Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties

Due to their simple and low-cost manufacturing process, matrix tablets are pharmaceutical dosage forms most widely used for the formulation of oral controlled release of drugs (1). These systems involve dispersion of one or more drugs in a support structure, which in the majority of the cases is of polymeric nature. Inert matrices are systems that consist of a polymeric porous solid network, which is insoluble and indigestible in the gastrointestinal tract (2). Drug release from these dosage forms occurs by drug diffusion through the pores of the matrices, including the initial pores and the pores that appear when the drug or soluble excipients have been dissolved (3). Ethylcellulose is a water insoluble polymer that has been used in the manufacturing of different solid oral dosage forms. It is used in the preparation of microcapsules and microspheres, as a barrier membrane in reservoirs, and as inert matrix-former for oral 115 Acta Pharm. 63 (2013) 115–129 Original research paper DOI: 10.2478/acph-2013-0005 Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties CELIA CIFUENTES1 ÁNGELA AGUILAR-DE-LEYVA1 ALI R. RAJABI-SIAHBOOMI2 ISIDORO CARABALLO1 1Department of Pharmacy and Pharmaceutical Technology University of Seville, 41012 Seville Spain 2Colorcon Inc. Global Headquarters Harleysville, PA 19438, USA Accepted October 3, 2012 Percolation theory has been applied to study the drug release behaviour in multicomponent inert matrices containing ethylcellulose as a matrix forming polymer. Global influence of major formulation factors such as polymer viscosity, polymer particle size, drug and filler solubility and porosity of the tablets in drug release kinetics has been studied for the first time. Batches containing three viscosity grades of Ethocel™, microcrystalline cellulose (MCC) and lactose as fillers, a lubricant and flow aid mixture and three drugs with different solubility have been manufactured. For some batches, compression pressure was varied in order to obtain matrices with five levels of initial porosity. The behaviour of inert matrices was explained based on the percolation ranges of the main components of the formulation. The effect of the porosity percolation threshold was observed and the existence of a tricoherent drug-polymer-filler system is hypothesized. Keywords: inert matrices, critical points, percolation theory, sustained release * Correspondence; e-mail: [email protected] sustained release dosage forms. The properties of ethylcellulose used as an excipient for obtaining tablets by direct compression has been extensively studied (4–6). Percolation theory studies the critical points of systems. These critical points correspond to the critical occupation probability in a mathematical lattice model. Nevertheless, when the concepts of percolation theory are applied to pharmaceutical tablets, these critical points correspond to the critical concentration of one or more components of the system, including tablet porosity (7). At these points, one component of the system experiences a geometrical phase transition and starts to spread over the whole system, having much greater influence on the properties of the entire system (8). Thus, according to the percolation theory, sudden changes are expected to happen in the properties of the system when a component reaches its percolation threshold. A key concept of the percolation theory is the concept of clusters. A cluster is defined as a group of neighbouring particles of the same component. It is considered coherent or percolating when the group of neighbouring particles extends from one side to the other sides of the system, i.e., it percolates through the entire system. Otherwise, the cluster is called finite or isolated (7, 9). This theory was first applied to the pharmaceutical sciences by Leuenberger and co- -workers in 1987 (9–14). Since then, concepts of the percolation theory have been applied to the study and design of different types of sustained release matrix systems, starting with inert matrices (3, 11–12). According to the percolation theory, an inert matrix must be formulated above the drug percolation threshold (the sum of the initial porosity of the system plus the volumetric fraction of the drug must percolate through the system, since in inert matrices the pores allow water uptake and drug release). Otherwise, a fraction of the drug would remain in the matrix surrounded by the insoluble excipient (11–12). On the other hand, the polymer must also be above its percolation threshold, where polymer networks percolate across the matrix system. In this way, disintegration of the matrix during the release process is avoided and the matrix controls drug release (15–16). Most of these percolation thresholds are around 30–35 % (V/V) (17). As previously reported, an understanding of the critical points of the formulation as well as their relationship with the main factors affecting the behaviour of the system will result in a more robust formulation and faster industrial development. Previous works have estimated the percolation thresholds of different drugs such as morphine hydrochloride (16), naltrexone hydrochloride (15) and dextromethorphan hydrobromide (18) and excipients like ethylcellulose (19) or acrylic polymers (20) in binary and ternary inert matrices. In the present work, typical multi-component controlled release matrix tablets have been studied using ethylcellulose as the matrix forming polymer, two common fillers of different solubility (microcrystalline cellulose and lactose), a lubricant and flow aid mixture and three drugs with different solubility (verapamil hydrochloride, carbamazepine and sodium diclofenac). The presence of these components introduces a higher level of complexity to the study of the behaviour of these systems with respect to the binary matrices previously reported in the literature. The aim of this study is to explain the release behaviour of the studied multicomponent ethylcellulose matrices, based on percolation ranges of the main components of the formulations, i.e., drug, filler and polymer, to provide a general explanation of the kinetic behaviour of the prepared matrices. 116 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. EXPERIMENTAL Materials The following materials were used for the preparation of matrix tablets: Ethocel™ 7 FP (particle size range 5–15 mm), Ethocel™ 10 FP (particle size range 3–15 mm), Ethocel™ 100 FP (particle size range 30–60 mm) (Colorcon Ltd.,UK) as matrix forming polymers, verapamil hydrochloride (mean particle size 120.5 ± 45.2 mm) (Recordatti, Italy), carbamazepine (mean particle size 138.9 ± 47.8 mm) (Fagron, Spain) and sodium diclofenac (mean particle size 212.4 ± 75.9 mm) (Fagron, Spain) as model drugs with different solubility. Solubility values as well as the true density of the drugs were taken from the literature: verapamile (45.4 g L–1 and 1.160 g cm–3); carbamazepine (0.12 g L–1 and 1.330 g cm–3) and sodium diclofenac (20.4 g L–1 and 1.022 g cm–3). Fast Flo lactose (Foremost, USA) and microcrystalline cellulose (MCC) (AVICEL PH 102), (FMC BioPolymer, Ireland) were used as fillers while magnesium stearate (Fagron, Spain) and colloidal silicon dioxide (AEROSIL 200) (Degussa, Germany) were employed as lubricants and flow aid, respectively. Tablet preparation Twenty batches of inert matrices containing 90 mg of verapamil hydrochloride, carbamazepine or diclofenac and an initial porosity between 0 and 5 % were prepared. Table I shows the composition of the studied formulations. Ten more batches containing 90 mg of carbamazepine and 40 % (m/m) of ethylcellulose (5 batches of EthocelTM 7 FP and 5 batches of EthocelTM 10 FP) were prepared with 5 different porosities, ranging from 0 to 25 % (V/V). Table I shows the composition of these formulations as well. All materials, with the exception of magnesium stearate, were blended for 10 minutes in a Turbula mixer (Willy A. Bachofen, Switzerland). Magnesium stearate was added and blended for an additional 2 minutes. A standard eccentric tablet machine (Bonals A-300, Spain) equipped with 9 mm flat punches was used to compress the mixtures in order to obtain tablets with a target weight of 300 mg. The powder blends were weighed and placed manually into the die of the compression machine (60 tablets per lot). The position of the upper punch was adjusted according to the initial porosity and crushing force of the tablets. Tablet characterization The mass of 10 tablets corresponding to each lot was measured using an electronic balance (Sartorius CP 2245). Thickness and diameter of 10 tablets of each lot were measured using a 25-mm digital micrometer (Comecta, SA). The volume of 10 tablets from each batch was calculated according to the following Eq. (1): V=pH(D/2)2(1) where Vis the tablet volume. Hand Dare tablet thickness and diameter, respectively. 117 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. 118 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. Table I. Composition (in percentage) of the different formulations prepared Carbamazepine Diclofenac Verapamil Ethocel 7 FP Ethocel 10 FP Ethocel 100 FP MCC Lactose SiO2Mg- -stearate FP V1a30.0 10.0 59.0 0.5 0.5 FP V2a30.0 20.0 49.0 0.5 0.5 FP V3a30.0 30.0 39.0 0.5 0.5 FP V4a30.0 40.0 29.0 0.5 0.5 FP V5a30.0 50.0 19.0 0.5 0.5 FP V6a30.0 10.0 59.0 0.5 0.5 FP V7a30.0 20.0 49.0 0.5 0.5 FP V8a30.0 30.0 39.0 0.5 0.5 FP V9a30.0 40.0 29.0 0.5 0.5 FP V10a30.0 50.0 19.0 0.5 0.5 FP C1a30.0 10.0 59.0 0.5 0.5 FP C2a30.0 20.0 49.0 0.5 0.5 FP C3a30.0 30.0 39.0 0.5 0.5 FP C4a30.0 40.0 29.0 0.5 0.5 FP C5a30.0 50.0 19.0 0.5 0.5 FP D1a30.0 10.0 59.0 0.5 0.5 FP D2a30.0 20.0 49.0 0.5 0.5 FP D3a30.0 30.0 39.0 0.5 0.5 FP D4a30.0 40.0 29.0 0.5 0.5 FP D5a30.0 50.0 19.0 0.5 0.5 7FP CAa30.0 40.0 29.0 0.5 0.5 7FP CBb30.0 40.0 29.0 0.5 0.5 7FP CCc30.0 40.0 29.0 0.5 0.5 7FP CDd30.0 40.0 29.0 0.5 0.5 7FP CEe30.0 40.0 29.0 0.5 0.5 10FP CAa30.0 40.0 29.0 0.5 0.5 10FP CBb30.0 40.0 29.0 0.5 0.5 10FP CCc30.0 40.0 29.0 0.5 0.5 10FP CDd30.0 40.0 29.0 0.5 0.5 10FP CEe30.0 40.0 29.0 0.5 0.5 aInitial porosity between 0 and 5 % (V/V). bInitial porosity between 5 and 10 % (V/V). cInitial porosity between 10 and 15 % (V/V). dInitial porosity between 15 and 20 % (V/V). eInitial porosity between 20 and 25 % (V/V). The initial porosity (e0) was determined using known volume and weight values according to the following Eq. (2): e0= (Vreal –Vtheor)/Vreal (2) where Vreal is the volume of the tablet and Vtheor is the theoretical volume of the tablet, determined as the sum of volumes of individual components (obtained dividing the mass of each component by its real density). Friability was determined for 20 tablets per lot using a friability tester Erweka TA (Germany) according to USP 26 (21). Crushing forces of the tablets were determined using 10 tablets per lot, with a crushing strength tester Sotax HT1 (Sotax, Switzerland). Drug release Dissolution studies were performed using the paddle method in a USP dissolution apparatus Sotax AT7 smart (Sotax) for 6 tablets from each batch. 900 mL of distilled water at 37 ± 0.5 °C was used as dissolution medium. The stirring speed was fixed at 50 rpm for verapamil hydrochloride and sodium diclofenac and at 100 rpm for carbamazepine. Samples of 5 mL were withdrawn at 0.25, 0.5, 1.0, 1.5, 2, 3, 6 and 8 h. The percentage of drug released was measured using a UV/VIS spectrophotometer Agilent 8453 (USA) at 276 nm for sodium diclofenac, 278 nm for verapamil hydrochloride and 284 nm for carbamazepine. Drug release data obtained for extended release matrices were fitted into the following equations: Eq. (3) (zero-order model), Eq. (4) proposed by Higuchi (22), Eq. (5) proposed by Korsmeyer et al. (23) and Eq. (6) proposed by Peppas and Sahlin (24). Q = k0t(3) Q=bt(4) Q = kktn(5) Q = kdtm+krt2m(6) where Qis the quantity of drug released at time t,k0is the zero order release rate constant in equation 3, bis the Higuchi release rate constant in equation 4, kkis the Korsmeyer kinetic constant in equation 5, tis the release time, nis the Korsmeyer time exponent that depends on the release mechanism and on the shape of the matrix tested (25), kdand krare the diffusion and relaxation rate constants, respectively, and mis the purely Fickian diffusion exponent for a device of any geometrical shape exhibiting controlled release. 119 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. Optimum values for the parameters included in each equation were determined by linear or nonlinear least-squares fitting methods using the SPSS 14.0 software. The determination coefficient (R2) was used to test the applicability of release models. Results of the dissolution studies were employed to estimate the percolation threshold of the different components of the formulations. Phase transition is related to the presence of the percolation threshold of one component of the formulation, and normally produces a significant change in release profiles (11, 26). RESULTS AND DISCUSSION Tablet characterization Results on the tablet diameter and thickness were used to determine the apparent volume of tablets. The weight and density of the different components of tablets were employed to calculate the volume fractions corresponding to each component and the initial porosity of the tablets manufactured. As an example, Table II shows the results (including the crushing strength and tablet friability data) for batches FP C1 to FP C5 (containing carbamazepine and EthocelTM 7 FP). As it can be seen, high hardness values were obtained. These values are considered adequate for obtaining controlled release systems, especially inert matrix systems that must not erode during the drug release process. On the other hand, friability values are adequate, with the exception of batch FP C1, containing only 10 % of ethylcellulose. This fact seems to indicate that the plasticity of ethylcellulose improves the cohesion strength of the tablet. Drug release profiles and release kinetics Thirty batches of tablets were prepared employing three different drugs, two different fillers, a lubricant mixture and three different viscosity grades of EthocelTM. Table I shows the composition and porosity level of each formulation. Results of dissolution studies are discussed below. Estimation of the ethylcellulose percolation threshold Release profiles as well as kinetic parameters of the different batches were employed to estimate the excipient percolation threshold of each formulation. A sudden change in kinetic parameters indicates a change in the release behaviour that could be indicative of a phase transition due to the presence of the percolation threshold of ethylcellulose. Considering the release profiles of the different batches, the percolation threshold of the matrix forming polymer can be situated between 20 and 30 % (m/m) for the matrices containing sodium diclofenac (batches FP D, Fig. 1b). Changes in kinetic parameters shown in Table III, including b(Higuchi equation) ranging from 7.2778 to 5.8827 min–0.5,k0(zero order equation) ranging from 0.6696 to 0.4867 min–1,kk(Korsmeyer model) ranging from 12.219 to 4.9828 min–nand kdand kr(Peppas and Sahlin model) ranging from 15.050 to 10.586 min–0.425 and –0.555 to –0.253 min–0.850, respectively, confirm the critical point observed in the release profiles. The behaviour of the EthocelTM 100 FP matrices (Fig. 2b 120 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. and Table III) is also consistent with this assumption. On the other hand, in the matrices prepared with EthocelTM 7 FP and verapamil hydrochloride (Fig. 2a), no clear change in release profiles can be observed. In the case of inert matrices containing carbamazepine and EthocelTM 7 FP (batches FP C), the lot FP C3 (30 % (m/m) of EthocelTM 7 FP) shows an anomalous behaviour (Fig. 1a). Therefore, the change in release behaviour attributable to the percolation of the polymer is observed between 20 and 40 % (m/m) of EthocelTM 7 FP. The results of kinetic studies confirm this statement, since a fast increase in kinetic constants can be observed when the concentration of ethylcellulose is lower than 40 % (m/m), indicating lower control of drug release. These data support the results previously discussed in this section. Therefore, as a general behaviour, these results suggest that the volumetric fraction of Ethocel, needed to obtain a percolating cluster of this polymer controlling the drug release, ranges from 21.7 (lot FP D2) to 31.8 % (V/V) (lot FP V8). 121 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. Table II. Formulation composition and physico-mechanical characteristics of carbamazepine and EthocelTM 7 FP inert matrices Batch FP C1 FP C2 FP C3 FP C4 FP C5 Massb(g) 0.299 ± 0.006a0.298 ± 0.005a0.298 ± 0.005a0.297 ± 0.004a0.297 ± 0.006a Drug (m/m) 0.30 0.30 0.30 0.30 0.30 Ethocel*(m/m) 0.10 0.20 0.30 0.40 0.50 Lactose (g) 0.590 0.490 0.390 0.290 0.190 Mg-stearate (g) 0.005 0.005 0.005 0.005 0.005 SiO2(g) 0.005 0.005 0.005 0.005 0.005 Volume (cm3)b0.215 ± 0.027a0.221 ± 0.031a0.228 ± 0.029a0.234 ±0.028a0.241 ± 0.026a Total porosity (%) 33.45 32.56 31.50 30.74 29.80 Initial porosity (%) 2.03 2.16 1.97 2.12 1.96 Carbamazepine (%, V/V) 31.42 30.40 29.53 28.61 27.84 Ethocel (%, V/V) 12.22 23.64 34.45 44.51 54.13 Lactose (%, V/V) 53.37 42.88 33.15 23.89 15.23 Mg stearate (%, V/V) 0.64 0.62 0.60 0.58 0.57 SiO2(%, V/V) 0.32 0.31 0.30 0.29 0.28 dens. Carbamazepine 1.330 1.330 1.330 1.330 1.330 dens. Ethocel 1.140 1.140 1.140 1.140 1.140 dens. Lactose 1.540 1.540 1.540 1.540 1.540 dens Mg-stearate 1.092 1.092 1.092 1.092 1.092 dens. SiO22.200 2.200 2.200 2.200 2.200 Hardness (N)c275 ± 0.120a289 ± 0.151a299 ± 0.163a281 ± 0.132a270 ± 0.137a Friability (%)c1.00 0.25 0.57 0.43 0.36 aStandard error of the mean. bn= 10, cn= 20 This fact means that matrices containing less than 21.7 % (V/V) Ethocel are expected to contain isolated clusters of this polymer, which will result in their lower ability to control drug release as well as higher probability of tablet disintegration during the drug release process. Effect of drug substances. Percolation of drug substances Although drug load was kept constant in weight, its volume fraction was not the same for the different drugs and batches, due to different densities of the substances employed as well as to the differences in volume of the tablets obtained. Sodium diclofenac Matrices containing sodium diclofenac showed the highest drug volume fraction, ranging from 33.5 to 37.5 % (V/V) of drug. Addition of the initial porosity of the matrices to the drug volume fraction leads to a total porosity range from 35.1 to 39.5 % (V/V). This could explain why sodium diclofenac matrices show the fastest drug release rate, despite lower water solubility of sodium diclofenac compared to the solubility of verapamil hydrochloride (20.4 vs. 45.4 g L–1) (Figs. 1a,b and 2a). Another argument that could explain faster release from sodium diclofenac matrices is the fact that diclofenac molecules have smaller volume than verapamil molecules, so that higher diffusivity could be expected. 122 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. Fig. 1. Dissolution profiles of batches containing: a) carbamazepine and EthocelTM 7 FP, b) diclofenac sodium and EthocelTM 7 FP. a) b) Verapamil hydrochloride The drug volume fraction for verapamil hydrochloride matrices ranges between 30.3 and 34.3 % (V/V). The total porosity, calculated adding the initial porosity of the matrices to the volumetric fraction of the drug, ranges from 33.1 to 38.1 % (V/V). Therefore, it can be expected that both sodium diclofenac and verapamil hydrochloride are situated above or very close to their percolation thresholds in the studied matrices (17). The obtained drug release profiles seem to confirm this hypothesis. Some of the batches released 100 % of drug during the release assay and even for the lots containing higher amounts of polymer, the release rate shows significant values, suggesting that the drug was not encapsulated by the polymer. Carbamazepine In matrices containing drugs with slow dissolution kinetics, the presence of a fast dissolution filler will mask the percolation threshold of the drug. Therefore, due to the low dissolution rate of carbamazepine and to the presence of lactose, as soluble filler in 123 C. Cifuentes et al.: Critical points in ethylcellulose matrices: Influence of the polymer, drug and filler properties, Acta Pharm. 63 (2013) 115–129. Table III. Kinetic parameters of the formulations prepared maintaining constant the initial porosity Higuchi Zero-order Korsmeyer Peppas and Sahlin Batch b R2k0R2kkn R2kdkrR2 FP V1 6.4425 0.821 0.633 0.9365 0.9753 0.876 0.983 1.755 0.635 0.983 FP V2 5.9139 0.9438 0.2116 0.8545 6.9033 0.349 0.957 4.448 –0.027 0.982 FP V3 3.3795 0.9787 0.1561 0.838 3.5715 0.549 0.969 4.857 0.105 0.975 FP V4 2.745 0.9668 0.1072 0.8214 3.7024 0.529 0.922 6.674 –0.141 0.965 FP V5 1.931 0.978 0.0767 0.8588 5.8415 0.358 0.96 3.714 –0.022 0.986 FP V6 7.6104 0.9723 0.9445 0.9235 3.3973 0.708 0.943 12.911 –0.378 0.976 FP V7 5.109 0.9574 0.4291 0.9003 10.257 0.317 0.858 7.667 –0.115 0.976 FP V8 3.009 0.9434 0.1343 0.7547 7.3082 0.355 0.949 4.161 0.2500 0.961 FP V9 3.2735 0.9542 0.2153 0.8191 8.5677 0.314 0.919 3.917 0.057 0.985 FP V10 3.4972 0.9557 0.2251 0.7859 10.145 0.298 0.967 5.413 –0.066 0.991 FP C1 3.9112 0.9339 0.2863 0.9926 0.6914 0.827 0.993 3.585 0.1050 0.978 FP C2 2.9056 0.9925 0.1191 0.9285 3.2413 0.472 0.969 2.987 0.0610 0.994 FP C3 2.7911 0.8155 0.1194 0.599 24.928 0.123 0.701 6.250 –0.160 0.92 FP C4 1.1898 0.9747 0.0484 0.8972 0.5091 0.655 0.893 1.342 0.0180 0.974 FP C5 1.0199 0.9719 0.0419 0.9153 0.6730 0.56 0.958 0.957 0.0270 0.973 FP D1 7.9813 0.9235 0.7356 0.9716 7.1277 0.481 0.755 15.800 –0.550 0.919 FP D2 7.2778 0.9584 0.6696 0.9264 12.219 0.365 0.77 15.050 –0.555 0.93 FP D3 5.8827 0.9678 0.4867 0.9761 4.9828 0.511 0.913 10.586 –0.253 0.943 FP D4 5.3814 0.9833 0.4455 0.968 4.9381 0.512 0.934 10.08 –0.211 0.939 FP D5 4.6077 0.941 0.3046 0.9743 2.4645 0.609 0.974 5.82 0.035 0.97