A novel aid for the preparation of pellets by extrusion/spheronization

Fridrun Podczeck

Fridrun Podczeck is professor of pharmaceutics at Sunderland School of Pharmacy, University of Sunderland, UK

Pharmaceutical Technology Europe

Pharmaceutical Technology Europe, Pharmaceutical Technology Europe-12-01-2008, Volume 20, Issue 12

Microcrystalline cellulose is the main excipient used in the industrial manufacture of pellets by extrusion/spheronization, but pellets containing this spheronizing aid do not readily disintegrate and are expensive.

Since the first report by Reynolds,1 the preparation of pellets by extrusion/spheronization has become a well-documented and accepted commercial process. Most pellet formulations on the market and most studies of this process are based on the use of microcrystalline cellulose (MCC) as a spheronization agent to ensure the process functions well. However, MCC cannot provide a formulation for all drugs. For many drugs it is also not possible to incorporate high-drug doses into MCC pellets. Even when pellets can be prepared, there could be problems regarding chemical instability of the drug, as reported, for example, for ranitidine.2

A disadvantage of using MCC is the high costs of the material, especially as large quantities of at least 50% w/w or more are required in a formulation, which also restrict the maximum dose of drug that can be incorporated to ensure a reasonably small size of a capsule or a tablet. A number of reports have demonstrated considerable product variability because of the source ('brand') and grade variability of MCC,3–6 with respect to the extrusion/spheronization process, and some cheaper brands have been suspected not to function very well.7

For the formulations to extrude and change shape during the spheronization process, it is necessary to add larger amounts of liquid binder to achieve a plastic wet mass. The liquid binder is preferably water or a mixture of water with ethanol, glycerol or self-emulsifying systems. The use of organic solvents is very restricted under industrial manufacturing conditions. Pellets based on MCC as a spheronization aid do not disintegrate and they release the drug rather slowly and often incompletely. However, the drug release is not slow enough to produce an extended release dosage form and the application of modified release film coatings is required. There are also reports on the influence of various processing parameters, such as standing time of the wet mass on the quality of the final pellets,8 which are of concern to the industry where material and product intermediate hold-ups cannot always be avoided.

One approach to provide alternative excipients to aid the process of extrusion/spheronization has been the modification of the standard grades of MCC by co-processing the wet cake with hydrophilic polymers such as polyvinyl pyrrolidone, hydroxypropyl methylcellulose (HPMC) or sodium carboxymethylcellulose;9 the latter having been shown to perform better in terms of increasing the drug level that can be incorporated into the pellets.10,11

In recent years, there has also been a series of publications showing the ability of a range of excipients to act in a similar capacity to MCC. Examples include carrageenan,12–14 powdered cellulose,15 chitosan,16,17 glyceryl monostearate,18 HPMC and hydroxy ethylcellulose (HEC),19 pectin,20,21 cross-linked pectin,22 polyethylene oxide,23 starch24 and starch-dextrin mixtures.25 However, none of these materials have as yet been subjected to the same degree of evaluation as MCC, and some have limitations; for example, HPMC and HEC require the use of isopropyl alcohol as the binder liquid,19 which may be useful for water-sensitive drugs, but is restrictive in practical manufacturing terms.

When studying glyceryl monostearate, only one of the four model drugs tested provided good quality pellets.18 Chitosan required the binder liquid to be 0.2 M acetic acid.17 The use of starch required more effort than simple addition of water by mixing,24 which is the method used for MCC. From the data provided in many of the above publications, it is difficult to judge the quality of the pellets in terms of their roundness and narrowness of particle size distribution, both being an essential requirement for the product to be manufactured under industrial conditions. Inspection of photographs provided in these publications would suggest that the pellets were rather 'rounded' (i.e., still being considerably elongated with rounded edges, rather than spherical). This would indicate that the alternative spheronizing aids studied lacked plasticity and would not be a universal replacement for MCC in the pharmaceutical industry.

The criteria that ensure an excipient will function as an aid to extrusion/spheronization are not known with certainty. Excipients need to have a binder liquid — usually water — holding capacity, especially when the wet mass is subjected to pressure.26,27 Podczeck et al. indicated that water retention capacity alone may not be the only factor involved in the performance of spheronization aids.11 The rheological properties of the system are also important, but as systems with a wide range of rheological properties will make satisfactory pellets,28 this remains to be clearly defined. From measurements of the rheological properties of the wet mass of MCC containing a large quantity of lactose monohydrate, MacRitchie et al. suggested that the elastic properties of the wet mass should also be at an optimum level.29

A pharmaceutically acceptable excipient, which has a high retention capacity for water, is colloidal silicon dioxide.30 Preliminary experiments with this material proved unsuccessful under a wide range of conditions. However, with the addition of a non-ionic surfactant to the aqueous liquid binder, it is possible to form good quality pellets using this material as an aid to extrusion/spheronization.31

Materials

Colloidal silicon dioxide (CSD) Ph. Eur. was obtained as Aerosil 200 (Degussa Ltd, UK). The fillers were lactose monohydrate Ph. Eur. (Borculo Whey Products, UK) and heavy magnesium carbonate Ph. Eur. (Thornton & Ross, UK). The model drug used was ephedrine hydrochloride Ph.Eur. (S&D Chemicals Ltd, UK). Purified polyoxy–35–castor oil Ph.Eur. (Cremophor ELP; BASF, Germany) was used as model non-ionic surfactant, Imwitor 742 (Condea Chemie GmbH, Germany) was used as a plasticizer and freshly distilled water was used as a liquid.

On the go...

Methods

Preparation of pellets. The appropriate quantities of powder and plasticizer, as set out in Table 1, were first dry mixed in a planetary mixer (Model A901E; Kenwood Chef, UK) for 5 min. A 5% w/w solution of Cremophor ELP in water was gradually incorporated into the mixture, which was mixed for a further 10 min, with scraping the sides and bottom of the bowl and mixing blades performed at 5 min intervals during the process. The amount of binder liquid solution required to produce round pellets with a narrow size distribution was obtained by trial and error. The wet mass was extruded with a radial extruder (Model 10 Extruder; G.B. Caleva Ltd, UK) and fitted with a screen of 1.5 mm thickness with 1.5 mm diameter holes. Quantities of 30 g of extrudate were spheronized for 10 min on a 125 mm cross-hatch plate in a spheronizer (Model 120; G.B. Caleva Ltd, UK). The speed of rotation was optimized in a systematic fashion for each formulation (Table 1). Considering the small batch size, the pellets were dried at room temperature (22–24 °C) and 50–55% RH by storing them on sieves with 0.355 mm mesh size, until a constant weight was obtained.

Table 1 Composition including optimized liquid binder concentration and optimum spheronization speed for pellets made with CSD.

Pellet size distribution. The size distribution of the pellets was determined using a mechanical sieve shaker (Type 47–300; Retsch, UK) fitted with a root–2 progression of British Standard Sieves (Endecotts, UK) between 500 and 4000 μm. The amount of pellets in the size fractions below 710 μm and above 2 mm were found negligible in most cases.

Pellet shape. The two-dimensional shape factor eR, first described by Podczeck and Newton,32 in its revised form,33 and the aspect ratio34 were determined by image analysis. The measurements were conducted with a Seescan image analyser (Seescan Solitaire 512; Seescan Imaging, UK) attached to a black and white camera (CCD-4 miniature video camera, Rengo Co. Ltd, Japan) and fitted to a zoom lens (18-108/2.5; Olympus Europe, Germany). The resolution was set to 18 μm/pixel and was calibrated with a graticule (Macro systems, UK). One hundred pellets from the 1.0–1.4 mm sieve fraction of each batch were randomly selected. The pellets were mounted on a surface painted with nonreflective black paint and illuminated from above with a twin cold light source adjusted at 180° to the surface (Olympus Highlight 3001; Olympus Europe, Germany) to minimize the formation of shadows. These conditions were followed to satisfy the requirements set out by Podczeck et al.33

Pellet mechanical strength. The load required to break 50 randomly selected pellets from the 1.0–1.4 mm size sieve fraction of each batch was determined using a physical testing instrument (CT-5; Engineering Systems, UK) fitted with a 5 kg load cell and using a cross-head speed of 1 mm/min. From the breaking load, the value of the surface tensile strength of the pellets could be calculated as described by Salako et al.35

Pellet density. The pellet density was measured with a helium pycnometer (MVP-1; Quanta Chrome Corp., NY, USA) as outlined in Ph.Eur. (Appendix XVII K, method 2.9.23). Five replicated measurements were done per batch. From a measurement of the helium particle density of the constituent powders and their proportion in the final pellets, the porosity of the pellets could be determined.

Disintegration and drug release from the pellets. The drug release from the pellets was assessed according to the Ph.Eur. dissolution apparatus II procedure, with a paddle speed of 50 rpm in 900 mL of distilled water at 37±0.5 °C for 30 min. The pellets were monitored and the time taken for them to disintegrate during the dissolution test was recorded. This was preferred over a standard disintegration test, which is more vigorous in agitation and may not reflect the conditions encountered during the dissolution test.

Six samples of 300 or 600 mg of pellets, depending on the drug content, were tested. The amount of drug released into the solution at 5 min time intervals was assessed spectrophotometrically at a wavelength of 271 nm with a UV spectrophotometer (Model CE272; Cecil Instruments, UK).

Results and discussion

Pellet preparation by extrusion/spheronization. Pellets could be produced from all formulations tested. One important feature of these formulations is that pellets can be prepared from CSD without the addition of any bulking agent or drug (Table 1, Formulations 11 and 12). This is also possible with MCC, but for CSD the binder liquid must contain the non-ionic surfactant and the optional addition of the plasticizer Imwitor 742 is possible. The amount of liquid binder solution added is more than twice the weight of the CSD powder and only slightly reduced by the addition of the plasticizer. CSD can take up water up to 40% of its own weight without losing its appearance as free-flowing powder. It is, therefore, not surprising that such a large amount of liquid binder is needed to transform the material into a wet mass with sufficient plasticity to enable extrusion. In the formulations containing ⅓ of CSD and ⅔ of drug and/or filler, the ratio between the CSD and the amount of liquid binder added remains at approximately 1:2 when soluble formulation components have been added and increases to 1:3 for the formulations containing only the insoluble filler magnesium carbonate. Whether or not the plasticizer has been added to the formulations seems to have no effect.

Another feature of the CSD formulations appears to be their sensitivity to the speed of rotation during the spheronization step (Table 1). CSD extrudates can be spheronized at a higher speed of rotation than formulations containing drug and/or filler in addition to CSD. However, the related linear peripheral velocity value of 9.7 m/s is considerably smaller than that commonly used for MCC containing formulations (23.6 m/s). Because of the relatively larger amount of liquid incorporated into the plain CSD formulations, the extrudate is heavier and less prone to elastic bouncing on the spheronization plate. For magnesium carbonate mixtures, the relative amount of liquid binder added is reduced to approximately half, but the high density of this excipient adds weight, preventing elastic bouncing of the extrudate. However, when the less dense drug and/or lactose monohydrate is added to CSD, the extrudates are very elastic and the speed of the spheronizer plate must be further reduced and tightly controlled. Sensitivity of formulations to the spheronization speed has been previously reported by Speirs36 for formulations containing large amounts of the disintegrant croscarmellose sodium. As can be seen from Table 1, the addition of a plasticizer in small quantities does not influence the optimum spheronization speed.

Pellet size distribution. When evaluating the results for the sieve analysis reported in Table 2, it should be borne in mind that a screen with a diameter and length of the die bores of 1.5 mm was used for the extrusion. This choice had been made to increase the length of the dies and, therefore, the pressure and densification of the extrudate. Initially, a smaller screen of 1 mm die length and thickness had been used, resulting in shark-skinned and very brittle extrudate that broke during the spheronization process. The change in the screen was successful in preventing shark skinning and indicates that the CSD formulations will perform even better when extruded with long die extruders. A wet mass extruded through 1.5 mm die bores should be expected to result in pellets of approximately 1.5 mm diameter or even more. Hence, considerably smaller pellet sizes are an indication of brittleness of the extrudate. This can be seen for pellets made from CSD only, as the modal fraction for these two formulations is the sieve fraction 0.71–1.0 mm. Pellets made from CSD and the drug are of a modal size range of 1.4–2.0 mm, and their extrudate is least brittle. Pellets made from CSD and either lactose monohydrate or magnesium carbonate are smaller (modal size fraction 1.0–1.4 mm), but the addition of drug to these formulations, again, results in an increase in the modal pellet size.

Table 2 Pellet properties

Pellet shape. The pellet shape has been assessed using two parameters. The aspect ratio is sufficient for industrial inprocess control of pellet formulations during manufacture and an upper limit of 1.2 has been stated for the use of pellets in, for example, capsule filling.37 The results presented in Table 2 indicate that ⅔ of the formulations produced would fill into capsules without any problems. The pellets made from CSD only, and those made from a mixture of equal parts of drug, lactose monohydrate and CSD in the powder mixtures, however, are just outside the recommended limit. The addition of the plasticizer did not improve the shape formation of these formulations, but a change of the filler from lactose monohydrate to magnesium carbonate resolves this issue (i.e., when formulating pellets using CSD as the spheronization aid, the selection of an appropriate filler can improve the product properties considerably).

Second, as conducted in research, the shape of the pellets was more scrutinized using the shape factor eR, which is largely reflecting the overall macro-surface properties of pellets as well as their elongation. A pellet might appear fairly round visually, but surface irregularities (e.g., bumps, flattening, etc.) will considerably reduce the value for this shape factor. A 'perfect' sphere with smooth surfaces would present a value of approximately 0.6.33 As can be seen from Table 2, none of the pellet batches has perfectly round and smooth pellets. However, as Figure 1 shows, most pellets are well rounded, yet their surface is rougher than that of MCC pellets. Figure 2 shows a larger set of pellets for batch ALEc_i, which was the least round batch of pellets, but which had a very narrow particle size distribution with more than 72% of all pellets in the modal fraction. The values of the shape factor eR seem, in some instances, related to the pellet porosity (i.e., the less porous the pellets, the better is their shape). This indicates that the use of long die extruders might further improve the pellet properties in terms of their shape.

Figure 1

Pellet mechanical strength. Pellets made only from CSD were very fragile and had insufficient strength for an accurate measurement. After addition of the drug and/or excipient to the formulation, the pellet strength was much improved (Table 2). However, the addition of the plasticizer resulted in a decrease in pellet strength and in two instances the pellets did not fracture but deformed plastically during the tests. Therefore, in terms of pellet strength the addition of the plasticizer is not recommended. Pellets without plasticizer, but containing the drug, have sufficient strength to undergo further processing such as film coating or capsule filling.

Figure 2

Disintegration and drug release from the pellets. Pellets disintegrated rapidly on contact with the dissolution liquid and after 10 min had disintegrated into their powder components. This is in contrast to MCC pellets, which disintegrate incompletely or not at all.38 Figure 3 shows the dissolution profiles for the six formulations containing the drug. Because the pellets disintegrated, the figure reflects the dissolution properties of the drug particles of the disintegrated pellets rather than the influence of the pellet formulations. Addition of the plasticizer increased the dissolution rate of the drug slightly in all cases. Also, pellets containing the insoluble filler, magnesium carbonate, had the slowest disintegration time and show slightly slower drug release rates.

Figure 3

It can be concluded that CSD can be used as a spheronization aid, provided that 5% of a non-ionic surfactant, such as Cremophor ELP, is added to the binder liquid water. The pellets would have better properties if the extrusion process results in a larger degree of densification, which could, for example, be achieved using a long die extruder. The extrudate is elastic and speed sensitive, and slow and well-maintained spheronization speeds are required. The addition of a plasticizer does not improve the performance of the formulations and is, hence, not essential. In contrast to MCC pellets, pellets made with CSD as a spheronization aid disintegrate rapidly when in contact with dissolution liquid and therefore ensure a better and more complete drug release from the pellets.

Acknowledgements

The author would like to acknowledge the assistance of Nuno Miguel Da Silva Carvalho with the practical work.

Fridrun Podczeck is Visiting Professor at the Department of Mechanical Engineering at University College London (UK).

References

1. A.D. Reynolds, Manuf. Chem. Aerosol News, 41(6), 40–44 (1970).

2. A. W. Basit, J.M. Newton and L.F. Lacey, Pharm. Dev. Technol., 4(4), 499–505 (1999).

3. J.M. Newton et al., Pharm. Technol. Int., 4(10), 52–59 (1992).

4. C.L. Raines, "Rheological Properties of Different Grades of Microcrystalline Cellulose," Ph.D. Thesis, University of London, UK (1990).

5. F. El Saleh et al., S.T.P. Pharma Sci., 10(5), 379–385 (2000).

6. O.M.Y Koo and P.W.S. Heng, Chem. Pharm. Bull., 49(11), 1383–1387 (2001).

7. M. Landin et al., Int. J. Pharm., 91(2), 123–131 (1993).

8. C.L. Raines et al., in R. Carter, Ed, Rheology of Food, Pharmaceuticals and Biological Materials with General Rheology (Elsevier Applied Science, London, UK 1990) pp 248–257.

9. H.W. Durand et al., US Patent 3,539,365 (1970).

10. I. Jover. F. Podczeck and J.M. Newton, J. Pharm. Sci., 85(7), 700–705 (1996).

11. F. Podczeck, P.E. Knight and J.M. Newton, Int. J. Pharm., 350(2), 145–154.

12. M. Bornhoft, M. Thommes and P. Kleinebudde, Eur. J. Pharm. Biopharm., 59(1), 127–131 (2005).

13. M. Thommes and P. Kleinebudde, Eur. J. Pharm. Biopharm., 63(1), 59–67 (2006).

14. M. Thommes and P. Kleinebudde, Eur. J. Pharm. Biopharm., 63(1), 68–75 (2006).

15. L. Alvarez et al., Eur. J. Pharm. Biopharm., 55(3), 291–295 (2003).

16. A.M. Agrawal, M.A. Howard and S.H. Neau, Pharm. Dev. Technol., 9(2), 197–217 (2004).

17. H. Steckel and F .Mindermann-Nogly, Eur. J. Pharm. Biopharm., 57(1), 107–114 (2004).

18. J.M. Newton et al., Pharm. Technol. Eur., 16(10), 21–27 (2004).

19. R. Chatlapalli and B.D. Rohera, Int. J. Pharm., 161(2), 179–193 (1998).

20. I. Tho, P. Kleinebudde and S.A. Sande, AAPS PharmSciTech., 2(4), Article 24 (2001).

21. I. Tho, P. Kleinebudde and S.A. Sande, AAPS PharmSciTech., 2(4), Article 27 (2001).

22. I. Tho, S.A. Sande and P. Kleinebudde, Chem. Eng. Sci., 60(14), 3899–3907 (2005).

23. M.A. Howard, S.H. Neau and M.J. Sack, Int. J. Pharm., 307(1) 66–76 (2006).

24. A. Dukic et al., Eur. J. Pharm. Biopharm., 66(1), 83–94 (2007).

25. S.A. Prieto et al., Eur. J. Pharm. Biopharm., 59(3), 511–521 (2005).

26. K.E. Fielden et al., Int. J. Pharm., 79(1–3), 46–60 (1992).

27. F. Podczeck and P.E. Knight, Pharm. Dev. Technol., 11(3), 263–274 (2006).

28. J.M. Newton et al., Eur. J. Pharm. Sci., 26(2), 176–183 (2005).

29. K.A. MacRitchie, J.M. Newton and R.C. Rowe, Eur. J. Pharm. Sci., 17(1–2), 43–50 (2002).

30. S.C. Owen in R.C. Rowe, P. Sheskey and S. Owen, Eds, Handbook of Pharmaceutical Excipients (Pharmaceutical Press, London, UK 2006) pp 188–191.

31. F. Podczeck, International Patent WO 2008/001140 A2 (2008).

32. F. Podczeck and J.M. Newton, J. Pharm. Pharmacol., 40(2), 82–85 (1994).

33. F. Podczeck, S R. Rahman and J.M. Newton, Int. J. Pharm., 192(2), 123–138 (1999).

34. P. Schneiderhöhn, Heidelb. Beitr. Miner. Petrogr., 4(2), 172–191 (1954).

35. M. Salako, F. Podczeck and J.M. Newton, Int. J. Pharm., 168(1), 49–57 (1998).

36. C. Speirs, US Patent 5,834,021 (1998).

37. R. Chopra et al., Eur. J. Pharm. Biopharm., 53(3), 327–333 (2002).

38. M. Schröder and P. Kleinebudde, Pharm. Sci., 1(1), 415–418 (1995)