- Email: [email protected]
Starch–dextrin mixtures as base excipients for extrusion–spheronization pellets
European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521 www.elsevier.com/locate/ejpb
Research paper
Starch–dextrin mixtures as base excipients for extrusion–spheronization pellets S. Almeida Prieto, J. Blanco Me´ndez, F.J. Otero Espinar* Departamento de Farmacia y Tecnologı´a Farmace´utica, Universidad de Santiago de Compostela, Santiago de Compostela, La Corun˜a, Spain Received 2 July 2004; accepted in revised form 20 September 2004 Available online 9 December 2004
Abstract Extrusion–spheronization pellets are generally produced with microcrystalline cellulose (MCC) as the principal excipient, giving rise to particles of very high quality. A number of alternative excipients have been proposed and evaluated, mostly other cellulose derivatives (e.g. different grades of Avicel), or mixtures of MCCs and other excipients. In the present study, we evaluated the possible use of starchCagglutinant mixtures as principal excipients for extrusion–spheronization pellets, with the aim of producing pellets with more suitable properties for certain types of release. We first characterized the different excipients in terms of morphometry and basic physical properties. Subsequently, torque-rheometry was used to characterize the rheology of wetted masses of the different excipients and excipient mixtures, with the aim of determining optimal amount of wetting agent (water). We also evaluated the water absorption and water retention capacities of each excipient. In view of the results obtained, we produced pellets with the different starchCagglutinant mixtures (but without drug), and used image analysis to characterize pellet morphology. Our results show that some of the mixtures—notably starch (corn starch or wheat starch)C20% white dextrin—gave high-quality pellets with good size and shape distributions. In addition, the properties of the different materials tested suggest that it may be possible to obtain pellets with very different properties. q 2004 Elsevier B.V. All rights reserved. Keywords: Pelletization; Extrusion–spheronization; Torque-rheometry; Starchs; Dextins; Image analysis
1. Introduction In recent years, pellets and pellet-based multiparticulate delivery systems have become increasingly important, in view of their numerous technological and biopharmaceutical advantages. The most widely used base excipient for pellet formulations is microcrystalline cellulose (MCC) [1,2], largely because wetted MCC powder masses give mixtures with appropriate rheological properties for extrusion–spheronization. If significant amounts of MCC are replaced by some other component, rheological properties typically become much more dependent on the precise amount of water added, complicating the production * Corresponding author. Address: Departamento de Farmacia e Tecnoloxı´a Farmace´utica, Facultade de Farmacia, Universidad de Santiago de Compostela, Campus Universitario sur s/n. 15782, Santiago de Compostela, La Corun˜a, Spain. Tel.: C34 981 563100x14878; fax: C34 981 547148. E-mail address: [email protected] (F.J. Otero Espinar). 0939-6411/$ - see front matter q 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ejpb.2004.09.010
process, and in particular making it more difficult to control final pellet size and shape. Over the years various cellulose derivatives have been proposed as alternatives to MCC, including diverse grades of Avicel (PH-101, PH-102, PH-103, PH-105, RC-501, RC-581, RC-591, CL-611), Emocel and Unimac (MG-100, MG-200) [3]. Jover et al. [4] reported use of Avicel 955 for pellets with high drug content. Chatlapalli and Rohera [5] used hydroxypropylmethylcellulose (HPMC) or hydroxyethylcellulose (HEC) as principal excipient, with isopropyl alcohol instead of water as wetting agent. One study by Linder et al. [6] used Elcema G 250 and Elcema P 100. Other authors have argued for pellet formulations in which MCC (or other cellulose derivatives) are partially or entirely replaced by other types of base excipient. Gazzaniga et al. [9] obtained pellets using a mixture of 8 parts bcyclodextrin and 1 part MCC. Other authors have obtained pellets using pectin [7,8] or b-cyclodextrin [10] without MCC. Zhou et al. [11–14] used various starch and
512
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
microcrystalline wax derivatives, though they prepared pellets by high-shear granulation, not extrusion–spheronization. Junnila et al. [15,16] obtained pellets with 67.5% MCC and 30% corn starch, though this high proportion of starch led to defects in pellet shape and surface texture. Subsequently this group proposed the use of waxy corn starch as coadjuvant for pellets produced by extrusion–spheronization, on finding that it is possible to replace up to 40% of MCC weight, and thus considerably reducing formulation costs. Of the methods used to evaluate the expected behaviour of excipients during granulation, torque-rheometry is widely used. This method has been used by various authors to characterize the behaviour of diverse MCC grades [17,18], and to relate behaviour to basic physical properties [19–22]. This technique has been used to optimize the proportion of excipient to wetting agent in granulation [23–25], extrusion/ spheronization [23] and other pelleting processes [26]. We have used this technique as the principal method of characterization of pharmaceutical formulations during the kneading stage. This technique has been used by several researchers with the same aim [19,27–30]. Effective characterization can be achieved on the basis of two basic parameters: mean torque and torque amplitude [31,32]. The amount of wetting agent giving maximum mean torque can be taken as an estimate of the optimal amount of water agent for extrusion [30]; similarly, mean torque is useful for comparing the expected behaviour of wetted masses with wetting agent amounts close to the optimal value. This latter approach was used for example by Faure et al. [33] for granulation. The aim of the present study was to evaluate starches and dextrins as substitutes for MCC in the preparation of pellets by extrusion–spheronization. These excipients have the advantage of being much cheaper than MCC, and additionally give pellets with different drug release properties.
2. Materials and methods 2.1. Materials The materials used for pellet production were Avicel PH-101 (FMC International), corn starch (CS) (Meritena 100), wheat starch (WS) (Meritena 200), waxy corn starch (i.e. pregelatinized cross-linked CS; WCS) (Merigel 341), yellow dextrin A-2000 (YD), white dextrin (WD), and talc (J. Escuder, Barcelona, Spain). All starches and dextrins were supplied free by Inigarbe S.L. (Padro´n, A Corun˜a, Spain). The wetting agent used was in all cases distilled water. 2.2. Methods 2.2.1. Characterization of powders Particle size was estimated on the basis of serial sieving (mesh size 20–500 mm) in an Alpine 200LS-AC sieve set. True densities were estimated with a Micropycnometer
MPY-2 helium pycnometer (Quantachrome Corporation, Syosset, NY, USA). Compressibility was determined on the basis of tap and bulk densities determined using a Hosokawa Powder Tester (Micron Corp.). Flow factor (FF) was determined with an ITP-RO-200 automatic rotational shear cell, at three different normal forces. All powders used were photographed under a scanning electron microscope (Leo 435 UP, Leo Electron Microscopy Ltd, UK), following shadowing. 2.2.2. Wetting studies To characterize the responses of the different materials and their mixtures to added water, we determined (a) water absorption capacity, (b) water retention capacity, and (c) wetted mass consistency, as detailed below. Water absorption capacity was estimated using a Lloyd Instruments LR 5K press fitted with a 5-N NLC Lloyd Instruments loadcell from which was suspended a small vessel containing a 5 g sample of the material to be tested. Water absorption capacity was estimated on the basis of increase in sample weight after maintaining the vessel in contact with water for 10 min [34]. Water retention capacity was estimated by the method of Tomer et al. [35,36]. Samples containing different amounts of water were prepared, then kneaded. The wetted mass was then introduced into a small tube (diameter approximately 5 mm) with a filter at its lower end. This tube was introduced into a centrifugation vial, then centrifuged (Sigma 2-15 for 30 min at 3079!g, 6020 rpm), with the water extracted being collected into the vial. Water retention capacity was calculated in two different ways: WRCðLLÞ Z Wpost =Wpre WRCðLSÞ Z Wpost =Wpowder where Wpost is the weight of water remaining after centrifugation, Wpre is the weight of water before centrifugation, and Wpowder is the weight of powder before addition of water. Plots of WRC(LS) and WRC(LL) against weight of water added were in all cases constructed on the basis of five replicate determinations for each data point. 2.2.3. Rheometry Torque-rheometry (Caleva Mixer Torque rheometer, Caleva, UK) was used (a) in stepwise multiple-addition assays to characterize the responses of the different wetted masses to addition of gradual wetting agent, and (b) to characterize wetted masses prior to extrusion–spheronization (measurements of consistency). For stepwise wateraddition assays, initial sample weight was 30 g, and water was added in 1.5-ml steps with 1 min mixing at each step. In some cases it was necessary to use a modified initial sample weight, either because a 30-g sample did not entirely cover the blades of the rheometer, or because swelling on addition of water meant that the sample
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
expanded out of the rheometer vessel. For characterization of wetted masses prior to extrusion–spheronization (only done for those mixtures that were used to obtain pellets), solid mixtures were prepared in a Turbula T2C mixer for 15 min, and the wetted masses were kneaded in a Kenwood Chef Classic orbital mixer with distilled water as wetting agent, again with mixing for 15 min. The wetted masses obtained were introduced into plastic bags that were hermetically sealed for 24 h so that the water distributed as homogeneously as possible. All consistency measurements by torque-rheometry were obtained in triplicate. 2.2.4. Extrusion–spheronization Extrusion was performed with a Caleva Model 10 extruder (6 rpm, hole diameter 1 mm). One hundred grams of the extrudate was then spheronized with a Caleva Model 120 spheronizer (2000 rpm, 15 min), and the resulting pellets were dried to constant weight in a Heraeus oven. Three replicate pellet batches were obtained for each formulation. 2.2.5. Image analysis Pellet size and shape were characterized using an Olympus SZ-CTV stereomicroscope linked to a JVC TK-530 black-and-white videocamera, with perpendicular illumination from above against a black background, provided by a Highlight 2000 coldlight (Olympus Europe). The form factors Vr and Vp [37] were estimated from the digital Images obtained with this system.
3. Results and discussion 3.1. Basic physical characteristics of the excipients Table 1 summarizes the principal physical characteristics of the products used in this study. Avicel PH-101 and WCS show the highest mean particle size, though with greater among-particle variation in size, and more irregular shape. Scanning micrographs (Fig. 1) show the typical elongated fibrous shape of the cellulose particles, and the irregular angular shape of the WCS particles. The remaining materials show lower mean particle size, with particles
513
more rounded and regular in shape. These morphological properties are of course expected to influence rheometric properties. All products showed similar true densities, except WCS for which values were rather lower. Avicel PH-101 showed the lowest tap and bulk densities. WCS showed the lowest compressibility and highest flow factor, indicating that it is the excipient showing the best flow properties of the six; though note that all six excipients show values indicative of poor flow properties. Despite this, these compressibility values are within the range proposed by Chatlapalli and Rohera [5] as acceptable for the production of dense low-friability pellets by extrusion–spheronization. The lower compressibility of WCS and Avicel compared to the other starches and dextrins may be related to the larger particle sizes of the former products, improving flow properties. 3.2. Water absorption and retention capacities of excipient mixtures The results of the water absorption assays are summarized in the raw at the bottom of Table 1. Marked differences are observed among the excipients: the Avicel PH-101 show very high water absorption capacity (as is well known), while yellow dextrin shows very low water absorption capacity. The other four excipients show less marked differences. The results of the water retention assays are summarized in Figs. 2 and 3. The capacity of a material to retain absorbed water depends on whether the water is present in free or bound form. Water retention capacities (WRCsZmoisture retention capacities) were determined as per Tomer and coworkers [35,36], though note that we used tubes of different sizes, so that our results are not directly comparable with theirs. In all cases, WRCs were determined under the maximum load that the pellet was able to bear without excipient appearing in the water recovered by centrifugation. The results for the microcrystalline celluloses and starches are summarized in Fig. 2. The values of WRC(LL) and WRC(LS) for Avicel PH-101 indicate that, in addition to its higher water absorption capacity, this excipient shows a much higher capacity to retain water than
Table 1 Basic physical properties of the excipients considered in the present study Excipients Size (meanGSD)a Real density (g/cm3)a Tapped density/bulk density (g/cm3)/(g/cm3) Compresibility (%) Hausner ratio Flow factor Water uptake (%) a
Corn starch
Wheat starch
Waxy corn starch
Yellow dextrin
White dextrin
Avicel pH-101
11.46G6.24 1.513G0.032 0.769/0.409
24.9G13.01 1.530G0.083 0.756/0.396
52.62G24.89 1.481G0.022 0.720/0.494
20.98G8.03 1.534G0.055 0.786/0.471
30.42G9.91 1.520G0.056 0.790/0.444
53.42G30.7 1.563G0.067 0.462/0.283
46.81 1.88 2.89 69.4
47.62 1.9 5.82 48.3
31.32 1.45 7.51 26.4
40.08 1.66 4.96 8
43.78 1.77 7.87 49.2
38.83 1.63 5.64 244.0
MeanGSD (for six determinations).
514
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 1. Scanning electron micrographs of the excipients considered in the present study.
the starches. Comparing the two starches, WS showed slightly higher water retention capacity than CS. The results for the starch–dextrin mixtures are summarized in Fig. 3. Mixtures with WCS are not included, since even with only 10% WCS no water was recovered on centrifugation (WRC(LL)Z100%), indicating that absorbed water is tightly bound in this excipient. Mixtures of CS or WS with yellow dextrin (YD) showed WRC values very similar to those for CS and WS alone, with the largest differences being for CS with 20% YD. Note that fewer valid data points were obtained for the YD mixtures, because YD has relatively high water solubility and therefore appears in the water extracted by centrifugation even when the absorbed water content of the pellet is low (thus invalidating the assay; see above). Mixtures of CS or WS with white dextrin (WD) showed completely different behaviour (Fig. 3). This excipient significantly improved the water retention capacity of both starches. In addition, mixtures with WD accepted higher water loadings than mixtures with YD (i.e. excipient did not appear in the water extracted by centrifugation even when absorbed water content was high, so that more valid data
Fig. 2. Water retention capacities of the different starches, and the Avicel pH-101.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
515
Fig. 3. Water retention capacities of the different starchCagglutinant mixtures (WD, white dextrin; YD, yellow dextrin).
points could be obtained). The shapes of the WRC(LS) curves obtained for mixtures with WD are also of interest: like the curve obtained for Avicel PH-101 (Fig. 2), these curves showed a well-defined peak, allowing clear identification of the amount of wetting agent giving peak water retention capacity (possibly related to the different stages undergone by the mass during wetting, and thus to the rheological behaviour of the mass). 3.3. Rheometric characterization of wetted masses of the individual excipients Torque-rheometry profiles for stepwise water-addition assays are shown for the different individual excipients in Fig. 4. These profiles are very useful for characterizing the different stages through which the mass passes during the wetting process [27,32]. In studies of MCCs, various authors have found that the optimal amount of wetting agent for extrusion–spheronization pellets is that which gives maximum torque, and thus capillary state.
The profiles obtained for Avicel PH-101 (Fig. 4) are relatively ‘wide’, with maximum torque of about 1.4 nm given by 1.2–1.4 ml/g of wetting agent. This result is similar to that obtained by other authors for Avicel PH-101 [18], and indicates that a mass with suitable consistency can be obtained over a fairly wide range of wetting agent amounts. In addition, it is worth noting that the wetting agent amounts giving maximum torque are those that give maximum water retention capacity (WRC(LS), Fig. 2). The profiles obtained for the two starches are much ‘narrower’ and more jagged, with maximum torque values significantly lower than those obtained for microcrystalline cellulose. In the case of WS, maximum torque amplitudes were very high, in some cases higher than maximum torque. These results indicate that masses obtained from starches alone have inadequate consistency for extrusion–spheronization. In addition, and particularly in the case of WS, the rheometric data indicate that water is not homogeneously distributed within the mass, and that the optimal amount of wetting agent is difficult to determine. Only small amounts
516
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 4. Torque-rheometry profiles for the two native starches (corn starch CS and wheat starch WS) and of Avicel, as obtained in stepwise water-addition assays. Closed symbols—mean torque; open symbols—mean torque amplitude.
of water were required to achieve the capillary state, in agreement with a previous study [38]. These authors found that excipients comprising smooth particles with uniform characteristics, and thus higher tap and bulk densities, require smaller amounts of water to reach maximum torque. Torque-rheometry profiles could not be obtained for WCS, YD or WD, because the torques reached were higher than the maximum measurable with the rheometer. These results indicate that pellet production using starches as base excipient requires the addition of an agglutinating agent. To explore this possibility, we next performed rheometric characterization of mixtures of starches with WCS, YD or WD. 3.4. Rheometric characterization of wetted masses of starch–agglutinant mixtures Torque-rheometry profiles are shown for the different starch–agglutinant mixtures in Figs. 5 and 6. In general, addition of dextrins or WCS to starches reduced the amount of wetting agent required to achieve maximum torque and the capillary state. In addition, the profiles became ‘wider’, indicating that the amount of wetting agent required to achieve the pendular or capillary state is less critical. Maximum torque values increased with increasing amounts of YD, WD or WCS, suggesting that with increasing amounts of these components water was bound more strongly in the mixture, giving rise to more compact and adhesive masses. More specifically, the addition of 10 or 20% WCS markedly increased mean torque values of both WS (Fig. 5)
and CS (Fig. 6), to values more than twice as high as those recorded for Avicel PH-101 with 20% WCS. These high torque values indicate wetted masses with high consistency, high mechanical resistance, and probably low plasticity; these properties may not be suitable for extrusion– spheronization. However, the relative position of torque amplitude suggests that water is homogenously distributed within these masses. The high mean torque values of these masses may be related to their high water retention capacity (WRC(LS)Z100%), indicating strong binding of water. The addition of 20% YD likewise markedly increased mean torque values, though to a lesser extent than WCS. The mean torque and torque amplitude values obtained for starchCYD (unlike those obtained for starchCWCS) indicate that water was not homogeneously distributed within the mass, with torque amplitude in some cases even higher than mean torque. Note also that maximum torque was obtained with a smaller amount of water in mixtures containing 20% YD compared to mixtures containing 10% YD. The addition of WD (especially 20% WD) again markedly increased mean torque values. The torque values obtained for WSCWD mixtures were similar to those obtained for Avicel PH-101, though maximum torque was obtained with smaller amounts of water. The rather ‘wide’ profiles and the relative positions of the torque amplitude peaks suggest that water is homogeneously distributed within these wetted masses. In this case, and as with Avicel PH-101, the range of added water giving maximum torque broadly coincides with that giving peak water retention capacity.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
517
Fig. 5. Torque-rheometry profiles for the different wheat starch (WS)Cagglutinant mixtures, as obtained in stepwise multiple-addition assays. Values shown are mean torque.
In general, and in view of the mean torque and torque amplitude curves obtained for the two starches alone (Figs. 4, 5 and 6), water seems to be more homogeneously distributed within CS wet masses than within WS wet masses. This suggests that CS should be more appropriate for extrusion–spheronization. 3.5. Extrusion–spheronization In view of the torque-rheometry profiles obtained for the different starchCagglutinant mixtures we selected the optimal amount of wetting agent for each mixture, and proceeded with extrusion–spheronization. The amount of wetting agent added was that which gave maximum torque (i.e. capillary state), reported by various authors to be optimal for extrusion–spheronization with MCCs;
additionally, we produced pellets using amounts of wetting agent close to this theoretical optimum. Note that no drug was included in the mixtures, and it should be borne in mind that drugs may modify the properties of the wetted mass, and thus affect its extrusion– spheronization. To evaluate the capacity of the excipient mixtures to form spherical pellets, the mixture was placed in an orbital kneader, and the entire amount of wetting agent was added at once. After kneading for 15 min, it was placed in a hermetically sealed vessel and left to equilibrate at room temperature for 24 h. We then determined the consistency of the mixture by torque-rheometry, and proceeded to extrusion–spheronization. In general, all mixtures assayed showed a strong decline in consistency with respect to the values obtained in the
518
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 6. Torque-rheometry profiles for the different corn starch (CS)Cagglutinant mixtures, as obtained in stepwise multiple-addition assays. Values shown are mean torque.
corresponding stepwise water-addition assays. This is basically attributable to the addition of all the wetting agent at once, and to the use of an orbital kneader, favouring the formation of small aggregates that behave like a granulate. Once the mixture is placed in the rheometer chamber and the blades of the rheometer begin to move, the aggregates are brought into movement but retain their integrity. As a result, the rheometer is basically measuring friction between aggregates, and between aggregates and the rheometer, rather than the true (and much higher) consistency of the wetted mass. The proof of this is that despite obtaining very low mean torques, pellets were successfully obtained from mixtures prepared with various amounts of wetting agent, not necessarily those showing highest mean torque in the multiple-addition assays. The quality of the pellets obtained depended both on the composition of the mixture and the amount of wetting agent. Table 2 summarizes these properties for the different
mixtures used, and lists shape factors for the pellets obtained. The amount of water needed to produce pellets was in all cases lower than that giving maximum torque (and thus capillary state and peak water retention). In fact, the optimal amount of water was better predicted by peak torque amplitude. This was particularly true of mixtures containing WD. The rightmost columns of Table 2 summarize the principal morphological characteristics of the pellets obtained. Pellets were also obtained with other mixtures not shown in Table 2, but these either had very small particle diameter (indicative of insufficient wetting), or formed large agglomerates (indicative of excessive wetting). Mixtures of native starches (CS or WS) with WCS gave pellets with deficient morphological characteristics (Table 2). However, in no case were these pellets spherical, with most lots showing a significant proportion of irregularly shaped pellets. The values of the form factors
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
519
Table 2 Morphometric characteristics of pellets obtained from the different starchCagglutinant mixtures Compositiona
Corn 1b 2b 3 4b 5 6 7 8 9 10 Wheat 11 12b 13 14 15b 16b 17 18 19 a b
Starch (%)
WCS (%)
85 85 75 75 75 75 75 75 75 75
10 10 20 20
YD (%)
WD (%)
20 20 20
85 10 85 10 75 20 75 20 75 20 75 20 75 75 100% Avicel PH-101
20 20 20
20 20
ml of water per g of powder
Vr (meanGSD) Popn 1
0.40 0.45 0.35 0.40 0.25 0.30 0.35 0.35 0.40 0.45
11.08G1.48 8.17G0.05 5.50G0.49 12.19G0.76 6.12G1.07 5.53G1.87 7.28G1.26 3.98G0.91 3.20G0.51 4.27G0.73
0.35 0.40 0.30 0.35 0.20 0.25 0.30 0.35 1.05
4.97G0.48 11.08G1.48 6.71G1.35 5.72G2.43 9.91G0.26 9.68G0.51 6.22G2.66 2.44G0.95 3.32G0.23
Vp (meanGSD) Popn 2
9.72G3.31 5.50G0.28
8.29G2.26
12.80G5.98
5.27G2.27
Popn 1
12.79G0.96 11.62G2.44 8.21G0.72 14.57G2.49 8.34G0.60 7.92G0.68 12.38G2.11 7.72G1.03 7.14G0.31 7.19G0.68 7.97G0.38 12.79G0.96 10.63G0.29 11.44G4.64 9.90G0.56 10.75G0.44 6.71G0.44 6.61G0.40 7.09G0.09
Popn 2
12.20G3.74
10.74G1.78 9.55G1.42 8.83G1.19
Feret (meanGSD)
1144.5G42.5 1101.3G111.2 1008.6G24.8 1252.6G193.2 762.9G120.5 1161.7G141.9 1538.7G358.4 939.0G155.7 998.1G127.4 1415.4G66.0
11.64G1.80
954.7G100.6 1144.5G42.5 1041.6G43.8 19.71G10.56 967.8G85.1 1021.0G85.6 1085.9G77.9 8.44G0.71 993.0G53.9 8.56G0.87 970.3G65.0 824.6G55.4
In all mixtures talc was added (5%) as lubricant agent. Arithmetic mean and SD, because a multigaussian fit was impossible.
Vr and Vp obtained for pellets prepared from mixtures with low water contents indicate predominantly cylindrical shape, with more or less angular edges. In general, increasing the amount of added water led to an increase in pellet size and in Vr (e.g. VrZ5.5 for formulation 3, 12.2 for formulation 4); higher Vr values are indicative of cylindrical shapes with high aspect ratio, or fusiform shapes. This is basically attributable to the rheological characteristics of these mixtures. Mixtures of native starches and WCS gave wet masses with high consistency and low plasticity which can be extruded, but which are not readily spheronized, so that particle shape remains markedly non-spherical. The observed increase in pellet size and aspect ratio with increasing amount of added water is attributable to the fact that wetter masses are more elastic and have higher consistency, so that the extrudates are longer and do not break so easily during spheronization, giving rise to cylindrical particles. Mixtures of CS or WS with YD showed very low mean torque values, so that pellets could only be obtained with mixtures containing a high proportion of YD (20%). Mixtures with only 10% YD had very low consistency, and could not be effectively extruded. The pellets obtained with CS or WS plus YD were not spherical: a high proportion of the pellets prepared with WS showed irregular shape, probably due to the heterogeneous distribution of wetting agent within the wetted mass, as indicated by the rheometry results. The pellets prepared with CS were somewhat better, but still markedly non-spherical.
Mixtures of CS or WS with WD likewise showed low mean torque, so that pellets could again only be obtained with mixtures containing 20% WD. However, the pellets obtained with these mixtures (whether CS, formulations 8–10, or WS, formulations 17 and 18) showed good sphericity. Indeed, many of the starchCWD formulations gave particles with Vr values close to those obtained for Avicel. The CSCWD mixtures (like the CSCWCS and CSCYD mixtures) gave increasing pellet size with increasing amount of wetting agent. This relationship was not so clearly evident in the case of the WSCWD mixtures, probably because of the different water absorption and retention capacities of these starches, and the more heterogeneous distribution of water within the wetted masses. Following our previously published morphological classification on the basis of Vr values [37], none of the mixtures tested gave spherical pellets except WSCWD and CSCWD. Pellets prepared, with similar shape and surface texture to pellets prepared with Avicel PH-101. Mixtures including WCS as agglutinant gave largely spherical pellet populations but with significant proportions of other cylindrical and fusiform shapes. Pellet properties were better for mixtures with 10% WCS than for mixtures with 20% WCS. In general, mixtures with YD or WCS gave worse pellet morphometric properties than mixtures with WD.
520
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
4. Conclusions The present results indicate that it is possible to prepare pellets with acceptable morphometric characteristics by extrusion–spheronization of mixtures in which the principal excipient is a starch, not a microcrystalline cellulose. The deficient rheological properties of WS and CS mean that their use as principal excipient requires the inclusion of other excipients that act as agglutinants and improve mixture consistency, resistance and plasticity. Torque-rheometry characterization of wet masses during stepwise addition of wetting agent, and studies of water retention capacity, were useful tools for comparing the behaviour of the different starch–agglutinant mixtures, and especially for selecting the optimal amount of wetting agent. However, consistency of equilibrated wet masses (prepared with the predicted optimal amount of wetting agent) was not an effective predictor of extrusion–spheronization efficacy, due to methodological differences with respect to the stepwise water-addition assay, and to the addition of wetting agent all at once rather than stepwise. Unlike microcrystalline cellulose mixtures (in which the optimal amount of wetting agent for extrusion– spheronization is that giving maximum torque), the optimal amount of wetting agent for the starch and starch–dextrin mixtures tested in the present study was lower than that giving maximum torque. In fact, the optimal amount of wetting agent was more accurately predicted by maximum torque amplitude. In general, CS mixtures needed slightly larger amounts of wetting agent than WS mixtures. The best (i.e. most spherical) pellet morphometries were obtained with WSC20% WD, or CSC20% WD. Pellets obtained with WSCWCS or CSCWCS (WCS at 10 or 20%) had poorer morphometries, but our results suggest that better results might be obtained with WCS at proportions of less than 10%.
References [1] A.D. Reynolds, A new technique for the production of spherical pellets, Manuf. Chem. Aerosol News June (1970) 40–43. [2] J.W. Connie, H.R. Hadley, Preparation of small solid pharmaceutical spheres, Drug Cosmet. Ind. April (1970) 38–41. [3] J.M. Newton, A.K. Chow, K.B. Jeeba, The effect of excipient source on spherical granules made by extrusion/spheronization, Pharm. Technol. Int. 4 (1992) 52–58. [4] I. Jover, F. Podczeck, J.M. Newton, Evaluation, by a statitically designed experiment, of an experimental grade of microcrystalline cellulose, avicel 955, as atechnology to aid the production of pellets with high drug loading, J. Pharm. Sci. 85 (1996) 700–705. [5] R. Chatlapalli, B.D. Rohera, Physicall characterization of HPMC and HEC and investigation of their use as pelletization aids, Int. J. Pharm. 161 (1998) 179–193. [6] H. Linder, P. Kleinebudde, Use of powdered cellulose for the production of pellets by extrusion/spheronization, J. Pharm. Pharmacol. 46 (1994) 2–7.
[7] I. Tho, E. Andersen, K. Dyrstad, P. Kleinebudde, S.A. Sande, Quantum chemical descriptors in the formulation of pectin pellets produced by extrusion/spheronization, Eur. J. Pharm. Sci. 16 (2002) 143–149. [8] I. Tho, S.A. Sande, P. Kleinebudde, Pectinic acid, a novel excipient for production of pellets by extrusion/spheronization: preliminary studies, Eur. J. Pharm. Sci. 54 (2002) 95–99. [9] A. Gazzaniga, M.E. Sangalli, G. Bruni, L. Zema, C. Vecchio, F. Giordano, The use of b-cyclodextrin as a pelletization agent in the extrusion/spheronization process, Drug Dev. Ind. Pharm. 24 (1998) 869–873. [10] M.E. Villar-Lopez, L. Nieto-Reyes, S. Anguiano-Igea, F.J. OteroEspinar, J. Blanco-Me´ndez, Formulation of triamcinolone acetonide pellets suitable for coating and colon targeting, Int. J. Pharm. 179 (1999) 229–235. [11] F. Zhou, C. Vervaet, J.P. Remon, Matrix pellets based on the combinations of waxes, starches and maltodextrins, Int. J. Pharm. 133 (1996) 155–160. [12] F. Zhou, C. Vervaet, J.P. Remon, Influence of processing on the characteristics of matrix pellets based on microcrystalline waxes and starch derivates, Int. J. Pharm. 147 (1997) 23–30. [13] F. Zhou, C. Vervaet, D.L. Massart, B. Massart, J.P. Remon, Optimization of the processing of matrix pellets based on the combination of waxes and starch using experimental design, Drug Dev. Ind. Pharm. 24 (1998) 353–358. [14] F. Zhou, C. Vervaet, M. Schelkens, R. Lefebre, J.P. Remon, Bioavailabity of ibuprofen from matrix pellets based on the combination of waxes and starch derivates, Int. J. Pharm. 168 (1998) 79–84. [15] R. Junnila, J. Heina¨ma¨ki, J. Yliryuusi, Effects of surface-active agent on the size, shape and hardness of MCC/maize starch pellets prepared by an extrusion–spheronization technique, STP Pharm. Sci. 8 (1998) 221–226. [16] R. Junnila, P. Palviainen, J. Heina¨ma¨ki, P. Mylla¨rinen, P. Forssell, J. Yliryuusi, Waxy corn starch: a potent cofiller in pellets produced by extrusion–spheronization, Pharm. Dev. Technol. 5 (2000) 67–76. [17] M.D. Parker, R.C. Rowe, Source variation in the wet massing (granulation) of some microcrystalline celluloses, Pow. Technol. 65 (1991) 273–281. [18] M.D. Parker, P. York, R.C. Rowe, Binder–substrate interactions in wet granulations 3: the effect of excipient source variation, Int. J. Pharm. 80 (1992) 179–190. [19] R.C. Rowe, G.R. Sadeghnejad, The rheology of microcrystalline cellulose powder/water mixer-measurement using a mixer torque rheometer, Int. J. Pharm. 38 (1987) 22–229. [20] P. Luukkonen, T. Schaefer, L. Helle´n, A.M. Juppo, J. Yliruusi, Rheological characterization of microcrystalline cellulose and silicified microcrystalline cellulose wet masses using a mixer torque rheometer, Int. J. Pharm. 188 (1999) 181–192. [21] P. Luukkonen, J. Rantanen, K. Ma¨kela¨k, E. Ra¨sa¨nen, J. Tenhunen, J. Yliruusi, Characterization of wet massing behaviour of silificied microcrystalline cellulose and a-lactose monohydrate using nearinfrared-spectroscopy, Pharm. Dev. Technol. 6 (2001) 1–9. [22] P. Luukkonen, T. Schaefer, F. Podczeck, M. Newton, L. Helle´n, J. Yliruusi, Characterization of microcrystalline cellulose and silicified microcrystalline cellulose wet masses using a powder rheometer, Eur. J. Pharm. Sci. 13 (2001) 143–149. [23] S.A. Schildcrout, Rheology of pharmaceutical granulations, J. Pharm. Pharmacol. 36 (1984) 502–505. [24] V. Janin, M.D. Parker, R.C. Rowe, Monitoring production scale granulation processes using an instrumented mixer torque rheometer, STP Pharm. 6 (1990) 233–237. [25] M. Hariharan, E. Mehdizadeh, The use of mixer torque rheometry to study the effect of formulations variables on the properties of wet granulations, Drug Dev. Ind. Pharm. 28 (2002) 253–263.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521 [26] J. Kristensen, T. Schaefer, P. Kleinebudde, Direct pelletization in a rotary processor controlled by torque measurements. Influence of process variables, Pharm. Dev. Tech. Ind. 5 (2000) 247–256. [27] M.D. Parker, R.C. Rowe, N.G. Upjohn, Mixer torque rheometry: a method for quantifying the consistency of wet granulations, Pharm. Technol. Int. September (1990) 50–64. [28] B.C. Hancock, P. York, R.C. Rowe, M.D. Parker, Characterization of wet masses using a mixer torque rheometer. 1. Effect of instrument geometry, Int. J. Pharm. 76 (1991) 239–245. [29] B.C. Hancock, P. York, R.C. Rowe, Characterization of wet masses using a mixer torque rheometer. 2. Mixing kinetics, Int. J. Pharm. 83 (1992) 147–153. [30] R. Chatlapalli, B.D. Rohera, Rheologycal characterization of diltiazem HCL/cellulose wet masses using a mixer torque rheometer, Int. J. Pharm. 175 (1998) 47–59. [31] B.C. Hancock, P. York, R.C. Rowe, An assessment of substrate– binder interactions in model wet masses. 1. Mixer torque rheometry, Int. J. Pharm. 102 (1994) 167–176. [32] R.C. Rowe, M.D. Parker, Mixer torque rheometry an update, Pharm. Technol. Eur. March (1994) 27–34.
521
[33] A. Faure, I.M. Grimsey, R.C. Rowe, P. York, M.J. Cliff, Process control in a high shear mixer–granulator using wet mass consistency: the effect of formulation variables, J. Pharm. Sci. 88 (1999) 191–195. [34] F.J. Otero-Espinar, M.B. Delgado-Charro, S. Anguiano-Igea, J. Blanco Me´ndez, Using a tensile test to study the bioadhesive and swelling properties of poluymers, in: A. Mun˜oz-Ruiz, H. Vromans (Eds.), Data Acquisition and Measurement Techniques, EEUU, Illinois, 1998, pp. 295–342. [35] G. Tomer, J.M. Newton, A centrifuge technique for the evaluation of the extent of water movement in wet powder masses, Int. J. Pharm. 188 (1999) 31–38. [36] G. Tomer, H. Patel, F. Podczeck, J.M. Newton, Measuring the water retention capacities (MRC) of different microcrystalline cellulose grades, Eur. J. Pharm. Sci. 12 (2001) 321–325. [37] S. Almeida-Prieto, J. Blanco-Me´ndez, F.J. Otero-Espinar, Image analysis of the shape of granulated powder grains, J. Pharm. Sci. 93 (2004) 621–634. [38] R. Chatlapalli, B.D. Rohera, Study of effect of excipient source variation on rheological behaviour of diltiazem HCL-HPMC wet masses using a mixer torque rheometer, Int. J. Pharm. 238 (2002) 139–151.
Research paper
Starch–dextrin mixtures as base excipients for extrusion–spheronization pellets S. Almeida Prieto, J. Blanco Me´ndez, F.J. Otero Espinar* Departamento de Farmacia y Tecnologı´a Farmace´utica, Universidad de Santiago de Compostela, Santiago de Compostela, La Corun˜a, Spain Received 2 July 2004; accepted in revised form 20 September 2004 Available online 9 December 2004
Abstract Extrusion–spheronization pellets are generally produced with microcrystalline cellulose (MCC) as the principal excipient, giving rise to particles of very high quality. A number of alternative excipients have been proposed and evaluated, mostly other cellulose derivatives (e.g. different grades of Avicel), or mixtures of MCCs and other excipients. In the present study, we evaluated the possible use of starchCagglutinant mixtures as principal excipients for extrusion–spheronization pellets, with the aim of producing pellets with more suitable properties for certain types of release. We first characterized the different excipients in terms of morphometry and basic physical properties. Subsequently, torque-rheometry was used to characterize the rheology of wetted masses of the different excipients and excipient mixtures, with the aim of determining optimal amount of wetting agent (water). We also evaluated the water absorption and water retention capacities of each excipient. In view of the results obtained, we produced pellets with the different starchCagglutinant mixtures (but without drug), and used image analysis to characterize pellet morphology. Our results show that some of the mixtures—notably starch (corn starch or wheat starch)C20% white dextrin—gave high-quality pellets with good size and shape distributions. In addition, the properties of the different materials tested suggest that it may be possible to obtain pellets with very different properties. q 2004 Elsevier B.V. All rights reserved. Keywords: Pelletization; Extrusion–spheronization; Torque-rheometry; Starchs; Dextins; Image analysis
1. Introduction In recent years, pellets and pellet-based multiparticulate delivery systems have become increasingly important, in view of their numerous technological and biopharmaceutical advantages. The most widely used base excipient for pellet formulations is microcrystalline cellulose (MCC) [1,2], largely because wetted MCC powder masses give mixtures with appropriate rheological properties for extrusion–spheronization. If significant amounts of MCC are replaced by some other component, rheological properties typically become much more dependent on the precise amount of water added, complicating the production * Corresponding author. Address: Departamento de Farmacia e Tecnoloxı´a Farmace´utica, Facultade de Farmacia, Universidad de Santiago de Compostela, Campus Universitario sur s/n. 15782, Santiago de Compostela, La Corun˜a, Spain. Tel.: C34 981 563100x14878; fax: C34 981 547148. E-mail address: [email protected] (F.J. Otero Espinar). 0939-6411/$ - see front matter q 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ejpb.2004.09.010
process, and in particular making it more difficult to control final pellet size and shape. Over the years various cellulose derivatives have been proposed as alternatives to MCC, including diverse grades of Avicel (PH-101, PH-102, PH-103, PH-105, RC-501, RC-581, RC-591, CL-611), Emocel and Unimac (MG-100, MG-200) [3]. Jover et al. [4] reported use of Avicel 955 for pellets with high drug content. Chatlapalli and Rohera [5] used hydroxypropylmethylcellulose (HPMC) or hydroxyethylcellulose (HEC) as principal excipient, with isopropyl alcohol instead of water as wetting agent. One study by Linder et al. [6] used Elcema G 250 and Elcema P 100. Other authors have argued for pellet formulations in which MCC (or other cellulose derivatives) are partially or entirely replaced by other types of base excipient. Gazzaniga et al. [9] obtained pellets using a mixture of 8 parts bcyclodextrin and 1 part MCC. Other authors have obtained pellets using pectin [7,8] or b-cyclodextrin [10] without MCC. Zhou et al. [11–14] used various starch and
512
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
microcrystalline wax derivatives, though they prepared pellets by high-shear granulation, not extrusion–spheronization. Junnila et al. [15,16] obtained pellets with 67.5% MCC and 30% corn starch, though this high proportion of starch led to defects in pellet shape and surface texture. Subsequently this group proposed the use of waxy corn starch as coadjuvant for pellets produced by extrusion–spheronization, on finding that it is possible to replace up to 40% of MCC weight, and thus considerably reducing formulation costs. Of the methods used to evaluate the expected behaviour of excipients during granulation, torque-rheometry is widely used. This method has been used by various authors to characterize the behaviour of diverse MCC grades [17,18], and to relate behaviour to basic physical properties [19–22]. This technique has been used to optimize the proportion of excipient to wetting agent in granulation [23–25], extrusion/ spheronization [23] and other pelleting processes [26]. We have used this technique as the principal method of characterization of pharmaceutical formulations during the kneading stage. This technique has been used by several researchers with the same aim [19,27–30]. Effective characterization can be achieved on the basis of two basic parameters: mean torque and torque amplitude [31,32]. The amount of wetting agent giving maximum mean torque can be taken as an estimate of the optimal amount of water agent for extrusion [30]; similarly, mean torque is useful for comparing the expected behaviour of wetted masses with wetting agent amounts close to the optimal value. This latter approach was used for example by Faure et al. [33] for granulation. The aim of the present study was to evaluate starches and dextrins as substitutes for MCC in the preparation of pellets by extrusion–spheronization. These excipients have the advantage of being much cheaper than MCC, and additionally give pellets with different drug release properties.
2. Materials and methods 2.1. Materials The materials used for pellet production were Avicel PH-101 (FMC International), corn starch (CS) (Meritena 100), wheat starch (WS) (Meritena 200), waxy corn starch (i.e. pregelatinized cross-linked CS; WCS) (Merigel 341), yellow dextrin A-2000 (YD), white dextrin (WD), and talc (J. Escuder, Barcelona, Spain). All starches and dextrins were supplied free by Inigarbe S.L. (Padro´n, A Corun˜a, Spain). The wetting agent used was in all cases distilled water. 2.2. Methods 2.2.1. Characterization of powders Particle size was estimated on the basis of serial sieving (mesh size 20–500 mm) in an Alpine 200LS-AC sieve set. True densities were estimated with a Micropycnometer
MPY-2 helium pycnometer (Quantachrome Corporation, Syosset, NY, USA). Compressibility was determined on the basis of tap and bulk densities determined using a Hosokawa Powder Tester (Micron Corp.). Flow factor (FF) was determined with an ITP-RO-200 automatic rotational shear cell, at three different normal forces. All powders used were photographed under a scanning electron microscope (Leo 435 UP, Leo Electron Microscopy Ltd, UK), following shadowing. 2.2.2. Wetting studies To characterize the responses of the different materials and their mixtures to added water, we determined (a) water absorption capacity, (b) water retention capacity, and (c) wetted mass consistency, as detailed below. Water absorption capacity was estimated using a Lloyd Instruments LR 5K press fitted with a 5-N NLC Lloyd Instruments loadcell from which was suspended a small vessel containing a 5 g sample of the material to be tested. Water absorption capacity was estimated on the basis of increase in sample weight after maintaining the vessel in contact with water for 10 min [34]. Water retention capacity was estimated by the method of Tomer et al. [35,36]. Samples containing different amounts of water were prepared, then kneaded. The wetted mass was then introduced into a small tube (diameter approximately 5 mm) with a filter at its lower end. This tube was introduced into a centrifugation vial, then centrifuged (Sigma 2-15 for 30 min at 3079!g, 6020 rpm), with the water extracted being collected into the vial. Water retention capacity was calculated in two different ways: WRCðLLÞ Z Wpost =Wpre WRCðLSÞ Z Wpost =Wpowder where Wpost is the weight of water remaining after centrifugation, Wpre is the weight of water before centrifugation, and Wpowder is the weight of powder before addition of water. Plots of WRC(LS) and WRC(LL) against weight of water added were in all cases constructed on the basis of five replicate determinations for each data point. 2.2.3. Rheometry Torque-rheometry (Caleva Mixer Torque rheometer, Caleva, UK) was used (a) in stepwise multiple-addition assays to characterize the responses of the different wetted masses to addition of gradual wetting agent, and (b) to characterize wetted masses prior to extrusion–spheronization (measurements of consistency). For stepwise wateraddition assays, initial sample weight was 30 g, and water was added in 1.5-ml steps with 1 min mixing at each step. In some cases it was necessary to use a modified initial sample weight, either because a 30-g sample did not entirely cover the blades of the rheometer, or because swelling on addition of water meant that the sample
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
expanded out of the rheometer vessel. For characterization of wetted masses prior to extrusion–spheronization (only done for those mixtures that were used to obtain pellets), solid mixtures were prepared in a Turbula T2C mixer for 15 min, and the wetted masses were kneaded in a Kenwood Chef Classic orbital mixer with distilled water as wetting agent, again with mixing for 15 min. The wetted masses obtained were introduced into plastic bags that were hermetically sealed for 24 h so that the water distributed as homogeneously as possible. All consistency measurements by torque-rheometry were obtained in triplicate. 2.2.4. Extrusion–spheronization Extrusion was performed with a Caleva Model 10 extruder (6 rpm, hole diameter 1 mm). One hundred grams of the extrudate was then spheronized with a Caleva Model 120 spheronizer (2000 rpm, 15 min), and the resulting pellets were dried to constant weight in a Heraeus oven. Three replicate pellet batches were obtained for each formulation. 2.2.5. Image analysis Pellet size and shape were characterized using an Olympus SZ-CTV stereomicroscope linked to a JVC TK-530 black-and-white videocamera, with perpendicular illumination from above against a black background, provided by a Highlight 2000 coldlight (Olympus Europe). The form factors Vr and Vp [37] were estimated from the digital Images obtained with this system.
3. Results and discussion 3.1. Basic physical characteristics of the excipients Table 1 summarizes the principal physical characteristics of the products used in this study. Avicel PH-101 and WCS show the highest mean particle size, though with greater among-particle variation in size, and more irregular shape. Scanning micrographs (Fig. 1) show the typical elongated fibrous shape of the cellulose particles, and the irregular angular shape of the WCS particles. The remaining materials show lower mean particle size, with particles
513
more rounded and regular in shape. These morphological properties are of course expected to influence rheometric properties. All products showed similar true densities, except WCS for which values were rather lower. Avicel PH-101 showed the lowest tap and bulk densities. WCS showed the lowest compressibility and highest flow factor, indicating that it is the excipient showing the best flow properties of the six; though note that all six excipients show values indicative of poor flow properties. Despite this, these compressibility values are within the range proposed by Chatlapalli and Rohera [5] as acceptable for the production of dense low-friability pellets by extrusion–spheronization. The lower compressibility of WCS and Avicel compared to the other starches and dextrins may be related to the larger particle sizes of the former products, improving flow properties. 3.2. Water absorption and retention capacities of excipient mixtures The results of the water absorption assays are summarized in the raw at the bottom of Table 1. Marked differences are observed among the excipients: the Avicel PH-101 show very high water absorption capacity (as is well known), while yellow dextrin shows very low water absorption capacity. The other four excipients show less marked differences. The results of the water retention assays are summarized in Figs. 2 and 3. The capacity of a material to retain absorbed water depends on whether the water is present in free or bound form. Water retention capacities (WRCsZmoisture retention capacities) were determined as per Tomer and coworkers [35,36], though note that we used tubes of different sizes, so that our results are not directly comparable with theirs. In all cases, WRCs were determined under the maximum load that the pellet was able to bear without excipient appearing in the water recovered by centrifugation. The results for the microcrystalline celluloses and starches are summarized in Fig. 2. The values of WRC(LL) and WRC(LS) for Avicel PH-101 indicate that, in addition to its higher water absorption capacity, this excipient shows a much higher capacity to retain water than
Table 1 Basic physical properties of the excipients considered in the present study Excipients Size (meanGSD)a Real density (g/cm3)a Tapped density/bulk density (g/cm3)/(g/cm3) Compresibility (%) Hausner ratio Flow factor Water uptake (%) a
Corn starch
Wheat starch
Waxy corn starch
Yellow dextrin
White dextrin
Avicel pH-101
11.46G6.24 1.513G0.032 0.769/0.409
24.9G13.01 1.530G0.083 0.756/0.396
52.62G24.89 1.481G0.022 0.720/0.494
20.98G8.03 1.534G0.055 0.786/0.471
30.42G9.91 1.520G0.056 0.790/0.444
53.42G30.7 1.563G0.067 0.462/0.283
46.81 1.88 2.89 69.4
47.62 1.9 5.82 48.3
31.32 1.45 7.51 26.4
40.08 1.66 4.96 8
43.78 1.77 7.87 49.2
38.83 1.63 5.64 244.0
MeanGSD (for six determinations).
514
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 1. Scanning electron micrographs of the excipients considered in the present study.
the starches. Comparing the two starches, WS showed slightly higher water retention capacity than CS. The results for the starch–dextrin mixtures are summarized in Fig. 3. Mixtures with WCS are not included, since even with only 10% WCS no water was recovered on centrifugation (WRC(LL)Z100%), indicating that absorbed water is tightly bound in this excipient. Mixtures of CS or WS with yellow dextrin (YD) showed WRC values very similar to those for CS and WS alone, with the largest differences being for CS with 20% YD. Note that fewer valid data points were obtained for the YD mixtures, because YD has relatively high water solubility and therefore appears in the water extracted by centrifugation even when the absorbed water content of the pellet is low (thus invalidating the assay; see above). Mixtures of CS or WS with white dextrin (WD) showed completely different behaviour (Fig. 3). This excipient significantly improved the water retention capacity of both starches. In addition, mixtures with WD accepted higher water loadings than mixtures with YD (i.e. excipient did not appear in the water extracted by centrifugation even when absorbed water content was high, so that more valid data
Fig. 2. Water retention capacities of the different starches, and the Avicel pH-101.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
515
Fig. 3. Water retention capacities of the different starchCagglutinant mixtures (WD, white dextrin; YD, yellow dextrin).
points could be obtained). The shapes of the WRC(LS) curves obtained for mixtures with WD are also of interest: like the curve obtained for Avicel PH-101 (Fig. 2), these curves showed a well-defined peak, allowing clear identification of the amount of wetting agent giving peak water retention capacity (possibly related to the different stages undergone by the mass during wetting, and thus to the rheological behaviour of the mass). 3.3. Rheometric characterization of wetted masses of the individual excipients Torque-rheometry profiles for stepwise water-addition assays are shown for the different individual excipients in Fig. 4. These profiles are very useful for characterizing the different stages through which the mass passes during the wetting process [27,32]. In studies of MCCs, various authors have found that the optimal amount of wetting agent for extrusion–spheronization pellets is that which gives maximum torque, and thus capillary state.
The profiles obtained for Avicel PH-101 (Fig. 4) are relatively ‘wide’, with maximum torque of about 1.4 nm given by 1.2–1.4 ml/g of wetting agent. This result is similar to that obtained by other authors for Avicel PH-101 [18], and indicates that a mass with suitable consistency can be obtained over a fairly wide range of wetting agent amounts. In addition, it is worth noting that the wetting agent amounts giving maximum torque are those that give maximum water retention capacity (WRC(LS), Fig. 2). The profiles obtained for the two starches are much ‘narrower’ and more jagged, with maximum torque values significantly lower than those obtained for microcrystalline cellulose. In the case of WS, maximum torque amplitudes were very high, in some cases higher than maximum torque. These results indicate that masses obtained from starches alone have inadequate consistency for extrusion–spheronization. In addition, and particularly in the case of WS, the rheometric data indicate that water is not homogeneously distributed within the mass, and that the optimal amount of wetting agent is difficult to determine. Only small amounts
516
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 4. Torque-rheometry profiles for the two native starches (corn starch CS and wheat starch WS) and of Avicel, as obtained in stepwise water-addition assays. Closed symbols—mean torque; open symbols—mean torque amplitude.
of water were required to achieve the capillary state, in agreement with a previous study [38]. These authors found that excipients comprising smooth particles with uniform characteristics, and thus higher tap and bulk densities, require smaller amounts of water to reach maximum torque. Torque-rheometry profiles could not be obtained for WCS, YD or WD, because the torques reached were higher than the maximum measurable with the rheometer. These results indicate that pellet production using starches as base excipient requires the addition of an agglutinating agent. To explore this possibility, we next performed rheometric characterization of mixtures of starches with WCS, YD or WD. 3.4. Rheometric characterization of wetted masses of starch–agglutinant mixtures Torque-rheometry profiles are shown for the different starch–agglutinant mixtures in Figs. 5 and 6. In general, addition of dextrins or WCS to starches reduced the amount of wetting agent required to achieve maximum torque and the capillary state. In addition, the profiles became ‘wider’, indicating that the amount of wetting agent required to achieve the pendular or capillary state is less critical. Maximum torque values increased with increasing amounts of YD, WD or WCS, suggesting that with increasing amounts of these components water was bound more strongly in the mixture, giving rise to more compact and adhesive masses. More specifically, the addition of 10 or 20% WCS markedly increased mean torque values of both WS (Fig. 5)
and CS (Fig. 6), to values more than twice as high as those recorded for Avicel PH-101 with 20% WCS. These high torque values indicate wetted masses with high consistency, high mechanical resistance, and probably low plasticity; these properties may not be suitable for extrusion– spheronization. However, the relative position of torque amplitude suggests that water is homogenously distributed within these masses. The high mean torque values of these masses may be related to their high water retention capacity (WRC(LS)Z100%), indicating strong binding of water. The addition of 20% YD likewise markedly increased mean torque values, though to a lesser extent than WCS. The mean torque and torque amplitude values obtained for starchCYD (unlike those obtained for starchCWCS) indicate that water was not homogeneously distributed within the mass, with torque amplitude in some cases even higher than mean torque. Note also that maximum torque was obtained with a smaller amount of water in mixtures containing 20% YD compared to mixtures containing 10% YD. The addition of WD (especially 20% WD) again markedly increased mean torque values. The torque values obtained for WSCWD mixtures were similar to those obtained for Avicel PH-101, though maximum torque was obtained with smaller amounts of water. The rather ‘wide’ profiles and the relative positions of the torque amplitude peaks suggest that water is homogeneously distributed within these wetted masses. In this case, and as with Avicel PH-101, the range of added water giving maximum torque broadly coincides with that giving peak water retention capacity.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
517
Fig. 5. Torque-rheometry profiles for the different wheat starch (WS)Cagglutinant mixtures, as obtained in stepwise multiple-addition assays. Values shown are mean torque.
In general, and in view of the mean torque and torque amplitude curves obtained for the two starches alone (Figs. 4, 5 and 6), water seems to be more homogeneously distributed within CS wet masses than within WS wet masses. This suggests that CS should be more appropriate for extrusion–spheronization. 3.5. Extrusion–spheronization In view of the torque-rheometry profiles obtained for the different starchCagglutinant mixtures we selected the optimal amount of wetting agent for each mixture, and proceeded with extrusion–spheronization. The amount of wetting agent added was that which gave maximum torque (i.e. capillary state), reported by various authors to be optimal for extrusion–spheronization with MCCs;
additionally, we produced pellets using amounts of wetting agent close to this theoretical optimum. Note that no drug was included in the mixtures, and it should be borne in mind that drugs may modify the properties of the wetted mass, and thus affect its extrusion– spheronization. To evaluate the capacity of the excipient mixtures to form spherical pellets, the mixture was placed in an orbital kneader, and the entire amount of wetting agent was added at once. After kneading for 15 min, it was placed in a hermetically sealed vessel and left to equilibrate at room temperature for 24 h. We then determined the consistency of the mixture by torque-rheometry, and proceeded to extrusion–spheronization. In general, all mixtures assayed showed a strong decline in consistency with respect to the values obtained in the
518
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
Fig. 6. Torque-rheometry profiles for the different corn starch (CS)Cagglutinant mixtures, as obtained in stepwise multiple-addition assays. Values shown are mean torque.
corresponding stepwise water-addition assays. This is basically attributable to the addition of all the wetting agent at once, and to the use of an orbital kneader, favouring the formation of small aggregates that behave like a granulate. Once the mixture is placed in the rheometer chamber and the blades of the rheometer begin to move, the aggregates are brought into movement but retain their integrity. As a result, the rheometer is basically measuring friction between aggregates, and between aggregates and the rheometer, rather than the true (and much higher) consistency of the wetted mass. The proof of this is that despite obtaining very low mean torques, pellets were successfully obtained from mixtures prepared with various amounts of wetting agent, not necessarily those showing highest mean torque in the multiple-addition assays. The quality of the pellets obtained depended both on the composition of the mixture and the amount of wetting agent. Table 2 summarizes these properties for the different
mixtures used, and lists shape factors for the pellets obtained. The amount of water needed to produce pellets was in all cases lower than that giving maximum torque (and thus capillary state and peak water retention). In fact, the optimal amount of water was better predicted by peak torque amplitude. This was particularly true of mixtures containing WD. The rightmost columns of Table 2 summarize the principal morphological characteristics of the pellets obtained. Pellets were also obtained with other mixtures not shown in Table 2, but these either had very small particle diameter (indicative of insufficient wetting), or formed large agglomerates (indicative of excessive wetting). Mixtures of native starches (CS or WS) with WCS gave pellets with deficient morphological characteristics (Table 2). However, in no case were these pellets spherical, with most lots showing a significant proportion of irregularly shaped pellets. The values of the form factors
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
519
Table 2 Morphometric characteristics of pellets obtained from the different starchCagglutinant mixtures Compositiona
Corn 1b 2b 3 4b 5 6 7 8 9 10 Wheat 11 12b 13 14 15b 16b 17 18 19 a b
Starch (%)
WCS (%)
85 85 75 75 75 75 75 75 75 75
10 10 20 20
YD (%)
WD (%)
20 20 20
85 10 85 10 75 20 75 20 75 20 75 20 75 75 100% Avicel PH-101
20 20 20
20 20
ml of water per g of powder
Vr (meanGSD) Popn 1
0.40 0.45 0.35 0.40 0.25 0.30 0.35 0.35 0.40 0.45
11.08G1.48 8.17G0.05 5.50G0.49 12.19G0.76 6.12G1.07 5.53G1.87 7.28G1.26 3.98G0.91 3.20G0.51 4.27G0.73
0.35 0.40 0.30 0.35 0.20 0.25 0.30 0.35 1.05
4.97G0.48 11.08G1.48 6.71G1.35 5.72G2.43 9.91G0.26 9.68G0.51 6.22G2.66 2.44G0.95 3.32G0.23
Vp (meanGSD) Popn 2
9.72G3.31 5.50G0.28
8.29G2.26
12.80G5.98
5.27G2.27
Popn 1
12.79G0.96 11.62G2.44 8.21G0.72 14.57G2.49 8.34G0.60 7.92G0.68 12.38G2.11 7.72G1.03 7.14G0.31 7.19G0.68 7.97G0.38 12.79G0.96 10.63G0.29 11.44G4.64 9.90G0.56 10.75G0.44 6.71G0.44 6.61G0.40 7.09G0.09
Popn 2
12.20G3.74
10.74G1.78 9.55G1.42 8.83G1.19
Feret (meanGSD)
1144.5G42.5 1101.3G111.2 1008.6G24.8 1252.6G193.2 762.9G120.5 1161.7G141.9 1538.7G358.4 939.0G155.7 998.1G127.4 1415.4G66.0
11.64G1.80
954.7G100.6 1144.5G42.5 1041.6G43.8 19.71G10.56 967.8G85.1 1021.0G85.6 1085.9G77.9 8.44G0.71 993.0G53.9 8.56G0.87 970.3G65.0 824.6G55.4
In all mixtures talc was added (5%) as lubricant agent. Arithmetic mean and SD, because a multigaussian fit was impossible.
Vr and Vp obtained for pellets prepared from mixtures with low water contents indicate predominantly cylindrical shape, with more or less angular edges. In general, increasing the amount of added water led to an increase in pellet size and in Vr (e.g. VrZ5.5 for formulation 3, 12.2 for formulation 4); higher Vr values are indicative of cylindrical shapes with high aspect ratio, or fusiform shapes. This is basically attributable to the rheological characteristics of these mixtures. Mixtures of native starches and WCS gave wet masses with high consistency and low plasticity which can be extruded, but which are not readily spheronized, so that particle shape remains markedly non-spherical. The observed increase in pellet size and aspect ratio with increasing amount of added water is attributable to the fact that wetter masses are more elastic and have higher consistency, so that the extrudates are longer and do not break so easily during spheronization, giving rise to cylindrical particles. Mixtures of CS or WS with YD showed very low mean torque values, so that pellets could only be obtained with mixtures containing a high proportion of YD (20%). Mixtures with only 10% YD had very low consistency, and could not be effectively extruded. The pellets obtained with CS or WS plus YD were not spherical: a high proportion of the pellets prepared with WS showed irregular shape, probably due to the heterogeneous distribution of wetting agent within the wetted mass, as indicated by the rheometry results. The pellets prepared with CS were somewhat better, but still markedly non-spherical.
Mixtures of CS or WS with WD likewise showed low mean torque, so that pellets could again only be obtained with mixtures containing 20% WD. However, the pellets obtained with these mixtures (whether CS, formulations 8–10, or WS, formulations 17 and 18) showed good sphericity. Indeed, many of the starchCWD formulations gave particles with Vr values close to those obtained for Avicel. The CSCWD mixtures (like the CSCWCS and CSCYD mixtures) gave increasing pellet size with increasing amount of wetting agent. This relationship was not so clearly evident in the case of the WSCWD mixtures, probably because of the different water absorption and retention capacities of these starches, and the more heterogeneous distribution of water within the wetted masses. Following our previously published morphological classification on the basis of Vr values [37], none of the mixtures tested gave spherical pellets except WSCWD and CSCWD. Pellets prepared, with similar shape and surface texture to pellets prepared with Avicel PH-101. Mixtures including WCS as agglutinant gave largely spherical pellet populations but with significant proportions of other cylindrical and fusiform shapes. Pellet properties were better for mixtures with 10% WCS than for mixtures with 20% WCS. In general, mixtures with YD or WCS gave worse pellet morphometric properties than mixtures with WD.
520
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521
4. Conclusions The present results indicate that it is possible to prepare pellets with acceptable morphometric characteristics by extrusion–spheronization of mixtures in which the principal excipient is a starch, not a microcrystalline cellulose. The deficient rheological properties of WS and CS mean that their use as principal excipient requires the inclusion of other excipients that act as agglutinants and improve mixture consistency, resistance and plasticity. Torque-rheometry characterization of wet masses during stepwise addition of wetting agent, and studies of water retention capacity, were useful tools for comparing the behaviour of the different starch–agglutinant mixtures, and especially for selecting the optimal amount of wetting agent. However, consistency of equilibrated wet masses (prepared with the predicted optimal amount of wetting agent) was not an effective predictor of extrusion–spheronization efficacy, due to methodological differences with respect to the stepwise water-addition assay, and to the addition of wetting agent all at once rather than stepwise. Unlike microcrystalline cellulose mixtures (in which the optimal amount of wetting agent for extrusion– spheronization is that giving maximum torque), the optimal amount of wetting agent for the starch and starch–dextrin mixtures tested in the present study was lower than that giving maximum torque. In fact, the optimal amount of wetting agent was more accurately predicted by maximum torque amplitude. In general, CS mixtures needed slightly larger amounts of wetting agent than WS mixtures. The best (i.e. most spherical) pellet morphometries were obtained with WSC20% WD, or CSC20% WD. Pellets obtained with WSCWCS or CSCWCS (WCS at 10 or 20%) had poorer morphometries, but our results suggest that better results might be obtained with WCS at proportions of less than 10%.
References [1] A.D. Reynolds, A new technique for the production of spherical pellets, Manuf. Chem. Aerosol News June (1970) 40–43. [2] J.W. Connie, H.R. Hadley, Preparation of small solid pharmaceutical spheres, Drug Cosmet. Ind. April (1970) 38–41. [3] J.M. Newton, A.K. Chow, K.B. Jeeba, The effect of excipient source on spherical granules made by extrusion/spheronization, Pharm. Technol. Int. 4 (1992) 52–58. [4] I. Jover, F. Podczeck, J.M. Newton, Evaluation, by a statitically designed experiment, of an experimental grade of microcrystalline cellulose, avicel 955, as atechnology to aid the production of pellets with high drug loading, J. Pharm. Sci. 85 (1996) 700–705. [5] R. Chatlapalli, B.D. Rohera, Physicall characterization of HPMC and HEC and investigation of their use as pelletization aids, Int. J. Pharm. 161 (1998) 179–193. [6] H. Linder, P. Kleinebudde, Use of powdered cellulose for the production of pellets by extrusion/spheronization, J. Pharm. Pharmacol. 46 (1994) 2–7.
[7] I. Tho, E. Andersen, K. Dyrstad, P. Kleinebudde, S.A. Sande, Quantum chemical descriptors in the formulation of pectin pellets produced by extrusion/spheronization, Eur. J. Pharm. Sci. 16 (2002) 143–149. [8] I. Tho, S.A. Sande, P. Kleinebudde, Pectinic acid, a novel excipient for production of pellets by extrusion/spheronization: preliminary studies, Eur. J. Pharm. Sci. 54 (2002) 95–99. [9] A. Gazzaniga, M.E. Sangalli, G. Bruni, L. Zema, C. Vecchio, F. Giordano, The use of b-cyclodextrin as a pelletization agent in the extrusion/spheronization process, Drug Dev. Ind. Pharm. 24 (1998) 869–873. [10] M.E. Villar-Lopez, L. Nieto-Reyes, S. Anguiano-Igea, F.J. OteroEspinar, J. Blanco-Me´ndez, Formulation of triamcinolone acetonide pellets suitable for coating and colon targeting, Int. J. Pharm. 179 (1999) 229–235. [11] F. Zhou, C. Vervaet, J.P. Remon, Matrix pellets based on the combinations of waxes, starches and maltodextrins, Int. J. Pharm. 133 (1996) 155–160. [12] F. Zhou, C. Vervaet, J.P. Remon, Influence of processing on the characteristics of matrix pellets based on microcrystalline waxes and starch derivates, Int. J. Pharm. 147 (1997) 23–30. [13] F. Zhou, C. Vervaet, D.L. Massart, B. Massart, J.P. Remon, Optimization of the processing of matrix pellets based on the combination of waxes and starch using experimental design, Drug Dev. Ind. Pharm. 24 (1998) 353–358. [14] F. Zhou, C. Vervaet, M. Schelkens, R. Lefebre, J.P. Remon, Bioavailabity of ibuprofen from matrix pellets based on the combination of waxes and starch derivates, Int. J. Pharm. 168 (1998) 79–84. [15] R. Junnila, J. Heina¨ma¨ki, J. Yliryuusi, Effects of surface-active agent on the size, shape and hardness of MCC/maize starch pellets prepared by an extrusion–spheronization technique, STP Pharm. Sci. 8 (1998) 221–226. [16] R. Junnila, P. Palviainen, J. Heina¨ma¨ki, P. Mylla¨rinen, P. Forssell, J. Yliryuusi, Waxy corn starch: a potent cofiller in pellets produced by extrusion–spheronization, Pharm. Dev. Technol. 5 (2000) 67–76. [17] M.D. Parker, R.C. Rowe, Source variation in the wet massing (granulation) of some microcrystalline celluloses, Pow. Technol. 65 (1991) 273–281. [18] M.D. Parker, P. York, R.C. Rowe, Binder–substrate interactions in wet granulations 3: the effect of excipient source variation, Int. J. Pharm. 80 (1992) 179–190. [19] R.C. Rowe, G.R. Sadeghnejad, The rheology of microcrystalline cellulose powder/water mixer-measurement using a mixer torque rheometer, Int. J. Pharm. 38 (1987) 22–229. [20] P. Luukkonen, T. Schaefer, L. Helle´n, A.M. Juppo, J. Yliruusi, Rheological characterization of microcrystalline cellulose and silicified microcrystalline cellulose wet masses using a mixer torque rheometer, Int. J. Pharm. 188 (1999) 181–192. [21] P. Luukkonen, J. Rantanen, K. Ma¨kela¨k, E. Ra¨sa¨nen, J. Tenhunen, J. Yliruusi, Characterization of wet massing behaviour of silificied microcrystalline cellulose and a-lactose monohydrate using nearinfrared-spectroscopy, Pharm. Dev. Technol. 6 (2001) 1–9. [22] P. Luukkonen, T. Schaefer, F. Podczeck, M. Newton, L. Helle´n, J. Yliruusi, Characterization of microcrystalline cellulose and silicified microcrystalline cellulose wet masses using a powder rheometer, Eur. J. Pharm. Sci. 13 (2001) 143–149. [23] S.A. Schildcrout, Rheology of pharmaceutical granulations, J. Pharm. Pharmacol. 36 (1984) 502–505. [24] V. Janin, M.D. Parker, R.C. Rowe, Monitoring production scale granulation processes using an instrumented mixer torque rheometer, STP Pharm. 6 (1990) 233–237. [25] M. Hariharan, E. Mehdizadeh, The use of mixer torque rheometry to study the effect of formulations variables on the properties of wet granulations, Drug Dev. Ind. Pharm. 28 (2002) 253–263.
S. Almeida Prieto et al. / European Journal of Pharmaceutics and Biopharmaceutics 59 (2005) 511–521 [26] J. Kristensen, T. Schaefer, P. Kleinebudde, Direct pelletization in a rotary processor controlled by torque measurements. Influence of process variables, Pharm. Dev. Tech. Ind. 5 (2000) 247–256. [27] M.D. Parker, R.C. Rowe, N.G. Upjohn, Mixer torque rheometry: a method for quantifying the consistency of wet granulations, Pharm. Technol. Int. September (1990) 50–64. [28] B.C. Hancock, P. York, R.C. Rowe, M.D. Parker, Characterization of wet masses using a mixer torque rheometer. 1. Effect of instrument geometry, Int. J. Pharm. 76 (1991) 239–245. [29] B.C. Hancock, P. York, R.C. Rowe, Characterization of wet masses using a mixer torque rheometer. 2. Mixing kinetics, Int. J. Pharm. 83 (1992) 147–153. [30] R. Chatlapalli, B.D. Rohera, Rheologycal characterization of diltiazem HCL/cellulose wet masses using a mixer torque rheometer, Int. J. Pharm. 175 (1998) 47–59. [31] B.C. Hancock, P. York, R.C. Rowe, An assessment of substrate– binder interactions in model wet masses. 1. Mixer torque rheometry, Int. J. Pharm. 102 (1994) 167–176. [32] R.C. Rowe, M.D. Parker, Mixer torque rheometry an update, Pharm. Technol. Eur. March (1994) 27–34.
521
[33] A. Faure, I.M. Grimsey, R.C. Rowe, P. York, M.J. Cliff, Process control in a high shear mixer–granulator using wet mass consistency: the effect of formulation variables, J. Pharm. Sci. 88 (1999) 191–195. [34] F.J. Otero-Espinar, M.B. Delgado-Charro, S. Anguiano-Igea, J. Blanco Me´ndez, Using a tensile test to study the bioadhesive and swelling properties of poluymers, in: A. Mun˜oz-Ruiz, H. Vromans (Eds.), Data Acquisition and Measurement Techniques, EEUU, Illinois, 1998, pp. 295–342. [35] G. Tomer, J.M. Newton, A centrifuge technique for the evaluation of the extent of water movement in wet powder masses, Int. J. Pharm. 188 (1999) 31–38. [36] G. Tomer, H. Patel, F. Podczeck, J.M. Newton, Measuring the water retention capacities (MRC) of different microcrystalline cellulose grades, Eur. J. Pharm. Sci. 12 (2001) 321–325. [37] S. Almeida-Prieto, J. Blanco-Me´ndez, F.J. Otero-Espinar, Image analysis of the shape of granulated powder grains, J. Pharm. Sci. 93 (2004) 621–634. [38] R. Chatlapalli, B.D. Rohera, Study of effect of excipient source variation on rheological behaviour of diltiazem HCL-HPMC wet masses using a mixer torque rheometer, Int. J. Pharm. 238 (2002) 139–151.