- Email: [email protected]
Drug diffusion front movement is important in drug release control from swellable matrix tablets
Drug Diffusion Front Movement Is Important in Drug Release Control from Swellable Matrix Tablets PAOLOCOLOMBO+, RUGGEROBETTINI~~, GINAMASSIMO~, PIER LUIGI CATELLANI~, PATRIZIASANTI~,AND NIKOLAOS A. PEP PAS^^ Received November 4, 1994, from the fpharmaceutical Deparfment, University of Parma, Via delle Scienze, 43100 Parma, h / y , and
#School of Chemical Engineering, 1283 CHME Building, Purdue University, West Lafayette, IN 47907- 1283. publication May 16, 1995@. Abstract 0 Swellable controlled release devices of buflomedil pyridoxalphosphate in hydroxypropyl methylcellulose were prepared, and their swelling and release behavior was investigated. The drug release as a function of time was investigated for various system parameters. Three distinct fronts were observed during the swelling and release processes, i.e., a swelling, a drug diffusion, and an erosion front. The drug diffusion front could be readily determined due to the drug's yellow color. The relative positions of the fronts and the drug release rate were studied as functions of the initial porosity and the molecular weight of the polymer carrier. It was shown that the drug diffusion front best describes the overall release behavior of the system. The fractional drug release was a strong function of the dissolved drug gel layer thickness, which separates the diffusion front from the erosion front. The effect of drug solubility was also investigated by altering the pH and the ionic strength of the dissolution medium. It was shown that as drug solubility increased, the undissolved drug gel layer thickness decreased, again showing the importance of the movement of the diffusion front in controlling the overall release.
Swelling-controlledrelease systems may be suitable for oral, buccal, and nasal applications. 1-3 These systems are based on swellable hydrogel matrix carriers which can be used also as components in mucoadhesive systems.2 However, the mechanisms of drug release from these systems continue to be a matter of particularly in the case of swellable matrix tablets. Previous studies have shown6s7that drug delivery from swelling-controlled release matrices is governed by the gel layer corresponding to the rubbery polymer formed by the glassy core during swelling by water or biological fluids. During drug delivery, the gel layer thickness first increases, then remains constant, and finally decreases as swelling and dissolution compete. The relative importance of these three steps is dictated by the characteristics of the polymer and the drug.8 The same phenomena are encountered in hydrophilic matrices formed by compression of hydrophilic microparticulate powders. Such systems will be henceforth designated as swellable matrix tablets. Again, in these systems if the gel layer thickness remains constant, there is a high probability of constant drug d e l i ~ e r y . ~ We have previously shownlO that the gel layer thickness is important in drug delivery and we have analyzed8its behavior for one model system. Recently, Mockel and Lippold5 have claimed that, in the case of zero-order release behavior from swellable matrix tablets, it is only the polymer dissolution (i.e., the erosion front movement) that controls the drug release rate. The statement is questionable, as it offers only a very limited understanding of the phenomenon. Dissolution medium penetration into a swellable matrix creates sharp boundaries (fronts) which separate various @Abstractpublished in Advance ACS Abstracts, July 1, 1995.
0 1995, American Chemical Society and American Pharmaceutical Association
Accepted for
thermodynamic states of the polymer or various phases of the matrix. Colombo et al.9 have previously measured the swelling and erosion fronts movement concurrently with drug release. Lee and Peppasll termed these boundaries as the swelling and dissolutioderosion fronts, and analyzed the mechanism of drug release from these matrices by photomicrography of the f r o n t ~ . ~These J~ researchersll have analyzed drug delivery in terms of the behavior of the gel layer thickness and other front-related parameters that may be important in this phenomenon. Consider swelling-controlled release systems consisting of a drug molecularly dissolved or dispersed a t high concentration in a polymer matrix. If the drug has a limited solubility in the swollen polymer matrix, it is probable that an undissolved drug front will be observed within the continuously swelling polymer gel layer. This observation was first made by Lee7 using theoretical models for the case of bioerodible systems. In addition, Peppas and collaborator^^^ had indicated that in swellable matrix tablets, drug dissolution might be responsible for an observed zero-order release mechanism. More recently, Lee14 observed a drug dissolution front (henceforth designated in the present work as the diffusion front) due to the limiting solubility phenomenon described above. Therefore, in general it may be expected that the presence of a drug sufficiently loaded so that it is above its solubility limit (regardless of whether the drug has a low or high solubility in water) and released from a swellable matrix table would create conditions such that the release would be controlled more by the drug's dissolution than by the polymer swelling. This does not mean that swelling will not be important (especially as it controls the associated polymer dissolution process). Instead, it points to the importance of the undissolved drug layer within the gel. In general, a simple identification of the fronts during swelling and dissolution of the matrix does not provide a complete understanding of the mechanisms of drug release from these systems. The only exception to this rule is the presence of front s y n c h r o n i ~ a t i o n , ~ J which ~ J ~ J ~has led to a series of new, zero-order release, swelling-controlledsystems.1° The goals of the present work were to study the matrix swelling and drug dissolution in swellable matrix tablets containing high drug loading, in order t o establish relationships between front position and drug release kinetics. Variables considered in this work included the polymer molecular weight as indicated by the sample's viscosity grade, matrix porosity, and pH and ionic strength of the dissolution medium.
Experimental Section Materials-Buflomedil pyridoxalphosphate (BPRD, Pirxane, Lisapharma S.p.A., Erba, Italy; mol wt 553.54, density 1.4 g/cm3) was used as a model drug, due to its range of desirable solubilities in water and buffered solutions. This drug has the following characteristic solubilities a t 37 "C: 65 g/100 mL in distilled water, 55 g/lOO mL in
0022-3549/95/3184-0991$09.00/0
Journal of Pharmaceutical Sciences / 991 Vol. 84, No. 8,August 1995
USP XXII phosphate buffer solution of pH 7.4 and ionic strength 0.09 M, 47 g/100 mL in phosphate buffer solution of pH 7.4 and ionic strength 0.36 M, and 100 mg/100 mL in USP XXII simulated gastric fluid of pH 1.2 and ionic strength 0.11 M. The polymer carrier used was hydroxypropyl methylcellulose (HPMC, Methocel, Colorcon, Orpington, UK, of viscosity grades K4M, K15M, and K100M). Cellulose acetate phthalate (CAP, Eastman Fine Chemicals, Kingsport, TN) was used as a binder. Talc (NF XVII) and magnesium stearate (NF XVII) were used as lubricants. Swellable Matrix Tablet Preparation-Granules were prepared via a wet granulation process by kneading in a mortar a mixture of 64.3% BPRD and 24.5% HPMC with 4.4% CAP dissolved in a 2:l acetone-alcohol solution (corresponding to a concentration of 6.9% w/v). The mixture was passed through a screen of 800 pm with a n oscillating granulator (model FGS, Erweka, Frankfurt, Germany). After drying, the granules were lubricated by adding 4.4% talc and 2.4% magnesium stearate. The true density of the powder mixture was measured using a helium pycnometer (Multivolume Pycnometer 1305, Micromeritics, Norcross, GA). The final mixture was compressed using a reciprocating tabletting machine (model EKO, Korsch, Berlin, Germany), equipped with flatfaced punches of 0.7 cm diameter, in order to manufacture tablets weighing 0.13 g with a thickness ranging from 2.6 to 3.1 mm. Tablets were prepared for all polymer used with a nominal (calculated) porosity of around 13%. In addition, matrices having a nominal porosity of 21% and 7.5% were prepared with Methocel K15M using different compression forces. D r u g Release and Fronts Movement-The position of the fronts and the release behavior of the prepared swellable matrix tablets were followed in a USP XXII dissolution apparatus 2 as a function of time. For the measurement of the front movement during drug release, a simple photographic method was employed because of the drug color. In a typical experiment, the cylindrical matrix was placed in a cell consisting of two transparent Plexiglas disks fixed on their bases by screws, as previously r e ~ 0 r t e d . lThe ~ water could enter the matrix only from the lateral side. Thus, the matrix was placed in this circular cell, and swelling and drug release were accomplished only from the lateral side of the matrix. The assembled device containing the matrix was introduced into the vessel of the dissolution apparatus and tested with the paddle rotation speed of 250 rpm in 1 L of dissolution medium a t 37 "C. The rotation used in these experiments was selected to obtain release without any boundary layer effects. At fixed time intervals, the device was removed from the dissolution apparatus and photographed by means of a video camera connected to a stereomicroscope, to follow the swelling phenomenon and the front movement. Drug release was followed by measuring the drug concentration spectrophotometrically a t 282 nm (model Spectracomp 602, Advanced Products, Milan, Italy). All experiments were done in triplicate.
Results and Discussion The BPRD used in the experiments conducted in this work is a yellow drug which, when dissolved in water at high concentration, gives a bright orange solution. This allowed us to identify clearly not only the boundaries between the glassy and rubbery states of the polymer but also the front observed between the undissolved and dissolved drug in the polymer gel state. Starting from the center of the matrix, during the swelling process (see Figures 1and 21, the following three fronts were observed: (i) a swelling front, identifying the boundary between the still glassy polymer (A) and its rubbery gel state (B);(ii) a diffusion front, indicating the boundary between the still undissolved (solid) drug (B) and the dissolved drug in the gel layer; and (iii) an erosion front, identifying the boundary between the matrix (C) and the dissolution medium (water or buffer). The positions of these fronts inside the matrix could be readily measured through the transparent top of the device during release time. It is evident that the movement of the three fronts can be used to calculate three important parameters of the swelling/dissolution process: (i) the rate of water 992 /Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
Erosion Front Diffusion Front Swelling Front
A Undissolved Drug, Glassy Polymer Layer B Undissolved Drug, Gel Layer C Dissolved Drug, Gel Layer
Figure 1-Schematic representation of erosion, diffusion, and swelling fronts, along with the relevant layers.
uptake, broadly associated with the position of the swelling front, rA; (ii) the rate of drug dissolution depending on the position of the diffusion front, r B ; and (iii) the rate of matrix erosion controlled by the erosion front position, rc. In previous studies by our group6JoJ8and by other^,^^-^^ polarized microscopy, photomicrography, or magnetic resonance imaging have been used to observe the swelling and erosion fronts. Yet, there have been no previous studies to determine the diffusion fronts observed in a matrix sufficiently loaded with a water soluble drug. There is only one limited study by Lee,14who investigated sodium diclofenac release during swelling of initially glassy poly(2-hydroxyethyl methacrylate) samples. In these studies, the appearance of the drug diffusion front was only coincidental due to the poor water solubility of the drug selected (see for example, the comments of Lee and Kim on p 232 of ref 19.) For most water soluble drugs, the limited drug concentration a t which the diffusion front disappears is rather low and coincides with the concentration of water a t which the polymer changes from glassy to rubbery. Several theoretical studies have indicated that such a front should be observed with selected d r ~ g s . ~ , ' ~ , The lack of previous studies on this subject is not surprising, as most micrographic techniques have a limitation in front observation. It is indeed difficult, if not impossible, to distinguish the drug diffusion front for most drugs. However, BPRD has the advantage that it is colored and has a solubility that can be easily controlled. Thus, we loaded a relatively high amount of drug. The matrix appeared pale yellow in dry form, yellow when wetted, and bright orange when the drug gave rise to concentrated solutions. In the present studies, the position of the fronts was measured and plotted versus time. The amount of drug released was also plotted as a function of time. Figure 3 shows typical drug release profiles of three matrices prepared with three HPMC grades having different molecular weights (as indicated by three different viscosity grades). The release profiles exhibited typical kinetics of hydrophilic matrices which could be well fittedz6by eq 1:
1v1,
Here, Mt is the amount of drug released a t time t , M , is the amount of drug released over a very long time, which corresponds in principle to the initial loading, t is the release time, k is a kinetic constant, and n is a diffusional exponent
minutes
15
210
minutes
450
minutes
Figure 2-Pictures of the matrix presenting the three fronts taken at three different times during swelling and release. 0.6 I
0
I
1
I
I
I
I
I
I
I
I
I
1
I
I
2
3
4
5
6
7
Time (h)
I
6
Figure 3-BPRD fraction released in distilled water at 37 "C from matrices and Methocel K100M (A) prepared with Methocel K4M (0),Methocel K15M (U), with a porosity of 13%. All values are in triplicate SEM. The solid lines represent the best fit of the experimental data to eq 1.
*
that depends on the release mechanism and the shape of the swelling device tested.26 For thin slabs, values of n = 0.5 indicate Fickian release, values of 0.5 < n < 1.0 indicate an anomalous (non-Fickian or coupled diffusiodrelaxation) drug release, whereas values of n = 1.0 indicate a case I1 (purely relaxation-controlled) drug release. Table 1 summarizes the values of n for all the samples tested. In general, and within statistical error, the release behavior was non-Fickian regardless of the method of sample preparation or testing, with values of n varying from 0.57 to 0.63. The drug release rates could be calculated from the slopes of the curves of Figure 3 , multiplied by M,. As shown in Figure 4, the drug release rates are very high in the first 30 min, but continue to drop significantly up to about 90 min. Beyond that time, the rate becomes slower, almost constant. The results of Figure 3 indicate that drug release can be influenced only by high molecular weight HPMC (Methocel K100M). This is probably because a t high molecular weights the polymer is entangled and the effective molecular diffusion area is reduced. No significant difference in release behavior was observed in systems prepared with Methocel K4M and K15M. This phenomenon has been observed before with
Table 1-Analysis of Release Data from Swellable Matrix Tablet Using Equation 1
Polymer Grade
Dissolution Medium
Methocel K4M Methocel K15M Methocel K100M Methocel K15M Methocel K15M Methocel K4M Methocel K4M Methocel K4M
Distilled water Distilled water Distilled water Distilled water Distilled water pH 7.4, p 0.09 M pH 7.4, p 0.36 M pH 1.2, p 0.11 M
a
Porosity
Diffusion Exponent n f 95% cia
Kinetic Constant k x I 03 (SP)
13 13 12 7.5 21 13 13 13
0.63 f 0.002 0.61 f 0.010 0.63 f 0.003 0.60 f 0.020 0.60 f 0.001 0.63 i0.010 0.60 f 0.009 0.57 f 0.010
0.84 1.03 0.77 1.09 1.23 0.87 1.12 1.76
Yo
Confidence intervals of three replications.
'w
' *
4
Time (h)
5
6
7
8
Figure 4-BPRD release rate as a function of time from systems prepared with Methocel K4M (0),Methocel K15M (U), Methocel K100M (A).
It can be interpreted by considering that soluble undissolved drug is present at high concentration close to the polymer transition region, i.e., near the swelling front. Yet it is well-known that macromolecular relaxations occur predominantly near that front. These, in turn, will be suppressed by the presence of undissolved drug, leading to minimal relaxation. Indeed, our analysis of the values of n of eq 1for these systems shows that, although non-Fickian, they are very close to the Fickian limit of n = 0.5. Lee14 has previously Journal of Pharmaceutical Sciences / 993 Vol. 84, No. 8, August 1995
I
3 F
2
l ac
-
M L
Table 2-Rate of Change of Front Positiona (Mean Values t SEM) Rate x lo5 ( m r n c ' ) Polymer Grade
Dissolution Medium
% Porosity
Swelling Front
Diffusion Front
Methocel K4M Methocel K15M Methocel KIOOM Methocel K15M Methocel K15M Methocel K4M Methocel K4M Methocel K4M
Distilled water Distilled water Distilled water Distilled water Distilled water pH 7. 4, ,u0.09 M pH 7.4, ,u 0.36 M pH1.2,,u00.11 M
13 13 12 7.5 21 13 13 13
6.77 f 0.01 7.95 -t 0.11 6.73 t 0.09 8.17t0.11 10.66 0.33 6.61 2 0.03 5.60 -t 0.08 7.52 t 0.34
5.99 i0.03 5.51 k 0.23 5.95 -t 0.07 5.92 -t 0.1 1 5.76 f 0.05 5.51 k 0.05 4.36 t 0.05 7.171f-0.34
l
+
Calculated from the slopes of the linear parts of the fronts positions versus time curves represented in Figures 5, 7, and 9.
1
3c
5
4 3 2
0
Lu
d-2
0
1
2
3
, 1-LLLL I
4
5
L
LL UILU
6
7
u
8
Time (h) 0
1
0
2
1
2
3
4
3
4
5
5
6
6
7
7
8
8
Time (h) Figure 5-Erosion (0),diffusion (O),and swelling (A) front positions versus time in systems prepared with Methocel K4M (panel a), Methocel K15M (panel b), and Methocel KIOOM (panel c). Values are in triplicate k SEM, and the release was performed in distilled water at 37 "C.
shown the same behavior by pointing out that the transport becomes more Fickian at higher drug loading. Understanding of the swelling behavior of the systems studied could be obtained from the detailed analysis of the movement of the three fronts. For example, Figure 5 shows the positions of the erosion, diffusion, and swelling fronts during the drug release process. The interface between the matrix and the dissolution medium at the beginning of the experiment is indicated by position 0. The standard errors of the mean are relatively small in all cases. Due to the immediate swelling of the matrix, the erosion front exhibited 994 / Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
Figure 6-Gel layer thickness versus time in systems prepared with Methocel K4M (O), Methocel K15M (0),and Methocel K100M (A) Values are in triplicate k SEM, and the release was performed in distilled water at 37" C
a rapid initial movement outward; for the lowest molecular weight polymer formulation, it seemed to approach an almost constant value in approximately 3 h. The Methocel KlOOM formulation did not seem to attain a constant value even after 7 h. The diffusion and swelling fronts exhibited a rapid initial change, followed by an almost linear decrease as they moved inward. As shown in Table 2, the rates of movement of the diffusion fronts were only slightly different for the three Methocel grade formulations. The relative movement of either the erosion and swelling fronts or the erosion and diffusion fronts indicated that, after an initial divergence, the fronts tended to move in the same direction. This phenomenon can be represented in terms of the gel layer thickness, which is defined as the difference between erosion and swelling front positions , rc - r A (see Figure 1). As shown in Figure 6, the gel layer thickness increased quickly early on and then slowed down. It is also interesting to note that the patterns of change of drug release rate (see Figure 4)follow inversely the dynamics of gel layer thickness (see Figure 6). To study the effect of matrix porosity on the front position and release behavior, matrices containing Methocel K15M with porosity from 7.5% to 21% were investigated. It was found that the amount of drug released increased as the porosity increased from 7.5%to 21% (see Table 1). This was expected, because tablets with higher porosity had a larger lateral area (at constant weight). Our dissolution experiments were done with a device that allowed water to enter only radially. Thus, tablets with larger thickness (higher porosity) exhibited larger amounts of released drug. Analysis of the front movement in these porous matrices indicated that there were differences between the two matri-
4
3
€
E
2
v
r
0 ..-c g Q c
c
e
1
o -1
-2 -3 -4
0
2
4
6
8
1
0
2
4
6
8
10
%me (h)
Figure 7-Erosion (0),diffusion (U),and swelling (A) front positions versus time in systems prepared with Methocel K15M and porosities of 7.5% (panel a) and 21% (panel b) Values are in triplicate i SEM, and the release was performed in distilled water at 37" G.
increased, the fraction of drug released decreased. The movement of the erosion front was slower (Figure 9b,c) in the case of matrix dissolution in a solution with pH 7.4 and ionic strength of 0.09 M. Therefore, at pH 7.4, different release rates were observed for systems exhibiting approximately the same gel layer thickness. To further investigate the importance of drug solubility on drug release profiles, the undissolved drug layer thickness was measured as the difference between diffusion and swelling front positions. This distance was smaller in samples swollen at pH 1.2 than at pH 7.4. Indeed, a plot of the undissolved drug layer thickness as a function of time showed the importance of drug solubility on drug release. In Figure 10, the experimental points are indicated along with three straight lines which represent the undissolved drug layer 0 2 4 6 0 10 thickness as calculated from subtraction of the curves fitting Time (h) the data of Figure 9. Thus, these three straight lines are equivalent to the best fitting. This observation should not be Figure 8-BPRD fraction released in distilled water at 37" C from matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media construed as the result of an influence of the pH on the matrix having different pH and ionic strengths: buffer solution pH 7.4, ,u = 0.09 M (0);dissolution process, because the polymer HPMC is not pHbuffer solution pH 7.4, p = 0.36 M (0);and simulated gastric fluid pH 1.2, p sensitive. Instead, this pH dependence is clearly because the 0.11 M (a). All values are in triplicate ? SEM. The solid lines represent the best drug used in the present study has a significantly different fit of the experimental data to eq 1. solubility in different media. To elucidate this last point, the data of Figure 10 of the ces, as shown in Figure 7 and in Table 2. The swelling front undissolved drug layer thickness versus time were normalized of the matrix with 21% porosity moved faster and reached by dividing them by the intercepts of the best fits of the data. the matrix center [corresponding to a position of 3.5 mm This was done in order to eliminate artifacts stemming from However, the rate of (radius of tablet)] in less than 8 h. the specific morphology of each matrix (porosity, granulomchange of the diffusion front remained practically unchanged etry, particle components distribution). The results of this for the two matrices. The swelling and diffusion fronts analysis are presented in Figure 11. Thus, the time-dependdiverged, whereas the erosion front continued to increase ent change in the undissolved drug layer thickness (i.e., the when the glassy core disappeared. Thus, the higher porosity slope of each of the three lines of Figure 11)was the smallest of the matrix facilitated the uptake of water, although the when the drug solubility was highest (100 g/lOO mL), i.e., at drug dissolution rate, expressed by the movement of the pH 1.2 (curve 3). In this case, the drug dissolved readily and diffusion front, remained unchanged. the diffusion front approached the swelling front. The timeMatrices containing Methocel K4M were also tested in two dependent change of the undissolved drug layer thickness was dissolution media of pH 1.2 and 7.4 and at two different ionic less prominent in the case of swelling in a pH 7.4 solution strengths (for experiments at pH 7.4). As shown in Figure 8, having an ionic strength of 0.09 M (curve 2), where the drug the amount of drug released was significantly higher at pH solubility is 55 g/lOO mL. When the drug solubility decreased 1.2 than at pH 7.4. Again, the front positions were measured even further, as in the case of a solution of pH 7.4 with an as functions of time and the results are shown in Figure 9 ionic strength of 0.36 M (curve 11, where the solubility is only and in Table 2. The analysis of the front movement revealed 47 g/lOO mL, the undissolved drug layer was the highest. This some interesting features: although the erosion front moveclearly indicates that, because of its low solubility, a sigment was not significantly affected by the pH of the dissolunificant amount of the drug remained undissolved in the tion medium, the diffusion and swelling fronts had a tendency tablet. to move faster when the system was swollen in a solution of pH 1.2 (Figure 9a) than in a solution of pH 7.4 (Figure 9c). Finally, the rate of change of the undissolved drug layer thickness was plotted in Figure 12 as a function of the drug The data of Figure 8 also show that, as the ionic strength 0.7
Journal of Pharmaceutical Sciences / 995 Vol. 84, No. 8, August 1995
2
0
0
4
2
6
10
8
3 2
4
Time (h)
6
lo
8
Figure 10-Undissolved drug layer thickness as a function of time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: buffer solution pH 7.4, p 0.09 M (0); buffer solution pH 7.4, p 0.36 M (0);and simulated gastric fluid pH 1.2, p 0.1 1 M (A). All values are in triplicate k SEM. The solid lines represent the undissolved drug layer thickness as calculated from subtraction of the curves fitting the data of Figure 9.
h
EE1
1
v
C
.o c .-
0
8
a
c.
C
-1
E -2 -31' 0
'
'
' 2
'
'
'
'
'
'
4
'
I
6
'
'
'
'
'
'
'
1
10
8
31
2
3
4
Time (h)
5
6
7
8
Figure 11-Normalized undissolved drug layer thickness as a function of time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: buffer solution pH 7.4, p 0.36 M (1); buffer solution pH 7.4, p 0.09 M (2); and simulated gastric fluid pH 1.2, p 0.11 M (3).
h
E €
1
I
2
v
L& 0
1
C
0 ..4 -
8 a c
C
0
-1
4
2
:
-2
:
.
4 3.5
-3 0
2
6
4
8
10
Time (h) Figure 9-Erosion (0),diffusion (U),and swelling (A) front positions versus time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: Simulated gastric fluid pH 1.2, p 0.1 1 M (panel a); buffer solution pH 7.4, p 0.36 M (panel b); and buffer solution pH 7.4, p 0.09 M (panel c). All values are in triplicate k SEM.
solubility. These data also include studies of swelling and release in distilled water, where the solubility is 65 g/100 mL. These data indicate that it is possible to establish a correlation between an equilibrium variable such as the drug solubility and a dynamic (time-dependent) parameter, like the undissolved drug layer thickness; the latter, in turn, is related to the drug release rate. In fact, the limited amount of water in the gel layer accentuated the difference in drug dissolution rate, which would otherwise have been difficult to predict with a conventional dissolution experiment. 996 / Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
2 2 1.5 . : 1
1
A
v
30
40
50
60
70
80
90
100
110
Solubility g/lOOml
Figure 12-Rate of undissolved drug layer thickness change versus drug solubility in matrices prepared with Methocel K4M (initial porosity 13%) and tested in different dissolution media: distilled water (A),buffer solution pH 7.4, p 0.09 M (a);buffer and simulated gastric fluid pH 1.2, p 0.11 M (r). solution pH 7.4, p 0.36 M (0);
Conclusions We can conclude that the diffusion front movement is the main parameter affecting the drug release rate observed. This, in turn, is strictly dependent on the matrix prepared, i.e., on drug solubility and type of polymer. When the diffusion front
moves faster due to increased solubility of the drug, for example, the drug release rate is higher. As a consequence, the dissolved drug gel layer thickness is more important in analyzing drug release from swellable matrix tablets than the previously defined gel layer thickness. Indeed, the former gives a much better representation of the continuous drug concentration gradient predominant in the swellable matrix. Therefore, the data presented here indicate that there are two parameters that influence drug release from such systems: (1)the diffusion front and (2) the dissolved drug gel layer thickness. In fact, because of the suppression of relaxation near the swelling front by undissolved drug, one may conclude that in these swellable matrix tablets drug dissolution and subsequent diffusion through the gel layer are much more important than swelling. Until the swelling and erosion fronts diverge, the release rate is higher and varies with time. When the erosion front movement begins to decrease, a tendency toward synchronization with the swelling front may occur and the release of the drug may become almost constant. For the systems studied here, phase erosion did not play a major role in drug delivery: when the erosion front moved in the same direction as the other fronts (erosion process), more than 50% of the drug was already delivered. This does not mean that erosion was not present, but that the transport through the gel layer was more relevant in the presence of a very soluble drug.
8. Harland, R. S.; Gazzaniga, A,; Sangalli, M. E.; Colombo, P.; Peppas, N. A. Pharm. Res. 1988, 5,488-493. 9. Colombo, P.; Gazzaniga, A,; Conte, U.; Sangalli, M. E.; La Manna, A. Proceed. Intern. Symp. Controlled Release Bioact. Muter. 1987. 14. 83-84. 10. Colombo, P.;Gazzaniga, A,; Caramella, C.; Conte, U.; La Manna, A. Acta Pharm. Technol. 1987, 33, 15-20. 11. Lee, P. I.; Peppas N. A. J . Controlled Release 1981,6,207-215. 12. Lee, P. I. Pharm. Res. 1993, 10, 980-985. 13. Gurny, R.; Doelker, E.; Peppas, N. A. Biomaterials 1982,3,273z. 14. Lee, P. I. J . Controlled Release 1985, 2, 227-288. 15. Conte, U.; Colombo, P.; Gazzaniga, A,; Sangalli, M. E.; La Manna, A. Biomaterials 1988, 9, 489-493. 16. Colombo, P.; Lee, P. I.; Peppas, N. A. Pharm. Res. 1990,7,431432. 17. Bettini, R.; Colombo, P.; Massimo, G.; Catellani, P. L.; Vitali, T. Eur. J . Pharm. Sci. 1994, 2, 213-219. 18. Peppas, N. A,; Franson, N. M. J . Polym. Sci. Polym. Phys. Ed. 1983, 21,983-997. 19. Lee, P. I.; Kim, C. J . Controlled Release 1991,16, 229-236. 20. Cappello, B.; Del Nobile, M. A.; La Rotonda, M. I.; Mensitieri, G.; Miro A.; Nicolais, L. I1 Farmaco 1994, 49, 809-818. 21. Mitchell, K.; Ford, J. L.; Armstrong, D. J.; Elliot, P. N. C.; Hogan, J. E.; Rostron, C. Znt. J. Pharm. 1993, 100, 143-154. 22. Weisenbereer. L. A.: Koenicr. J. L. Macromolecules 1990., 23,. 2445-2453. ' 23. Rajabi-Siahboomi, A. R.; Bowtell, R. W.; Mansfield, P.; Henderson, A.; Davies, M. C.; Melia, C. D. J . Controlled Release 1994, 31, 121-128. 24. Ashraf, M.; Iuorno, V. L.; Coffin-Beach, D.; Anderson Evans, C.; Auesbureer. L. L. Pharm. Res. 1994. 11. 733-737. 25. PoGhchck, A. Y.; Zaikov, G. E.; Petropoulos, J. H. Int. J . Polym. Muter. 1993, 19, 1-14. 26. Sinclair, G. W.; Peppas, N. A. J . Membr. Sci. 1984, 17, 329nn
-I
""*
531.
27. Ford, J. L.; Rubinstein, M. H.; Hogan, J. E. Znt. J . Pharm. 1985, 24, 327-338.
References and Notes 1. Colombo, P. Adu. Drug. Del. Reu. 1993, 11, 37-57. 2. Gurny, R.; Peppas, N. A. In Bioadhesive Drug Delivery Systems; Lenaerts, V., Gurny, R., Eds.; CRC Press, Boca Raton, FL, 1990; pp 153-168. 3. Ryden, L.; Edman, P. Znt. J. Pharm. 1992, 83, 1-10, 4. Pham, A. I.; Lee, P. I. Proceed. Intern. Symp. Controlled Release Bioact. Mater. 1993, 20, 220-221. 5 . Mockel, J. E.; Lippold, B. C. Pharm. Res. 1993, 10, 1066-1070. 6. Peppas, N. A. J . Membr. Sci. 1980, 7, 241-253. 7. Lee, P. I. J . Membr. Sci. 1980, 7, 225-275.
Acknowledgments This work was presented in part a t the 21st International Conference on Controlled Release of Bioactive Materials, Nice, France, 1994. It was supported by grant No. 93.02913.CT03 from the bilateral Italy-USA program of the Consiglio Nazionale delle Ricerche (CNR) and grant No. CRG 940084 from the North Atlantic Treaty Organization (NATO). The authors wish to thank Maria Paola Chiesi for her help in the data collection.
JS94063OX
Journal of Pharmaceutical Sciences / 997 Vol. 84, No. 8,August 1995
#School of Chemical Engineering, 1283 CHME Building, Purdue University, West Lafayette, IN 47907- 1283. publication May 16, 1995@. Abstract 0 Swellable controlled release devices of buflomedil pyridoxalphosphate in hydroxypropyl methylcellulose were prepared, and their swelling and release behavior was investigated. The drug release as a function of time was investigated for various system parameters. Three distinct fronts were observed during the swelling and release processes, i.e., a swelling, a drug diffusion, and an erosion front. The drug diffusion front could be readily determined due to the drug's yellow color. The relative positions of the fronts and the drug release rate were studied as functions of the initial porosity and the molecular weight of the polymer carrier. It was shown that the drug diffusion front best describes the overall release behavior of the system. The fractional drug release was a strong function of the dissolved drug gel layer thickness, which separates the diffusion front from the erosion front. The effect of drug solubility was also investigated by altering the pH and the ionic strength of the dissolution medium. It was shown that as drug solubility increased, the undissolved drug gel layer thickness decreased, again showing the importance of the movement of the diffusion front in controlling the overall release.
Swelling-controlledrelease systems may be suitable for oral, buccal, and nasal applications. 1-3 These systems are based on swellable hydrogel matrix carriers which can be used also as components in mucoadhesive systems.2 However, the mechanisms of drug release from these systems continue to be a matter of particularly in the case of swellable matrix tablets. Previous studies have shown6s7that drug delivery from swelling-controlled release matrices is governed by the gel layer corresponding to the rubbery polymer formed by the glassy core during swelling by water or biological fluids. During drug delivery, the gel layer thickness first increases, then remains constant, and finally decreases as swelling and dissolution compete. The relative importance of these three steps is dictated by the characteristics of the polymer and the drug.8 The same phenomena are encountered in hydrophilic matrices formed by compression of hydrophilic microparticulate powders. Such systems will be henceforth designated as swellable matrix tablets. Again, in these systems if the gel layer thickness remains constant, there is a high probability of constant drug d e l i ~ e r y . ~ We have previously shownlO that the gel layer thickness is important in drug delivery and we have analyzed8its behavior for one model system. Recently, Mockel and Lippold5 have claimed that, in the case of zero-order release behavior from swellable matrix tablets, it is only the polymer dissolution (i.e., the erosion front movement) that controls the drug release rate. The statement is questionable, as it offers only a very limited understanding of the phenomenon. Dissolution medium penetration into a swellable matrix creates sharp boundaries (fronts) which separate various @Abstractpublished in Advance ACS Abstracts, July 1, 1995.
0 1995, American Chemical Society and American Pharmaceutical Association
Accepted for
thermodynamic states of the polymer or various phases of the matrix. Colombo et al.9 have previously measured the swelling and erosion fronts movement concurrently with drug release. Lee and Peppasll termed these boundaries as the swelling and dissolutioderosion fronts, and analyzed the mechanism of drug release from these matrices by photomicrography of the f r o n t ~ . ~These J~ researchersll have analyzed drug delivery in terms of the behavior of the gel layer thickness and other front-related parameters that may be important in this phenomenon. Consider swelling-controlled release systems consisting of a drug molecularly dissolved or dispersed a t high concentration in a polymer matrix. If the drug has a limited solubility in the swollen polymer matrix, it is probable that an undissolved drug front will be observed within the continuously swelling polymer gel layer. This observation was first made by Lee7 using theoretical models for the case of bioerodible systems. In addition, Peppas and collaborator^^^ had indicated that in swellable matrix tablets, drug dissolution might be responsible for an observed zero-order release mechanism. More recently, Lee14 observed a drug dissolution front (henceforth designated in the present work as the diffusion front) due to the limiting solubility phenomenon described above. Therefore, in general it may be expected that the presence of a drug sufficiently loaded so that it is above its solubility limit (regardless of whether the drug has a low or high solubility in water) and released from a swellable matrix table would create conditions such that the release would be controlled more by the drug's dissolution than by the polymer swelling. This does not mean that swelling will not be important (especially as it controls the associated polymer dissolution process). Instead, it points to the importance of the undissolved drug layer within the gel. In general, a simple identification of the fronts during swelling and dissolution of the matrix does not provide a complete understanding of the mechanisms of drug release from these systems. The only exception to this rule is the presence of front s y n c h r o n i ~ a t i o n , ~ J which ~ J ~ J ~has led to a series of new, zero-order release, swelling-controlledsystems.1° The goals of the present work were to study the matrix swelling and drug dissolution in swellable matrix tablets containing high drug loading, in order t o establish relationships between front position and drug release kinetics. Variables considered in this work included the polymer molecular weight as indicated by the sample's viscosity grade, matrix porosity, and pH and ionic strength of the dissolution medium.
Experimental Section Materials-Buflomedil pyridoxalphosphate (BPRD, Pirxane, Lisapharma S.p.A., Erba, Italy; mol wt 553.54, density 1.4 g/cm3) was used as a model drug, due to its range of desirable solubilities in water and buffered solutions. This drug has the following characteristic solubilities a t 37 "C: 65 g/100 mL in distilled water, 55 g/lOO mL in
0022-3549/95/3184-0991$09.00/0
Journal of Pharmaceutical Sciences / 991 Vol. 84, No. 8,August 1995
USP XXII phosphate buffer solution of pH 7.4 and ionic strength 0.09 M, 47 g/100 mL in phosphate buffer solution of pH 7.4 and ionic strength 0.36 M, and 100 mg/100 mL in USP XXII simulated gastric fluid of pH 1.2 and ionic strength 0.11 M. The polymer carrier used was hydroxypropyl methylcellulose (HPMC, Methocel, Colorcon, Orpington, UK, of viscosity grades K4M, K15M, and K100M). Cellulose acetate phthalate (CAP, Eastman Fine Chemicals, Kingsport, TN) was used as a binder. Talc (NF XVII) and magnesium stearate (NF XVII) were used as lubricants. Swellable Matrix Tablet Preparation-Granules were prepared via a wet granulation process by kneading in a mortar a mixture of 64.3% BPRD and 24.5% HPMC with 4.4% CAP dissolved in a 2:l acetone-alcohol solution (corresponding to a concentration of 6.9% w/v). The mixture was passed through a screen of 800 pm with a n oscillating granulator (model FGS, Erweka, Frankfurt, Germany). After drying, the granules were lubricated by adding 4.4% talc and 2.4% magnesium stearate. The true density of the powder mixture was measured using a helium pycnometer (Multivolume Pycnometer 1305, Micromeritics, Norcross, GA). The final mixture was compressed using a reciprocating tabletting machine (model EKO, Korsch, Berlin, Germany), equipped with flatfaced punches of 0.7 cm diameter, in order to manufacture tablets weighing 0.13 g with a thickness ranging from 2.6 to 3.1 mm. Tablets were prepared for all polymer used with a nominal (calculated) porosity of around 13%. In addition, matrices having a nominal porosity of 21% and 7.5% were prepared with Methocel K15M using different compression forces. D r u g Release and Fronts Movement-The position of the fronts and the release behavior of the prepared swellable matrix tablets were followed in a USP XXII dissolution apparatus 2 as a function of time. For the measurement of the front movement during drug release, a simple photographic method was employed because of the drug color. In a typical experiment, the cylindrical matrix was placed in a cell consisting of two transparent Plexiglas disks fixed on their bases by screws, as previously r e ~ 0 r t e d . lThe ~ water could enter the matrix only from the lateral side. Thus, the matrix was placed in this circular cell, and swelling and drug release were accomplished only from the lateral side of the matrix. The assembled device containing the matrix was introduced into the vessel of the dissolution apparatus and tested with the paddle rotation speed of 250 rpm in 1 L of dissolution medium a t 37 "C. The rotation used in these experiments was selected to obtain release without any boundary layer effects. At fixed time intervals, the device was removed from the dissolution apparatus and photographed by means of a video camera connected to a stereomicroscope, to follow the swelling phenomenon and the front movement. Drug release was followed by measuring the drug concentration spectrophotometrically a t 282 nm (model Spectracomp 602, Advanced Products, Milan, Italy). All experiments were done in triplicate.
Results and Discussion The BPRD used in the experiments conducted in this work is a yellow drug which, when dissolved in water at high concentration, gives a bright orange solution. This allowed us to identify clearly not only the boundaries between the glassy and rubbery states of the polymer but also the front observed between the undissolved and dissolved drug in the polymer gel state. Starting from the center of the matrix, during the swelling process (see Figures 1and 21, the following three fronts were observed: (i) a swelling front, identifying the boundary between the still glassy polymer (A) and its rubbery gel state (B);(ii) a diffusion front, indicating the boundary between the still undissolved (solid) drug (B) and the dissolved drug in the gel layer; and (iii) an erosion front, identifying the boundary between the matrix (C) and the dissolution medium (water or buffer). The positions of these fronts inside the matrix could be readily measured through the transparent top of the device during release time. It is evident that the movement of the three fronts can be used to calculate three important parameters of the swelling/dissolution process: (i) the rate of water 992 /Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
Erosion Front Diffusion Front Swelling Front
A Undissolved Drug, Glassy Polymer Layer B Undissolved Drug, Gel Layer C Dissolved Drug, Gel Layer
Figure 1-Schematic representation of erosion, diffusion, and swelling fronts, along with the relevant layers.
uptake, broadly associated with the position of the swelling front, rA; (ii) the rate of drug dissolution depending on the position of the diffusion front, r B ; and (iii) the rate of matrix erosion controlled by the erosion front position, rc. In previous studies by our group6JoJ8and by other^,^^-^^ polarized microscopy, photomicrography, or magnetic resonance imaging have been used to observe the swelling and erosion fronts. Yet, there have been no previous studies to determine the diffusion fronts observed in a matrix sufficiently loaded with a water soluble drug. There is only one limited study by Lee,14who investigated sodium diclofenac release during swelling of initially glassy poly(2-hydroxyethyl methacrylate) samples. In these studies, the appearance of the drug diffusion front was only coincidental due to the poor water solubility of the drug selected (see for example, the comments of Lee and Kim on p 232 of ref 19.) For most water soluble drugs, the limited drug concentration a t which the diffusion front disappears is rather low and coincides with the concentration of water a t which the polymer changes from glassy to rubbery. Several theoretical studies have indicated that such a front should be observed with selected d r ~ g s . ~ , ' ~ , The lack of previous studies on this subject is not surprising, as most micrographic techniques have a limitation in front observation. It is indeed difficult, if not impossible, to distinguish the drug diffusion front for most drugs. However, BPRD has the advantage that it is colored and has a solubility that can be easily controlled. Thus, we loaded a relatively high amount of drug. The matrix appeared pale yellow in dry form, yellow when wetted, and bright orange when the drug gave rise to concentrated solutions. In the present studies, the position of the fronts was measured and plotted versus time. The amount of drug released was also plotted as a function of time. Figure 3 shows typical drug release profiles of three matrices prepared with three HPMC grades having different molecular weights (as indicated by three different viscosity grades). The release profiles exhibited typical kinetics of hydrophilic matrices which could be well fittedz6by eq 1:
1v1,
Here, Mt is the amount of drug released a t time t , M , is the amount of drug released over a very long time, which corresponds in principle to the initial loading, t is the release time, k is a kinetic constant, and n is a diffusional exponent
minutes
15
210
minutes
450
minutes
Figure 2-Pictures of the matrix presenting the three fronts taken at three different times during swelling and release. 0.6 I
0
I
1
I
I
I
I
I
I
I
I
I
1
I
I
2
3
4
5
6
7
Time (h)
I
6
Figure 3-BPRD fraction released in distilled water at 37 "C from matrices and Methocel K100M (A) prepared with Methocel K4M (0),Methocel K15M (U), with a porosity of 13%. All values are in triplicate SEM. The solid lines represent the best fit of the experimental data to eq 1.
*
that depends on the release mechanism and the shape of the swelling device tested.26 For thin slabs, values of n = 0.5 indicate Fickian release, values of 0.5 < n < 1.0 indicate an anomalous (non-Fickian or coupled diffusiodrelaxation) drug release, whereas values of n = 1.0 indicate a case I1 (purely relaxation-controlled) drug release. Table 1 summarizes the values of n for all the samples tested. In general, and within statistical error, the release behavior was non-Fickian regardless of the method of sample preparation or testing, with values of n varying from 0.57 to 0.63. The drug release rates could be calculated from the slopes of the curves of Figure 3 , multiplied by M,. As shown in Figure 4, the drug release rates are very high in the first 30 min, but continue to drop significantly up to about 90 min. Beyond that time, the rate becomes slower, almost constant. The results of Figure 3 indicate that drug release can be influenced only by high molecular weight HPMC (Methocel K100M). This is probably because a t high molecular weights the polymer is entangled and the effective molecular diffusion area is reduced. No significant difference in release behavior was observed in systems prepared with Methocel K4M and K15M. This phenomenon has been observed before with
Table 1-Analysis of Release Data from Swellable Matrix Tablet Using Equation 1
Polymer Grade
Dissolution Medium
Methocel K4M Methocel K15M Methocel K100M Methocel K15M Methocel K15M Methocel K4M Methocel K4M Methocel K4M
Distilled water Distilled water Distilled water Distilled water Distilled water pH 7.4, p 0.09 M pH 7.4, p 0.36 M pH 1.2, p 0.11 M
a
Porosity
Diffusion Exponent n f 95% cia
Kinetic Constant k x I 03 (SP)
13 13 12 7.5 21 13 13 13
0.63 f 0.002 0.61 f 0.010 0.63 f 0.003 0.60 f 0.020 0.60 f 0.001 0.63 i0.010 0.60 f 0.009 0.57 f 0.010
0.84 1.03 0.77 1.09 1.23 0.87 1.12 1.76
Yo
Confidence intervals of three replications.
'w
' *
4
Time (h)
5
6
7
8
Figure 4-BPRD release rate as a function of time from systems prepared with Methocel K4M (0),Methocel K15M (U), Methocel K100M (A).
It can be interpreted by considering that soluble undissolved drug is present at high concentration close to the polymer transition region, i.e., near the swelling front. Yet it is well-known that macromolecular relaxations occur predominantly near that front. These, in turn, will be suppressed by the presence of undissolved drug, leading to minimal relaxation. Indeed, our analysis of the values of n of eq 1for these systems shows that, although non-Fickian, they are very close to the Fickian limit of n = 0.5. Lee14 has previously Journal of Pharmaceutical Sciences / 993 Vol. 84, No. 8, August 1995
I
3 F
2
l ac
-
M L
Table 2-Rate of Change of Front Positiona (Mean Values t SEM) Rate x lo5 ( m r n c ' ) Polymer Grade
Dissolution Medium
% Porosity
Swelling Front
Diffusion Front
Methocel K4M Methocel K15M Methocel KIOOM Methocel K15M Methocel K15M Methocel K4M Methocel K4M Methocel K4M
Distilled water Distilled water Distilled water Distilled water Distilled water pH 7. 4, ,u0.09 M pH 7.4, ,u 0.36 M pH1.2,,u00.11 M
13 13 12 7.5 21 13 13 13
6.77 f 0.01 7.95 -t 0.11 6.73 t 0.09 8.17t0.11 10.66 0.33 6.61 2 0.03 5.60 -t 0.08 7.52 t 0.34
5.99 i0.03 5.51 k 0.23 5.95 -t 0.07 5.92 -t 0.1 1 5.76 f 0.05 5.51 k 0.05 4.36 t 0.05 7.171f-0.34
l
+
Calculated from the slopes of the linear parts of the fronts positions versus time curves represented in Figures 5, 7, and 9.
1
3c
5
4 3 2
0
Lu
d-2
0
1
2
3
, 1-LLLL I
4
5
L
LL UILU
6
7
u
8
Time (h) 0
1
0
2
1
2
3
4
3
4
5
5
6
6
7
7
8
8
Time (h) Figure 5-Erosion (0),diffusion (O),and swelling (A) front positions versus time in systems prepared with Methocel K4M (panel a), Methocel K15M (panel b), and Methocel KIOOM (panel c). Values are in triplicate k SEM, and the release was performed in distilled water at 37 "C.
shown the same behavior by pointing out that the transport becomes more Fickian at higher drug loading. Understanding of the swelling behavior of the systems studied could be obtained from the detailed analysis of the movement of the three fronts. For example, Figure 5 shows the positions of the erosion, diffusion, and swelling fronts during the drug release process. The interface between the matrix and the dissolution medium at the beginning of the experiment is indicated by position 0. The standard errors of the mean are relatively small in all cases. Due to the immediate swelling of the matrix, the erosion front exhibited 994 / Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
Figure 6-Gel layer thickness versus time in systems prepared with Methocel K4M (O), Methocel K15M (0),and Methocel K100M (A) Values are in triplicate k SEM, and the release was performed in distilled water at 37" C
a rapid initial movement outward; for the lowest molecular weight polymer formulation, it seemed to approach an almost constant value in approximately 3 h. The Methocel KlOOM formulation did not seem to attain a constant value even after 7 h. The diffusion and swelling fronts exhibited a rapid initial change, followed by an almost linear decrease as they moved inward. As shown in Table 2, the rates of movement of the diffusion fronts were only slightly different for the three Methocel grade formulations. The relative movement of either the erosion and swelling fronts or the erosion and diffusion fronts indicated that, after an initial divergence, the fronts tended to move in the same direction. This phenomenon can be represented in terms of the gel layer thickness, which is defined as the difference between erosion and swelling front positions , rc - r A (see Figure 1). As shown in Figure 6, the gel layer thickness increased quickly early on and then slowed down. It is also interesting to note that the patterns of change of drug release rate (see Figure 4)follow inversely the dynamics of gel layer thickness (see Figure 6). To study the effect of matrix porosity on the front position and release behavior, matrices containing Methocel K15M with porosity from 7.5% to 21% were investigated. It was found that the amount of drug released increased as the porosity increased from 7.5%to 21% (see Table 1). This was expected, because tablets with higher porosity had a larger lateral area (at constant weight). Our dissolution experiments were done with a device that allowed water to enter only radially. Thus, tablets with larger thickness (higher porosity) exhibited larger amounts of released drug. Analysis of the front movement in these porous matrices indicated that there were differences between the two matri-
4
3
€
E
2
v
r
0 ..-c g Q c
c
e
1
o -1
-2 -3 -4
0
2
4
6
8
1
0
2
4
6
8
10
%me (h)
Figure 7-Erosion (0),diffusion (U),and swelling (A) front positions versus time in systems prepared with Methocel K15M and porosities of 7.5% (panel a) and 21% (panel b) Values are in triplicate i SEM, and the release was performed in distilled water at 37" G.
increased, the fraction of drug released decreased. The movement of the erosion front was slower (Figure 9b,c) in the case of matrix dissolution in a solution with pH 7.4 and ionic strength of 0.09 M. Therefore, at pH 7.4, different release rates were observed for systems exhibiting approximately the same gel layer thickness. To further investigate the importance of drug solubility on drug release profiles, the undissolved drug layer thickness was measured as the difference between diffusion and swelling front positions. This distance was smaller in samples swollen at pH 1.2 than at pH 7.4. Indeed, a plot of the undissolved drug layer thickness as a function of time showed the importance of drug solubility on drug release. In Figure 10, the experimental points are indicated along with three straight lines which represent the undissolved drug layer 0 2 4 6 0 10 thickness as calculated from subtraction of the curves fitting Time (h) the data of Figure 9. Thus, these three straight lines are equivalent to the best fitting. This observation should not be Figure 8-BPRD fraction released in distilled water at 37" C from matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media construed as the result of an influence of the pH on the matrix having different pH and ionic strengths: buffer solution pH 7.4, ,u = 0.09 M (0);dissolution process, because the polymer HPMC is not pHbuffer solution pH 7.4, p = 0.36 M (0);and simulated gastric fluid pH 1.2, p sensitive. Instead, this pH dependence is clearly because the 0.11 M (a). All values are in triplicate ? SEM. The solid lines represent the best drug used in the present study has a significantly different fit of the experimental data to eq 1. solubility in different media. To elucidate this last point, the data of Figure 10 of the ces, as shown in Figure 7 and in Table 2. The swelling front undissolved drug layer thickness versus time were normalized of the matrix with 21% porosity moved faster and reached by dividing them by the intercepts of the best fits of the data. the matrix center [corresponding to a position of 3.5 mm This was done in order to eliminate artifacts stemming from However, the rate of (radius of tablet)] in less than 8 h. the specific morphology of each matrix (porosity, granulomchange of the diffusion front remained practically unchanged etry, particle components distribution). The results of this for the two matrices. The swelling and diffusion fronts analysis are presented in Figure 11. Thus, the time-dependdiverged, whereas the erosion front continued to increase ent change in the undissolved drug layer thickness (i.e., the when the glassy core disappeared. Thus, the higher porosity slope of each of the three lines of Figure 11)was the smallest of the matrix facilitated the uptake of water, although the when the drug solubility was highest (100 g/lOO mL), i.e., at drug dissolution rate, expressed by the movement of the pH 1.2 (curve 3). In this case, the drug dissolved readily and diffusion front, remained unchanged. the diffusion front approached the swelling front. The timeMatrices containing Methocel K4M were also tested in two dependent change of the undissolved drug layer thickness was dissolution media of pH 1.2 and 7.4 and at two different ionic less prominent in the case of swelling in a pH 7.4 solution strengths (for experiments at pH 7.4). As shown in Figure 8, having an ionic strength of 0.09 M (curve 2), where the drug the amount of drug released was significantly higher at pH solubility is 55 g/lOO mL. When the drug solubility decreased 1.2 than at pH 7.4. Again, the front positions were measured even further, as in the case of a solution of pH 7.4 with an as functions of time and the results are shown in Figure 9 ionic strength of 0.36 M (curve 11, where the solubility is only and in Table 2. The analysis of the front movement revealed 47 g/lOO mL, the undissolved drug layer was the highest. This some interesting features: although the erosion front moveclearly indicates that, because of its low solubility, a sigment was not significantly affected by the pH of the dissolunificant amount of the drug remained undissolved in the tion medium, the diffusion and swelling fronts had a tendency tablet. to move faster when the system was swollen in a solution of pH 1.2 (Figure 9a) than in a solution of pH 7.4 (Figure 9c). Finally, the rate of change of the undissolved drug layer thickness was plotted in Figure 12 as a function of the drug The data of Figure 8 also show that, as the ionic strength 0.7
Journal of Pharmaceutical Sciences / 995 Vol. 84, No. 8, August 1995
2
0
0
4
2
6
10
8
3 2
4
Time (h)
6
lo
8
Figure 10-Undissolved drug layer thickness as a function of time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: buffer solution pH 7.4, p 0.09 M (0); buffer solution pH 7.4, p 0.36 M (0);and simulated gastric fluid pH 1.2, p 0.1 1 M (A). All values are in triplicate k SEM. The solid lines represent the undissolved drug layer thickness as calculated from subtraction of the curves fitting the data of Figure 9.
h
EE1
1
v
C
.o c .-
0
8
a
c.
C
-1
E -2 -31' 0
'
'
' 2
'
'
'
'
'
'
4
'
I
6
'
'
'
'
'
'
'
1
10
8
31
2
3
4
Time (h)
5
6
7
8
Figure 11-Normalized undissolved drug layer thickness as a function of time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: buffer solution pH 7.4, p 0.36 M (1); buffer solution pH 7.4, p 0.09 M (2); and simulated gastric fluid pH 1.2, p 0.11 M (3).
h
E €
1
I
2
v
L& 0
1
C
0 ..4 -
8 a c
C
0
-1
4
2
:
-2
:
.
4 3.5
-3 0
2
6
4
8
10
Time (h) Figure 9-Erosion (0),diffusion (U),and swelling (A) front positions versus time in matrices prepared with Methocel K4M (initial porosity 13%) and tested in dissolution media having different pH and ionic strengths: Simulated gastric fluid pH 1.2, p 0.1 1 M (panel a); buffer solution pH 7.4, p 0.36 M (panel b); and buffer solution pH 7.4, p 0.09 M (panel c). All values are in triplicate k SEM.
solubility. These data also include studies of swelling and release in distilled water, where the solubility is 65 g/100 mL. These data indicate that it is possible to establish a correlation between an equilibrium variable such as the drug solubility and a dynamic (time-dependent) parameter, like the undissolved drug layer thickness; the latter, in turn, is related to the drug release rate. In fact, the limited amount of water in the gel layer accentuated the difference in drug dissolution rate, which would otherwise have been difficult to predict with a conventional dissolution experiment. 996 / Journal of Pharmaceutical Sciences Vol. 84, No. 8, August 1995
2 2 1.5 . : 1
1
A
v
30
40
50
60
70
80
90
100
110
Solubility g/lOOml
Figure 12-Rate of undissolved drug layer thickness change versus drug solubility in matrices prepared with Methocel K4M (initial porosity 13%) and tested in different dissolution media: distilled water (A),buffer solution pH 7.4, p 0.09 M (a);buffer and simulated gastric fluid pH 1.2, p 0.11 M (r). solution pH 7.4, p 0.36 M (0);
Conclusions We can conclude that the diffusion front movement is the main parameter affecting the drug release rate observed. This, in turn, is strictly dependent on the matrix prepared, i.e., on drug solubility and type of polymer. When the diffusion front
moves faster due to increased solubility of the drug, for example, the drug release rate is higher. As a consequence, the dissolved drug gel layer thickness is more important in analyzing drug release from swellable matrix tablets than the previously defined gel layer thickness. Indeed, the former gives a much better representation of the continuous drug concentration gradient predominant in the swellable matrix. Therefore, the data presented here indicate that there are two parameters that influence drug release from such systems: (1)the diffusion front and (2) the dissolved drug gel layer thickness. In fact, because of the suppression of relaxation near the swelling front by undissolved drug, one may conclude that in these swellable matrix tablets drug dissolution and subsequent diffusion through the gel layer are much more important than swelling. Until the swelling and erosion fronts diverge, the release rate is higher and varies with time. When the erosion front movement begins to decrease, a tendency toward synchronization with the swelling front may occur and the release of the drug may become almost constant. For the systems studied here, phase erosion did not play a major role in drug delivery: when the erosion front moved in the same direction as the other fronts (erosion process), more than 50% of the drug was already delivered. This does not mean that erosion was not present, but that the transport through the gel layer was more relevant in the presence of a very soluble drug.
8. Harland, R. S.; Gazzaniga, A,; Sangalli, M. E.; Colombo, P.; Peppas, N. A. Pharm. Res. 1988, 5,488-493. 9. Colombo, P.; Gazzaniga, A,; Conte, U.; Sangalli, M. E.; La Manna, A. Proceed. Intern. Symp. Controlled Release Bioact. Muter. 1987. 14. 83-84. 10. Colombo, P.;Gazzaniga, A,; Caramella, C.; Conte, U.; La Manna, A. Acta Pharm. Technol. 1987, 33, 15-20. 11. Lee, P. I.; Peppas N. A. J . Controlled Release 1981,6,207-215. 12. Lee, P. I. Pharm. Res. 1993, 10, 980-985. 13. Gurny, R.; Doelker, E.; Peppas, N. A. Biomaterials 1982,3,273z. 14. Lee, P. I. J . Controlled Release 1985, 2, 227-288. 15. Conte, U.; Colombo, P.; Gazzaniga, A,; Sangalli, M. E.; La Manna, A. Biomaterials 1988, 9, 489-493. 16. Colombo, P.; Lee, P. I.; Peppas, N. A. Pharm. Res. 1990,7,431432. 17. Bettini, R.; Colombo, P.; Massimo, G.; Catellani, P. L.; Vitali, T. Eur. J . Pharm. Sci. 1994, 2, 213-219. 18. Peppas, N. A,; Franson, N. M. J . Polym. Sci. Polym. Phys. Ed. 1983, 21,983-997. 19. Lee, P. I.; Kim, C. J . Controlled Release 1991,16, 229-236. 20. Cappello, B.; Del Nobile, M. A.; La Rotonda, M. I.; Mensitieri, G.; Miro A.; Nicolais, L. I1 Farmaco 1994, 49, 809-818. 21. Mitchell, K.; Ford, J. L.; Armstrong, D. J.; Elliot, P. N. C.; Hogan, J. E.; Rostron, C. Znt. J. Pharm. 1993, 100, 143-154. 22. Weisenbereer. L. A.: Koenicr. J. L. Macromolecules 1990., 23,. 2445-2453. ' 23. Rajabi-Siahboomi, A. R.; Bowtell, R. W.; Mansfield, P.; Henderson, A.; Davies, M. C.; Melia, C. D. J . Controlled Release 1994, 31, 121-128. 24. Ashraf, M.; Iuorno, V. L.; Coffin-Beach, D.; Anderson Evans, C.; Auesbureer. L. L. Pharm. Res. 1994. 11. 733-737. 25. PoGhchck, A. Y.; Zaikov, G. E.; Petropoulos, J. H. Int. J . Polym. Muter. 1993, 19, 1-14. 26. Sinclair, G. W.; Peppas, N. A. J . Membr. Sci. 1984, 17, 329nn
-I
""*
531.
27. Ford, J. L.; Rubinstein, M. H.; Hogan, J. E. Znt. J . Pharm. 1985, 24, 327-338.
References and Notes 1. Colombo, P. Adu. Drug. Del. Reu. 1993, 11, 37-57. 2. Gurny, R.; Peppas, N. A. In Bioadhesive Drug Delivery Systems; Lenaerts, V., Gurny, R., Eds.; CRC Press, Boca Raton, FL, 1990; pp 153-168. 3. Ryden, L.; Edman, P. Znt. J. Pharm. 1992, 83, 1-10, 4. Pham, A. I.; Lee, P. I. Proceed. Intern. Symp. Controlled Release Bioact. Mater. 1993, 20, 220-221. 5 . Mockel, J. E.; Lippold, B. C. Pharm. Res. 1993, 10, 1066-1070. 6. Peppas, N. A. J . Membr. Sci. 1980, 7, 241-253. 7. Lee, P. I. J . Membr. Sci. 1980, 7, 225-275.
Acknowledgments This work was presented in part a t the 21st International Conference on Controlled Release of Bioactive Materials, Nice, France, 1994. It was supported by grant No. 93.02913.CT03 from the bilateral Italy-USA program of the Consiglio Nazionale delle Ricerche (CNR) and grant No. CRG 940084 from the North Atlantic Treaty Organization (NATO). The authors wish to thank Maria Paola Chiesi for her help in the data collection.
JS94063OX
Journal of Pharmaceutical Sciences / 997 Vol. 84, No. 8,August 1995