution medium in all the
examine the structure of the uncoated and the EC- preliminary experiments of the plasticizer effect and
coated beads. Some of the sample beads were cut in the thermal treatment time effect. In the pH-effect
liquid nitrogen for cross-section observation. Before study, mediums of 0.1 N HCl (aq) and 0.2 M
SEM observation, all samples were coated with gold KH 2 PO 4 –NaOH (aq ) (buffered at pH56.8) were used
by using an ion sputter coater under vacuum. The as simulated gastric and intestinal solutions, respec-
swelling behaviors of those beads were observed tively.
with optical stereo microscope (Olympus SZ1145-
TR-PT). Dry beads were fixed on the bottom of a
glass dish with a piece of double sided tape. Water 2.8. Animal study
was dropped into the dish just to cover the beads in
order to swell them for swollen state observation. In order to demonstrate the sustained drug release
and compare the drug bioavailability, two kinds of
2.7. Dissolution study selected EC-coated drug loaded beads (their formula-
tion and preparation conditions will be described
Dissolution of diclofenac sodium from EC-en- later in Section 3.6) and a dose of drug powder as a
capsulated beads with drug content of 60 mg was reference were tested in-vivo. All the dosages con-
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 251

placed in hard capsules (No. 3), were delivered

orally to the lower throat of the rabbit with the aid of
a pair of long sticks and a mouth opener. About 5 ml
of water was given afterwards to help the rabbit to
swallow the capsules. Blood samples of 1 ml were
taken from rabbit ears following drug administration
at predetermined times. The blood sampling con-
tinued for 12 and 24 h for subjects with powder
dosage and with EC-coated beads, respectively.
Except for water, no food was given during the trial.
A washout of 10 days was alowed between the first,
the second, and the third trial in one subject.
The drug concentration in plasma was analyzed by
a HPLC method. The HPLC consisted of a Jasco 980
pump, a Jasco 975 UV detector, and a Hewlett
Packard 3365 Series-II PC-based data station. A
Lichrospher RP-100 (4.63250 mm) column (Hew-
lett Packard) was used, and the mobile phase was
acetonitrile / methanol / dilute acetic acid (1%)512 /
55 / 33 (v / v / v). The flow rate of mobile phase was 1
ml / min, and the detector wavelength was set at 276
nm. The whole blood taken from the rabbit was
transferred into a centrifuge tube containing 0.1 ml
heparin sodium solution (100 IU) and then cen-
trifuged at 3000 rev. / min for 10 min. The clear
upper layer of plasma was taken out, placed in
another centrifuge tube, added with 0.1 ml of 1 N
HCl (aq) , and then vortexed for 20 s. Extraction of
drug was done by adding 2 ml dichloromethane to
the tube, vortexing for 1 min, and then centrifuging
at 3000 rev. / min for 10 min. The drug-containing
organic solvent phase was pipetted from the bottom
layer and placed in a test tube, and the solvent was
removed by blowing a nitrogen stream over it. A
mobile phase of 0.5 ml with internal standard
(piroxicam, 5 mg / ml) was then added to the tube,
Fig. 1. The dimension of the Wurster-type fluidized-bed used in and the tube was vortexed for 1 min to reconstitute
this study (unit:mm).
the drug. A sample of 20 ml was taken for HPLC
analysis. In calibration, a fresh plasma sample was
added with a known amount of diclofenac sodium
tained 60 mg diclofenac sodium. Three male New and then prepared in the same way. The relative
Zealand white rabbits weighing between 2.5 and 3.5 height of the peaks of diclofenac sodium to the
kg were used as the animal models. A three-subject internal standard in the chromatograph was used to
Latin square crossover design was applied to carry calculate the drug concentration. A linear regression
out the study to avoid variation due to time and was performed for the calibration curve and the
subject-to-subject difference. The animals were square of the correlation coefficient was better than
fasted for 24 h before each trial, but water was 0.998. Drug recovery was better than 91% as the
allowed. The beads or drug powders, which were plasma concentration ranged from 0. 1 to 4 mg / ml.
252 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

3. Results and discussion

3.1. Appearance of uncoated and coated beads

The porous beads were successfully made by the

suspension polymerization method, loaded with drug,
and then coated in the fluidized-bed spray coating
apparatus. Samples were taken from different formu-
lations and operating conditions for SEM observa-
tion. Some typical images are shown in Fig. 2 to
illustrate the structure of the beads. Fig. 2a shows the
picture of a drug-loaded porous bead. The porous
structure is clearly shown, and some white spots on
the surface are crystals of diclofenac sodium. After
coating, the rough surface of the porous bead was
covered by a continuous and smooth film. Fig. 2b
shows the appearance of a coated bead, Fig. 2c
shows a cross-section of a coated bead, and Fig. 2d
shows a close-up of the coating and demonstrates
how the rough structure is covered. The thickness of
the film ranges from 15 to 40 mm in general. It was
confirmed that all the particles can be covered by a
layer of coating for all the formulation conditions in
the spray-coating process. Obviously, flaws or cracks
formed on the surface of coated film may signifi-
cantly affect the drug release rate.
Fig. 2e shows coated beads with cracks on the
surface, and Fig. 2f shows a close-up of a crack from
the side view. It can be seen that the coating splits
from top to bottom along the crack so that the mass
transport of drug and medium can be severely
affected. Formation of flaws or cracks on the coating
film was also suggested in Rowe’s work [16]. Cracks
on the surface can be found in beads from almost all
the operating conditions, but the possibility can be
minimized if the formulation of the coating materials
and the operating conditions of coating process were
properly selected.
The swelling behaviors of uncoated and coated Fig. 2. Scanning electron micrographs: (a) drug-loaded porous
hydrogel bead, (b) coated bead, (c) cross-section of a coated bead,
beads were observed by OM, and some of the results (d) side-view of the coating film, (e) coated bead with cracks on
are shown in Fig. 3. Fig. 3a,b shows micrographs of the surface, and (f) side-view of a crack.
an uncoated bead in dry and swollen states, respec-
tively. It can be seen that the dimension of the
swollen bead is 12% larger than that of the dry one.
The response of this dimensional change is very fast. almost achieved within 1 min. A capillary force is
During the experiments, significant change could be responsible for the fast water uptake. The air within
observed within 10 s after water was dropped onto the porous structure was quickly squeezed out and
the bead and the equilibrated swollen dimension was formed bubbles around the bead (Fig. 3b). The drug
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 253

Fig. 4. The effect of the amount of TEC content on the release

profiles of EC coated beads. Samples were thermal treated at 608C
for 4 h, and the corresponding information of them is given in
Table 1 (Runs 1–4).

kept after swelling and the dimensional increase of

the bead is within 3% at equilibrium. A well-coated
bead may take about 25–30 min to reach a stable
dimension during these tests. This result indicates
that the coating film is strong and tough enough to
overcome the swelling stress created by the swollen
core inside. The permeation of water is limited and
the drug release rate can be controlled by the barrier
properties of this film, accordingly.
Fig. 3e shows an OM picture of a coated bead
with cracks on the surface, and Fig. 3f shows the
picture of the corresponding swollen bead. The
Fig. 3. Optical micrographs: (a) uncoated hydrogel bead when dry, coating film tends to peel-off from the surface of the
(b) uncoated hydrogel bead swollen by water, (c) coated bead bead upon swelling, and the coating may completely
when dry, (d) coated bead swollen by water, (e) coated bead with
separate from the bead in some cases. That the
cracks on the surface when dry, (f) coated bead with cracks on the
surface swollen by water. swelling response is very fast in these cases (within
1–2 min) indicates that the transport through the
cracks plays a significant role. This result suggests
that drug within the bead may be quickly dumped
loading process can benefit from the fast response of into the extracting medium and the sustained release
swelling to minimize the processing time. of drug fails in this situation. Proper selection of the
The micrographs of a well-coated bead in dry and formulation and operation condition in order to
swollen states are shown in Fig. 3c and Fig. 4d, minimize the formation of flaws or cracks on the
respectively. The integrity of the coating film can be coated beads is of great importance.
254 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

3.2. Effect of the kind and amount of plasticizer higher recovery than the other three of the same
series so that the core / shell ratio was lowered to 4
The EC-coated beads with DMS or TRI as plasti- and a thicker coating film might form.
cizer gave the same release profile as those given by The dissolution profiles for the coated beads from
uncoated beads or EC-coated beads without plasti- Runs 1–4 (TEC) and 6–9 (DBS) with a 4-h post-
cizer. In these cases, adjustment of the amount of drying thermal treatment are shown in Fig. 4 and
plasticizer and thermal treatment time did not show Fig. 5, respectively. The profile for uncoated beads
any effect on the release profiles. All the loaded drug or coated beads without any plasticizer is also shown
tended to release completely within 20–50 min. in these two graphs as a reference. These samples
During the dissolution studies, the coating membrane were made by using the same amount of Aquacoat
peeled from the beads and a suspension of fine but with different amounts of plasticizer. The coated
particles was noticed in the dissolution vessel. The beads with plasticizer of 20% (based on ethylcellul-
peeling-off phenomena was also confirmed by OM ose) or less gave almost the same drug release profile
observation (Section 3.1). It is obvious that the as those obtained from the uncoated beads or coated
coating has flaws or cracks on the surface and is too ones without any plasticizer. At least 25% plasticizer
brittle to overcome the swelling stress exerted by the should be added to the coating formulation in order
inner swellable core. The formulation without any to significantly modify the release behavior from that
plasticizer and with DMS or TRI as a plasticizer in of the uncoated beads. Increasing the amount of
the pseudo-latex ethylcellulose based coating is not plasticizer can reduce the extent of initial burst
suitable for the design of a sustained release system because it enhances the formation of a better con-
consisting an inner core of swellable hydrogel bead. tinuous film. The outcome was demonstrated for the
Only the EC-coated beads with TEC and DBS could coatings with 20, 30, and 35% TEC (Fig. 4) or
produce satisfactory results of sustained release 20–35% DBS (Fig. 5). Similar results were obtained
quality, especially, when the coated beads were in many previously reported articles [15–17,20]. The
properly thermally treated. Hence, these formulations release profile for coated beads with 25% TEC in the
were tested in later studies.
The amount of plasticizer determines the glass
transition temperature of ethylcellulose and affects
the membrane formation process. The effect of
adding a different amount of TEC and DBS into
coating was evaluated. The composition of coating
solutions and relevant data for coated beads in this
study are summarized in Table 1.
The total recovery of the product (coated beads)
from the raw materials used (uncoated beads, coating
polymer, and plasticizer) was higher than 85% but
lower than 91%. The weight of the coated beads
increased by about 22–35% from that of the un-
coated beads depending on the amount of the
polymer and the kind of plasticizer used in the
formulation. If 1.11 parts of Aquacoat to 1 part of
beads was used (Runs 1–4 and 6–9), the averaged
weight increase was 22.660.3% and the resulting
core / shell ratio was about 5 (except those from Run
2) for the coatings added with TEC, and were
Fig. 5. The effect of the amount of DBS content on the release
26.060.5% and about 4, respectively, for the coat- profiles of the EC coated beads. Samples were thermal treated at
ings added with DBS. Although Runs 1–4 used the 608C for 4 h, and the corresponding information of them is given
same operation condition, Run 2 gave a slightly in Table 1 (Runs 6–9).
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 255

Fig. 4 is an exceptional one with thick coatings and

can not be compared with other cases mentioned
previously.
The release profiles showed a relatively constant
slow release after an initial burst when TEC was
added to the coating materials. In contrast, those
profiles showed a first initial burst, a second stage of
moderate linear release, and a third stage of slow
release when DBS was added. The initial burst may
be caused by imperfect coating or cracks on the
coating surface, and the later moderate or slow
release is due to the diffusional resistance of drug
through the film of those well-coated beads. Al-
though it is not clearly understood at this moment,
the results suggest that the kinetics of the film
controlled drug release after the initial burst may be
quite different when coating film contained a differ-
ent plasticizer. The solubility of the plasticizer in
water and the polymer-plasticizer interaction may Fig. 6. The effect of the core / shell ratio on the release profiles of
account for such deviation. the EC coated beads. Samples were thermal treated at 608C for 4
h, and the corresponding information of them is given in Table 1
(Runs 3 and 5).
3.3. Effect of core /shell ratio

The core / shell ratio can be adjusted by changing The time for thermal treatment was found to be a
the amount of coating materials delivered and is crucial parameter in the film forming process and
influenced by the recovery in the spray-coating would vary with the drug release profile [20,21].
process. The thickness of the coating film is directly In this study, we kept the thermal treatment
determined by the core / shell ratio, and it increases temperature at 608C because this temperature might
as the ratio decreases. Although it is not our major be appropriate as suggested by published reports
purpose to study the thickness effect on the film [20,21]. Treatment time was varied for the coated
coating, a comparison of the release profiles of the samples with 20, 25, 30, and 35% of TEC or DBS.
coated beads from Runs 3 and 5 gives a clue about In the release profiles of the beads coated by film
the effect. The initial burst is much reduced as the with 20% of TEC or DBS, a negligible difference
core / shell ratio is changed from 5 to 3 as shown in was noticed for various thermal treatment times and
Fig. 6. Increasing the amount of coating materials or all the active ingredient released within 1 h. The
coating thickness can increase the resistance of mass effect of treatment time becomes more significant as
transport and minimize the flaws or cracks on the the amount of TEC or DBS increases.
coating film; therefore, the release curve and the Under the effect of various thermal treatment time,
initial burst of coating with lower core / shell ratio is the drug release behaviors of the beads coated with a
below that of coating with higher core / shell ratio. higher content of TEC or DBS are shown in Fig. 7
and Fig. 8, respectively. Fig. 7 shows the release
3.4. Effect of the time of post-drying thermal profile of the diclofenac sodium from the coated
treatment beads obtained from Run 5, where 30% TEC was
added to the film coating. The initial burst dimin-
As that we have discussed in the introduction ished as the thermal treatment time increased up to
section, post-drying thermal treatment is usually 4–6 h, but the trend reversed as the treatment time
necessary for the dispersed particles of ethylcellulose further increased. There is an optimal thermal treat-
latex to melt into a homogenous and continuous film. ment time (4–6 h) which leads to smallest initial
256 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

beads with HPMC seal coat have also been reported

[20]. Microscopic observation also confirmed that
4–6 h thermal treatment had minimized the fraction
of beads with cracks on the surface. Each release
curve shows a relatively constant slope with time
(solid line on each curve) after the initial burst that
indicates a near-zero order release, which is the
characteristics of a diffusion-controlled membrane
reservoir system when a constant concentration
difference across the membrane can be maintained.
The T 60 (time for 60% release) calculated from each
curve ranged from a time shorter than 0.2 h to longer
than 14 h, and the longest T 60 was also resulted from
the one with 4–6 h thermal treatment.
The effect of thermal treatment time on the film
coating plasticized by DBS is quite different from
that by TEC. Fig. 8 shows the treatment time effect
on the release profile of diclofenac sodium from the
Fig. 7. The effect of thermal treatment time (at 608C) on the coated beads obtained from Run 9, where 35% DBS
release profiles of the EC coated beads with 30% TEC. Other
coating information can be found from Run 5 of Table 1. The
was added. Except the one been thermally treated for
symbols and dashed lines show the data of experimental results, 1 h, all the others performed similarly. The DBS
and the solid lines indicate the linear range of each curve. content was high enough so that the film was
possibly well-coated for almost all of the beads. The
initial burst could be ignored for all cases studied
burst and the coated film may attain a premium here, and each release curve shows linear release up
quality. Similar trends in the effect of treating time at to about 60%. The portion of linear release also
508C on ibuprofen release from Aquacoat-coated indicates that the zero-order release occurred due to
the diffusional resistance of film coating when the
concentration difference across the coating film was
kept constant. The T 60 of the one with 1-h thermal
treatment was 1 h and those of the others were all
around 2 h. On the other hand, the T 90 of all the
cases ranged from 8.5 to 14 h. The averaged release
rate within T 60 was about 15 times that of the period
between T 60 and T 90 . Because the span of linear
release was too short and the release rate of the later
40% was comparably much smaller than the initial
rate to give a long-term constant release rate, the
coating with DBS as a plasticizer is not tested
further.

3.5. Effect of adding HPMC

Ethylcellulose is a water-insoluble polymer, and

its degree of swelling is small (,4% by weight for
Fig. 8. The effect of thermal treatment time (at 608C) on the pure EC (unpublisshed data)). However, the per-
release profiles of the EC coated beads with 35% DBS. Other meability for most water soluble drugs depends on
coating information can be found from Run 9 of Table 1. the hydration of the membrane as the water channel
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 257

within an EC membrane is the major pathway for thermal treatment time leads to a decrease in the
drugs to pass through [22]. Adding a hydrophilic release rate. The one with 24-h thermal treatment has
polymer into an EC coating film may increase the the slowest release rate in this study, and its release
degree of swelling of the membrane so that the drug profile is very close to a Fickian-type kinetics, that
permeation rate can be modified. In addition, a more is, the fractional release is proportional to the square
swellable coating film may reduce the stress from the root of time, as confirmed by a linear regression
swollen core on the coating so that breaking of the analysis (regression coefficient r 2 .0.999 for the
coating film can be minimized even if there are some cumulative release of DS from 0 to 60%). It seems
small flaws or cracks. HPMC is a commonly used that the addition of HPMC can further minimize the
additive in the film coating process in the pharma- formation of cracks on the surface of coated beads or
ceutical industry [18]. In this study, HPMC was used reduce the breaking of the coating film so that the
as a model hydrophilic polymer to replace part of the burst effect disappears; however, the reason for the
ethylcellulose in the coating materials. The com- change in release kinetics is also not clear and it may
position of the coating solution used is listed as Run be related to the redistribution of drug between the
10 in Table 1. The ratio of EC to HPMC in the core and the coating film upon coating and sub-
coating was 9:1 (by weight), the amount of TEC was sequent treatment.
30% of the coating polymers, and the resulting core /
shell ratio of the coated beads equaled 3. 3.6. Effect of pH in the dissolution study
Fig. 9 shows the effect of thermal treatment time
on the release profile of diclofenac sodium from the After the preliminary study of plasticizer and
coated beads with HPMC in the coating film. The thermal treatment time effects, beads from two kinds
curves in this figure are quite different from those we of preparation conditions were chosen for later trials.
have discussed above. The initial burst and the linear One of them was from Run 5, and the other from
range of the release disappear, and the release rate Run 10. The release profile of the former represented
decreases with time in each profile. Increasing the a near zero-order kinetics after a short burst, and that
of the later represented a Fickian-type kinetics. In the
study of pH-effect and the animal study, thermal
treatment was done after the coated beads had been
stored for 3 months after the coating was completed,
and the finished samples were tested within one and
a half months after the thermal treatment. The beads
from Run 5 were treated at 608C for 4 h, and those
from Run 10 were treated at the same temperature
for 24 h. To facilitate later discussion, we assign the
former one as SR-A and the later one as SR-B (SR
stands for sustained release).
Fig. 10 shows the release profiles of those two
kinds of drug-loaded beads in the simulated gastric
fluid, water, and simulated intestinal fluid. Because
diclofenac sodium is only slightly soluble in acid, the
release rate of the drug from both kinds of beads in
simulated gastric fluid was very slow and the final
release in 24 h was only about 9% of the loaded
drug. Not much difference could be found for both
SR-A and SR-B in simulated gastric fluid. Both
Fig. 9. The effect of thermal treatment time (at 608C) on the
release profiles of the EC / HPMC (9 / 1) coated beads with 30% SR-A and SR-B showed similar release profiles in
TEC. Other coating information can be found from Run 10 of pure water as those obtained previously indicating
Table 1. that the reproducibility is acceptable. SR-A showed
258 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

was about the same as that in pure water, but that for
SR-B in simulated intestinal fluid could reach as high
as 95% of the loaded drug. The deviation of the
release profiles in simulated intestinal fluid from
those in pure water may result from the difference of
ionic strength of the solution since the pH difference
between pure water and the simulated intestinal fluid
is rather small. Hydrophilic HPMC could be affected
more than hydrophobic EC so that the release rate
and the final release in 24 h for SR-B in simulated
intestinal fluid were both much higher than those
obtained in pure water.

3.7. Animal study

An animal study was carried out to find out the

drug release behaviors in-vivo for SR-A, SR-B, and
a dose of pure drug powder (as a reference) of the
same drug content (60 mg). In Fig. 11a–c, the
plasma drug concentrations as a function of time for
the powder dose, SR-A, and SR-B, respectively are
shown. There are two maximum plasma concen-
trations (Cmax ) in each curve. The data of these
maximum plasma concentrations, time of maximum
(T max ), area under curve (AUC), and averaged
plasma drug concentration (Cavg ) during the experi-
ments are summarized in Table 2.
Apparently, both SR-A and SR-B showed obvious
sustained drug release in-vivo. The T max 2 for both
SR dosages were later than the one for a powder
dose, the Cmax for both SR dosages were much lower
than those for a powder dose. A relatively stable
plasma drug concentration could be maintained for a
long period of time since it was higher than 0.4 and
Fig. 10. The effect of pH on the release profiles for: (a) EC coated
beads with 30% TEC and (b) EC / HPMC (9 / 1) coated beads with lower than 1.5 mg / ml during the experiments (24 h)
30% TEC. Other coating information can be found from Runs 5 for both SR dosages. In contrast, the plasma drug
and 10, respectively, of Table 1 (symbols indicate the averaged concentration for the powder dose reached the
data, n 56). highest value at 5.7 mg / ml and dropped very fast to
below 0.5 mg / ml at 12 h after the dosage was given.
an initial burst and then a linear range up to more The long-term sustained release might also sacrifice
than 60% release, while SR-B showed a Fickian-type bioavailability since the AUC for both SR-A and
release in water. The final release in water was about SR-B were lower than for the powder dose. Never-
80% of the loaded drug for both of them. The release theless, the results demonstrate that coated porous
rates of the drug from both kinds of beads in the beads can provide the function pertaining to the
simulated intestinal fluid were both higher than the sustained drug release of diclofenac sodium, and
rates in pure water, but the one from SR-A leveled future application of this proposed method for de-
off at later stage with shorter linear range. The final veloping other sustained drug release dosages is
release in 24 h for SR-A in simulated intestinal fluid feasible.
 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260 259

4. Conclusions

In this study, swellable porous hydrogel beads

have been used as drug carriers and the drug release
rate can be moderated by coating the drug-loaded
beads with aqueous pseudo-latex ethylcellulose
through a fluidized-bed spray coating technique.
Among the tested plasticizers, triethyl citrate (TEC)
and dibutyl sebacate (DBS) can aid the formation of
good quality coating film, while dimethyl sebacate
(DMS) and triacetin (TRI) fail to do so. The coated
film made from proper formulation is strong and
tough enough to withhold the swelling stress from
the inner core. Adding TEC could result in a near-
zero order kinetics after an initial burst in the release
profile, while adding DBS showed a first initial burst,
a second stage of linear release (zero-order), and a
third stage of slow release. The initial burst in
release profile can be minimized by increasing the
amount of plasticizers or adding hydrophilic hy-
droxylpropyl methylcellulose (HPMC). A 4- to 6-h
thermal treatment at 608C is found to be the optimal
condition for coating with TEC as a plasticizer.
Adding hydrophilic hydroxylpropyl methylcellulose
(HPMC) in the TEC plasticized EC-latex coating
changes the release profile to a Fickian-type release
and the release rate is reduced by increasing the
thermal treatment time up to 24 h. Two kinds of
coated beads (SR-A and SR-B) have been tested
in-vivo and the results show positively sustained
release behaviors in comparison with a reference of
powder dose.
It is confirmed micrographically that the burst
effect in release profiles is due to the presence of
cracks on the coating surface. Smooth and crack-free
Fig. 11. Drug concentration in the plasma as a function of time coating film can be made with proper selection of the
for: (a) powder dose (dashed line indicates estimated data), (b) kind and amount of plasticizer and later thermal
SR-A, and (c) SR-B. (n53).
treatment time.

Table 2
Summary of the results from in-vivo study
Dosage form Cmax 1 (mg / ml) T max 1 (h) Cmax 2 (mg / ml) T max 2 (h) AUC (mg-h / ml) Cavg a (mg / ml)
Powder dose 4.32 5.5 5.68 7 25.5 ( 0→12 h) 2.13
SR-A 1.28 5 1.45 10 20.8 ( 0→24 h) 0.87
SR-B 1.03 7 0.95 10 15.2 ( 0→24 h) 0.63
a
Cavg , AUC / 12 for powder dose and AUC / 24 for SR-A and SR-B.
260 Y. Sun et al. / Journal of Controlled Release 47 (1997) 247 – 260

Although this study demonstrates that the meth- [8] O. Okay and C. Gurun, Synthesis and formation mechanism
of porous 2-hydroxyethyl methacrylate-ethylene glycol di-
odology is feasible, there are no sound fundamental
methacrylate copolymer beads. J. Appl. Polym. Sci. 46
data on solute transport and good understanding of (1992) 401–410.
the mass transport mechanisms through the coating [9] D. Horak, F. Lednicky, V. Rehak and F. Svec, Porous
film to predict the experimental results. Extensive polyHEMA beads prepared by suspension polymerization in
experimental trials are still required for design of aqueous medium. J. Appl. Polym. Sci. 49 (1993) 2041–
2050.
such a sustained release formulation with desirable
[10] H.S. Hall and R.E. Pondell, The Wurster process. In: A.F.
drug release rate. Currently, we are in a process of Kydonieus (Ed.), Controlled Release Technologies: Methods,
correlating the permeability of drug through the free Theory, and Applications, Vol. II, CRC Press, Boca Raton,
film of coating materials with different formulations, FL, 1980, pp. 133–154.
therefore, fundamental data of transport properties [11] K. Lehmann, Fluid-bed spray coating. In: M. Donbrow (Ed.),
Microcapsules and Nanoparticles in Medicine and Pharmacy,
can be accumulated for future design of such con-
CRC Press, Boca Raton, FL, 1991, pp. 73–97.
trolled release systems. [12] FMC Corporation, Product Information, Aquacoat /
Ethylcellulose aqueous dispersion NF, Philadelphia, PA,
1996.
Acknowledgments [13] R. Bodmeier and O. Paeratakul, The distribution of plasticiz-
ers between aqueous and polymer phases in aqueous colloi-
dal polymer dispersions. Int. J. Pharm. 103 (1994) 47–54.
The support from the National Science Council of [14] A.O. Okhamafe and P. York, Analysis of the permeation and
the Republic of China through the grant of NSC-85- mechanical characteristics of some aqueous-based film coat-
2331-B-155-001 is greatly appreciated. The authors ing systems. J. Pharm. Pharmacol. 35 (1983) 409–415.
would also like to thank Dr. Eileen Chang-Chein of [15] F.W. Goodhart, M.R. Harris, K.S. Murthy and R.U. Nesbitt,
An evaluation of aqueous film-forming dispersions for
Standard Chemical and Pharmaceutical Co. for valu-
controlled release. Pharm. Technol. 8 (1984) 64–71.
able discussions and kind assistance in many ways. [16] R.C. Rowe, The effect of the molecular weight of ethyl
cellulose on the drug release properties of mixed films of
ethyl cellulose and hydroxypropyl methylcellulose. Int. J.
References Pharm. 29 (1986) 37–41.
[17] D.B. Hutching and A. Sakr, Influence of pH and plasticizers
on drug release from ethylcellulose pseudolatex coated
[1] P.B. Deasy, Microencapsulation and Related Drug Processes, pellets. J. Pharm. Sci. 83 (1994) 1386–1390.
Marcel Dekker, New York, 1984. [18] A. Wade and P.J. Weller (Eds.), Handbook of Pharmaceutical
[2] M. Donbrow (Ed.), Microcapsules and Nanoparticles in Excipients, American Pharmaceutical Association, Washing-
Medicine and Pharmacy, CRC Press, Boca Raton, FL, 1991. ton DC, 1994.
[3] W.R. Good, Diffusion of water soluble drugs from initially [19] M.R. Harris, I. Ghebre-Sellassie and R.U. Nesbitt, A water-
dry hydrogels. In: R.J. Kostelnik (Ed.), Polymeric Delivery based coating process for sustained release. Pharm. Technol.
Systems, Gordon and Breach, New York, 1976, pp. 139–153. 10 (1986) 102–107.
[4] S.H. Gehrke and P.I. Lee, Hydrogels for drug delivery [20] R. Bodmeier and O. Paeratakul, The effect of curing on drug
systems. In: P. Tyle (Ed.), Specialized Drug Delivery Sys- release and morphological properties of ethylcellulose pseu-
tems, Marcel Dekker, New York, 1990, pp. 333–392. dolatex-coated beads. Drug Dev. Ind. Pharm. 20 (1994)
[5] P.I. Lee, Effect of non-uniform initial drug concentration 1517–1533.
distribution on the kinetics of drug release from glassy [21] D.E. Hutching, B. Kuzmak and A. Sakr, Processing consid-
hydrogel matrices. Polymer 25 (1984) 973–978. erations for an EC Latex coating system: influence of curing
[6] K.F. Mueller, S.J. Heiber and W.L. Plankl, Process for time and temperature. Pharm. Res. 11 (1994) 1474–1478.
preparing hydrogels as spherical beads of large size, U.S. [22] Y. Narisawa, H. Yoshino, Y. Hirakawa and K. Noda, Po-
Patent No. 4,224,427, 1981. rosity-controlled ethylcellulose film coating. II. Spontaneous
[7] A. Jayakrishnan and C. Thanoo, Suspension polymerization porous film formation in the spraying process and its solute
of 2-hydroxyethyl methacrylate in the presence of polymeric permeability. Int. J. Pharm. 104 (1994) 95–106.
diluents: a novel route to spherical highly porous beads for
biomedical applications. J. Biomed. Mater. Res. 24 (1990)
913–927.