Permeation of Drugs in Chitosan Membranes

A. N. L. ROCHA, T. N. C. DANTAS, J. L. C. FONSECA, M. R. PEREIRA

Departamento de Quı́mica, Universidade Federal do Rio Grande do Norte, Campus Universitário, Lagoa Nova, Natal,
RN 59078-970, Brazil

Received 26 March 2001; accepted 4 June 2001

Published online 24 January 2002

ABSTRACT: The permeabilities of isoniazid and amitriptyline hydrochloride in chi-

tosan membranes were investigated. Drug concentration was changed from 0.1 to
1.0% while membrane thickness was varied from 40 to 150 ␮m. The release rate was
measured in water at 30 ⫾ 0.1°C by spectrophotometric determination. The drugs
presented quite different permeabilities, which were related to their molecular
weights; the permeabilities did not change with thickness or drug concentration for the
ranges investigated. © 2002 John Wiley & Sons, Inc. J Appl Polym Sci 84: 44 – 49, 2002; DOI
10.1002/app.10185

Key words: chitosan membranes; isoniazid; amitriptyline hydrochloride; permeabil-

ity; membranes; drug delivery systems; UV–vis spectroscopy

INTRODUCTION usefulness of polymeric materials arises from the

tremendously wide variations that can be ob-
Polymer permeability is the basis for a number of tained in their properties through the variation of
important applications in various areas of phar- the nature and/or concentration of comonomers,
maceutical interest, such as tablet coatings,1 hae- crosslinkers, and plasticizers.
modialysis, and wound dressing.2 In addition to Various types of polymeric membranes may be
these applications, a considerable amount of re- used in this field. In general, they can be classi-
search is concerned with the use of polymers as fied according to the release mechanism as (1)
agents for the controlled release of drugs from hydrophobic, nonporous membranes, (2) micro-
various types of formulated products, for exam- porous membranes, and (3) water-swollen, hydro-
ple, tablets, implants, and adhesives strips. Evi- philic membranes (hydrogels). For the first type,
dence of the high degree of interest in the design the release process involves the consecutive pro-
of such dosage forms is provided by the number of cess of drug partition between the core formula-
reviews3–5 and books6 –9 that have been concerned
tion and the membrane, diffusion of the drug in
with these subjects.
solution across the latter, and subsequent parti-
A controlled-release dosage form consists es-
tion of the drug into an aqueous environment. For
sentially of a drug-containing device that permits
the second, it involves transfer of the dissolved
the release of the drug at a predetermined rate
drug through discrete, liquid-filled pores. Parti-
when the dosage form is placed in the body. The
tion of the drug between the core formulation and
the liquid in the pores of the membrane must
Correspondence to: M. R. Pereira (pereira@linus.quimica. occur before membrane transport can proceed.
ufrn.br). Finally, for the third type, permeability is
Contract grant sponsors: CNPq; CAPES; PPPg-UFRN.
Journal of Applied Polymer Science, Vol. 84, 44 – 49 (2002)
strongly affected by the solubility of the diffusant
© 2002 John Wiley & Sons, Inc. in the aqueous continuous phase of these systems.

44
 PERMEATION OF DRUGS IN CHITOSAN MEMBRANES 45

Figure 1 Absorbance as a function of time for permeation experiments with isonia-

zide solutions at different concentrations: 0.1% (squares), 0.3% (circles), 0.5% (up
triangles), 0.7% (down triangles), and 1.0% (diamonds). L ⫽ 35 ⫾ 3 ␮m.

These membranes can be considered intermedi- EXPERIMENTAL

ate between porous and nonporous ones.
Much of the previous work on controlled-re- Materials
lease drug-delivery systems has used polydimeth-
ylsiloxane10 because of its biocompatibility and Chitosan was supplied by Polymer Ltd. (For-
taleza, Brazil). Its deacetytilation degree was
high permeability to hydrophobic drugs or poly-
about 80 mol %. The solutes used for the trans-
urethanes10 due to the possibility of different co-
port experiment were amitriptyline hydrochloride
polyether– urethane combinations.
[weight-average molecular weight (Mw ⫽ 313.57,
In this study, the potential use of chitosan in
pH ⫽ 5.5– 6.5 in aqueous solution, ␭ ⫽ 263 nm]
controlled-release drug-delivery devices was in- and isoniazid (M w ⫽ 137, pH ⫽ 5.0 – 6.0 in aque-
vestigated. Chitosan [␤-(1– 4)-2-amino-2-deoxy-D- ous solution, ␭ ⫽ 239 nm). These were purchased
glucose], derived from chitin by deacetylation, is a from Sigma (St. Louis, MO). We chose these
natural polycationic polymer that possesses valu- drugs, taking into account their molecular
able properties as a biomaterial for biomedical weights, water solubility, pH, and suitability in
applications.11 The film-forming property of chi- ultraviolet (UV) absorption.
tosan offers many applications for various mem-
brane separations fields. The selectivity of the
membrane is a critical parameter in membrane Membrane Preparation
separation, being affected by factors such as We prepared the chitosan solution by dissolving
membrane molecular weight,12 chain flexibility,13 chitosan in an aqueous 2% acetic acid solution at
thickness,14 and preparation conditions,15 which ambient temperature with stirring overnight.
have been investigated. To obtain insight into the The concentration of chitosan in the acid solution
process of molecular separation, we prepared chi- was changed according to the desired membrane
tosan membranes of different thicknesses and thickness. The solution was filtered with a G4
used two drugs of different molecular weights. glass filter and allowed to stand for about half a
46 ROCHA ET AL.

Figure 2 Absorbance as a function of time for the permeation of amitriptyline at

different solution concentrations: 0.1% (squares), 0.3% (circles), 0.5% (up triangles),
0.7% (down triangles), and 1.0% (diamonds). L ⫽ 35 ⫾ 3 ␮m.

day to remove air bubbles. The solution was then trophotometer model, and then it was returned to
cast onto a glass plate and placed in a drying air the receiving solution. All the experiments were
oven at 50°C for 24 h. The dry film was immersed done in duplicate.
in a 5% aqueous solution of NaOH for 2 h. The Permeability (P) values were determined with
chitosan membrane was repeatedly washed with the model proposed by Crank16 for flow through a
water and placed on a extensor for drying. The membrane described as following: one face of the
dry membrane thickness was measured with a membrane ( x ⫽ 0) was kept at constant concen-
digital micrometer Check-line (model DCF-900). tration c 1 , the other ( x ⫽ L) was kept at concen-
tration c 2 , and the membrane was initially at a
Permeation Experiments uniform concentration c 0 .
According to our experimental arrangement,
The diffusion cell consisted of two cylindrical half-
we could assume that the membrane was initially
cells 230 cm3 in volume. The chitosan membrane
at zero concentration (c 0 ⫽ 0) and that the con-
was placed between the compartments and was
centration at one face was much higher than the
not supported. The membrane area was 8.55 cm2.
concentration on the emerging face (i.e., c 1 Ⰷ c 2 )
Each compartment was stirred continuously by
and, therefore, c 1 ⫺ c 2 ⬵ c 1 . In this case, the total
externally mounted constant-speed synchronous
amount of diffusing substance (Q t ), which has
motors. The diffusion cell was placed in a water
passed through the membrane in time t is given
bath maintained at 30 ⫾ 0.1°C. The chitosan
by
membranes were initially immersed in water for
12 h before the experiment.
The feed solution was prepared by dissolution
of the solute in water at different concentrations.
Qt ⫽
Dc 1
L 冉t⫺
L2
6D 冊 (1)

The receiving solution was distilled water. Sam- where D is the diffusion coefficient and L is the
ples of 2.5 cm3 of the receiving solution were membrane thickness.
taken at various time intervals, and the solute Because the amount of drug released was mea-
concentration was analyzed by a Varian UV spec- sured spectroscopically, Q t is given by:
 PERMEATION OF DRUGS IN CHITOSAN MEMBRANES 47

where K is the partition coefficient given by

c1 c2
K⫽ ⫽ (5)
C1 C2

where C 1 and C 2 are the concentrations on each

side of the cell. When eq. (5) is substituted in eq.
(4) and then in eq. (3) it follows that

PC 1␧bS c 1L␧bS
A共t兲 ⫽ t⫺ (6)
VL 6V

from which P can be calculated with the slope of

the curve of A(t) against t in the steady state.

RESULTS AND DISCUSSION

The results from the permeation experiments

with isoniazide and amitriptyline, L ⫽ 35 ⫾ 3
␮m, are plotted in Figures 1 and 2, respectively.
Apparently, there was no occurrence of time lag.
The time lag is usually related to a build-up pe-
riod necessary to establish an equilibrium at the
interface between the more concentrated side
(C 1 ) and the membrane. The absence of a time lag
indicates, therefore, that for these experiments,
the equilibrium seemed to be instantaneously es-
tablished. This behavior can be understood if we
Figure 3 Permeability as a function of concentration
remember that the membranes used were already
for (■) isoniazide and (䊐) amitriptyline.
swollen with water because they were immersed
in it for 12 h before the experiments. The swollen
membrane had a high water content, which facil-
VA itated the permeation of water-soluble solutes
Qt ⫽ (2) like isoniazide and amitriptyline. The absence of
␧bS
a time lag indicates, as a consequence, that either
the mechanism involved in the permeability was
where V is the half cell volume, A is the absor- related to a microporous membrane or the combi-
bance, S is the membrane area, b is the cell path nation of thickness (L) and diffusion coefficient
length, and ␧ is the extinction coefficient. Substi- (D) led to values of L 2 /6D that were negligible
tuting eq. (2) in eq. (1), we get relative to the time scale of the experiment.
Although drug flux increased with solution
Dc 1␧bS c 1L␧bS concentration in both cases, as indicated by the
A共t兲 ⫽ t⫺ (3) increasing slopes of the resultant curves, perme-
VL 6V
ability did not change as the concentration in-
creased, as shown in Figure 3. From this, one can
Because we do not have the c 1 value, the dif- confirm that the drugs were liberated at a con-
fusion coefficient cannot be calculated. Instead, stant rate for periods of time up to 6 h. This also
we can determine the permeability coefficient, gives us two indications: (1) the diffusion process
which can be related to the diffusion coefficient by can be said to be Fickian, and (2) the partition
coefficient between water and chitosan was inde-
P ⫽ KD (4) pendent of concentration for this range of concen-
48 ROCHA ET AL.

Figure 4 Absorbance as a function of time for isoniazide, with different chitosan film
thicknesses: (Œ) 149 ␮m, (F) 129 ␮m, (■) 94 ␮m, and () 34 ␮m.

trations. Seo et al.17 reported the dependence of in permeability shows that chitosan membranes
permeant concentration and permeability for chi- could be quite specific in terms of these two drugs.
tosan membranes. According to the authors, the If this selectivity is defined as the ratio between
permeation should increase with increases in C 1 the permeability to isoniazide and the permeabil-
until membrane saturation is reached (c 1 ). In our ity to amitriptyline, the specificity is around 3
experiment, this increase in permeability was not ⫻ 102, which indicates that the membrane was
observed. Therefore, we can conclude that the ini- very selective in relation to amitriptyline.15
tial concentration used (0.1%) was already above
the saturation concentration of our membranes.
The Fickian character of the diffusion is also CONCLUSIONS
depicted by Figure 4, which shows the absorbance
in the less concentrated compartment as a func- The diffusion of isoniazide and amitriptyline in
tion of time for membranes with different thick- chitosan membranes obeyed all the ideal equa-
nesses, as well as Figure 5, which shows the cal- tions for permeation used in the modeling, which
culated values of permeability for the same mem- were based on a solution-diffusion mechanism.
branes. Permeability was not a function of The absence of a time lag for swollen membranes
membrane thickness, as expected for the case of was probably due to the small membrane thick-
Fickian diffusion. The permeability of isoniazide ness (L) used that led to values of L 2 /6D that
was also much higher than that of amitriptyline, were negligible relative to the time scale of the
which could be explained through the analysis of experiment (t). Both drugs were liberated at a
their molecular dimensions, which are directly constant ratio. Their permeabilities were strongly
related to their molecular weights. It is clear that dependent on their molecular dimensions: isonia-
the permeation of amitripytiline did not occur in zide (the compound with the lower molecular
the same extent as in the case of isoniazide, which weight) had a much higher value of permeability.
is hydrodynamically explained by larger Stokes’s The membranes, therefore, were highly selective
radius in the case of amitriptyline, the compound in relation to these two drugs. From these two
with highest molecular weight.17 This difference properties, a constant ratio and a high selectivity,
 PERMEATION OF DRUGS IN CHITOSAN MEMBRANES 49

REFERENCES

1. Lehmann, K.; Dreha, D. Int J Pharm Tech Prod

Manuf 1981, 2, 31.
2. Wendt, R. P.; Klein, E.; Bresler, E. H.; Holland,
F. F.; Serino, R. M.; Villa, H. J Membr Sci 1979, 5,
23.
3. Davis, S. S. Pharm Int 1981, 2, 41.
4. Langer, R. S.; Hsieh, D. S. T. Org Coat Plant Chem
1981, 44, 63.
5. Langer, R. S.; Peppas, N. A. Biomaterials 1981, 2,
201.
6. Sustained and Controlled Release Drug Delivery
Systems; Robinson, J. R., Ed.; Marcel Dekker: New
York, 1978.
7. Drug Delivery Systems; Juliano, R. L., Ed.; Oxford
University Press: Oxford, England 1980.
8. Controlled Release Technologies: Methods, Theory
and Applications; Kydonieus, A. F., Ed.; CRC: Boca
Raton, FL, 1980; Vols. I and II.
9. Controlled Release of Bioactive Materials; Baker,
R., Ed.; Academic: New York, 1980.
10. Zentner, G. M.; Cardinal, J. R.; Kim, S. W. J Pharm
Sci 1978, 67, 10.
11. Chandy, T.; Sharma, C. Biomat Art Cells Artif Org
1990, 1, 18.
12. Chen, H.-D. H. R. H. Carbohydr Polym 1996, 29,
353.
13. Chen, W.-C. L. R. H.; Tsaih, M. L. Carbohydr
Figure 5 Permeability as a function of film thickness Polym 1996, 31, 141.
for isoniazide. 14. Singh, A. R. R. D. K. J Membr Sci 1999, 155,
107.
15. Matsuyama, H.; Shiraishi, H.; Kitamura, Y. J Appl
we conclude that chitosan membranes can poten- Polym Sci 1999, 73, 2715.
tially be used in a controlled-release system. 16. Crank, J. The Mathematics of Diffusion; Oxford
University Press: New York, 1985.
The authors thank Professor Francisco G. de Azevedo 17. Seo, T.; Ohtake, H.; Unishi, T.; Ijima, T. J Appl
for his support and encouragement. Polym Sci 1995, 58, 633.