Evaluation of roll compaction as a preparation method for hydroxypropyl cellulose-based matrix tablets.

Roll compaction was applied for the preparation of hydroxypropyl cellulose (HPC)-based sustained-release matrix tablets. Matrix tablets made via roll compaction exhibited higher dosage uniformity and faster drug release than direct-compacted tablets. HPC viscosity grade, roll pressure, and milling speed affected tablet properties significantly. Roll compaction seems to be an adequate granulation method for the preparation of HPC-based matrix tablets due to the simplicity of the process, less handling difficulty from HPC tackiness as well as easier particle size targeting. Selecting the optimum ratio of plastic excipients and the particle size of starting materials can however be critical issues in this method.

Hydrophilic polymer-based matrix tablets are one of the most widely used controlled-release dosage forms due to their good compatibility, ease of preparation, and economic merit (Heller, Helwing, Baker, and Tuttle;[1] Kudela,[2] Omidian, and Park;[3] Sriamornsak, Thirawong, and Korkerd;[4] Sriamornsak, Thirawong, Weerapol, Nunthanid, and Sungthongjeen[5]). When the polymers in matrix tablets are hydrated, a viscous gel layer is formed around the tablet, thus drug release can be controlled through this gel barrier. The most simple preparation method of matrix tablets is direct compaction. However, granulation is often required before tableting owing to the inadequate physical properties of the starting powder for direct tableting, for example poor flowability. In spite of its inconvenience compared to direct compaction, tableting via granulation is worthy to do, since granulation can obviously increase the effectiveness of the subsequent tableting process by providing better flowability, reducing dust, and preventing segregation. In consequence, the final tablet properties such as dosage as well as crushing strength uniformity can be also improved. Wet granulation is the commonly performed granulation method for this purpose.

Dry granulation by roll compaction can be another approach which has unique advantages over wet granulation. Powder blend is compacted to ribbons by pressure between two-counter rotating rolls, and these ribbons are milled into granules. The produced granules are compacted to tablets in most cases. This method is continuous, simple, cost-saving, environment-friendly, and particularly attractive for moisture and/or heat-sensitive materials, because neither liquid binder nor drying step is necessary (Kleinebudde;[6] Teng, Qiu, and Wen[7]). For those reasons, the use of roll compaction has been growing intensively in recent years. Using hydrophilic polymers in wet granulation often causes handling problems because of their swelling property and tackiness in water. In roll compaction those difficulties can be avoided. Moreover, particle size, one of the crucial factors determining the drug release behavior of matrix tablets (Saeio, Pongpaibul, Viernstein, and Okonogi[8]), can be more easily targeted and controlled in roll compaction than in wet granulation, thanks to the process simplicity.

Several studies report about roll compaction as advantageous process step for the formulation of matrix tablets. Kawashima et al., (Kawashima, Takeuchi, Kino, Niwa, Lin, and Sekigawa[9]), prepared sustained-release matrix tablets of acetaminophen with low-substituted hydroxypropyl cellulose (HPC) via slugging and roll compaction. The drug prepared by roll compaction was applied practically to produce a controlled-release matrix tablet and it was possible to control the drug release by changing the roll compaction pressure. Sheskey et al. (Sheskey, Cabelka, Robb, and Boyce[10]), employed roll compaction to prepare Hydroxypropylmethylcellulose (HPMC) matrix tablets. Roll compaction enhanced material flow and the effct of roll pressure on tablet properties as well as product recycle was found. They also studied the effects of roll compaction equipment scale-up on HPMC matrix tablet properties (Sheskey, Sackett, Maher, Lentz, Tolle, and Polli[11]). Although HPC matrix tablets are widely used for controlled release, relatively few investigations are assessing roll compaction within the manufacturing process. This study was therefore aimed to evaluate roll compaction as a preparation method for HPC-based sustained-release matrix tablets. Caffeine was selected as a model drug possessing short half-life (3-5 h) and the goal was to prepare sustained-release matrix tablets for 8-12 h which is the ideal duration of the caffeine effect in daily life (Tan, Zhao, Moochhala, and Yang[12]). First, the effects of HPC viscosity grade and granule particle size in matrix tablets were investigated. Furthermore, the influence of roll compaction process variables were elucidated and optimized. The matrix tablets prepared by direct tableting and roll compaction were then compared.

Materials and Methods

Materials

Hydroxypropyl cellulose of three different viscosity grades: HPC-L (6-10 mPa.s), HPC-M (150-400 mPa.s), and HPC-H (1000-4000 mPa.s) were provided by Nippon Soda Co. Ltd., (Tokyo, Japan).[13] Caffeine anhydrous was supplied from BASF AG (Ludwigshafen, Germany). Microcrystalline cellulose (Avicel® PH 101) was purchased from FMC, USA. Magnesium stearate was supplied by Novartis, Switzerland. Median particle size of each material measured using laser diffraction was as follows: HPC-L 65 μm, HPC-M 95 μm, HPC-H 135 μm, caffeine anhydrous 45 μm, microcrystalline cellulose 106 μm, and magnesium stearate 32 μm.

Preparation of powder blend

Powder blends were prepared with HPC (20% w/w), caffeine (60%), microcrystalline cellulose (19.5%), magnesium stearate (0.5%), in a turbula mixer (Bachofen, Switzerland). Mixing time was 10 min at 25 rpm.

Roll compaction and milling

In each run 800 g of powder blend was roll-compacted using Chilsonator? IR 220 (Fitzpatrick, Belgium) equipped with smooth rim rolls 22 cm in diameter. The Chilsonator? IR 220 system works with one fixed roll and one floating roll. Compaction pressure is kept constant by changing the roll gap according to the powder flow. For the subsequent milling, tableting, and analysis, the ribbons produced at roll gap 1 ± 0.1 cm were selectively used to assure uniform ribbon thickness and the uniform density of ribbons as well as resulting granules. The ribbons were milled into granules by FitzMill? L1A (Fitzpatrick, Belgium) equipped with bar rotor and 2 mm screen sieve. Roll speed, horizontal feed screw speed and vertical feed screw speed were kept constant in all experiments to 6 rpm, 35 rpm and 150 rpm, respectively. Different roll pressure and milling speed were applied in each step as described in the following section.

Optimization of selected variables

The optimization was performed as two steps: first, optimization of HPC grade and granule particle size composing matrix tablets, second, optimization of roll pressure and milling speed. A 32 factorial design was employed. The levels of each variable are described in Table 1 and Table 2. In each step a total number of nine experiments were carried out. The statistical analysis of data was performed by ANOVA using the software MINITAB? (Minitab Inc., USA). The effects and interactions of variables were calculated and their significances were evaluated at the confidence levels of 90, 95 and 99%.

Table 1

Levels of HPC viscosity grade and granule particle size composing matrix tablets

Table 2

Levels of roll pressure and milling speed

Step 1: HPC viscosity grade and granule particle size composing matrix tablets

The objective of this step was to find the optimal HPC viscosity grade and the most desirable granule particle size to achieve the best flowability of produced granules as well as sustained release of caffeine for 8-12h from the resulting matrix tablets. The effects of HPC grades have been rarely investigated unlike the influence of HPC content in the formulation, therefore it was one of the examined factors in this study. Another input variable studied was the granule particle size composing matrix tablets. It is also one of critical parameters determining tablet properties (Alebiowu, Khanna, and Singh;[14] Campos-Aldrete, and Villafuerte-Robles;[15] Golomb, and Fisher;[16] Saeio, Pongpaibul, Viernstein, and Okonogi[8]), which is a dramatically changing characteristic in the granulation process. The response variables observed were the flowability of produced granules and the drug release of the resulting matrix tablets. Flowability is usually the most common reason of granulation and drug release is a primary concern in the development of sustained-release matrix tablets. Three grades of HPC and three size fractions of roll-compacted granules were tested [Table 1]; 20 bar of roll pressure and 800 rpm of milling speed were applied.

Step 2: Roll pressure and milling speed

Roll pressure and milling speed were optimized to maximize the yield of target granule particle size found in the first step. Roll pressure has been reported as a particularly influential parameter on the size of produced granules (Ingelbrecht, and Remon;[17] Sheskey, and Hendren;[18] Weyenberg, Vermeire, Vandervoot, Remon, and Ludwig[19]). Also, milling speed can affect granule size by changing the opening size of the screen (Fitzpatrick Company Europe).[20] For these reasons, roll pressure and milling speed were investigated with the highest priority over other process parameters. Three levels of roll pressure and milling speed were evaluated [Table 2].

Tableting

Tablets (round, flat, 8 mm in diameter, 300 ± 1 mg weight) were prepared from non-granulated and granulated sample using Zwick? 1478 universal material tester (Zwick GmbH, Germany). The sample was fed manually into the die cavity and compacted uniaxially by a stationary upper punch and a movable die; 7kN of compaction force and 200 mm/s of speed were applied.

Characterization of powder and granules

Bulk, tap density and Carr's index

Bulk and tap densities (ρbulk and ρtab) were measured according to Ph. Eur. 2000 using an apparatus STAV® 2003 (Engelsmann, Germany). The Carr's index calculated according to Equation 1 was used to compare flowability. All measurements were performed in triplicate.

True density

True densities were determined by using a helium pycnometer (AccuPyc? 1330, Micromeritics, USA) with a cell volume of 10 mL. The mean of five parallel measurements was taken.

Particle size and distribution measurement

Practicle size and distribution were determined using laser diffraction (Mastersizer? X, Malvern Instruments, UK). The mean of three measurements was taken. Dry dispersion method was applied for all samples. Mass median diameter (d50) and span value were used to compare particle size and distribution, respectively. Span is the width of distribution based on the 10%, 50% and 90% quantile. A smaller span indicates a narrower particle size distribution.

Determination of ribbon tensile strength

The test was conducted using Universal Force Tester® FMT 310 (Alluris GmbH and Co, Germany). A ribbon piece of 10 x 10 mm was placed in the test plate and the upper stamp was driven down towards the ribbon with 10 mm/min of speed. The tensile strength (N/mm2) was calculated from the breaking force measured (N). The mean of ten measurements was taken.

Evaluation of tablets

Tensile strength

The crushing strength of tablets (n=6) was measured using a tablet hardness tester (Tablet tester 8M®, Dr. Schleuniger Pharmatron AG, Switzerland). The tensile strength of tablet was calculated according to Equation 2.

Is the tensile strength (N/mm2), F the breaking force (N), h the thickness (mm), and D is the diameter (mm), Fell, and Newton.[21]

Porosity

The porosity of tablet (ε, %) was calculated from the tablet weight (M, g), tablet volume (V, cm3), tablet thickness (cm), and the true density of powders (ρ, g/cm3) using Equation 3.

Weight uniformity

Tablets (n=20) were accurately weighed using an analytical balance (AG 204, Mettler Toledo, Switzerland). Coefficient of weight variation (%) was used to evaluate weight uniformity.

Drug release

Dissolution test was performed according to USP paddle method II using an automatic dissolution tester (Sotax AT7®, Sotax, Switzerland) in 900 mL of pH 6.8 phosphate buffer solution for 20 h at 37 ± 0.5°C. Caffeine concentration was measured spectrophotometrically at 274 nm. The mean of six tablets was used as the result.

In the comparison between direct-compacted and roll-compacted matrix tablets under optimized conditions, the dissolution data was analyzed by fitting to zero order (Equation 4), first order (Equation 5), Higuchi (Equation 6), and Korsmeyer-Peppas (Equation 7) equations to clarify drug release mechanism. For Korsmeyer-Peppas equation, first 60% of drug release data was applied.

T is the time, M and M0 are the amount of drug released at T and T0, respectively. K is a constant, F is the fraction of drug released at time t, and n is the release exponent. Korsmeyer-Peppas equation is often applied to describe the drug release behavior from polymeric systems when the mechanism is not well-known or when more than one type of release is involved. The n value in the Korsmeyer-Peppas equation is used to indicate different release mechanisms. In the case of cylindrical tablets, 0.45 ≤ n corresponds to a diffusion control (Fickian release), 0.45 < n ≤ 0.89 to non-Fickian or anomalous, n > 0.89 to relaxational transport (Sriamornsak, Thirawong, and Korkerd;[4] Sriamornsak, Thirawong, Weerapol, Nunthanid, and Sungthongjeen[5]).

The difference in drug release pattern between the direct-compacted and roll-compacted matrix tablets with optimized parameters was compared using the similarity factor (f2) determined by Equation 8. For curves to be considered similar, f2 values should be close to 100. Generally, f2values greater than 50 (50-100) ensure sameness of the two curves (Jeffrey, and Henry[22]).

n is the number of dissolution sample times, R and T are the individual percentages dissolved of reference (direct-compacted matrix tablets) and test dissolution profile (the matrix tablets produced via roll compaction) at each time point (t), respectively.

Swelling and erosion

The swelling and erosion studies were performed by monitoring tablet weight. A tablet was accurately weighed and immersed in the dissolution medium under the same dissolution condition as described above. After 12 h, the swollen tablet was weighed again. The tablet was then dried in an oven at 40°C until constant weight was obtained. The final weight of the dried tablet was determined. The percentage swelling due to liquid uptake was calculated according to Equation 9 and the swelling rate was expressed as percentage swelling per hour. The degree of erosion was calculated using Equation 10. The experiments were carried out in triplicate.

W the initial weight of the tablet, W the weight of swollen tablet, W the final weight of the dried tablet.

Results and Discussion

Effect of HPC viscosity grade and granule particle size composing matrix tablets

Flowability

The starting powder containing different grades of HPC showed fair to passable flowability indicated by Carr's index from 20.6–22.0%. After roll compaction the flowability improved up to good or excellent range corresponding to Carr's index from 7.0–15.4% [Table 3].

Table 3

Flowability of starting powder and roll-compacted granules

The granule particle size strongly influenced flowability. Carr's index significantly (P < 0.05) decreased with increasing particle size. HPC viscosity grade did not show statistically significant effect on flowability.

Drug release

Drug release profiles from the resulting matrix tablets are illustrated in Figure 1.

Figure 1

Effect of HPC viscosity grade and particle size of roll-compacted granules on drug release

Sustained release for 8 h or longer was achieved in all matrix tablets regardless of HPC viscosity grade and granule particle size of the starting material. As expected, the drug release decreased with increasing HPC viscosity grade. This result might be attributed to the different swelling behavior of each viscosity grade as revealed in earlier studies (Campos-Aldrete, and Villafuerte-Robles;[15] Panomsuk, Hatanaka, Aiba, Katayama, and Koizumi;[23] Vázquez, Casalderrey, Duro, Gómez-Amoza, Martínez-Pacheco, Souto, and Concheiro;[24] Viridén, Wittgren, and Larsson[25]), reporting the change of drug dissolution behavior caused by different swelling properties depending on the polymer viscosity. The rate of swelling in water, hydration and dissolution of HPC increases as the viscosity of HPC is lowered. Therefore, drug release rate generally increases (Nippon Soda Co., LTD, HPC Technical Data 62-01).[13] The result of the current study was in good agreement with the report. The tablets containing HPC-L, the lowest viscosity grade, demonstrated the most rapid dissolution; almost 100% of drug was released after 8 h, followed by the tablets containing HPC-M, and HPC-H was the slowest.

The drug release pattern differed when the granule particle size of the starting material to prepare tablets was varied. As shown in Figure 1, the tablets composed of coarser granules showed faster release in all HPC grades. The faster disintegration of tablets with increasing particle size could be an explanation. The coarser particles possessed reduced cohesion due to the smaller surface area. As the result, faster disintegration as well as more rapid drug release was observed in the matrix tablets made of larger particles. However, the effect of particle size became less noticeable in higher viscosity grades, HPC-M and HPC-H, than in HPC-L-containing tablets. The reason could be the difference of the predominant release-controlling factor depending on the HPC viscosity. According to the study of Peppas (Peppas[26]), in fast-swellable matrix systems release kinetics is controlled predominantly by the pore network rather than the polymer. It can be therefore assumed that in higher viscosity resulting in slower swelling of polymer, the release kinetics became more dependent on the polymer rather than pore system which is closely related to particle size. That can well explain why in the present study the more distinguishable effect of particle size was observed in the HPC-L-containing tablets than HPC-M or HPC-H-based formulations. This explanation is also supported by Campos-Aldrete and Villafuerte-Robles.[15] They found that the effect of particle size and HPMC viscosity was clearer when lower content of HPMC was incorporated, but this effect was then minimized as the HPMC content increased up to 20-30%. In our study, the viscosity grade was varied while the polymer content was kept constant as 20%. In both the studies particle size effect showed similar behavior; either in low content or in low viscosity of polymer, particle size had a considerable effect on drug release but when polymer content or viscosity increased, this effect was masked. In the current study, the relative high content of HPC (20%) in the formulations is suggested to mask the effect of particle size.

Based on the results obtained, using HPC-L and roll-compacted granules of 500-710 μm was suggested as the most suitable combination to achieve the goal (the best flowability as well as sustained drug release for 8-12 h). Thus, for further experiments HPC-L was selected as the matrix polymer and 500-710 μm was set as the target granule particle size in roll compaction.

Effect of roll pressure and milling speed

Table 4 presents the particle size, the span value, and the yield of 500-710 μm fraction of the starting powder and the granules prepared at different roll pressures and milling speeds. The effects of roll pressure and milling speed on the particle size and the yield of 500-710 μm are summarized in Table 5.

Table 4

Particle size and yield of 500-710 μm fraction of starting powder and roll-compacted granules

Table 5

Effects of process variables on the particle size and the yield of 500-710 μm

Higher roll pressure generated harder ribbons. The tensile strength of ribbons compacted at 10 bar of roll pressure was 14 ± 1.9 N/mm2, 30 ± 2.7 N/mm2 at 20 bar, and 33 ± 3.6 N/mm2 at 30 bar, respectively. Table 4 and Table 5 show that larger granules were produced at higher roll pressure (P < 0.05). This result confirmed that higher roll pressure resulted in coarser granules by increasing ribbon strength which was more resistant to milling, as also discussed by other researchers (Ingelbrecht, and Remon;[16] Sheskey, and Hendren;[17] Weyenberg, Vermeire, Vandervoot, Remon, and Ludwig[18]). All roll-compacted runs had a smaller span value (from 2.4 to 4.3) than that of the starting powder blend (6.3), indicating that the particle size distribution became narrower after granulation. The portion of desired size fraction was highly increased in roll-compacted granules (between 22.6 and 26.3%) compared to the starting powder (5.3%). Milling speed also affected particle size significantly (P < 0.05). Higher milling speed might lead to finer particles due to the smaller shallow angle of the screen which the produced granules have to pass through. The greatest yield of target particle size was obtained when medium level of roll pressure and high level of milling speed were applied. The interaction between roll pressure and milling speed showed a significant (P < 0.05) effect on the yield of target particle size. The correlation found between roll pressure, milling speed, and the yield of target particle size is illustrated in Figure 2.

Figure 2

Effect of roll pressure and milling speed on the yield of target particle size

Twenty bar of roll pressure and 800 rpm of milling speed were selected as the optimized parameters to maximize the yield of target particle size as suggested in Figure 2. As a result, the yield of 42.2% for 500-710 μm was obtained.

Comparison between direct-compacted and roll-compacted matrix tablets

The matrix tablets prepared by roll compaction under optimized conditions were compared with direct-compacted matrix tablets with regard to tensile strength, weight uniformity, porosity, swelling and erosion [Table 6], and drug release profile [Figure 3].

Table 6

Properties of matrix tablets made by direct compaction and roll compaction

Figure 3

Drug release from direct-compacted and roll-compacted matrix tablets

The tablets made via roll compaction exhibited significantly (Student's t-test, P < 0.01) lower tensile strength, as commonly observed in roll compaction, known as ′loss of tabletability′. Two factors are considered as the main mechanisms of this phenomenon; particle size enlargement by granulation and work-hardening of materials due to limited binding potential (Kleinebudde;[6] Malkowska, and Khan[27]). Also, in our study the enlarged particle size by roll compaction seemed to be a main reason for the reduced tablet strength. The particle size of starting powder was 93 μm, while that of roll-compacted granules was 500-710 μm. This increased particle size might lead to lower tablet strength because of reduced surface area available for bonding.

To elucidate if work-hardening has also attributed to the reduced tablet tensile strength, tablets were prepared with starting powder and roll-compacted granules of the same particle size. Three sieve cuts were examined and the results are presented in Table 7.

Table 7

Tensile strength of tablets made by direct compaction and roll compaction

As seen in Table 7, the tablets made by roll compaction showed no significant (P > 0.05) difference in tensile strength compared to direct-compacted tablets when the initial particle size was the same. Thus, we concluded that the lower tablet tensile strength after roll compaction might be caused by mainly particle size enlargement rather than work-hardening. This observation could be because of the relative low content of Avicel®(19.5%) in the formulation, which is a highly plastic material known to be sensitive to work-hardening. If more various and higher roll pressures were applied, it would become more clear whether work-hardening considerably occurred or not. In the formulation used and under the roll pressure employed in the present study (20 bar), the effect of work-hardening was not obvious. Further optimization of formulation has not been conducted in the current study because the final tablets made by roll compaction exhibited sufficient strength, although it was lower than that of direct-compacted tablets. However, the result obtained shows that in roll compaction the use of highly plastic excipients should be more carefully optimized than in direct tableting or wet granulation to minimize undesired loss of tablet strength. According to the literature, brittle materials are less sensitive than plastic substances not only to work-hardening but also to particle size difference. Wu and Sun[28] found that the tabletability of brittle materials is relatively independent of particle size than plastic deforming substances, since the extensive fragmentation of brittle particles can minimize the effect of initial particle size difference. The influence of particle size on the tabletability observed in this study may diminish if the Avicel® in the formulation were replaced with other brittle excipients. Depending on the drug used as well as the final purpose, a specified study on the critical ratio of commonly used excipients would be necessary and useful. Furthermore, the particle size of starting materials seems to be especially important for successful roll compaction. For plastically deforming materials, smaller particle size of raw materials is suggested as a better choice by some studies (Herting, and Kleinebudde[2930]). However, in case of brittle materials, ′brittle/plastic transition′ (Franks, and Lange;[31] Roberts, and Rowe;[32] Wu, and Sun;[28] Yong, Smith, Dhir, and Kendall[33]), also needs to be considered. Smaller particle size might not be preferable in all cases, since below the material-specific critical size even brittle materials can behave plastically. Selecting particle size can therefore be a crucial issue in roll compaction not only for plastic materials but also in brittle substances. Due to the unique aspect of tableting via roll compaction- exposing starting materials to double compaction unlike direct compaction or wet granulation- a full understanding and prediction on the compaction behavior of materials would be essential.

The uniformity of dosage forms are generally expressed by either weight variation or content uniformity. The smaller the weight variation, higher the content is expected. As shown in Table 6, the roll-compacted matrix tablets exhibited smaller coefficient of weight variation (0.3%) than direct-compacted tablets (0.9%) indicating higher weight uniformity. This result was as expected confirming the advantage of granulation over direct compaction. Particle size enlargement by roll compaction resulted in better flowability. It might lead to more uniform feeding of materials in die cavity during tableting, and consequently smaller weight variation in final tablets was achieved.

Swelling, erosion and drug release

Irrespective of manufacturing methods, all matrix tablets showed prolonged caffeine release for 8-12 h. Roll-compacted matrix tablets had faster dissolution than direct-compacted tablets [Figure 3]. According to the similarity factor calculated (f2 = 35.98), this difference was found to be significant.

Altered swelling behavior by particle size might cause a difference in the drug release rate. As presented in Table 6, roll-compacted matrix tablets showed significantly higher swelling property (P < 0.05) than direct-compacted tablets. The matrix tablets made of roll-compacted granules had higher porosity (P < 0.01) because of the enlarged particle size. This more porous structure of tablets might allow faster water penetration resulting in quicker swelling and dissolution of HPC matrix. As a result, faster drug release was observed in the roll-compacted tablets. This finding was consistent with the study of Alebiowu et al.,[14] and also partially in agreement with the result of Kawashima et al.,[9]. The latter group describes that the matrix tablets containing bigger particles produced at higher roll compaction pressure or slugging pressure exhibited increased drug release rate with increasing median pore radius of the tablet. This observation was explained by Washburn's equation (Equation 11) describing that the water-penetrating speed (d) into a capillary is proportional (Equation 12) to the radius of the capillary (median pore radius).

Where ι is the length of penetration, r is the capillary radius, γ is the liquid surface tension, η is the liquid viscosity, t is time and θ is the contact angle.

Our result can also be interpreted by applying Washburn's equation. The roll-compacted matrix tablets (compacted at 20 bar of roll pressure) possessed bigger pores than direct-compacted tablets (0 bar of roll pressure was applied) due to the enlarged particle size. Therefore, based on Equation 11 and Equation 12, water penetration speed became higher in the roll-compacted matrix tablets contributed by the greater r value. Therefore, tablet disintegration might be facilitated by the rapid water penetration. Also, the weaker interparticle bonding of the roll-compacted matrix tablets indicated by the lower tensile strength [Table 6] could be another reason for faster tablet disintegration. Finally, swelling occurred faster and the drug release rate was higher in the roll-compacted matrix tablets. Because of the hydrophilicity of HPC, erosion was observed in all matrix tablets. After 12 h, similar percent of weight loss was found in all matrix tablets as the same polymer and content was incorporated in all cases.

According to the dissolution data analysis [Table 8], for all matrix tablets the best fit was obtained with Higuchi's model. The n value calculated from the Korsmeyer-Peppas equation was smaller than 0.5, indicating that diffusion is the predominant drug release mechanism. From the results it was noticed that only the drug release rate was changed by roll compaction without affecting the release mechanism.

Table 8

Dissolution data analysis

Conclusion

HPC-based matrix tablets for sustained release could be prepared by roll compaction with acceptable properties. Faster drug release, higher weight uniformity, and reduced tablet strength were observed compared to direct-compacted matrix tablets. HPC viscosity grade, roll pressure, and milling speed were significantly influential on the tablet properties by altering particle size and swelling behavior. From the result it could be considered that roll compaction can be an adequate preparation method for HPC-based matrix tablets in terms of its overall process simplicity than wet granulation as well as the feasibility to control and optimize crucial factors more easily, such as particle size, than direct compaction or wet granulation.