Drug controlled release from structured bioresorbable films used in medical devices—A mathematical model

Drug Controlled Release From Structured Bioresorbable FilmsUsed in Medical Devices—A Mathematical Model

Meital Zilberman, Alon Malka

Department of Biomedical Engineering, Faculty of Engineering, Tel-Aviv University, Tel-Aviv 69978, Israel

Received 13 December 2007; revised 1 April 2008; accepted 17 June 2008Published online 5 September 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.31200

Abstract: A mathematical model for predicting drug release profiles from structured

bioresorbable films was developed and studied. These films, which combine good mechanical

properties with desired drug release profiles, are designed for use in various biomedical

applications. Our structured polymer/drug films are prepared using a promising technique for

controlling the drug location/dispersion in the film. The present model was used for predicting

drug release profiles from two film types that is films in which the drug is located on the

surface (A-type) and films in which the drug is located in the bulk (B-type). The model is based

on Fick’s 2nd law of diffusion and assumes that the drug release profile from the films is

affected by the host polymer’s characteristics, the drug location/dispersion in the film and the

drug’s characteristics. This semiempirical model uses the weight loss profile of the host

polymers as well as the change in their degree of crystallinity with degradation. Our study

indicates that the model correlates well with in vitro release results, exhibiting a mean error of

less than 7% for most studied cases. It also shows that the host polymer’s degradation has a

greater effect on the drug release profile than the degree of crystallinity. This new model

exhibits a potential for simulating the release profile of bioactive agents from structured films

for a wide variety of biomedical applications. ' 2008 Wiley Periodicals, Inc. J Biomed Mater Res Part

B: Appl Biomater 89B: 155–164, 2009

Keywords: drug release; bioresorbable films; mathematical model; medical device;

dexamethasone

INTRODUCTION

Bioresorbable drug-eluting films can be used as basic ele-

ments in various biomedical implants, such as tracheal

stents1 and coatings for orthopaedic implants.2 Poly(a-hydroxy acid) films loaded with water-soluble and water-

insoluble drugs have been developed and studied for vari-

ous applications.3–8 For example, tetracycline was released

from poly(L-lactic acid) (PLLA) barrier films, and the sys-

tem was studied for potential use in periodontal therapy,3

gentamicin was released from poly(a-hydroxy acid) films

for potential use in local treatment of bone infection,2,4,5

dexamethasone (DM) was released from various poly(a-hydroxy acid) films used as basic elements in tracheal

films,1 and sirolimus was released from a bilayer PLLA-

PLGA film in order to simulate its release from a stent.7

Bioactive agents with a relatively high molecular weight,

such as albumin and the malaria vaccine (synthetic poly-

peptide), were released from PLGA films.6,8 New combina-

tions of materials, other than the traditional poly(a-hydroxy

acid), were tried as host polymers9–13 in order to exhibit

proper mechanical properties with the desired release pro-

file from polymeric bioresorbable films. For example, a

combination of alginate and polyethylene glycol and a graft

copolymer of vinyl alcohol and lactic/glycolic acid were

tried in paclitaxel eluting films for stent applications,9,10

and segmented poly(ether-ester-amide) films and cross-

linked gelatin films were loaded with bioactive agents such

as metronidazole for dental applications.11,12

Solution casting of polymers is a well-known method

for preparing polymer films. To incorporate drugs by this

method, the polymer is dissolved in a solvent and mixed

with the drug prior to casting. The solvent is then evapo-

rated and the polymer/drug film is created. We have previ-

ously reported a method for controlling drug location/

dispersion in the film.14,15 To review briefly, bioresorbable

polymeric films containing drugs were prepared using the

solution casting technique, accompanied by a postprepara-

tion isothermal heat treatment. In this process, the solvent

evaporation rate determines the kinetics of drug and poly-

mer solidification and thus, the drug dispersion/location

within the film. Solubility effects in the starting solution

also contribute to the postcasting diffusion processes and

occur concomitantly to the drying step. In general, two

Correspondence to: M. Zilberman (e-mail: [email protected])

' 2008 Wiley Periodicals, Inc.

155

types of polymer/drug film structures were created and

studied for all matrix polymer types, as presented in Figure

1 for PLLA and poly(DL-lactic acid) (PDLLA) films loaded

with DM:

a. A polymer film with large crystalline drug particles

located on its surface, as presented in the upper micro-

graphs of Figure 1. This structure, derived from a

dilute solution of both polymer and drug, was obtained

using the slow solvent evaporation rate, which enables

prior drug nucleation and growth on the polymer solu-

tion surface. This skin formation is accompanied by a

later polymer core formation/solidification. This struc-

ture was named the ‘‘A-type.’’

b. A polymer film with small drug particles and crystals

distributed within the bulk, as presented in the lower

micrographs of Figure 1. This structure, derived from a

concentrated solution, was obtained using the fast sol-

vent evaporation rate and resulted from drug nucleation

and segregation within a dense polymer solution. Solid-

ification of drug and polymer occurred concomitantly.

In semicrystalline-based films, such as PLLA/DM, the

drug is located in amorphous domains of a semicrystal-

line matrix, around the spherulites.14 This structure

was named the ‘‘B-type.’’

We have studied the morphology and formation process

of these structured films extensively, and a detailed model

describing the structuring of these films is described else-

where.14 Our study indicates that both A-type and B-type

structures can be developed for both amorphous and semi-

crystalline polymeric films loaded with DM and also with

other drugs, such as cortisone14 and gentamicin.2 The type

of polymer affects the film morphology but has almost no

effect on the drug location/dispersion within the polymeric

film. The latter is determined mainly by the kinetics of film

formation.2,14,15

Mathematical models for controlled release of bioactive

agents from bioresorbable matrices are based on diffusion

aspects, on the structural characteristics of the matrix poly-

mer and its degradation and swelling, and on the micro-

environmental pH changes inside the polymer matrix pores

that are due to the degradation products. Prediction of the

drug release profile using a model can obviously be very

useful in the device’s design phase, as the model enables

fast evaluation and tuning of the various parameters for

achieving an optimal release profile, while reducing labora-

tory tasks to a minimum.16,17

Early drug delivery models described either systems

based on nondegradable matrices or surface-eroding sys-

tems.18 Several more recent models described bulk-eroding

systems, in which the drug is physically immobilized. For

example, Siepman and Gopferich19 quantified drug release

from slab-shaped PLLA and poly(DL-lactic-co-glycolicacid) (PDLGA) matrices. Their model was based on Higu-

chi’s classical pseudo steady-state equation for oversatu-

rated, planar, nondegrading polymeric films, where the

permeability of the drug within the polymer matrix was

assumed to increase with time due to cleavage of the poly-

mer’s bonds. Another possibility for simulating the effect

of erosion on the diffusion process is to use a diffusion

coefficient which increases with time. Various theories,

therefore related the drug diffusion coefficient inside a

degradable polymer directly to its molecular weight, as

Figure 1. Polarized light micrographs of DM-loaded PLLA (semicrystalline) and PDLLA (amorphous)

films: upper micrographs, A-type; lower micrographs, B-type.

156 ZILBERMAN AND MALKA

Journal of Biomedical Materials Research Part B: Applied Biomaterials

short chains offer less restriction to drug diffusion than

long chains. This was one of the main assumptions in the

models of Charlier et al.20 for predicting the release rate of

Mifepristone from 50/50 PDLGA bulk-eroding films of var-

ious molecular weights, and Faisant et al.21 who examined

the release rate of 5-fluorouracil from 50/50 PDLGA

microspheres. Both used an additional assumption regard-

ing the polymer’s first order degradation kinetics in order

to calculate an appropriate time-dependent diffusion coeffi-

cient, which they used in Fick’s laws of diffusion. Zhang

et al.22 addressed three mechanisms for drug release in a

microspheric matrix, namely dissolution of the drug from

the polymer matrix, diffusion of the dissolved drug, and

erosion of the matrix.

However, our promising technique for controlling the

drug location/dispersion in the film raised the need for a

suitable model which could predict the release profile of var-

ious drugs from these structured films, and therefore, shorten

the design process of biomedical implants based on these

drug-eluting films. This need was the driving force for the

current research, in which a suitable mathematical model

was developed. The hypotheses of our study were: (1) A

model based on Fick’s laws will be able to provide good

prediction of the drug release profile from our structured

drug-eluting films, and (2) The drug release profile from the

films is affected by the host polymer’s characteristics (degra-

dation profile and degree of crystallinity), by the drug loca-

tion/dispersion in the film, and by the drug characteristics.

MATERIALS AND METHODS

Case Definition and Model Assumptions

A mathematical model for predicting drug release from our

structured drug-eluting films was developed in this study,

using Matlab 6.1. The release profiles that were predicted

using this model were compared with experimental results

for certain films loaded with DM.

The Experimental System. Polymer films (0.12–0.15

mm thick) consisting of poly(a-hydroxy acid) and DM (5%

w/w in each film) were prepared by a three-step solution

processing method as follows: The components were mixed

in a solvent at room temperature until polymer dissolution.

Both dilute and concentrated solutions were prepared for

the various polymer/DM systems. The solutions were then

cast into Petri dishes, and solvent drying was performed

under atmospheric pressure at room temperature. A slow

solvent evaporation rate was used for dilute solutions in

order to obtain A-type films, whereas a fast solvent evapo-

ration rate was used for concentrated solutions in order to

obtain B-type films. Finally, an isothermal heat treatment at

a temperature higher than the glass transition temperature

of the polymer (60–908C) was performed for 1 h in a

vacuum oven. This heat treatment enabled the disposal of

residual solvent. The process of film formation is described

in greater detail elsewhere.15

The films were immersed in sterile water at 378C for 20

weeks in order to determine the DM release kinetics. Sam-

ples of 1.5 mL were collected every week and were

replaced with sterile water. A schematic representation of

the experimental system is presented in Figure 2. The DM

content in each sample was measured using high perform-

ance liquid chromatography (HPLC), DX 500 (Dionex

Corp., Sunnyvale, CA). An AD-20 ultraviolet detector

(Dionex Corp., Sunnyvale, CA) was used to monitor ab-

sorbance at 254 nm. Five samples were examined for each

film type. The examined films are as follows:

Films based on amorphous host polymers:

1. 50/50 PDLGA, A and B types.

2. 85/15 PDLGA, A and B types.

Figure 2. The experimental system at the initial point: A-type film and B-type film floated on water

in a Petri dish. The diffusion occurs in the Z direction. [Color figure can be viewed in the online

issue, which is available at www.interscience.wiley.com.]

157DRUG CONTROLLED RELEASE FROM STRUCTURED BIORESORBABLE FILMS


Films based on semicrystalline host polymers:

1. PLLA, A and B types.

2. Polydioxanone (PDS), B type.

3. 10/90 PDLGA, B type.

The weight loss profile of the various host polymers and

their degree of crystallinity versus time are presented in

Figure 3(a,b, respectively). They were taken from one of

our recent publications.15

Assumptions Used When Deriving the Model’s Equa-

tions.

1. The drug’s diffusion coefficient within the polymeric

film is only time-dependent, and the basic mathemati-

cal equation which describes the diffusion behavior is

Fick’s 2nd law (partial differential equation). Also,

there is no drug concentration profile in the aqueous

medium, and there is homogenic distribution of drug

on the surface of the film.

2. The diffusion coefficient (D) inside the film is affected

by the bulk erosion of the host polymer and its degree

of crystallinity (which also changes with time). The

weight loss of the polymeric film increases with time,

resulting in increasing film porosity and D. In contra-

distinction, the increase in the polymer’s degree of

crystallinity with degradation15 results in a decrease

in D.3. The polymeric chains create a porous film with a

homogenous structure. The film’s pore size increases

with weight loss, but the structure is homogenous.

4. The size and shape of the drug particles do not change

with time.

5. During the advanced stage of polymer degradation,

monomers, and oligomers diffuse out. This diffusion

occurs in parallel to the drug diffusion. No reciprocal

effect exists between the diffusion of the drug particles

and the diffusion of the monomers.

6. There is drug flux only toward the film’s interface with

water (z direction).7. The external environment provides perfect sink condi-

tions for the released drug.

8. The drug particles within the film are released only

when they find a continuous path to the surface and

their release occurs solely by diffusing through the

water.

9. The drug particles diffuse out only through amorphous

regions and not through dense crystalline regions.

Model Equations

According to assumption (1), the mathematical equation

which describes the diffusion behavior in our system is

Fick’s 2nd law of diffusion. The basic form of the equation

is:

r � ðDrCÞ ¼ @C

@tð1Þ

where D is the diffusion coefficient (cm2/s), ! is the Lap-

lace operator, C is the drug concentration (mole/cm3) and tis time (s).

Using assumption 1 we can simplify Eq. (1) to:

D � r2C ¼ @C

@tð2Þ

Using assumption 6 and the Laplace operator, the final

diffusion equation is:

D@2C

@Z2¼ @C

@tð3Þ

Appropriate initial and boundary conditions should be

used in order to solve the diffusion Eq. (3). Assuming an

initial uniform drug concentration (C0) in the bulk of the

Figure 3. The time-dependent parameters which affect the diffusion

coefficient of the drug in the film: (a) The polymer’s weight loss,

(b) The polymer’s degree of crystallinity. The host polymer type is

indicated.



polymeric film or on its surface leads to the following ini-

tial condition:

C ¼ C0 at t ¼ 0; Z ¼ Zfilm for A-type films ð4aÞC ¼ C0 at t ¼ 0; 0 � Z < Zfilm for B-type films ð4bÞ

Second initial condition: there is no drug in the medium

(water) at the starting point:

C ¼ 0 at t ¼ 0; Zfilm < Z � L ð5Þ

This is a closed system, and therefore, drug particles

cannot leave the system. Hence, the flux at the boundaries

is equal to 0, and the boundary conditions are:

@C

@Zðt; Z ¼ 0Þ ¼ 0 ð6Þ

@C

@Zðt; Z ¼ LÞ ¼ 0 ð7Þ

Using the above initial and boundary conditions, the

differential Eq. (3) was resolved via the ‘‘pdepe’’ Matlab

function, yielding C(Z,t) which describes how the drug con-

centration changes with time, and how it changes along the

Z-axis.The released drug particles reach the surface of the film

(Z 5 Zfilm), and the cumulative drug particles can therefore

be ‘‘counted’’ as follows:

mðtÞ ¼Z t

0

d

dt0CðZ ¼ Zfilm; t

0ÞAdt0 ð8Þ

where m(t) is the cumulative drug released (moles), t, t0—Time (s), CðZ ¼ Zfilm; t

0Þ—area concentration of drug

(mole/cm2), and A is the film area (cm2).

Estimation of the Diffusion Coefficient. The diffusion

coefficient (D) in our model is divided into two regions,

corresponding to the Z-axis division. One region is inside

the film at 0 � Z � Zfilm, where the diffusion coefficient is

time depended during the film degradation and can be writ-

ten as Df(t). The second region is in the water at Zfilm \ Z� L, where the diffusion coefficient is constant and can be

written as Dw. Hence, D is expressed as follows:

D ¼ DfðtÞ at 0 � Z � Zfilm ð9Þ

D ¼ Dw at Zfilm < Z � L ð10Þ

Because the diffusion medium in both cases is water,

the following theoretical equation for diffusion in water

can be used23:

Dw ¼ KBT

2lR0

ð11Þ

where Dw is the theoretical diffusion coefficient of a spher-

ical particle in water (cm2/s), KB is Boltzmann’s constant

5 1.38 3 10216 (g cm2/s2 K), T is the absolute tempera-

ture (3108K 5 378C), l is the water’s viscosity 5 0.01 (g/

cm s), and R0 is the radius of a spherical particle (cm).

In our study, R0 was estimated using light and electron

microscopy.

In our model we had to slightly change Eq. (11)

because of physical restriction on the drug particles’

motion. In A-type films the drug particles sense mainly

the water and not the film in the surrounding and the non-

ideal (nonspherical) shape of the drug particles is there-

fore the dominant restriction. Equation (13) actually

shows how far are the particles from spherical shape. In

B-type films, the drug particles sense mainly the polymer

chains of the film, which create paths through which the

drug particles move. In B-type films, Eq. (14) elucidates

the ability of the drug particles to pass through the porous

structure.

A restriction factor (Fr) was therefore added to Eq. (14)

as follows:

Dw ¼ KBT

2lR0Fr

ð12Þ

where:

Fr ¼ ðsurface area of particleÞ=ðsurface area of equivalentsphereÞ in A-type films

ð13ÞFr ¼ average pore size=R0 in B-type films ð14Þ

The initial diffusion coefficient inside the film (Df0) is

based on Eq. (12), but we should consider the effects of

the film characteristics as follows:

Df0 ¼ DwP0FsFt ð15Þ

where Df0 is the initial diffusion coefficient inside the film

(cm2/s), and P0 is the initial porosity of the film (%). P0

indicates the volume of water inside the film, which was

evaluated experimentally and used in our model. Fs is a

surface factor ([1) that we used in A-type films. It gives

an indication for the initial binding (physical interactions

such as hydrogen bonds) between the drug and the film sur-

face. The value of Fs is higher for smaller (weaker) interac-

tions. When there are relatively weak interactions between

the host polymer and the drug molecules (in A-type films),

the diffusion coefficient is closer to that of the drug in water.

Ft is a tortuosity factor (\1) used for B-type films. It indi-

cates the initial tortuosity of the film, which can indicate the

ratio of shortest and the actual path of particles. Fs and Ft

values were evaluated using our experimental data.15

Ft ¼ L1=L2 ð16Þ



where L1 is the distance between particle location and film

surface (Z 5 Zfilm), and L2 is the length of the actual parti-

cle path inside the film. The factors Fr, Fs, and Ft for the

various films used in this study are presented in Table I.

The time-dependent diffusion coefficient in the bulk

[Df(t)] starts from its initial value Df0 [calculated in Eq.

(15)] and changes with time due to the polymer’s weight

loss (bulk erosion) and changes in its degree of crystallin-

ity, as stated in assumption 2. Df(t) is evaluated as

follows:

DfðtÞ ¼ Df0 þ fðDw � Df0Þ3WðtÞ3ð100%� CRSTLðtÞÞgð17Þ

where W(t) is the film’s weight loss (%), and CRSTL(t) isthe polymer’s degree of crystallinity (%). Both parameters

were measured for all polymers used in this study, and the

results are presented in Figure 3. Because diffusion of drug

particles occurs through the amorphous regions of the semi-

crystalline polymer rather than through the dense crystal-

line regions (assumption 9), we used [100% 2 CRSTL(t)]in Eq. (17).

The degree of crystallinity was measured as follows: the

heat of fusion (DHm) was determined by differential scan-

ning calorimetry using an indium-calibrated TA Instru-

ments DSC 2010 differential scanning calorimeter (DSC).

The measurements were carried out on 10 mg samples

under N2 atmosphere, heating the samples from 308C to

2508C (above their melting points), using a heating rate of

108C/min. The analysis was performed using TA Universal

TABLE I. Film Types Used for Obtaining the Experimental Dataand Their Calculated Factors Used in the Model

The Polymer Film Type Fr Fs Ft

50/50 PDLGA A 2 30 –

B 40 – 0.33

85/15 PDLGA A 9 80 –

B 30 – 0.67

PDS B 20 – 0.67

10/90 PDLGA B 50 – 0.67

PLLA A 7 50 –

B 20 – 0.33

Figure 4. DM release profile from amorphous polymer films containing 5% (w/w) DM. The experi-

mental results (mean —— and upper and lower limits -------) are compared with the predicted

results (—l—): (a) A-type (85/15 PDLGA)/DM, (b) B-type (85/15 PDLGA)/DM, (c) A-type (50/50

PDLGA)/DM, (d) B-type (50/50 PDLGA)/DM. [Color figure can be viewed in the online issue, which isavailable at www.interscience.wiley.com.]



Analysis Software. The degree of crystallinity, CRST(t),was calculated by the relationship:

%CRST ¼DHm

DHF

3100 ð18Þ

where DHm and DHF are the heats of fusion of the sample

(a semicrystalline material) and the perfect crystal, respec-

tively. DHF (PLLA) 5 93.6 J/g,24 DHF (PGA) 5 191.2 J/g

and DHF (PDS) 5 141.2 J/g.25

Mean errors for the predicted release profiles were eval-

uated as follows: for each data point, when the model’s

prediction is within the error of the experimental results,

the model’s error is considered as 0. When the model’s

prediction is not in the range of the experimental’s error,

there appears an error (in %) for this point. The average

error is the average of the various errors of all points.

RESULTS AND DISCUSSION

As mentioned earlier, the experimental data used to vali-

date this model were taken from one of our recent stud-

ies.15 Two types of films were used, A-type with drug

located on the surface of the polymer film and B-type with

drug located in the bulk. Amorphous and semicrystalline

polymers were used for both types of drugs (Table I). The

predicted DM release profile was compared with the exper-

imental release profile for each type of structured film. The

results for the amorphous based films are presented in Fig-

ure 4 and the results for the semicrystalline-based films are

presented in Figure 5. A very good fit was obtained for all

studied films (less than 7% error), except for the 10/90

PDLGA/DM based film, where the predicted DM release

during the first 7 weeks was higher than the experimental

DM release [Figure 5(b)]. In A-type films the release pro-

file exhibited a burst release of �30% followed by a

release profile, which was determined by the degradation

rate of the host polymer, whereas the B-type films exhib-

ited a release profile governed by the degradation rate of

the host polymer with a low burst release. Our model dem-

onstrated a good fit for both types of films. These results

thus support our first hypothesis about the good predictabil-

ity of a model based on Fick’s laws.

The advantage of building a model for drug delivery

systems is the ability to elucidate the effect of the system’s

Figure 5. DM release profile from semicrystalline polymer films containing 5% (w/w) DM. The ex-

perimental results (mean —— and upper and lower limits -------) are compared with the predicted

results (—l—): (a) B-type PDS / DM, (b) B-type (10/90 PDLGA)/DM, (c) A-type PLLA/DM, (d) B-

type PLLA/DM. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]



parameters (each one separately) on the release profile of

the bioactive agent. In this regard, the effect of the host

polymer’s degradation rate and degree of crystallinity (for

semicrystalline polymets) were studied. The effect of the

degradation rate on the release profile was examined using

the measured weight loss profiles of various host polymers

[Figure 3(a)]. The results for the 85/15 PDLGA based films

are presented as an example, in Figure 6. After the burst

release, if degradation of the host polymer is not consid-

ered, the predicted amount of drug released is much lower

than the experimental release or the predicted release when

using the full model. These results are reasonable, because

degradation of the host polymer results in higher porosity,

which enables a faster drug release rate. The release pro-

files from the B-type films are affected by degradation

more than the release profiles from the A-type films (Fig-

ures 4 and 5), because most drug particles are located

within the bulk of the former. Indeed, our model shows

that the difference between the experimental results and the

predicted results (when degradation is neglected) is higher

for B-type films [Figure 6(b)] than for A-type films [Figure

6(a)].

The degree of crystallinity of the studied semicrystal-

line polymers as a function of degradation time is pre-

sented in Figure 3(b). During the first 12 weeks of

degradation, the degree of crystallinity increased signifi-

cantly as follows: for PDS from 53 to 83%, for 10/90

PDLGA from 43 to 59%, and for PLLA from 53 to 69%.

It is widely accepted that poly(a-hydroxy acids) are

degraded by simple hydrolysis. Hydrolysis begins in the

amorphous phase of the polymer, because of the relatively

easy penetration of water into these domains.26 Thus, the

overall degree of crystallinity of the polymer film

increases during degradation, mainly due to the erosion of

amorphous domains. The increase in the degree of crystal-

linity with time may also occur due to further crystalliza-

tion of low molecular weight chains. It should be noted

that water acts as a softener, giving rise to that additional

crystallization.

The effect of the host polymer’s degree of crystalliniy

on the release profile of DM from B-type PDS film is pre-

sented in Figure 7 and is compared with the effect of deg-

radation. When the polymer’s degree of crystallinity and its

change with degradation time is not considered, the pre-

dicted released quantity of drug is higher than the experi-

mental release or the predicted release when using the full

model. These results are reasonable, because the polymeric

crystalline domains are dense and do not allow diffusion of

drug through them. Therefore, an increase in the degree of

crystallinity with degradation time acts as an obstacle for

drug release, which slows the rate of diffusion. As men-

tioned earlier, the polymer’s degree of crystallinity is

affected by its degradation, and it is therefore not possible

Figure 6. DM release profile from 85/15 PDLGA films containing5% (w/w) DM, showing the effect of the host polymer’s weight loss

during drug release. The experimental results (mean —— and upper

and lower limits -------) are compared with the predicted results

(—l—) and with the predicted results when neglecting weight loss(—~—): (a) A-type films, (b) B-type films. [Color figure can be

viewed in the online issue, which is available at www.interscience.

wiley.com.]

Figure 7. DM release profile from PDS films containing 5% (w/w)

DM, showing the effects of the host polymer’s weight loss and

degree of crystallinity and its change during drug release. The ex-perimental results (mean —— and upper and lower limits -------) are

compared with the predicted results ( ) and with the predicted

results when neglecting crystallinity ( ) and when neglecting

weight loss ( ). [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.]



to examine the effect of each of these parameters on the

drug release profile separately. Our model indicates that the

effect of degradation on the drug release profile is more

important than the effect of the degree of crystallinity

(Figure 7). These results support our second hypothesis that

the drug release profile from the films is affected by the

host polymer’s degradation profile and its degree of crys-

tallinity and also elucidates the relative contribution of

each.

SUMMARY AND CONCLUSIONS

The aim of this study was to develop a mathematical

model for predicting drug release profiles from struc-

tured bioresorbable films designed to be used in vari-

ous biomedical applications. Our structured polymer/

drug films are prepared using a promising technique

for controlling the drug location/dispersion within the

film and can be loaded with either water-soluble or

water-insoluble bioactive agents. Use of the suggested

model affords a good and rapid evaluation of the

release profile, enabling further economical in vitro/invivo release studies.

The model is based on Fick’s 2nd law of diffusion and

assumes that the drug release profile from the films is

affected by the host polymer’s characteristics, by the drug’s

location/dispersion in the film and by the drug’s character-

istics. The model uses the weight loss profile of the host

polymers as well as their change in degree of crystallinity.

It also uses empirical parameters such as restriction factor,

porosity, tortuosity factor, and a surface factor, which gives

an indication for possible physical binding between drug

molecules and the host polymer.

In this study, the model was used for predicting DM

release profiles from structured films of both types, with

the drug located on the surface of the film (A-type), and

with the drug located in the bulk (B-type). We demon-

strated that the model correlates well with in vitro release

results, exhibiting a mean error of less than 7% for most

studied cases. Furthermore, two types of host bioresorbable

polymers were used, amorphous, and semicrystalline, and

the model demonstrated that the host polymer’s degradation

has a greater effect on the drug release profile than the

degree of crystallinity. This new model exhibits a potential

for simulating the release profile of bioactive agents from

structured films for a wide variety of biomedical applica-

tions.

REFERENCES

1. Zilberman M, Eberhart RC, Schwade ND. In-vitro study ofdrug loaded bioresorbable films and support structures. J Bio-mater Sci Polym Ed 2002;13:1221–1240.

2. Aviv M, Berdicevsky I, Zilberman M. Gentamicin-loaded bio-resorbable films for prevention of bacterial infections associ-

ated with orthopedic implants. J Biomed Mater Res Part A2007;83:10–19.

3. Park YJ, Lee YM, Park SN, Lee JY, Ku Y, Chung CP, LeeSJ. Enhanced guided bone regeneration by controlled tetra-cycline release from poly(L-lactide) barrier membranes. JBiomed Mater Res 2000;51:391–397.

4. Lee KB, Kang SB, Kwon IC, Kim YH, Choi K, Jeong SY.Controlled release and in vitro antimicrobial activity of genta-micin from poly-a-hydroxy acid films. Proceedings of the24th International Symposium on Controlled Release of Bio-active Materials 1997. pp 571–572.

5. Mauduit J, Bulkh N, Vert M. Gentamicin/poly(lactic acid)blends aimed at sustained release local antibiotic therapyadministrated per-operatively. III The case of gentamicinsulfate in films prepared from high and low molecularweight poly(DL-lactic acids). J Controlled Release 1993;25:43–49.

6. Shah SS, Cha Y, Pitt CG. Poly(glycolic acid-co-DL-lacticacid): Diffusion or degradation controlled drug delivery?J Controlled Release 1992;18:261–270.

7. Wang X, Venkatraman SS, Boey FYC, Loo JSC, Tan LP.Controlled release of sirolimus from a multilayered PLGAstent matrix. Biomaterials 2006;27:5588–5595.

8. Dorta MJ, Santovena A, Liabres M, Farina JB. Potentialapplications of PLGA film-implants in modulating in-vitrodrugs release. Int J Pharm 2002;248:149–156.

9. Livnat M, Beyar R, Seliktar D. Endoluminal hydrogel filmsmade of alginate and polyethylene glycol: Physical characteri-atics and drug-eluting properties. J Biomed Mater Res A2005;75A:710–722.

10. Westedt U, Wittmar M, Hellwig M, Hanefeld P, Greiner A,Schaper AK, Kissel T. Paclitaxel releasing films consisting ofpoly(vinyl alcohol)-graft-poly(lactide-co-glycolide) and theirpotential as biodegradable stent coatings. J Controlled Release2006;111:235–246.

11. Cetin EO, Buduneli N, Athhan E, Kurilmaz L. In-vitro studiesof a degradable device for controlled-release of meloxicam.J Clin Periodontol 2005;32:773–777.

12. Quaglia F, Vignole MC, De Rosa G, La Rotonda MI, MaglioG, Palumbo R. New segmented copolymers containing poly(e-caprolactone) and etheramide segments for the controlledrelease of bioactive compounds. J Controlled Release 2002;83:263–271.

13. Tarvainen T, Karjalainen T, Malin M, Perakorpi K, TuominenJ, Seppala J, Jarvinen K. Drug release profiles from and deg-radation of a novel biodegradable polymer, 2,2-bis(2-oxazo-line) linked poly(e-caprolactone). Eur J Pharm Sci 2002;16:323–331.

14. Zilberman M, Schwade N, Meidell R, Eberhart R. Structureddrug loaded bioresorbable films for support structures. J Bio-mater Sci: Polym Ed 2001;12:875–892.

15. Zilberman M. Dexamethasone loaded bioresorbable filmsused in medical support devices: Structure, degradation,crystallinity and drug release. Acta Biomater 2005;1:615–625.

16. Narasimhan B, Peppas A. The role of modeling studies in thedevelopment of future controlled-release devices. In: Park K,editor. Controlled Drug Delivery: Challenges and Strategies.Washington, DC: American Chemical Society; 1997. pp 529–558.

17. Tongwen X, Binglin H. Mechanism of sustained drug releasein diffusion-controlled polymer matrix-application of percola-tion theory. Int J Pharm 1998;170:139–149.

18. Gopferich A. Mechanisms of polymer degradation and ero-sion. Biomaterials 1996;17:103–114.

19. Siepmann J, Gopferich A. Mathematical modeling of bioerod-ible polymeric drug delivery systems. Adv Drug Delivery Rev2001;48:229–247.



20. Charlier A, Leclerc B, Couarraze G. Release of mifepristonefrom biodegradable matrices: Experimental and theoreticalevaluations. Int J Pharm 2000;200:115–120.

21. Faisant N, Siepmann J, Benoit JP. PLGA-based microparticles:Education of mechanisms and a new, simple mathematicalmodel quantifying drug release. Eur J Pharm Sci 2002;15:355–366.

22. Zhang M, Yang Z, Chow L, Wang C. Simulation of drugrelease from biodegradable polymeric microspheres withbulk and surface erosions. J Pharm Sci 2003;92:2040–2056.

23. Cussler EL. Diffusion Mass Transfer in Fluid Systems, 2ndEd. Cambridge, United Kingdom: Cambridge UniversityPress; 1977.

24. Celli MA, Scandola M. Thermal properties and physical agingof Poly-L-lactic acid. Polymer 1992;33:2699–2710.

25. Ishikiriyama K, Pyda M, Zhang G, Forschner T, Grebowicz J,Wunderlich B. Heat capacity of poly-p-dioxanone. J Macro-mol Sci Phys Ed 1998;B37:27–44.

26. Pistner H, Bendix DR, Muhling J, Reuther JF. Poly(L-lactide):A long-term degradation study in vivo. III. Analytical Charac-terization. Biomaterials 1993;14:291–298.



Documents

Drug controlled release from structured bioresorbable films used in medical devices—A mathematical model