Graft copolymers of ethyl methacrylate on waxy maize starch derivatives as novel excipients for matrix tablets: Drug release and fronts movement kinetics

1 Graft copolymers of ethyl methacrylate on waxy maize starch derivatives as novel excipients for matrix tablets: drug release and fronts movement kinetics J.A. Marinich, C. Ferrero*, M.R. Jiménez-Castellanos Departamento de Farmacia y Tecnología Farmacéutica, Facultad de Farmacia, Universidad de Sevilla, C/ Profesor García González nº 2, 41012 Sevilla, Spain * Corresponding author. Phone number: +34-954557218, Fax number: +34-954556085; email address: [email protected] 2 Abstract A previous paper [1] deals with the physicochemical and technological characterization of novel graft copolymers of ethyl methacrylate (EMA) on waxy maize starch (MS) and hydroxypropylstarch (MHS). The results obtained suggested the potential application of these copolymers as excipients for compressed non-disintegrating matrix tablets. Therefore, the purpose of the present study was to investigate the mechanism governing drug release from matrix systems prepared with the new copolymers and anhydrous theophylline or diltiazem HCl as model drugs with different solubility. The influence of the carbohydrate nature, drying procedure and initial pore network on drug release kinetics was also evaluated. Drug release experiments were performed from free tablets. Radial drug release and fronts movement kinetics were also analysed and several mathematical models were employed to ascertain the drug release mechanisms. The drug release markedly depends on the drug solubility and the carbohydrate nature, but is practically not affected by the drying process and the initial matrix porosity. A faster drug release is observed for matrices containing diltiazem HCl compared with those containing anhydrous theophylline, in accordance with the higher drug solubility and the higher friability of diltiazem matrices. In fact, although diffusion is the prevailing drug release mechanism for all matrices, the erosion mechanism seems to have some contribution in several formulations containing diltiazem. A reduction in the surface exposed to the dissolution medium (radial release studies) leads to a decrease in the drug release rate, but the release mechanism is not essentially modified. The nearly constant erosion front movement confirms the behaviour of these systems as inert matrices where the drugs are released mainly by diffusion through the porous structure. Keywords : Ethyl methacrylate-waxy maize starch copolymers; Anhydrous theophylline; Diltiazem hydrochloride; Matrix tablets; Drug release; Fronts movements. 3 1. Introduction Among the different approaches experimented for obtaining sustained-release delivery systems, matrix tablets, particularly if obtained by direct compression, still appear as one of the most efficient and interesting from both the economic and the process development points of view [2]. A variety of polymers is employed as matrix-forming excipients whose nature and characteristics may play an important role and significantly influence the behaviour of these devices. The controlling effect of a polymer material on drug release depends on its physicochemical properties and the way it is mixed during the manufacture of the system. To be more specific the effect is due to the polymer molecular properties, such as the nature of the monomer, type and degree of substitution and whether the polymer is mixed dry or dissolved [3-4]. Three main types of polymers may be used as drug delivery modulators: natural, synthetic and semi-synthetic. Among semi-synthetic polymers, modified starches have been proposed as direct compression excipients for controlled release matrices [5-9]. For instance, substituted amylose leads to hydrophilic matrix systems resistant to biodegradation [10] whereas starch acetate and graft copolymers with vinyl monomers yield hydrophobic matrices [8, 11]. Therefore, native starch properties are substantially modified and the mechanism involved on drug release differs depending on the type of polymer used. In a previous work [1], a novel generation of copolymers combining waxy maize starch (amylose content < 1%) derivatives (MS, MHS) and ethyl methacrylate (EMA) were synthesized by free-radical polymerization and alternatively dried by vacuum oven (OD) or freeze-drying (FD) techniques. The physicochemical and technological properties of these materials were thoroughly evaluated and their performance compared with the raw starches. The grafting of EMA on the carbohydrates backbone introduced hydrophobicity and leaded to amorphization and changes in particle size and morphology that affected the densification behaviour of the original carbohydrates. Graft copolymerization also improved the compactibility and mechanical resistance of native starches, suggesting the potential value of these copolymers as excipients for compressed non-disintegrating matrix tablets. In addition to the polymer nature, drug properties such as polymorphism, degree of cristallinity, particle size, solubility and amount in the pharmaceutical dosage form can 4 influence the release kinetic [12]. From all these variables, the solubility characteristics of the active are particularly important when designing extended-release dosage forms, as they can strongly influence the overall release profile. In fact, an excessively high or an extremely low solubility may give rise to formulation problems. Indeed, it is widely accepted that for both inert and swellable systems, diffusion, preceded by dissolution, may represent the key parameter for controlling drug release [13]. Thus, in the present study two model drugs were selected, anhydrous theophylline and diltiazem HCl, because both are suitable candidates for controlled release formulations [2, 14] and have different solubility. For the above reasons, the aim of this work was to identify the mechanism governing drug release from matrix systems prepared with the novel copolymers of ethyl methacrylate and waxy maize starch derivatives, using as model drugs anhydrous theophylline (a slightly water soluble drug) and diltiazem hydrochloride (a freely water soluble drug). Attention has also been focused on the influence of the carbohydrate nature, drying process and matrix porosity on the mechanistic aspects of drug release from the tested matrices. 2. Materials and methods 2.1. Materials Copolymers synthesised by free radical copolymerization of ethyl methacrylate (EMA) and waxy maize starch (MS) or waxy maize hydroxypropylstarch (MHS) were selected for the study. The preparation of the grafted copolymers has been described in detail by Marinich et al. [1]. The products obtained (waxy maize starch-ethyl metacrylate -MSEMA- and waxy maize hydroxypropylstarch-ethyl methacrylate -MHSEMA-) were dried either in a vacuum oven (OD copolymers) or freeze dried (FD copolymers). The starch-based copolymers (MSEMA) were crushed at 10000 r.p.m. in a knives mill (Retsch ZM 200, Haan, Germany) to obtain powdery samples. Anhydrous theophylline (T) (Roig Farma, Barcelona, Spain, batch 0212030) and diltiazem HCl (D) (Roig Farma, Barcelona, Spain, batch 05F1704) were selected as model drugs. Anhydrous theophylline is slightly soluble in water (100-1000 ml/g) and diltiazem HCl is freely soluble in water (1-10 ml/g) [15]. The batches used have a mean particle size of 162 (96) µm for T and 202 (129) µm for D (determined by sieve analysis). Stearic acid (Estearina ® , Roig Farma, Barcelona, Spain, batch 90003410) was selected as lubricant. Before use, the materials were stored at constant relative humidity (40%) and room temperature (20 ºC). 5 2.2. Methods 2.2.1. Mixtures preparation Drug (anhydrous theophylline or diltiazem HCl) (24%, w/w) and polymer (75%, w/w) were mixed for 15 min using a double-cone mixer (Retsch, Haan, Germany) at 50 r.p.m. After addition of stearic acid (1%, w/w), the mixing procedure was continued for another 5 min. 2.2.2. Preparation of tablets The different mixtures were compacted into tablets using an instrumented [16] single punch tablet machine (Bonals AMT 300, Barcelona, Spain) running at 30 cycles/min. The powders were manually fed into the die (12 mm) to obtain flat-faced compacts of 500 mg weight at a fixed crushing force (140-150 N). Compression data were collected from four tabletting cycles and statistically analysed by one-way analysis of variance (ANOVA) using SPSS ® 14 software. Post-ANOVA analysis was carried out according to Bonferroni’s multiple comparison tests. Results were quoted as significant when p<0.05. 2.2.3. Standard physical test of tablets The physical testing of tablets was performed 24h after production to allow for stress relaxation. The tablet average weight and the standard deviation (SD) were obtained from 20 individually weighed (Sartorius CP224S, Göttingen, Germany) tablets according to European Pharmacopoeia [15]. The thickness of 10 tablets was measured individually using an electronic micrometer (Mitutoyo MDC-M293, Tokyo, Japan). The crushing force [15] of 10 tablets was determined by diametral loading with a Schleuninger-2E tester (Greifensee, Switzerland). Tablet friability [15] was calculated as the percentage weight loss of 20 tablets after 4 min at 25 r.p.m. in an Erweka TA (Heusenstamm, Germany) friability tester. Disintegration testing [15] was performed at 37 ºC in deionized water (800 ml), using an Erweka ZT3 (Heusenstamm, Germany) apparatus without discs. The disintegration times reported are averages of six determinations. 6 2.2.4. Mercury porosimetry measurements Mercury porosimetry runs were undertaken using an Autopore IV 9510 (Micromeritics, Madrid, Spain) porosimeter with a 3 cm 3 penetrometer. A quantity of sample was included in order to obtain 20-90% of mercury intrusion. Working pressures covered the range 0.1- 60000 psi and the mercury solid contact angle and surface tension were assumed to be 130º and 485 mN/m, respectively. Blank runs were also undertaken to correct the data for compressibility effects using the same run conditions and penetrometer type than for the real sample. Total porosity and pore size distribution were determined, in duplicate, for each tablet tested. 2.2.5. Drug release study Release experiments (six tablets) were performed in an automatic dissolution apparatus 2 (Erweka DT 600 HH, Heusenstamm, Germany) [15] as a function of time (12 h). Deionized and deareated water (900 ml) maintained at 37±0.5 ºC was used as dissolution medium and tablets were tested with a paddle rotation speed of 50 r.p.m. Filtered samples (2.8 ml) were withdrawn at specified time intervals via a peristaltic pump (Hewlett Packard 89079AX, Böblingen, Germany). Drug release was monitored continuously on a spectrophotometer (Agilent Technologies 8453 UV-vis, China) at λ = 272 nm for theophylline and λ = 235 nm for diltiazem. In a second series of experiments, special devices [17] were used in order to obtain a rigorous radial release. The tablets were locked between two transparent Plexiglass ® discs by means of four stainless steel screws. The upper disc was carved with concentric circles (from 8 to 20 mm of diameter), so that the tablet could be placed just in the centre (Figure 1). The assembled devices (three replicates) were introduced into the vessels of the dissolution apparatus and tested for 12 h in the same conditions as previously. For both studies, drug release data (M t /M ∞ ≤ 0.6) were analysed according to Higuchi [18] (1), Korsmeyer et al. [19] (2) and Peppas and Sahlin [20] (3) equations: 2 1 tk M M t = ∞ (1) n t tk M M'= ∞ (2) 7 m r m d t tktk M M 2 += ∞ (3) where M t /M ∞ is the drug released fraction at time t (the drug loading was considered as M ∞ ), k and k’ are kinetic constants characteristic of the drug/polymer system, t is the release time, n is the release exponent that depends on the release mechanism and the shape of the matrix tested [21], k d and k r , are the diffusion and relaxation rate constants, respectively, m is the purely Fickian diffusion exponent for a device of any geometrical shape which exhibits controlled release. The optimum values for the parameters present in each equation were determined by linear or nonlinear least-squares fitting methods with SPSS ® 14 software. Besides the adjusted coefficient of determination (r 2adj ), the Akaike Information Criterion (AIC) was used to test the applicability of the release models. The AIC can be defined as p n SSR nAIC ×+       ×= 2ln (4) where n is the number of dissolution data points, SSR is the sum of the squared residuals and p is the number of the parameters of the model. When comparing several models for a given set of data, the model associated with the smallest value of AIC is regarded as giving the best fit [12, 22]. 2.2.6. Fronts movement study The experiment was carried out, in duplicate, in the same conditions as the radial release studies (37 ºC and 50 r.p.m.). However, in order to enhance the visual detection of the fronts, methylene blue (0.004%, w/v) was added to the dissolution medium (900 ml deionized and deareated water). At defined time intervals (10, 30, 60, 90, 120, 180, 240, 360, 480, 600, 720 min), the Plexiglass ® devices were removed from the dissolution apparatus and photographed by means of a camera Sony ® DSC-F717 (Tokyo, Japan) with a 10× digital zoom. Focal distance was kept constant during all measurements. The photographs were analysed by computer using Corel Draw X3 ® software as previously reported [23]. The interface between the matrix and the dissolution medium at the beginning of the experiment (initial diameter) was referred to as position 0. The inward front movement was represented by a negative value, while the outward movement was indicated by a positive one. 8 3. Results and discussion 3.1. Preparation of tablets Although a thorough study on compression characteristics of the copolymers has been reported in a previous paper [1], it is known that the addition of drug and lubricant to direct compression tablet matrices could produce substantial changes in compaction profiles [8, 24-25]. For this reason, some compression parameters [26-27] obtained from the different mixtures are summarised in Table 1. Concerning the influence of the carbohydrate nature, the applied pressure (P) required to obtain tablets from the mixtures with a crushing force of 140-150 N is significantly (p<0.05) larger for MHSEMA than MSEMA. Similar tendency is observed for OD products compared with FD derivatives (p<0.05). This behaviour would be related with the higher ability to plastic deformation reported for MSEMA and FD derivatives [1]. On the other hand, theophylline mixtures show better compression properties than diltiazem mixtures, in agreement with the plastic deformation behaviour of theophylline [24, 28]. Furthermore, the smaller theophylline particles would provide a higher total area for bonding than the larger diltiazem particles [29]. It is also important to mention that the type of drug has a remarkable influence on friction properties. As a consequence of the adhesion to the surfaces of diltiazem, their mixtures show worse lubrication ratio (R), ejection force (F e ) and apparent net work (W an ) values than theophylline mixtures (p<0.05). 3.2. Standard physical test of tablets Results from the physical testing of tablets obtained from the different mixtures are also illustrated in Table 1. All tablets satisfy the requirements specified in European Pharmacopoeia [15] related to weight uniformity test and the tablet thickness ranges between 4.2-4.3 mm, so a decrease of this dimension is found compared with tablets prepared from the pure materials [1]. The crushing force test confirms the values of 140-150 N for all tablets and the friability percentages are lower than 1% [15]. The higher friability values obtained for mixtures containing diltiazem agree with their lower binding capacity and the higher friction suffered by these tablets during the ejection process. However, the addition of both the drug and stearic acid improves the friability compared with tablets elaborated only with copolymers [1]. 9 None of the tablets disintegrated after 30 min. Tablets obtained from theophylline formulations and MSEMA/diltiazem mixtures maintained their physical integrity after the test whereas tablets from MHSEMA/diltiazem mixtures suffered some attrition, probably due to a combined effect of the matrix friability and the hydrophilic character of this copolymer. 3.3. Mercury porosimetry measurements As the porous structure is a useful tool in the prediction of water and drug diffusivity [8, 11, 30-31], results from mercury intrusion-extrusion porosimetry are compiled in Table 2. Matrices containing anhydrous theophylline are characterised by higher porosities and mean pore diameters than tablets containing diltiazem hydrochloride. This behaviour correlates well with the higher pressures applied to obtain tablets from diltiazem formulations. The unimodal pore size distribution profiles (Figure 2) resemble those reported for the copolymers [1] and, according to IUPAC guidelines definitions [32], the systems under study contain mesopores. However, the addition of theophylline or diltiazem and stearic acid diminishes the porosity and mean pore diameter compared with tablets containing only copolymers [1]. This behaviour is in agreement with the reduction in the tablet thickness and is clearer for matrices containing diltiazem. The higher contribution of smaller pores for these tablets is also reflected in the median pore diameters shifting to smaller values. 3.4. Drug release study Figure 3 illustrates the drug release profiles from matrices prepared from the different mixtures. The studies are performed over a period of 12 h and a higher percentage of drug release is observed for matrices containing MHSEMA copolymers compared with the tablets produced from MSEMA derivatives. This behaviour is noticed for both theophylline and diltiazem mixtures and could be explained on the basis of the higher hydrophilic character of MHSEMA copolymers. Echeverria et al. [3] reported also a relationship between the higher hydrophilia of HS copolymers (potato derivatives) and the faster drug release. Concerning the drying process, differences are only found for MHSEMA-diltiazem formulations, where the matrices elaborated with the freeze-dried derivative show slightly higher drug release. 16 [24] M.G. Vachon, D. Chulia, The use of energy indices in estimating powder compaction functionality of mixtures in pharmaceutical tableting, Int. J. Pharm. 177 (1999) 183- 200. [25] P. Di Martino, E. Joiris, S. Martelli, Particle interaction of lubricated or unlubricated binary mixtures according to their particle size and densification mechanism, Il Farmaco. 59 (2004) 747-758. [26] E. Doelker, Physique de la compression. Intérêt et limite des machines instrumentées pour l’optimisation de la formulation, Pharm. Acta Helv. 53 (1978) 1-7. [27] M.J. Järvinen, M.J. Juslin, Evaluation of force-displacement measurements during one-sided powder compaction in a die; the influence of friction with die wall and of the diameter of punches and die on upper and lower punch pressure, Powder Technol. 28 (1981) 115. [28] K.M. Picker, The use of carrageenan in mixture with microcrystalline cellulose and its functionality for making tablets. Eur. J. Pharm. Biopharm., 48 (1999) 27-36. [29] C. Sun, D.J.W. Grant, Effects of initial particle size on the tableting properties of L- lysine monohydrochloride dihydrate powder, Int. J. Pharm. 215 (2001) 221-228. [30] K.A. Mehta, M.S. Kislalioglu, W. Phuapradit, A.W. Malick, N.H. Shah, Effect of formulation and process variables on porosity parameters and release rates from a multi unit erosion matrix of a poorly soluble drug, J. Control. Release. 63 (2000) 201- 211. [31] J. Pajander, B. Van Veen, O. Korhonen, R. Lappalainen, J. Ketolainen, Liquid boundary movements in cylindrical and convex hydrophobic matrix tablets: Effects on tablet cracking and drug release, Eur. J. Pharm. Biopharm. 64 (2006) 167-172. [32] B.D. Zdravkov, J.J. Čermák, M. Šefara, J. Janků, Pore classification in the characterization of porous materials: A perspective, Cent. Eur. J. Chem. 5 (2007) 385- 395. [33] X.C. Fu, G.P. Wang, W.Q. Liang, M.S.S. Chow, Prediction of drug release from HPMC matrices: effect of physicochemical properties of drug and polymer concentration, J. Control. Release 95 (2004) 209-216. 17 [34] I. Bravo-Osuna, C. Ferrero, M.R. Jiménez-Castellanos, Drug release behaviour from methyl methacrylate-starch matrix tablets: Effect of polymer moisture content, Eur. J. Pharm. Biopharm. 69 (2008) 285-293. [35] H. Kojima, K. Yoshihara, T. Sawada, H. Kondo, K. Sako, Extended release of a large amount of highly water-soluble diltiazem hydrochloride by utilizing counter polymer in polyethylene oxides (PEO)/polyethylene glycol (PEG) matrix tablets, Eur. J. Pharm. Biopharm. 70 (2008) 556-562. [36] M. Casas, C. Ferrero, M.R. Jiménez-Castellanos, Graft tapioca starch copolymers as novel excipients for controlled-release matrix tablets, Carbohyd. Polym. 80 (2010) 71- 77. [37] J. Alias, I. Goñi, M. Gurruchaga, Enzymatic and anaerobic degradation of amylose based acrylic polymers, for use as matrices for drug release, Polym. Degrad. Stabil. 92 (2007) 658-666. 18 Table 1. Main compression parameters and tablet test results for the different mixtures: maximum applied upper punch pressure (P), lubrication ratio (R), maximum ejection force (F e ), Juslin’s apparent net work (W an ), average weight, thickness, crushing force (CF), friability (F) and disintegration time (t d ). Values in parentheses represent the standard deviation. Mixtures P (MPa) R F e (N) W an (J) Weight (mg) Thickness (mm) CF (N) F (%) t d (min) OD-MSEMAT 133.29 (0.63) 0.785 (0.016) 184.9 (12.5) 14.0 (0.3) 499 (1) 4.214 (0.011) 146 (6) 0.56 > 30 FD-MSEMAT 108.88 (1.10) 0.786 (0.012) 186.2 (10.7) 12.5 (0.1) 500 (2) 4.246 (0.005) 149 (7) 0.57 > 30 OD-MHSEMAT 147.01 (3.77) 0.745 (0.009) 201.3 (9.8) 15.7 (0.3) 500 (2) 4.302 (0.009) 150 (6) 0.46 > 30 FD-MHSEMAT 125.61 (3.25) 0.772 (0.017) 154.5 (6.1) 14.0 (0.2) 499 (1) 4.281 (0.010) 150 (11) 0.48 > 30 OD-MSEMAD 176.41 (5.82) 0.610 (0.006) 1739.4 (73.1) 15.9 (0.6) 500 (3) 4.212 (0.012) 147 (8) 0.66 > 30 FD-MSEMAD 144.47 (1.98) 0.585 (0.008) 1562.5 (63.6) 14.6 (0.1) 501 (4) 4.323 (0.014) 147 (8) 0.90 > 30 OD-MHSEMAD 202.90 (0.63) 0.635 (0.006) 1481.5 (40.5) 18.7 (0.1) 503 (2) 4.281 (0.009) 148 (7) 0.48 > 30 FD-MHSEMAD 169.53 (1.72) 0.612 (0.004) 1459.4 (20.2) 17.2 (0.2) 499 (2) 4.316 (0.007) 148 (9) 0.67 > 30 19 Table 2. Parameters characterising the porous structure of the matrix systems, calculated by mercury intrusion-extrusion porosimetry. Values in parentheses represent the standard deviation (n = 2). Mixtures Porosity (%) Mean pore diameter (nm) Median pore diameter (nm) OD-MSEMAT 22.4 (0.0) 33.5 (0.1) 805 (2) FD-MSEMAT 22.8 (0.6) 30.5 (0.7) 677 (31) OD-MHSEMAT 23.2 (0.2) 32.6 (0.0) 827 (3) FD-MHSEMAT 23.5 (0.2) 32.2 (0.0) 826 (9) OD-MSEMAD 18.6 (0.3) 28.4 (0.3) 563 (8) FD-MSEMAD 20.1 (0.1) 27.7 (0.4) 507 (2) OD-MHSEMAD 19.2 (0.2) 27.2 (0.0) 604 (2) FD-MHSEMAD 20.2 (0.0) 28.5 (0.1) 600 (6) 20 Table 3. Mathematical modelling and drug release kinetics from EMA copolymers-based tablets. k, Higuchi kinetic constant; n, release exponent; k’, Korsmeyer kinetic constant; k d , Peppas diffusion kinetic constant; k r , Peppas relaxation kinetic constant; r 2 adj , adjusted coefficient of determination; AIC = Akaike Information Criterion. Higuchi equation Korsmeyer equation Peppas equation Mixtures k (min -1/2 ) Fit factors n k’ (min -n ) Fit factors k d (min -0.44 ) k r (min -0.88 ) Fit factors OD-MSEMAT 0.015 r 2adj = 0.9947 AIC = -236.9 0.48 0.019 r 2adj = 0.9973 AIC = -179.2 0.029 1.8 10 -4 r 2adj = 0.9998 AIC = -331.1 FD-MSEMAT 0.015 r 2adj = 0.9939 AIC = -232.9 0.48 0.019 r 2adj = 0.9968 AIC = -174.4 0.030 -2.0 10 -4 r 2adj = 0.9998 AIC = -322.4 OD-MHSEMAT 0.022 r 2adj = 0.9986 AIC = -209.7 0.51 0.022 r 2adj = 0.9980 AIC = -148.4 0.036 -0.8 10 -4 r 2adj = 0.9998 AIC = -257.1 FD-MHSEMAT 0.025 r 2adj = 0.9981 AIC = -185.4 0.53 0.022 r 2adj = 0.9969 AIC = -125.8 0.041 -1.4 10 -4 r 2adj = 0.9998 AIC = -230.7 OD-MSEMAD 0.048 r 2adj = 0.9973 AIC = -107.0 0.51 0.048 r 2adj = 0.9962 AIC = -78.8 0.081 -9.4 10 -4 r 2adj = 0.9999 AIC = -151.9 FD-MSEMAD 0.049 r 2adj = 0.9974 AIC = -106.8 0.51 0.048 r 2adj = 0.9963 AIC = -79.0 0.083 -9.7 10 -4 r 2adj = 0.9999 AIC = -157.9 OD-MHSEMAD 0.079 r 2adj = 0.9996 AIC = -70.1 0.65 0.037 r 2adj = 0.9995 AIC = -53.6 0.080 32.9 10 -4 r 2adj = 0.9998 AIC = -73.1 FD-MHSEMAD 0.115 r 2adj = 0.9998 AIC = -48.0 0.68 0.049 r 2adj = 0.9981 AIC = -30.9 0.139 23.4 10 -4 r 2adj = 0.9997 AIC = -46.3 21 Table 4. Mathematical modelling and radial drug release kinetics from EMA copolymers-based tablets. k, Higuchi kinetic constant; n, release exponent; k’, Korsmeyer kinetic constant; k d , Peppas diffusion kinetic constant; k r , Peppas relaxation kinetic constant; r 2 adj , adjusted coefficient of determination; AIC = Akaike Information Criterion. Higuchi equation Korsmeyer equation Peppas equation Mixtures k (min -1/2 ) Fit factors n k’ (min -n ) Fit factors k d (min -0.44 ) k r (min -0.88 ) Fit factors OD-MSEMAT 0.008 r 2adj = 0.9987 AIC = -313.4 0.49 0.009 r 2adj = 0.9993 AIC = -213.9 0.012 -0.4 10 -5 r 2adj = 0.9998 AIC = -365.1 FD-MSEMAT 0.008 r 2adj = 0.9992 AIC = -323.8 0.50 0.008 r 2adj = 0.9990 AIC = -202.9 0.012 1.1 10 -5 r 2adj = 0.9998 AIC = -368.0 OD-MHSEMAT 0.007 r 2adj = 0.9996 AIC = -347.7 0.49 0.008 r 2adj = 0.9987 AIC = -197.2 0.010 4.8 10 -5 r 2adj = 0.9997 AIC = -355.1 FD-MHSEMAT 0.008 r 2adj = 0.9997 AIC = -349.7 0.54 0.006 r 2adj = 0.9993 AIC = -210.0 0.010 8.1 10 -5 r 2adj = 0.9996 AIC = -343.2 OD-MSEMAD 0.022 r 2adj = 0.9972 AIC = -197.3 0.48 0.027 r 2adj = 0.9976 AIC = -147.2 0.038 -20.2 10 -5 r 2adj = 0.9999 AIC = -279.6 FD-MSEMAD 0.021 r 2adj = 0.9979 AIC = -203.9 0.47 0.028 r 2adj = 0.9982 AIC = -153.2 0.036 -12.4 10 -5 r 2adj = 0.9998 AIC = -252.7 OD-MHSEMAD 0.018 r 2adj = 0.9994 AIC = -263.6 0.48 0.021 r 2adj = 0.9995 AIC = -199.6 0.026 12.0 10 -5 r 2adj = 0.9993 AIC = -258.8 FD-MHSEMAD 0.022 r 2adj = 0.9997 AIC = -244.3 0.50 0.023 r 2adj = 0.9997 AIC = -190.5 0.032 10.0 10 -5 r 2adj = 0.9999 AIC = -275.4 22 Figure legends Figure 1. Photograph of the matrix tablet (OD-MSEMAD at t = 0h) fitted into the Plexiglass  device used for radial drug release and fronts movement experiments. Figure 2. Pore size distribution profile for OD-MSEMAT matrices. Figure 3 . Release profiles (over 12 h) of anhydrous theophylline (open symbols) and diltiazem hydrochloride (closed symbols) from formulated tablets of MSEMA and MHSEMA copolymers. Figure 4 . Radial release profiles (over 12 h) of anhydrous theophylline (open symbols) and diltiazem hydrochloride (closed symbols) from formulated tablets of MSEMA and MHSEMA copolymers. Figure 5 . Water uptake ( ○ ), complete wetting (□) and erosion ( ) fronts positions over time for matrices containing anhydrous theophylline and MSEMA (a) or MHSEMA (b) copolymers. OD derivatives are represented by open symbols and FD derivatives by close ones. Figure 6 . Water uptake ( ○ ) and erosion ( ) fronts positions over time for matrices containing diltiazem hydrochloride and MSEMA (a) or MHSEMA (b) copolymers. OD derivatives are represented by open symbols and FD derivatives by close ones. 23 Figure 1 24 Figure 2 Pore size Diameter (nm) 101001,00010,000100,000 Log Differential Intrusion (mL/g) 0.000.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 OD-MSEMAT 25 Figure 3 0 20 40 60 80 100 120 0 100 200 300 400 500 600 700 Time (min) % drug dissolved OD-MSEMAT FD-MSEMAT OD-MHSEMAT FD-MHSEMAT OD-MSEMAD FD-MSEMAD OD-MHSEMAD FD-MHSEMAD