three-layer matrix tablets and simple approach of …431 three-layer matrix tablets and simple...

431

Three-layer matrix tablets and simple approach of drug release programming

N. Al-Zoubi1*, S. Malamataris2

1Department of Pharmaceutical Sciences and Pharmaceutics, Faculty of Pharmacy, Applied Science University, Amman, Jordan2Department of Pharmaceutical Technology, School of Pharmacy, University of Thessaloniki, Thessaloniki, Greece

*Correspondence: [email protected]

A face-centered central composite experimental design was applied in programming the sustained drug release from three-layer matrix tab-lets. Xanthan gum (XG), sodium alginate (SA) and their 1:1 mixture were employed as the matrix former controlling the release of diltiazem HCl (DH). Mass fraction of DH in the intermediate layer (X1) and percentage of XG in the matrix former of the intermediate and outer layers (X2 and X3, respectively) were the independent experimental variables (formulation factors). Cumulative percent release at 2 and 12 h (rel2h, rel12h), shape parameter of the release profiles in Weibull function (b), and exponent in the power law model of Peppas (n), were selected as dependent vari-ables and related to the formulation factors via multiple linear regression analysis using second order polynomial equations including two-factor interaction terms. Simplified equations were derived and response surface analysis enabled the formulation factor effects and interactions to be visualized. It was found that different shapes of release profiles can be obtained corresponding to Weibull shape parameter (b) between 0.311 and 1.247. In general, increased incorporation of DH in the intermediate layer or of XG in the outer layers reduced the drug release because of restricted and delayed exposure to the dissolution medium or formation of a stronger diffusion barrier, respectively. Highly significant linear correlation (r = 0.894, p < 0.001) was found between the values of b and the exponent in the power law model of Peppas (n). Good agreement between the predicted, on the basis of the simplified equations (regression models), and the experimental values of three control formulations confirmed the validity of the suggested models in programming the release behavior by the proposed three-layer tablet system within the experi-mental domain.

Key words: Three-layer matrix tablets – Xanthan gum – Sodium alginate – Drug release – Weibull equation – Face center cube design – Multiple regression.

J. DRUG DEL. SCI. TECH., 18 (6) 431-437 2008

Sustained-release oral dosage forms have become popular for the administration of many drugs because they give more consistent blood levels and the tablet matrices of water-swellable polymers (hy-drophilic colloids) are the simplest and least expensive systems [1]. Their mechanism of release control is based on the polymer swelling due to hydration by the gastrointestinal fluids and formation of a high viscosity gel layer, which retards the diffusion of dissolved drug [2]. A major disadvantage of these diffusion-controlled matrix devices is the non-linear release and multi-layered matrix tablets comprising an intermediate layer containing most of the active drug(s), and two outer barrier layers have been suggested as zero-order sustained re-lease systems [3-6]. Furthermore, various release profiles other than constant such as delayed release, pulsatile or multimodal delivery profiles may be achieved by the design of multi-layer tablets differing in the geometry, the composition or combination (arrangement) of layers [7]. Since the system of multi-layered matrix tablets is a recognized flexible technology for modifying release of drugs and statistical experimental design methodologies are systematic efficient tools for optimizing pharmaceutical formulations, their combined application was thought to be of interest. In the present work, a system of three-layer matrix tablet is evaluated as a simple approach of drug release programming, by altering the drug (diltiazem HCl, DH) distribution and the composition of the hydrophilic matrix former (xanthan gum, XG, and sodium alginate, SA) in the intermediate and outer layers. Statistical significance of the main effects and two-way interactions is quantified in order to derive simplified equations as a tool of drug release programming.

I. MATERIALS AND METHODS1. Materials Diltiazem HCl fine powder (particle size < 180 µm) was kindly offered by the United Pharmaceutical Manufacturing (UPM) Co., Amman, Jordan. Xanthan gum (80.2% < 180 µm, 19.8% 180-400 µm) and sodium alginate (90.4% < 180 µm, 9.6% 180-400 µm) were pur-chased from Sigma, USA, and BDH, UK, respectively. The viscosity of 0.67% w/v polymer solution in distilled water at 25°C and shear rate 30 s-1 was determined using the Cup and bob method on Physica MCR 301 rheometer (Anton Paar, Austria) and was found to be 0.396 and 0.012 Pa.s for xanthan gum and sodium alginate, respectively. All materials were used as received.

2. Methods 2.1. Preparation of three-layer tablets Fifteen batches of different three-layer matrix experimental tablets comprising a 400-mg intermediate layer and two 200-mg outer layers of identical composition were prepared. All the tablets contained a fixed amount (300 mg) of diltiazem HCl (DH) distributed in the inter-mediate and the outer layers by using appropriate physical mixtures with xanthan gum (XG), sodium alginate (SA) or a binary mixture of them (1:1). The composition of the tablets is described in Table I as the mass fraction of DH incorporated in the intermediate layer (X1) acquiring levels 0.2, 0.6 and 1.0 and the percentage of XG in the matrix former of the intermediate (X2) and outer (X3) layers, acquiring levels 0, 50 and 100%. Physical mixing was applied in a small mortar with a spatula for 15 min. For the preparation of tablets, a 13-mm flat-faced punch and die set and a hand-operated hydraulic press (Shimadzu,

432

Japan) were used. Two hundred milligrams of accurately weighted mixture of outer layer was transferred into the die cavity and slightly compressed by hand for uniform spreading. The upper punch was lifted and 400 mg of intermediate layer mixture was placed over the first outer layer and again slightly compressed for uniform spreading. Another 200 mg of outer layer mixture was poured and then 10 MPa pressure was applied for 30 s, resulting in saturated (zero porosity) three-layer compacts (matrix tablets).

2.2. In vitro drug release The release rate of DH from the matrix tablets was determined using USP Apparatus II paddle dissolution system (Pharma Test PTW 2, Hainburg, Germany), at 100 rpm, with 900 mL of distilled water as dissolution medium. The paddle instead of the basket system was used in order to avoid possible interference with the swelling of the three-layer matrix tablets and the release process, since the tablets are relatively large and contain highly-swellable xanthan gum [8]. At certain time intervals, 10-mL samples were taken and the volume of water was replaced. The samples were filtered through 0.45-µm cellu-lose acetate syringe filter and the concentration of DH dissolved was determined, after suitable dilution, by UV spectroscopy (Spectronic 601, Milton Roy, USA) at a wavelength corresponding to maximum absorbance (237 nm). All tests were performed in triplicate and the percent drug release was determined from the mean concentration. The effects of formulation variables on the release profile were evaluated with the aid of MS Excel by fitting the cumulative Weibull distribution function to the release results employing the linearized form [9 -11]:

log[-ln(1-m)] = b log (t - Ti) - log a

where m is the cumulative drug release at time t, Ti is the lag time before the onset of dissolution or release process that in most cases will be zero, b is the shape parameter and a is the time scale of the process. From a and b the time for 63.2% release, td, can be calculated [12]: td = a1/b. This Weibull function was selected as being capable of dealing with dissolution profiles corresponding to an initial release phase followed by either a faster or slower one. More specifically, when b = 1, the curve is exponential and Equation 1 reduces to the simple first order model. For b > 1, the release profile is sigmoid, while for b < 1 it is parabolic corresponding to fast initial release slowing

gradually. In extreme cases, where b → 0, Equation 1 yields a straight line and when b → ∞, the curve degenerates to a step function [13]. The linearized form is equivalent and provides almost identical results to non-linear fit [14]. In order to characterize the release mechanism, the power law model of Peppas was fitted by non-linear regression to the release data (for the first 60% dissolved) [15]:

Mt/M∞ = Kp. tn

where Mt/M∞ represents the fractional release of drug at time t, Kp is the release rate constant and the exponent n is indicative of the release mechanism. In the example of cylindrical tablets, a value of n ≈ 0.45 indicates Fickian diffusion, while higher values of n (between 0.45 and 0.9) indicate non-Fickian diffusion and a value of n = 0.9 is indicative of erosion controlled and zero-order release kinetics. The non-linear regression was applied using the program Sigma Plot 10.0 for Windows (Systat Software, San Jose, CA, USA) and in all experiments of the present work time (t) was expressed in identical quantity (e.g. time in hours/1 h). Thus, the dimensionality of a and Kp do not depend on b and n, respectively, and this enabled td to be expressed in hour units.

2.3. Experimental design and statistical analysis A face centered cubic experimental design (a central composite design with a = 1) reducing the number of runs (15 instead of 27 for a 33 full factorial design) was applied [16]. Moreover, this design allows sequential experimentation by first using a full 23 factorial design and then seven additional experimental points at intermediate levels of the independent experimental variables (formulation factors) in order to study the curvature. The levels (low, medium and high) of the formulation factors X1 (0.2, 0.6 and 1.0), X2 and X3 (0, 50 and 100%) are presented in Table I, for the fifteen experimental batches of three-layer matrix tablets. The responses characterizing: i) the initial and final stages of drug release (% release at 2 and 12 h, rel2h and rel12h), ii), the overall release profile (shape parameter in the Weibull function, b) and iii) the release mechanism (exponent in the power law model of Peppas, n), were related with the independent experimental variables (formulation factors) by applying multiple linear regression and fitting of second order polynomial equations including two-factor interaction terms:

Table I - Composition and drug release parameters of the experimental three-layer matrix tablets [cumulative percent release at 2 and 12 h (rel2h, rel12h), correlation coefficients (r), shape parameter (b) and time for 63.2 % release (td) in the Weibull release function, and release rate constant (Kp) and exponent (n) in the power law model of Peppas].

Tablet composition rel2h (%)

rel12h (%)

Weibull parameters Power law parameters

X1 X2 X3 b td (h) r n Kp (h-n) r

123456789101112131415

1.01.01.01.00.20.20.20.20.6 0.60.61.00.20.60.6

100100 0 0100100 0 0 50100 0 50 50 50 50

100 0100 0100 0100 0 50 50 50 50 50100 0

7.916.1 4.815.921.921.936.236.622.416.132.410.220.8 9.029.7

17.591.513.283.570.162.997.0100.092.555.593.880.488.534.692.9

0.3111.1730.3981.2080.6650.7860.9541.1761.1780.8191.1270.9551.0380.8721.247

5186.1 6.42969.1 7.7 17.1 11.4 3.9 3.8 5.9 17.5 5.1 13.9 6.8 17.3 4.5

0.9620.9870.9720.9940.9790.9990.9910.9980.9970.9980.9990.9560.9920.9880.998

0.3671.2280.3031.0980.4980.6730.6580.9161.0280.7070.9451.1690.9200.6371.056

0.0620.0780.0400.0790.1460.1380.2230.1840.1100.0940.1420.0450.1140.1010.139

0.9520.9980.9900.9970.9970.9990.9990.9980.9990.9980.9990.9950.9990.9880.999

X1: mass fraction of diltiazem hydrochloride (DH) incorporated in the intermediate layer. X2 and X3: percentage of xanthan gum in the mixture of hy-drophilic polymers used as matrix former of the intermediate and the outer layers, respectively.

Eq. 1

Eq. 2

Three-layer matrix tablets and simple approach of drug release programmingN. Al-Zoubi, S. Malamataris


433

Y = A0 + A1X1 + A2X2 + A3X3 + A4X12 + A5X2

2 + A6X3

2 + A7X1X2 + A8X1X3 + A9X2X3 + E

where A0 is an intercept, A1 to A9 are the coefficients of the respective experimental variables (formulation factors) and their interaction terms, and E is an error term. The obtained polynomial equations (full models) were simplified applying a backward elimination procedure (p to remove = 0.051) and those of highest adjusted coefficient of determination, R2

adj (lowest standard error of estimates, SEE) were selected. Response surface methodology was also applied in order to visualize the effects of the formulation factors on the selected release parameters. SPSS 14.0 program (SPSS, Inc., Chicago, IL, USA) was used for the statistical analysis. The chosen experimental design and the suggested simplified equations were validated by employing three test data points. Three-layer tablets of composition corresponding to these test points were prepared and evaluated on the basis of the selected release parameters. Subsequently, the experimental results were compared with values predicted from the simplified polynomial equations.

II. RESULTS AND DISCUSSION1. Drug release The release profiles for the eight experimental formulas comprising the 23 full factorial design (corner points of the face centered cubic design) and for the seven additional formulas (representing the middle points of the six faces and the center) are shown in Figures 1 and 2, respectively. The results of the release parameters (rel2h, rel12h, b and n) for all the experimental tablets are summarized in Table I. Figures 1 and 2 and Table I show that the drug release became slower with the increased incorporation of DH in the intermediate layer and with the increased percentage of XG, particularly in the outer layers. Tablets containing all the DH in the intermediate layer, X1 = 1.0, and only XG in the outer layers, X3 = 100%, gave very slow release (rel12h = 17.5 and 13.2%, for formulas 1 and 3, respectively). On the contrary, formula 8 with low level of DH mass-fraction in the intermediate layer, X1 = 0.2, and without XG in the matrix former of the outer layers, X3 = 0%, gave very fast release (rel12h = 100.0%). Furthermore, Figures 1 and 2 show a great difference in the shape of the release profiles, corresponding to Weibull shape parameter (b) values between 0.311 and 1.247 (Table I). More specifically the shape of the release profiles varies from parabolic (formulas 1, 3, 6, 7, 8, 9, 11, 13 and 15, Figures 1 and 2) to sigmoid with an accelerated initial release followed by a decelerating phase (formulas 2 and 4, Figure 1) or with a decelerating initial release followed by an accelerated phase (formulas 5, 10 and 14, Figures 1 and 2). Almost linear profile corres-ponds to tablets containing all the DH in the intermediate layer, X1 = 1.0, and binary mixture (1:1) of XG and SA as matrix former in the intermediate and outer layers as well (X2 = X3 = 50%, formula 12 in Figure 2). Also, Table I shows that the shape parameter in the Weibull func-tion (b) decreases, like the release parameters rel2h and rel12h, with the increased percentage of XG (X3) or the decreased presence of SA in the outer layer. SA should result in rapid erosion accompanied by accelerated dissolution during the initial release phase and in values of b > 1. On the contrary XG should be responsible for swelling without erosion resulting in values of b < 0.67. Furthermore, Table I shows that the changing manner of the release exponent (n) with the formulation variables is similar to that of b. 2. Multiple regression analysis and drug release modeling Simplified equations derived by multiple regression analysis and backward elimination for the selected release parameters (rel2h, rel12h, b and n) are:

rel2h (%) = 38.027 - 0.196X2 - 0.121X3 - 21.377(X1)2

+1.229 × 10-3(X3)2 + 0.208(X1).(X2) - 0.112(X1).(X3)

rel12h (%) = 92.834 + 0.643X3 - 4.003 × 10-3(X2)

2 - 4.186 × 10-3(X3)

2 + 0.379(X1).(X2) - 0.991(X1).(X3)

b = 0.989 + 0.955X1 + 2.339 × 10-3X3 - 0.724(X1)2

- 4.51 × 10-5(X2)2 - 2.13 × 10-5(X3)

2 + 3.727 × 10-3(X1).(X2) - 8.31 × 10-3(X1).(X3)

n = 0.873 + 3.966 × 10-3X2 + 5.834 × 10-3X3 + 0.267(X1)2

- 7.10 × 10-5(X2)2 - 6.27 × 10-5(X3)

2 + 3.732 × 10-3(X1).(X2) - 7.64 × 10-3(X1).(X3)

Eq. 3

Figure 1 - Release profiles for the eight experimental formulas comprising the full 23 factorial design (corner points of the face-center cube design). Error bars represent the standard deviation of three replicates.

Figure 2 - Release profiles for the seven additional experimental formulas of the applied factorial design (representing the middle points of the six sides and the center). Error bars represent the standard deviation of three replicates.

Eq. 4

Eq. 5

Eq. 6

Eq. 7



434

All these equations involve either linear or quadratic main effect terms of the formulation factors (X1, X2, X3), justifying their inclusion in the experimental design. Interaction terms are also involved, ex-cept of the (X2).(X3) interaction. The values of R2

adj, and the standard error of estimates (SEE) as well as the ANOVA results (F values and significance level, p) are given in Table II for the full models and the simplified equations. The adjusted coefficient of determination, R2

adj, and F values are higher, while the SEE and the p-values are lower for the simplified equations than for the full models of all the selected release parameters, which justifies the use of simplified equations. Significance of the effects of each formulation factor and of interactions on the drug release parameters (expressed by the significance of the corresponding terms in equation 4 to 7), is given in Table III. More specifically, Table III shows the existence of a significant quadratic effect of drug incorporation in the intermediate layer (X1) on rel2h only (p < 0.001) and significant effects of XG presence in the intermediate layer (X2) on rel2h (linear, p < 0.001) and on rel12h as well (quadratic, p = 0.001). A significant linear effect of XG presence in the outer layer (X3) exists on rel12h only (p = 0.015), while the linear effect on rel2h and the quadratic effects on both rel2h and rel12h are si-gnificant at a probability level lower than 0.05 (p = 0.133, 0.081 and 0.064, respectively). Also, Table III shows that both the linear and quadratic effects of X1 and X3 on the shape parameter b are significant, but only those of X1 at 0.05 probability level (p = 0.009 and 0.010 for X1, while p = 0.160 and 0.152 for X3). The effect of XG percentage in the intermediate layer (X2) on the shape parameter b is quadratic and highly significant (p < 0.001). Furthermore, Table III shows that only the quadratic but not the linear effect of XG percentage in the intermediate layer (X2) on the exponent n is significant at 0.05 probability level (p = 0.004 and 0.082, for the quadratic and linear effect, respectively) while both the linear and quadratic effects of XG percentage in the outer layer (X3) are significant (p = 0.020 and 0.008). The effect of DH-distribution (X1) on n is only quadratic (p = 0.015). Regarding the interactions, Table III shows that the effects of (X1).(X2) and (X1).(X3) interactions are significant on all the release parameters (p = 0.003 and 0.049 for rel2h, p = 0.011 and < 0.001 for rel12h, p = 0.004 and <0.001 for b, and p = 0.020 and < 0.001 for n).

Further discussion of the effects of each formulation factor and of interactions on the selected release parameters follows together with the presentation of the response surfaces derived on the basis of the simplified equations (shown in Figures 3 to 6).

3. Response surface presentation and analysis Figures 3A and 4A show that rel2h and rel12h, characterizing the initial and final stages of drug release, respectively, remarkably de-crease with the increased incorporation of diltiazem HCl (DH) in the intermediate layer (X1). This is in agreement with Figures 1 and 2 and can be explained by the restricted and delayed exposure of DH to the dissolution medium. Also, Figures 3A and 4A show that at low level of DH-incorporation in the intermediate layer (X1 = 0.2) both rel2h and rel12h remarkably decrease with increasing xanthan gum (XG) content in the matrix former of the intermediate layer (X2), while at high level of X1 (X1 = 1), the same increase in X2 leads to negligible changes. However, the decrease in rel2h and rel12h with increasing X2 might be explained by the capability of xanthan gum to form stronger gels and thus having a stronger ability of release retardation than sodium

Table II - Adjusted coefficient of determination, R2adj, standard error of

estimate (SEE) and ANOVA results (F values and significance p) for the full second order models and the simplified equations (reduced models) of the selected responses.

Re-sponse

Second order

R2adj SEE ANOVA results

F p

rel2h

rel12h

b

n

fullreduced

fullreduced

fullreduced

fullreduced

0.8720.9110.8410.8980.9540.9640.9150.939

3.3282.78011.1758.9290.0620.0550.0840.071

11.58524.7559.21025.73033.18154.90417.76931.944

0.007< 0.0010.012

< 0.0010.001

< 0.0010.003

< 0.001

Table III - Statistical significance of the formulation factor effects on the selected release parameters expressed by the terms in the simplified release equations.

Release parameter

Linear Quadratic Interactions

X1 X2 X3 X1 X2 X3 X1X2 X1X3 X2X3

rel2h

rel12h

bn

--

0.009-

< 0.001--

0.082

0.1330.0150.1600.020

< 0.001-

0.0100.015

-0.001

< 0.0010.004

0.0810.0640.1520.008

0.0030.0110.0040.020

0.0490.0110.0040.020

----


Figure 3 - Response surfaces showing the effects of the formulation factors, X1, X2 and X3, on rel2h.



435

alginate [17]. Furthermore, the different manner of rel2h and rel12h changing with X2, at different levels of X1 indicates the existence of a competitive interaction between X1 and X2. A possible explanation of this competitive interaction is the fixed weight of the intermediate layer (400 mg) and the combination of release enhancement due to a decrease in the polymer content (from 340 to 100 mg) with release reduction due to the increased incorporation of DH in the intermediate layer (from 60 to 300 mg). Figures 3B and 4B show a synergistic interaction between X1 and X3 on rel2h and rel12h, which is also expressed in Table I by the slight change of rel2h and rel12h due to increase in X3 at low levels of X1 (formulas 7 and 8), while at high levels of X1 both rel2h and rel12h remarkably decrease by a similar increase in X3 (formulas 3 and 4). The decrease in rel2h and rel12h with the increased percentage of XG in the matrix former of the outer layer (X3) can be explained by the formation of a stronger diffusion barrier of the XG polymer compa-red to SA, while the interaction between X1 and X3 can be explained similarly to the interaction between X1 and X2 described above. As the amount of DH in the intermediate layer increases, the amount of DH in the outer layer decreases and the amount of matrix former in the outer layer increases and hence the effect of the composition of the matrix former becomes more pronounced. Figures 3C and 4C show that, at medium level of X1 (X1 = 0.6), both rel2h and rel12h decrease by the increase in X2 and X3, and that no interaction can be observed between the two variables. Figure 5A-C visualizes the effects of the formulation factors on the shape of the total release profile (expressed as b). Figure 5A shows that, for binary (1:1) mixture of XG and SA in the outer layers (X3 = 50%) and for any level of DH-distribution (X1), b increases in a qua-dratic manner with the decrease in XG in the intermediate layer (X2) or the increase in SA. However, this increase is smaller at high level of X1 (X1 = 1.0) followed by a slight decrease. This is an indication

Figure 5 - Response surfaces showing the effects of the formulation factors, X1, X2 and X3, on the shape of the release profile (b, in the Weibull function).

Figure 4 - Response surfaces showing the effects of the formulation factors, X1, X2 and X3, on rel12h.

Figure 6 - Response surfaces showing the effects of the formulation factors, X1, X2 and X3, on the release exponent (n, in the power law model of Peppas).



436

of competitive interaction between X1 and X2, which is attributed to the different manner of rel2h and rel12h changing with X2 mentioned above. Therefore, the increase in b with the decrease in X2 should be related to the erosion due to the presence of SA. Figure 5B shows the increase in b with the decrease in XG presence in the outer layer (X3), at any level of DH distribution (X1), and the strong interaction between X1, and X3 as distortion of the response surface or greater increase in b at higher level of X1. A possible ex-planation of the change in the shape of the release profile with com-position of the outer layer (X3) and the interaction between X1 and X3 is again the fixed weight of the intermediate (400 mg) and the outer layers (2 × 200 mg) or the increased presence of hydrophilic polymer in the outer layer. The outer layers erode quickly in the example of SA (low level of X3) and as it erodes the dissolution of drug from the intermediate layer accelerates leading to high b values. As the percen-tage of XG in the outer layer increases, its swelling increases while its erosion is reduced and the release lowers and decelerates (leading to low b values). As X1 decreases from 1.0 to 0.2 the amount of DH in the outer layers increases from 0 to 240 mg while the amount of matrix-former decreases from 400 to 160 mg leading to minimization of the matrix-former effect (X3). Figure 5C shows the combined effect of X2 and X3 on b, at medium level of X1 (X1 = 0.60), or that b decreases by the increase in X2 and X3, as is already shown in Figures 5A and B, and that no interaction can be observed between the two variables. Figure 6 illustrates the effects of the formulation factors on the release mechanism expressed as the exponent (n) of the power law model of Peppas. Figure 6A shows the quadratic manner of n increase with the decrease in X2 or the increase in SA in the intermediate layer presumably due to increased erosion. Also, Figure 6A shows that at higher levels of X1 the change of n with X2 is minimized due to the lower polymer content, while n increases with X1 at a high level of X2 (from 0.70 to 0.95), but slightly decreases at a low level of X2 (from 0.94 to 0.89), probably due to quicker erosion in the presence of SA. As mentioned above, a value of n = 0.9 is indicative of erosion-con-trolled release. Therefore, the increase in n with X1 at a high level of X2 is probably caused by the greater effect of DH on erosion when it is combined with XG, comparatively to the readily erodible SA. The changes of n with X1 and X3, at medium level of X2 (X2 = 50%) are presented in Figure 6B and are quite similar to those of b regarding the effect of each factor and the interaction between them (Figure 5B). More specifically, with XG in the outer layers (X3 = 100%), n decreases from 0.75 to 0.54 with the increase in X1 from 0.2 to 1.0 indicating a shift of the release mechanism towards diffusion. This agrees with and confirms the explanation of low and decelerating release by diffusion through strong jelly outer layers composed of XG at high levels of X1 and X3. On the contrary, at a low level of X3, n increases greatly with the increase in X1, and it is attributed to increased erosion due to incorporation of DH in the middle layer when the matrix former of the outer layer is composed of SA (readily erodible). Also, Figure 6B shows that n increases with the decrease in X3, at any level of X1, like the example of b in Figure 5B. The changes of n with X2 and X3, at medium level of X1 (X1 = 60%), shown in Figure 6C, agree with those shown in Figures 6A and B and already described for the two factors, while no interaction between them is seen in the response surface. Similarity between the response surfaces of b and n, Figures 5B-C and 6B-C as well as between the terms in the simplified equations (Equations 6 and 7) is in agreement with the reported existence of linear relation between b and n and the use of b for characterization of the drug release mechanism [18]. Dealing with the determined values, plotted in Figure 7, a highly significant relation (r = 0.894, p < 0.001) was found between b and n, which is described by the equation:

n = 0.8888b - 0.0105

Also, a relation would be expected between the release rate constant in the power law model of Peppas, Kp, calculated by nonlinear regres-sion of the portion from 0 to 60% release, and the reciprocal of time required to achieve 63.2% cumulative release, 1/td, of the Weibull function. Dealing with the experimental data, plotted in Figure 8, good correlation (r = 0.9389, p < 0.001) was obtained only after exclusion of formulas with low rel12h (< 75%). The correlation is described by the equation:

Kp = 0.8308/td - 0.0255

Eq. 8

Figure 7 - Plot of shape parameter (b, in the Weibull function) vs. the release exponent (n, in the power law model of Peppas).

Eq. 9

Figure 8 - Plot of the reciprocal of time for 63.2% release (td-1, calculated

from parameters a and b in the Weibull function) versus the release rate constant (Kp, in the power law model of Peppas).

4. Validation of the simplified equations Predicted (P) and experimental (E) values of the selected release parameters (rel2h, rel12h, b and n) for the three tablet formulations employed as test points are given in Table IV. They show good agreement between the theoretical (predicted) and the experimental values confirming the validity of the suggested simplified equations



437

Table IV - Predicted (P) and experimental (E) values of the selected release parameters (rel2h, rel12h, b and n) for three tablet formulations employed as test points.

Composition rel2h (%) rel12h (%) b n

X1 X2 X3 P E P E P E P E

0.40.80.5

367540

325960

26.313.922.8

31.712.123.7

96.769.687.8

92.972.292.4

1.1990.9231.098

1.2210.9581.194

1.0020.9231.025

1.0450.9310.954


in predicting the release behavior for other three-layer tablets of composition within the experimental domain.

*

Three-layer tablets comprising XG and SA as matrix former can achieve various release profiles (parabolic, sigmoid and linear) depending on the drug distribution and composition of the layers. They offer a simple approach for release programming based on a face-centered cubic experimental design (three formulation factors at three levels).

REFERENCES

1. Takeuchi H., Kawashima Y. - Aqueous based coatings and mi-croparticles in matrix tablet formulations. - In: Aqueous polymeric coatings for pharmaceutical dosage forms, J.W. McGinity Ed., Marcel Dekker Inc., New York, 1997, pp. 549-570.

2. Peppas N.A., Buresa A.P., Leobandunga W. Ichikawab H. - Eur. J. Pharm. Biopharm., 50 (1), 27-46, 2000.

3. Conte U., Colombo P., La Manna A. - Compressed barrier layers for constant drug release in swellable matrix tablets - Proceedings of 6th International Conference on Pharmaceutical Technology, Paris, 2-4 June, 1992, Vol. II, pp. 102-111.

4. Conte U., Maggi L., Giunchedi P., Sangalli M.E., La Manna A. - Multi-layer hydrophilic matrix tablets as controlled delivery systems - Proceedings of the Second European Symposium on Controlled Drug Delivery, Noordwijk Aan Zee, The Netherlands, 1-3 April, 1992, pp. 117-120.

5. Chidambaram N., Porter W., Flood K., Qiu Y. - Formulation and characterization of new layered diffusional matrices for zero-order sustained-release. - J. Control. Rel., 52, 149-158, 1998.

6. Qiu Y., Chidambaram N., Flood K. - Design and evaluation of layered diffusion matrices for zero-order sustained-release. - J. Control. Rel., 51, 123-130. 1998.

7. Abdul S., Poddar S. - A flexible technology for modified release of drugs: multi layered tablets. - J. Control. Rel., 97, 393-405, 2004.

8. Dürig T., Fassihi R. - Evaluation of floating and sticking extended release delivery systems: an unconventional dissolution test. - J. Control. Rel., 67, 37-44, 2000.

9. Bonferoni M., Rossi, S., Ferrari F., Bertoni M., Bolhuis G.,

Caramella C. - On the employment of l carrageenan in a matrix system. III. Optimization of a l carrageenan-HPMC hydrophilic matrix. - J. Control. Rel., 51, 231-239, 1998.

10. Casault S., Slater G. - Systematic characterization of drug release profiles from finite-sized hydrogels, Physica A, 387, 5387-5402, 2008.

11. Langenbucher F. - Linerarization of dissolution rate curves by the Weibull distribution. - J. Pharm. Pharmacol., 24, 979-981, 1972.

12. Costa P., Souza Lobo J.M. - Modeling and comparison of dis-solution profiles. - Eur. J. Pharm. Sci., 13, 123-133, 2001.

13. Kachrimanis K., Malamataris S. - Release of ibuprofen from spherical crystal agglomerates prepared using the solvent-change technique in the presence of Eudragit polymers and from corresponding compacts. - STP Pharma Sci., 10 (5), 387-393, 2000.

14. Dokoumetzidis A., Macheras P. - A century of dissolution research: from Noyes and Whitney to the Biopharmaceutics Classification System. - Int. J. Pharm., 321, 1-11, 2006.

15. Peppas N., Sahlin J. - A simple equation for description of solute release. III. Coupling dissolution and relaxation. - Int. J. Pharm., 57, 169-172, 1989.

16. Myers R., Montgomery D. - Response surface methodology: process and product optimization using designed experiments. - John Wiley and Sons Inc., New York, 2nd ed., 2002, pp. 335-336.

17. Badwan A., Al-Remawi M., Sheikh Salem M. - Universal controlled-release composition comprising xanthan gum and sodium alginate. - European Patent 1510205 A1, 2005.

18. Papadopoulou V., Kosmodis K., Vlachou M., Macheras P. - On the use of the Weibull function for the discernment of drug release mechanisms. - Int. J. Pharm., 309, 44-50, 2006.

ACKNOWLEDGMENTS

This research work received financial support from the Research Council, Applied Science University, Amman, Jordan.

MANUSCRIpT

Received 18 June 2008, accepted for publication 3 September 2008.



three-layer matrix tablets and simple approach of …431 three-layer matrix tablets and simple...

Documents