UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of formulation variables and compression on drug release from beads coated using a pseudolatex… Singh, Akaash 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-0243.pdf [ 6.49MB ]
Metadata
JSON: 831-1.0090285.json
JSON-LD: 831-1.0090285-ld.json
RDF/XML (Pretty): 831-1.0090285-rdf.xml
RDF/JSON: 831-1.0090285-rdf.json
Turtle: 831-1.0090285-turtle.txt
N-Triples: 831-1.0090285-rdf-ntriples.txt
Original Record: 831-1.0090285-source.json
Full Text
831-1.0090285-fulltext.txt
Citation
831-1.0090285.ris

Full Text

EFFECTS OF FORMULATION VARIABLES AND COMPRESSION ON DRUG RELEASE FROM BEADS COATED USING A PSEUDOLATEX DISPERSION OF ETHYLCELLULOSE by Akaash Singh B.A.Sc , The University of Toronto, 1991 A THESIS SUBMITTED IN THE PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Faculty of Pharmaceutical Sciences Division of Pharmaceutics and Biopharmaceutics We accept this thesis as^pnforming to the required standard The University of British Columbia December 2001 © Akaash Singh, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract A modified, drug release system was developed in which a drug-layered, inert substrate (sugar spheres, NF) was coated with a membrane formed from a pseudolatex dispersion of ethylcellulose (Surelease®). This system has the same drug release profile regardless of encapsulation or compression into a disintegrating or a non-disintegrating tablet. The study examined the effects of substrate size, coating level and formulation with various external excipients on drug release from the compressed spheres. A preliminary investigation of the effects of some compression variables was also conducted. Compression of the coated spheres resulted in non-disintegrating tablets in which the ethylcellulose coating fused to form an apparent matrix. At the 10% Surelease® coating level, the rate of drug release from encapsulated spheres increased as sphere size decreased, however, release from the two intermediate sizes (20-25 and 30-35 mesh size) were similar. When compressed into non-disintegrating tablets, drug Release was fastest for tablets compressed from the largest spheres (14-18 mesh size). Release profiles from the three smaller sizes were similar (i.e. within 10%). When drug release was compared between encapsulated and compressed spheres of the same size, release was faster for tablets of the largest size, slower for tablets of the smallest size (45-60 mesh size) and similar for tablets of the two intermediate sizes. Little or no control of drug release was seen for disintegrating tablets made from 10% Surelease® coated spheres regardless of excipient used. When 30-35 mesh spheres were coated at Surelease® levels of 20 and 30%, drug release rate from encapsulated spheres was observed to decrease as coating level ii increased. Upon compression, drug release was fastest at the 10% coating level however, release profiles were similar at the 20 and 30% coating levels. Drug release profiles of encapsulated and compressed spheres were similar at each coating level. Compression of the 20% and 30% ethylcellulose coated, 30-35 mesh spheres with different excipients resulted in faster release of the model drug as compared to encapsulated spheres at the corresponding coating level indicating damage to the control release membrane. Apparent drug release for all of the formulations slowed as coating level increased regardless of diluent used. The most effective of the three diluent materials examined was Starch 1500® which produced tablets with the slowest release. Examination of compression factors such as tablet porosity, work of compression and peak offset time yielded inconclusive results. However, the tensile strength behaviour of the spheres with respect to compressional force at the three coating levels appeared to deviate from the trends observed for the uncoated and drug-layered spheres. This apparent deviation supports the concept of change in bonding mechanism but again appears inconclusive relative to the effects on drug release. As a direct match in dissolution release profiles could not be found for the disintegrating formulations with the encapsulated and compressed spheres at the same ethylcellulose coating level, a mixed system was proposed. Ten percent Surelease® coated, 30-35 mesh size spheres were suggested for the hard gelatin capsule and the non-disintegrating tablet while the disintegrating tablet would be made from 30% Surelease® coated, 30-35 mesh size spheres compressed with Starch 1500®. iii Table of Contents Abstract ii Table of Contents iv List of Tables viii List of Figures x List of Abbreviations xiv Acknowledgements xv 1. Introduction 1 2. Background 3 2.1 Oral Controlled Drug Release 3 2.1.1. Overview of Multiparticulate Controlled-Release 3 2.1.2. Multiparticulate Sugar Substrate 5 2.2 Coating Methods 6 2.2.1. Solution-Based Coating 6 2.2.2. Pseudolatex Dispersion Film Coating 7 2.3 Coating Materials 9 2.3.1. Materials for Drug Layer 9 2.3.1.1. Model Drug 9 2.3.1.2. Povidone 10 2.3.1.3. Hydroxypropylmethylcellulose 11 2.3.1.4. Other Materials 12 2.3.2. Ethylcellulose 12 2.4. Sphere Compaction 13 2.4.1. Compression Event 13 2.4.2. Compression Parameters 15 2.4.2.1. Work of Compression and Peak Offset Time 15 2.4.2.2. Tensile Strength and Porosity 17 2.4.3. Mechanical Properties of (Uncoated and Coated) 18 Compressed Spheres 2.5 Tabletting Excipients 26 2.5.1. Diluents 26 2.5.2. Disintegrants 28 2.5.3. Lubricant 29 2.6. Drug Release from Spheres 30 2.6.1. Dissolution Testing • 30 2.6.2. Models for Drug Release 31 2.6.2.1. Membrane Diffusion 32 2.6.2.2. Matrix Diffusion 34 iv Table of Contents continued 2.6.3. Factors Influencing Drug Release from Spheres 35 2.6.3.1. Sphere Diameter 38 2.6.3.2. Coating 39 2.6.3.3. Compression with External Excipients 44 2.7. Data Evaluation 50 2.7.1. Data Modelling 50 2.7.2. Similarity Factor 51 3. Objectives and Specific Aims 55 4. Experimental 56 4.1. Materials 56 4.1.1. Model Drug 56 4.1.2. Substrate 56 4.1.3. Coating Materials 56 4.1.4. Tablet and Capsule Materials 57 4.2. Equipment 57 4.2.1. Coating Equipment 57 4.2.2. Balances 58 4.2.3. Dissolution Equipment 59 4.2.4. U V Spectrophotometer Systems 59 4.2.5. Compression Equipment 59 4.2.6. Miscellaneous Equipment 61 4.3. Methods 61 4.3.1 Drug Layering: Method Development and Final Method 61 4.3.2. Application of Ethylcellulose Membrane 63 4.3.3. Preparation of Various Dosage Units 64 4.3.3.1. Capsules 64 4.3.3.2. Non-disintegrating Tablets 64 4.3.3.3. Disintegrating Tablets 64 4.3.4. Dissolution Testing 65 4.3.5. Assay of Model Drug Content 66 4.3.5.1. U V Assay Method 66 4.3.5.2. Validation of Assay Method 67 4.3.6. Investigation of Dissolution Related Errors 67 4.3.6.1. Solvent Evaporation 67 4.3.6.2. Apparent Drug Gains and Losses Due to 68 Insoluble Materials 4.3.7. Formulation Development 69 4.3.7.1. Initial Formulations (25% Excipient Level) 69 4.3.7.2. Formulations Containing 50% Excipients 70 4.3.7.3. Evaluation of an Alternate Disintegrant 70 4.3.7.4. Evaluation of Alternate Brittle Diluents 71 v Table of Contents continued 4.3.7.5. Comparison of Variability in Dissolution 72 Profiles Between Tablet Preparations (Individual vs. Blend Sampling) 4.3.7.6. Final Formulations 72 4.3.8. Compression Studies 73 4.3.8.1. True Density Determination 73 4.3.8.2. Tablet Compression 73 4.3.8.3. Hardness Testing 73 4.3.8.4. Compression Data Evaluation 74 5. Results 75 5.1. Coating Method Development 75 5.1.1. Drug Layer Coating Solution Formulation Development 75 5.1.2. Ethylcellulose Application Development 76 5.2. Validation of U V Assay of Model Drug 79 5.3. Investigations of Dissolution Related Errors 81 5.3.1. Solvent Evaporation 81 5.3.2. Apparent Drug Gains and Losses Due to Insoluble 82 Materials 5.4. Formulation Development 84 5.4.1. Initial Formulations (25% Excipient Level) 84 5.4.2. Formulations Containing 50% Excipients 85 5.4.3. Evaluation of Alternate Disintegrant 86 5.4.4. Evaluation of Alternate Brittle Diluents 86 5.4.5. Comparison of Variability in Dissolution Profiles 87 Between Tablet Preparations (Individual vs. Blend Sampling) 5.5. Effects of Substrate Size on Drug Release from Encapsulated and 88 Compressed Spheres 5.6. Effects of Compression with Excipients on Drug Release 92 5.7. Effects of Ethylcellulose Coating Level on Drug Release from 93 Encapsulated and Compressed Spheres 5.8. Effects of Ethylcellulose Coating Level and Compression with 98 Excipients on Drug Release 5.9. Preliminary Compression Studies 102 5.10. Comparison of Profiles for Proposed System 103 6. Discussion 104 6.1. Coating Method Development 104 6.1.1. Drug Layer Coating Solution Formulation Development 104 6.1.2. Ethylcellulose Application Development 106 6.2. Validation of U V Assay of Model Drug 107 6.3. Investigations of Dissolution Related Errors 108 6.3.1. Solvent Evaporation 108 VI Table of Contents continued 6.3.2. Apparent Drug Gains and Losses Due to Insoluble 108 Materials 6.4. Formulation Development 109 6.4.1. Initial Formulations (25% Excipient Level) 109 6.4.2. Formulations Containing 50% Excipients 111 6.4.3. Evaluation of Alternate Disintegrant 111 6.4.4. Evaluation of Alternate Brittle Diluents 112 6.4.5. Comparison of Variability in Dissolution Profiles 113 Between Tablet Preparations (Individual vs. Blend Sampling) 6.5. Effects of Substrate Size on Drug Release from Encapsulated and 113 Compressed Spheres 6.6. Effects of Compression with Excipients on Drug Release 116 6.7. Effects of Ethylcellulose Coating Level on Drug Release from 118 Encapsulated and Compressed Spheres 6.8. Effects of Ethylcellulose Coating Level and Compression with 120 Excipients on Drug Release 6.9. Preliminary Compression Studies 122 6.10. Proposed System 123 7. Summary and Conclusions 127 7.1. Method and Formulation Development 127 7.2. Effects of Substrate Size on Drug Release from Encapsulated and 127 Compressed Spheres 7.3. Effects of Compression of 10% Ethylcellulose Coated Spheres 128 with Excipients on Drug Release 7.4. Effects of Ethylcellulose Coating Level on Drug Release from 129 Encapsulated and Compressed Spheres 7.5. Effects of Ethylcellulose Coating Level and Compression with 129 Excipients on Drug Release 7.6. Proposed System 130 8. Future Work 131 9. References 133 Appendices 148 A l Preliminary Compression Data 148 A2 Formulation Development Figures 169 vu List of Tables Table 1 Coating parameters (Initial and After Method Development) 61 Table 2 Typical formulation for Preparation of C P M Layered Spheres and 62 10% Surelease coated spheres Table 3 Dissolution Parameters 65 Table 4 Final Formulations for the Disintegrating Tablets 72 Table 5 Drug Layer Solution Formulation Development 75 Table 6 Comparison of Yields for 30% Ethylcellulose Application 77 Table 7 C P M Linearity Data for HP and Ocean Optics U V 79 Spectrophotometers Table 8 Drug Recovered from Dissolution Vessels Containing Various 82 Combinations of Excipients and Disintegrating Tablet Formulations Table 9 Results of Disintegration Testing for Alternate Brittle Diluent 86 Table 10 Estimated Similarity Factors for Encapsulated, 10% EC Coated, 89 C P M Layered Spheres of Various Substrate Size (Controls) Table 11 Estimated Similarity Factors for Compressed, 10% EC Coated, 90 C P M Layered Spheres of Various Substrate Size Table 12 Estimated Similarity Factors for Encapsulated and Compressed, 91 10% EC Coated, C P M Layered Spheres (Variable Substrate Size) Table 13 Estimated Similarity Factors for Encapsulated, C P M Layered, 30-35 94 Mesh Size Spheres at Various EC Coating Levels Table 14 Estimated Similarity Factors for Compressed, C P M Layered, 30-35 95 Mesh Size Spheres at Various EC Coating Levels Table 15 Estimated Similarity Factors for Encapsulated and Compressed, 96 C P M Layered, 30-35 Mesh Size Spheres at Various Coating Levels Table 16 Regression Coefficients or % Drug Released vs. Square Root of 97 Time for Compressed, EC Coated, C P M Layered Spheres Table 17 Estimated Similarity Factors for Disintegrating Tablet Formulations 99 Containing 20% EC Coated, C P M Layered, 30-35 Mesh Size Spheres viii List of Tables continued Table 18 Estimated Similarity Factors for Disintegrating Tablet Formulations 100 Containing 30% EC Coated, C P M Layered, 30-35 Mesh Size Spheres Table 19 Estimated Similarity Factors for Starch 1500® Formulations and 101 Encapsulated and Compressed, EC Coated, C P M Layered, 30-35 Mesh Size Spheres Table 20 Estimated Similarity Factors for Each Formulation for the Various 103 Ethylcellulose Coating Levels Table 21 Proposed System 124 IX List of Figures Figure 1 Schematic of Film (Solution-Based Coating) 6 Figure 2 Formation of Pseudolatex Film 7 Figure 3 Structure of Chlorpheniramine Maleate 10 Figure 4 Structure of Povidone 11 Figure 5 Structure of Ethylcellulose 13 Figure 6 Schematic of Sphere in Dissolution Fluid 33 Figure 7 Schematic of Coating Column 58 Figure 8 Dissolution Profiles for 30% EC Coated, C P M Layered, 30-35 78 Mesh Size Spheres Coated by 1 and 2 Step Coating Applications Figure 9 U V Spectrum of C P M in Water 79 Figure 10 Typical Beer's Law Plot of Chlorpheniramine Maleate 80 Figure 11 Dissolution Solvent Losses Due to Evaporation Over a 24- 81 Hour Period Figure 12 Absorbance of Water soluble Components of Starch 1500® at 83 262 nm Figure 13 Drug Release Profiles for Tablets Containing 10%-EC Coated, 84 C P M Layered, 20-25 Mesh Size Spheres (75%), Diluent (24%) and Stearic Acid (1%) Figure 14 Drug Release Profiles for Tablets Containing 10%-EC Coated, 85 C P M Layered, 30-35 Mesh Size Spheres (50%), Diluent (46%), Croscarmellose Sodium (3%) and Stearic Acid (1%) Figure 15 Dissolution Profiles of 10% EC Coated, C P M Layered; 30-3 5 87 Mesh Size Spheres Compressed with Sorbitol-Mannitol CSD and Other Excipients: (A) Tablets Prepared Individually and (B) Tablets Prepared from a Formulation Blend Figure 16 Drug Release Profiles for Encapsulated, 10% EC Coated, 89 C P M Layered Spheres of Various Substrate Size (Controls) x List of Figures continued Figure 17 Drug Release Profiles for Compressed, 10% EC Coated, C P M 90 Layered Spheres of Various Substrate Size Figure 18 Dissolution Profiles for Encapsulated and Compressed 10% 91 EC Coated, C P M Layered Spheres of Various Substrate Size Figure 19 Drug Release Profiles for 10% EC Coated, C P M Layered, 30- 92 35 Mesh Size Spheres Tabletted with Various Excipients Figure 20 Drug Release Profiles for Encapsulated, 30-35 Mesh Size 94 Spheres at Various EC Coating Levels Figure 21 Drug Release Profiles for Compressed, 30-35 Mesh Size 95 Spheres at Various EC Coating Levels Figure 22 Drug Release Profiles for Encapsulated and Compressed, 30- 96 35 Mesh Size Spheres at Various EC Coating Levels Figure 23 Drug Release for Compressed, EC Coated Spheres - % 97 Released vs. Square Root of Time Plot Figure 24 Dissolution Profiles for Disintegrating Tablet Formulations 99 Containing 20% EC Coated, 30-35 Mesh Size Spheres Figure 25 Dissolution Profiles for Disintegrating Tablet Formulations 100 Containing 30% EC Coated, 30-35 Mesh Size Spheres Figure 26 Dissolution Profiles for Disintegrating Tablet Formulations 101 Figure 27 Tensile Strength of Tablets Compressed from 30-35 Mesh 102 Size Spheres Coated with Various Coating Materials and EC Coating Levels Figure 28 Dissolution Profiles for Dosage Forms of Proposed System 103 Figure Al -1 Porosity and Peak Offset Time for 10% EC Coated Spheres of 148 Various Mesh Size Figure A1-2 Work of Compression and Tensile Strength for 10% EC 149 Coated Spheres of Various Mesh Size Figure Al -3 Porosity for 30-35 Mesh Size Spheres with Various Coating 150 Materials and EC Coating Levels XI List of Figures continued Figure A l - 4 Peak Offset Time for 30-35 Mesh Size Spheres with Various 151 Coating Materials and EC Coating Levels Figure A l - 5 Work of Compression for 30-35 Mesh Size Spheres with 152 Various Coating Materials and EC Coating Levels Figure A l - 6 Porosity for Diluent Materials and Formulations Containing 153 10% EC Coated, 30-35 Mesh Size Spheres Figure A l -7 Peak Offset Time for Diluent Materials and Formulations 154 Containing 10% EC Coated, 30-35 Mesh Size Spheres Figure A l - 8 Work of Compression for Diluent Materials and Formulations 155 Containing 10% EC Coated, 30-35 Mesh Size Spheres Figure A1-9 Tensile Strength for Diluent Materials and Formulations 156 Containing 10% EC Coated, 30-35 Mesh Size Spheres Figure Al-10 Porosity for Diluent Materials and Formulations Containing 157 20% EC Coated, 30-35 Mesh Size Spheres Figure Al-11 Peak Offset Time for Diluent Materials and Formulations 158 Containing 20% EC Coated, 30-35 Mesh Size Spheres Figure Al-12 Work of Compression for Diluent Materials and Formulations 159 Containing 20% EC Coated, 30-35 Mesh Size Spheres Figure Al-13 Tensile Strength for Diluent Materials and Formulations 160 Containing 20% EC Coated, 30-35 Mesh Size Spheres Figure Al-14 Porosity for Diluent Materials and Formulations Containing 161 30% EC Coated, 30-35 Mesh Size Spheres Figure Al-15 Peak Offset Time for Diluent Materials and Formulations 162 Containing 30% EC Coated, 30-35 Mesh Size Spheres Figure Al-16 Work of Compression for Diluent Materials and Formulations 163 Containing 30% EC Coated, 30-35 Mesh Size Spheres Figure Al-17 Tensile Strength for Diluent Materials and Formulations 164 containing 30% EC coated, 30-35 mesh size spheres Figure Al-18 Porosity for Disintegrating Tablet Formulations 165 Figure Al-19 Peak Offset Time for Disintegrating Tablet Formulations 166 Xll List of Figures continued Figure A l -20 Work of Compression for Disintegrating Tablet Formulations 167 Figure A1-21 Tensile Strength for Disintegrating Tablet Formulations 168 Figure A2-1 Drug Release for 10% Surelease® Coated Spheres (75%) 169 Compressed with Avicel (24%) and Stearic Acid (1%) Xll l List of Abbreviations CDER Center for Drug Evaluation and Research, US Food and Drug Administration CSD co-spray dried C P M chlorpheniramine maleate EC ethylcellulose f2 similarity factor A f2 estimated similarity factor HP Hewlett Packard HPMC hydroxypropylmethylcellulose L V D T linear variable-differential transformer M C C microcrystalline cellulose ® registered trademark US FDA United States Food and Drug Administration USP United States Pharmacopoeia xiv Acknowledgements The completion of this work could not have been accomplished without the aid and support of certain individuals. Included in this group are members of my research committee members: Dr. Helen Burt, Dr. Thomas Chang, Dr. Kishor Wasan and Dr. Ron Reid (chair). Gratitude is given to Colorcon Inc. for financial support over the course of this study. Special acknowledgement must be given to Ms. Lee Shean Er who provided statistical advice for the compression studies through the Statistical Consulting and Research Laboratory (SCARL), Department of Statistics, University of British Columbia. Special thanks is given to Dr. Bob Miller for all of his patience and gentle mentoring. Finally, warmest thanks are given to my family and friends who have provided endless emotional support and understanding throughout this experience. Thanks for always being there when I needed you and rocking my world as needed. xv 1. Introduction Multiparticulate systems used in oral drug administration are dosage forms comprised of many small units, each containing a fixed quantity of the required drug and whose release of drug is independent of any other unit. The nature of the materials used and the methods of preparation give rise to a wide variety of units thus permitting flexibility and control over the observed drug release. Additionally, a desired release profile can be tailored by combining different types of units (i.e. units releasing drug differently) into one dosage unit. Membrane coated and uncoated spheres offer a means of controlled release through a variety of release mechanisms and can be formulated into suspensions, capsules and tablets (Ghebre-Sellassie, 1989). Currently, encapsulated, polymer-coated, drug-layered, multiparticulate systems have gained wide acceptance as a reliable means of controlling drug release. Encapsulation provides protection to the rate-controlling membrane on individual units during normal handling. There has been some interest in compressing membrane coated spheres into tablets which retain the release characteristics of the encapsulated dosage form. Ideally, the tabletted spheres would disintegrate into the intact, individual units after administration and the observed drug release profile would be unchanged from that of the encapsulated dosage form (Bodmeier, 1997). However, attempts to compress such multiparticulate units into tablets have been largely unsuccessful because of compression-induced damage to the membranes and changes to the drug release profile (Chiao and Price, 1994; Vilivalam and Adeyeye, 1994; Sarisuta 1 and Punpreuk, 1994; Lopez-Rodriguez et al., 1993; Torrado-Santiago et al., 1995 and Miller et al., 1999). The primary goal of this study was to develop a multiparticulate system based on ethylcellulose coated, drug-layered spheres where the drug release profile was similar regardless of encapsulation of the spheres or compression into disintegrating and non-disintegrating tablets. Ideally, the proposed spheres could be used in the preparation of any of these three dosage forms. Additionally, several compression parameters for each material were studied to investigate possible associations between the compression behaviour and the utility of the materials with respect to dissolution. The dissolution work formed the major component of the study. The effects of substrate size on drug release from encapsulated and compressed spheres were examined at a uniform membrane coating level to establish the most appropriate size. The effects of compression with external excipients on drug release were then investigated. Finally, the effects of membrane coating level on drug release with respect to all three dosage forms were examined. Preliminary compression work was conducted on all individual materials and mixtures of materials to investigate any possible trends in the compression behaviour. 2 2. Background 2.1. Oral Con trolled Drug Release 2.1.1. Overview of Multiparticulate Controlled-Release Spheres can be made by extrusion-spheronization (Vervaet et al., 1995; O'Connor and Schwartz, 1989; and Zimm et al., 1996), drug layering (Hutchings and Sakr, 1994 and Li et al., 1995) and compression (Kaewvichit and Tucker, 1994 and Batra et al., 1994). Briefly, extrusion-spheronization consists of wet granulation of the components, extrusion and finally spheronization of the extrudate (Vervaet et al., 1995). Dry components of the spheres are blended together and mixed with solvent, typically water, to achieve the proper consistency for extrusion. This wet material is then forced through a perforated plate (extrusion). The final sphere diameter is often dependent on the size of the openings on the plate but in some cases, the final diameter may also depend on the nature of the formulation itself. In the spheronization step, the cylindrical extrudate is broken into smaller particles which are smoothed to produce a spherical shape. Drug layered spheres are generally prepared by applying layers of coatings onto inert spherical cores. The drug is dispersed within one or more of the layers. A rate-controlling membrane is often applied after drug application to retard drug release. Spheres made by compression are more aptly described as cylindrical spheres or even discs. These spheres are prepared by blending the necessary components and compressing them to form a small sphere. The size and dimensions of the sphere will depend on the tooling used in the compression. 3 Spheres can be compressed into tablets which can release drug at a controlled rate (Chiao and Price, 1994; Vilivalam and Adeyeye, 1994; Lopez-Rodriguez et al., 1993; Torrado-Santiago et al., 1995 and Miller et al., 1999). There has been interest in the literature regarding tabletting of controlled-release spheres. Traditionally, sustained-release spheres are encapsulated in gelatin capsules. Tablets which are made from such spheres and retain the release characteristics offer potential advantages over capsules (Bodmeier, 1997). The major advantage is cost. Manufacture of capsules tends to be more expensive than the manufacture of tablets. Also, capsules can be more easily tampered with than tablets and thus require additional steps (and cost) in their manufacture and packaging to minimize this risk. Another advantage is in the area of dietary limitations. Gelatin is produced mainly from pork and beef sources making it unsuitable for people whose religion (e.g. Judaism, Islam and Hinduism) or dietary philosophies (e.g. vegetarianism) forbid ingestion of such animal products. However, compaction of the spheres when preparing the tablets may cause some complications. The most commonly observed problem with tabletting of sustained-release spheres is the change in the drug release profile (Chiao and Price, 1994; Vilivalam and Adeyeye, 1994; Sarisuta and Punpreuk, 1994; Lopez-Rodriguez et al., 1993; Torrado-Santiago et al., 1995 and Miller et al., 1999). This effect has been observed in both uncoated and coated spheres. Tabletting of matrix type spheres (coated and uncoated spherical granules and polymer matrix) (Chiao and Price, 1994; Vilivalam and Adeyeye, 1994; Lopez-Rodriguez et al., 1993; Torrado-Santiago et al., 1995; Maganti and Celik, 1993; Maganti and Celik, 1994 and Bansal et al., 1993) and coated drug layered spheres (Miller et al., 1999 and Beckett et al., 1996) has been investigated in order to understand the behaviour of spheres 4 during the compaction event. Many researchers have tried to minimize the effects of compaction on the spheres by including an external diluent (e.g. microcrystalline cellulose) in the formulation. In most cases however, changes in the drug release profile were still observed, though not to the same extent as seen without the addition of the diluent (Torrado-Santiago et al, 1995; Yao et al., 1997 and Yao et al., 1998). 2.1.2. Multiparticulate Sugar Substrate This material is typically used as inert cores in multiparticulate, sustained-release formulations. The active drug can be layered onto the surface of the spheres and then a controlled-release membrane can subsequently be applied or a drug and polymer matrix layer can be applied (Wade and Weller, 1994). Sustained-release can be attained by combining appropriate amount of coated spheres with different release rates such that the targeted drug release profile is achieved. Sugar spheres have also been referred to as nonpareil spheres, beads and seeds. Sugar spheres NP are produced by continuous sugar coating of sugar crystals. Starch is used in the coating process as an anti-tack agent. The final batch is passed through a series of screens to separate out the individual size fractions. Similar size fractions are combined such that any given fraction received from the manufacturer may contain spheres produced from several coating batches (Wade and Weller, 1994). The sucrose content of the final product is 62.5 - 91.5 % (dried basis) (USP, 1999). Sugar spheres were selected as the substrate in this project because the material is inert, compressible, mainly spherical and available in different sizes. The sizes used in this study were: 14-18, 20-25, 30-35 and 45-60 mesh fractions. This corresponds 5 approximately to mean diameter ranges of 1000-1400 u,m, 710-850 p.m, 500-600 u.m and 250-355 urn respectively (Crompton and Knowles Corp., 1995). The similarity in sphere diameters within any designated mesh fraction provides batch-to-batch uniformity of surface area and permits some control over the thickness of the applied film. 2.2. Coating Methods 2.2.1. Solution-Based Film Coating The drug layer was formed by solution based coating technology. This process is also referred to as solvent coating as organic solvents are typically used. However, water can be used as the solvent depending on the details of the individual coating method. The film-forming polymer additive coating surface Figure 1: Schematic of Film (Solution-Based Coating) dissolved, film-forming polymer exists as long, freely moving chains within the coating solution. Additives such as plasticisers, stabilisers, dyes, opacifiers and drug are also dissolved or dispersed in the solution. When the atomised solution is deposited onto the coating surface, the chains spread out. As more droplets are deposited, the polymer chains intertwine to form a continuous film on the coating surface. The rate of solvent 6 evaporation will affect the arrangement of the chains. Slow solvent evaporation provides an opportunity for the chains to arrange into a somewhat ordered fashion. The resulting film is relatively dense. Conversely, rapid evaporation leads to a more random arrangement and a less dense film. Use of water as the solvent can result in a somewhat dense film because of the relatively slow rate of evaporation. Any additives will exist between the polymer chains (see Figure 1) (Wheatley and Steuernagel, 1997). 2.2.2. Pseudolatex Dispersion Film Coating Latex and pseudolatex films are formed differently from solution based coating films. Firstly, the film-forming polymer exists as discrete solid particles in the aqueous vehicle. Plasticisers, polymer stabilisers and other polymer soluble additives are contained within the particle. Water soluble additives may also be required and these are dissolved in the vehicle. Latex or pseudolatex film formation is depicted in Figure 2. Atomised Droplet deposition contact (c) Deformation (d) Coalescence and fusion Figure 2: Formation of Pseudolatex Film 7 coating dispersion is deposited as droplets onto the coating surface as seen in Figure 2(a). Water is evaporated off but the particles are still independent of each other. As evaporation continues, the particles are brought into close contact as seen in Figure 2(b). Further evaporation leads to deformation of the particles, Figure 2(c), and finally, fusion or coalescence of the particles into a continuous film, Figure 2(d) (Hogan, 1995). Curing or additional drying time may be necessary once the spheres are coated depending on many factors including the nature of the dispersion used in the coating and the coating parameters. This additional drying time accelerates a slower secondary phase of coalescence of the ethylcellulose particles. Guo, Robertson and Amidon suggested a variation on the film formation (Guo et al., 1993). Based on their studies in which permeability was measured through pseudolatex films prepared below and above 100°C, the researchers postulated that the film dried and coalesced progressively from the top to the bottom. When a drying temperature greater than 100°C is used, water vapour escapes rapidly leaving behind observable "pinholes" in the film. Latex and pseudolatex films differ only in the starting material. Latex dispersions are prepared from emulsion polymerization of the monomers. Pseudolatex dispersions are prepared from polymers which are already at the desired chain length and substitution level. In one representative process, the polymer is dissolved in an organic solvent along with other stabilisers. Water (containing other additives) is added gradually with agitation until it becomes the vehicle. The organic solvent is removed from the emulsion (again with continuous agitation) leaving behind solid or semi-solid, spherical particles of polymer (VanderhofTet al., 1979). 8 Variations in the structure of the film can result from many factors. For example, Sun and others examined permeabilty and mechanical properties of ethylcellulose films prepared from casting (i.e. pouring the dispersion into a rimmed plate) and from spraying onto a surface (Sun et al., 1999). The researchers found that the sprayed films were more porous and had a greater permeabilty (to water vapour) than the cast films. The porous nature was attributed the method of preparation in which air bubbles could be easily trapped as droplets of dispersion coating were deposited during film formation. Curing of the sprayed films resulted in a less brittle film and lower permeability. This study illustrates the need to be careful when extrapolating cast-film permeability to films formed from spraying. 2.3. Coating Materials 2.3.1. Materials for Drug Layer 2.3.1.1. Model Drug Chlorpheniramine maleate (CPM) was selected as the model drug. Its structure is depicted in Figure 3. This drug is very water soluble at 160 mg/mL of water @ 25°C and hence, "sink conditions" can easily be maintained during dissolution testing. "Sink conditions" refers to a state in which the concentration of drug in the bulk dissolution fluid is always less than one-tenth of the drug's solubility in that medium. This provides a continuous concentration gradient on either side of the control release membrane and also permits some simplification of first-order drug release mathematical models (Abdou, 1989). Aqueous chlorpheniramine maleate is easily detected by U V spectrophotometry 9 CHCOOH CHCOOH Figure 3: Structure of Chlorpheniramine Maleate and absorbs strongly @ 261 nm (molar absorptivity of 5.76 x 10"3 mL-mor'-cirT1). The aqueous drug is also very stable. Ninety-seven percent recovery was observed after 1 week @ 95°C in pH 7 buffer and 102% recovery in pH 6 and 8 buffer was observed after 3 months @ 25°C and 350 foot candles as determined by paper chromatography (Eckhart and McCorkle, 1978). Chlorpheniramine maleate has been used as a model drug in other in vitro release studies (Mathir et al., 1997, Wesseling and Bodmeier, 1999 and Tang et al., 2000). 2.3.1.2. Povidone Polyvinylpyrrolidone or povidone has a variety of uses in pharmaceutical formulation. In this study, it was used as the film-forming polymer in the drug layer. Its primary purpose was to provide a means for the model drug to be attached and held to the surface of the substrate (i.e. sugar sphere). Also, a plasticiser is not necessary to impart further flexibility to the film. An hydro-alcoholic solution of this material was used as the coating solution for the drug layering process. The alcoholic content of the vehicle reduced the coating application time. 10 The structure of the monomer, vinylpyrrolidone, appears in Figure 4. The different grades of povidone are designated by a K-value which describes its viscosity in aqueous solution relative to that of water. The viscosity will depend on the aqueous concentration (which is factored into the K-value calculation) and the molecular weight of the polymer. Grades with a K-value of 30 have an (approximate) molecular weig ht of 5 x 104 Da. Other available grades range between K-values of 12 and 120 corresponding to approximate molecular weights of 2500 and 3 x 106 Da respectively. Povidone is freely soluble in water, ethanol and a variety of other solvents. No incompatibilities were identified with its use in this study (Wade and Weller, 1994). N O CH C H 2 " —'n Figure 4: Structure of Povidone 2.3.1.3. Hydroxypropylmethylcellulose Hydroxypropylmethylcellulose (HPMC) was examined as a possible film-forming polymer for the drug layer. It is typically used as a tablet binder and in film coating. HPMC is a chemically modified product of cellulose refined from cotton waste and wood 11 pulp. It is soluble in cold water producing a viscous solution. Different viscosity grades are commercially available. Lower viscosity grades are used in aqueous film coating solutions (Wade and Weller, 1994). This material was selected initially as the film-former for the drug layer because it can be applied via an aqueous solution. However, it was later discarded as excessive tack was observed when model drug was included in the coating solution. 2.3.1.4. Other Materials Talc was used as an anti-tack agent during HPMC coating. It is a purified, hydrated magnesium silicate. It is mined from naturally occurring talc deposit and then processed and purified. The final product may contain small, variable, amounts of aluminium silicate and iron. It is practically insoluble in water (Wade and Weller, 1994). Solvent vehicles used were glass-distilled water and a water-alcohol mixture. Water was the preferable solvent as it is more environmentally-friendly, less toxic (should trace amounts remain in the spheres after preparation) and relatively inexpensive. These would be important considerations should such a system be adapted to large-scale production. Advantages of the aqueous-ethanolic vehicle include increased drying rate of the film and reduction in agglomerate formation. 2.-3.2. Ethylcellulose Surelease® is a commercially available pseudolatex dispersion of ethylcellulose in water and contains 25% w/w total solids. Figure 5 depicts the structure of ethylcellulose. The dispersion already contains any necessary plasticisers (e.g. dibutyl sebacate), stabilisers (e.g. oleic acid and ammonia) and antitack agents (e.g. fumed silica). This 12 material need only be diluted to the appropriate solids content with distilled water immediately prior to coating application (Moore, 1989). Ethylcellulose is a water insoluble derivative of cellulose typically used in modified-release films. In the dispersion, ethylcellulose exists as small, spherical particles within the aqueous vehicle. Other insoluble materials are distributed throughout the particles such that the particles are discrete, compositionally identical units. Surelease® was selected for the ethylcellulose coating as it is a representative commercial preparation and is designed to provide the type of modified-release coating required for this study. 2.4. Sphere Compaction 2.4.1. Compression Event In general, compaction of powders or spheres proceeds as follows: die fill, volume reduction, deformation, consolidation, decompression, ejection and post compression (Miller, 1996). Die fill involves the filling of the die cavity with the required amount CH 2 OCH 2 CH 3 A o n Figure 5: Structure of Ethylcellulose 13 (usually by volume) of the powder or spheres. When the upper punch initially contacts the loosely arranged material, the pressure forces the particles or spheres to move into a more compact arrangement. Further pressure from the punches leads to deformation of the particles or spheres. At low forces, this deformation is generally elastic in which case the particles will return to their original shapes if the pressure is removed. At any point, brittle fracture may occur. In such cases, particles or spheres may physically break rather than deform as pressure is applied. Reaction to the applied force depends on the nature of the materials being compressed. As the compression pressure continues to approach the peak or maximum applied pressure, consolidation of the compact begins. In this phase, the materials can undergo further brittle fracture or plastic deformation in which permanent changes to the shape or structure of particles occur. Adhesion (bonding between unlike materials) and cohesion (bonding between like materials) also occurs during this time. The amount of time during which the compressed material is subjected to the maximum applied force is termed the dwell time and its length can affect the degree of adhesion or cohesion between particles in the compact. The decompression stage begins with the upward motion of the upper punch and hence a decline in the amount of applied force. Elastic recovery during this time results in axial expansion of the compact. This may result in lamination of the compact if the elastic recovery is stronger than the adhesive or cohesive bonds formed during consolidation. As the compact is ejected from the die cavity (ejection phase), the radial expansion due to elastic recovery may result in capping. Post compression can range from minutes to months after compression. Further elastic recovery, release of heat or slow changes to the material due to residual energy from the compression event (e.g. recrystallization, degradation of a material) may occur. 14 2.4.2. Compression Parameters Parameters determined during the compression event include peak pressure, compact porosity, work of compression, compression time and a unique parameter defined as peak offset time. Peak offset time is a function of the rate and extent of a material's reaction to force at the peak force levels. During decompression, values for work of decompression and for the Young's modulus (related to elasticity) are determined. These values can be used to calculate other parameters which can then be used as a basis of comparison for different formulations. The above parameters can be calculated by methods developed by Dwivedi, Oates and Mitchell (Oates and Mitchell, 1989: Oates and Mitchell, 1990; Dwivedi et a l , 1991; and Dwivedi et al., 1992). Comparisons of different formulations are often done by comparing various parameters related to the compression event (in-die parameters) and characteristics of the tablet (compact parameters). Compression characteristics include work of compression, elasticity, work of decompression and total work of compression. Parameters characterizing the compact itself (micromeritic properties) include tensile strength (or breaking strength), diametric strain and porosity. 2.4.2.1. Work of Compression and Peak Offset Time The work of compression (Wc) is the total amount of mechanical energy imparted by the machine to the powder bed up to and including the consolidation phase of the compression event. This energy can be calculated by measuring the forces applied by the upper and lower punches and the distances through which they move. This was mathematically expressed by Maganti and Celik (1993) as: 15 w c = r = c J x= ( ( « p ) F dx + \ * up up J = X x= 0 Equation 1 where F u / J and F/p the forces applied by the upper and lower punches respectively xup and xip the distances moved by the upper and lower punches respectively. xmax(up) and xmax(lp) = the distance for the upper and lower punch respectively Oates and Mitchell used a variation of Equation 1 to calculate the work of compression on a rotary tablet press taking into account the deformation of the punches and the machine (Oates and Mitchell, 1989). In their study the work of compression was estimated as the total work done by the press less the amount of work done on the machine (i.e. any deformation). Since the geometry of the contact of upper and lower rollers with the punches are identical and the pressure is balanced, Equation 1 can be simplified to a single integral term incorporating the force and differential displacement. Since pressure can be expressed as force per unit area, the force term can be replaced by the product of pressure and cross-sectional area. Similarly, since volume is the product of the cross-sectional area and the displacement, the displacement differential can be replaced by the product of the volume differential and the cross-sectional area. Replacing the force and displacement factors with pressure and volume factors yields Equation 2 which is another simplified way when the maximum load is reached. x=0 is the point at which the porosity of the compact is equal to the initial porosity of the tablet blend. 16 of expressing the work of compression. The work of compression is integrated over the compression phase of the tabletting event. The other in-die parameter examined was peak offset time. This parameter is defined as the time interval between the time of peak pressure and the time of position of dead centre, the point at which the central axis of the press rollers and punches are in alignment. It has been found that this time interval varies depending on the material and may be a intrinsic property of the material (Dwivedi et al., 1991). 2.4.2.2. Tensile Strength and Porosity The tensile strength (!TS) is the amount of radially applied force needed to break the compact (Maganti and Celik, 1993) and is determined following compression. In other words, the tensile strength is a measure of the strength of the compact and how much force it can withstand and remain intact. This can be measured by commercially available hardness testers, preferably those which linearly apply force. It is mathematically represented by the following equation: Equation 2 where P = compression pressure and V volume of the material being compressed 2F c Equation 3 where Fc = force required to break the compact diameter of the compact 17 He = thickness of the compact The porosity is the percentage of the volume of the compact due to void spaces (Ansel et al., 1995) and is calculated by equation 4: V -V _ apparent true „ „ P = — ^ x 100 Equation 4 ^apparent where VappareM - volume based on the outer dimensions and Vtrue = true volume of the compact (i.e. the actual volume occupied by the matter of the compact) In terms of density, equation 4 can be expressed as: P = { P ,^ r apparent x 100 Equation 5 P,rue J where Papparem and ptruil are the apparent and true densities of the compact respectively. The relative amount of void space in a compact as well as the known nature of the material can indicate the arrangement of particles. 2.4.3. Mechanical Properties of Compressed (Uncoated and Coated) Spheres Various researchers have investigated the mechanical properties of compressed spheres. Often, mechanical properties such as tensile strength, compression force and work of compression are evaluated when spheres are tabletted. Maganti and Celik studied some of the effects of compaction of uncoated spheres as compared to powder blends of the same formulation (Maganti and Celik, 1993). The model drug (propranolol hydrochloride) and microcrystalline cellulose powders were combined and some of the blend was made into spheres by peptization (mesh size range of 18 to 35). Two additional formulations, one including lactose and the other including 18 dicalcium phosphate, were blended. Spheres were made from each of these formulations by pelletization. Spheres were dried to a moisture content matching that of the original powder blend. Heckel plots, total work of compression, average power consumption, tensile strength and elastic recovery (axial and radial) of the compacts made from the powder blend were compared to those of compacts compressed from the spheres. The maximum applied pressure required to produce a compact with a 3% in-die porosity was lower for the spheres than that for the corresponding powder formulation. A Heckel plot follows the Athy-Heckel equation which relates the log of the inverse of porosity with the applied pressure (i.e. log [E]"1 = KP + A) (Maganti and Celik, 1993 and Watt, 1988). The slope of the linear portion of the plot, K, is related to the inverse of the yield strength of the material and reflects the total ability (plastic and elastic) of the material to deform. The y-intercept of the plot, A, represents the degree of packing at low pressures during particle rearrangement. Such plots can provide information regarding the consolidation behaviour of the materials of the powder bed and allow for a means to compare different materials. Using the Heckel plots of the microcrystalline cellulose and drug formulations, the slope of the linear portion of the profile for the powder was less than that of the spheres. The researchers theorized that changing the shape and size of the particles may affect compaction properties such as degree of bonding. The amount of energy required to form intraparticle bonds would also be reflected in the tensile strength and total work of compression. The amount of energy absorbed by the powders during compression (i.e. the total work of compression) was greater than that absorbed by the spheres. In addition, the tensile strengths of the compacts made from spheres were much lower (0.5 to 0.6 MPa) than those for the compacts made from powders (8 to 11 MPa). However, the 19 average power consumption with respect to applied pressure was greater for the spheres than for the powders. Normally, a material which has a higher average power consumption will produce a stronger compact. The researchers suggested that this unexpected result might be due to the shorter contact time between the upper punch and the material (due to the absence of particle rearrangement which was observed in the case of the powders). To further study the consolidation mechanism of the materials, Maganti and Celik prepared compacts of the powder blends and spheres using different punch velocities since the rate of force application can affect the consolidation of a material (Maganti and Celik, 1993). A difference in the yield pressure (from the Heckel plots) and the tensile strength of the compacts made from powders was observed suggesting a time-dependent deformation during compaction. However, the two parameters were not significantly different at the different punch velocities for the compacts made from spheres. This again suggested that the compacts were formed by brittle fragmentation. The addition of lactose or dicalcium phosphate into the powder formulation resulted in a need for higher pressure in order to form compacts with the desired in-die porosity. This observation suggested that the compressibility of the microcrystalline drug blend was decreased by addition of the third component. The compacts made from the powder blend containing lactose exhibited a higher total work of compression and tensile strength than the powder blend containing dicalcium phosphate. However, the compacts made from the spheres (all formulations) exhibited similar total work of compression and tensile strength values. The researchers suggested that the powder blends underwent time dependent deformation during compaction while the spheres exhibited brittle 20 fragmentation. It was also observed that the compacts made from spheres exhibited a higher elastic recovery than those made from powders. The higher elastic recovery suggested that many of the bonds formed during compaction did not survive and hence the tensile strength was low. The drug release rates of the spheres and sphere compacts were very similar and displayed a very fast rate of drug release (95% of drug released within 5 minutes). The tablets made from the powder blends however, released drug more slowly (95% of drug released at 60 minutes) (Maganti and Celik, 1993). Maganti and Celik went on to study the effects of compaction on coated spheres (Maganti and Celik, 1994). Spheres made from a model drug (propranolol hydrochloride), microcrystalline cellulose, and either lactose or dicalcium phosphate were coated with Surelease® at coating levels of 10, 15 and 20 %w/w. All resulting compacts disintegrated within 64 s. The compacts made from uncoated spheres exhibited 100% friability while the friability of compacts of coated spheres increased with increasing coating level. Heckel plots of the compaction data revealed that the slope of the linear portion of the plot increased with coating level for both formulations. The slopes for the two formulations at identical coating levels were very similar. This implied that the yield strength (proportional to the inverse of the slope of the linear portion of the Heckel plot) decreased with increasing coating level. The maximum applied pressure needed to form compacts of the desired in-die porosity (3%) decreased as coating level increased. According to the researchers, this decrease indicated that the higher the coating level, the greater the compressibility of the spheres. However, this increase in compressibility with coating level was not accompanied by increases in tensile strength. Though the tensile strength of the compacts made from coated spheres was greater than that for the uncoated 21 spheres, the tensile strength decreased with increasing amount of coating. In addition, the total work of compaction appeared to decrease with increasing coating levels. The researchers theorized that the observed decrease in tensile strength may be due to the binding properties of Surelease®. The coating may be acting as an adhesive and at the lowest coating level studied, bonding of the compact may have been a combination of adhesive bonds between the coating, cohesive bonds between the coating and the substrate and substrate-substrate bonds (adhesive and cohesive). As coating levels increase, the ratio of adhesive bonds between the coatings on the spheres to cohesive bonds increased and thus, the tensile strength was reduced due to lower cohesive properties of the compact. The elastic recovery of the compacts made from coated spheres increased with increasing coating level. However, the elastic recovery of the compacts made from uncoated spheres was between that of the compacts made from 10% and 15% coated spheres. The researchers proposed that this observation was an effect of the elastic properties of the uncoated spheres and the film coating. Maganti and Celik also studied the effects of punch velocity on compacts made from coated spheres (Maganti and Celik, 1994). The observed yield pressure (from the Heckel plots) at the faster punch velocity (100 mm/s) was smaller than that for the slower punch velocity (1 mm/s) while the tensile strength was lower at the slower punch velocity from compacts with the same coating level. The elastic recovery was also larger for compacts produced with the faster punch speed for the same coating level. These results may have been due to a lower amount of plastic deformation which is known to be time dependent. Since the faster punch velocity occurs in a shorter amount of time as compared to a slower punch velocity, less inter-particle bonding and plastic deformation 22 was produced during compaction and hence the differences in tensile strength yield pressure and elastic recovery. Also, a material will resist further densification if the rate at which the load is applied exceeds the rate at which the material can react to the force. Elastic energy which is not used for bonding is stored as deformation energy during compaction and its release during decompression and ejection can cause rupture of weak particle-particle bonds. The total work of compaction at any given coating level was higher for the faster punch velocity. The researchers attributed this to the high energy input needed for fragmentation and elastic deformation of the spheres. Miller and others investigated the mechanical properties of tablets compressed from sugar spheres, sucrose, NF and corn starch, NF (Oates and Miller, 1996). The study was performed to gain an understanding of the behaviour of compacted spheres. Tablets were prepared from sucrose (powder), corn starch (powder), an 80:20 (powder) mixture of sucrose and starch, sugar spheres and ground sugar spheres. The tablets were evaluated for tensile strength, porosity, work of compression, Young's modulus and diametric strain. An increase in tensile strength with increasing peak pressure was observed for all formulations. The tablets produced from the pristine and ground spheres had higher tensile strength values than those of the constituent powders. The researchers suggested that the higher strength of the tablets made from spheres (pristine and ground) was due to the manufacturing process in which powdered sucrose and starch were layered onto a sucrose crystal in the presence of water. Upon evaporation of the water, bridges were formed leading to greater strength in the resulting spheres. Porosity was observed to decrease with increasing peak pressure for all the materials. The rate of change in porosity with respect to increasing peak pressure was lowest for sucrose and the sucrose-starch 23 mixture. Pristine and ground spheres exhibited higher rates of porosity reduction at low pressures but any porosity changes were minimal above peak pressures of 125 MPa. Starch had a high initial porosity which reduced to near zero at 175 MPa. Unfortunately, this decrease in porosity was not accompanied by a great degree of interparticulate bonding as observed by the low tensile strength of the starch compact. The work of compression increased with increasing peak pressure. The lowest works of compression were observed for tablets made from sucrose, the sucrose-starch mixture and ground spheres. The work of compression for tablets made from spheres was consistently higher than the three other formulations but was less than that for tablets made from starch. The researchers suggested that the higher work of compression for tablets made from spheres was due to the greater amount of energy needed to compress the particles and fill the large void spaces. The high work of compression observed for the starch tablets was consistent with the high porosity change over the pressure range studied. The Young's moduli for the tablets prepared from starch and from spheres were low indicating a lack of brittleness in those materials. The Young's moduli were high for the ground spheres, sucrose and sucrose-starch mixture indicating that these materials had a more brittle nature. The change in Young's moduli for the pristine and ground spheres suggested that grinding affects the brittleness of the material. Diametric strain indicates the degree of deformation the tablet can undergo before failure due to diametric stress. The diametric strain for the starch and sphere tablets was highest for all of the formulations suggesting a more flexible material. The sucrose-starch mixture, ground spheres and sucrose had lower values for diametric strain suggesting a more brittle material. 24 Miller and others continued their work using coated spheres (Miller et al., 1999). Sugar spheres, NF, were coated with an aqueous solution of a dye (model compound) and hydroxypropylmethylcellulose (HPMC). The spheres were then coated with an aqueous etheylcellulose dispersion (Surelease®) at coating levels of 5 and 10%. The dissolution profiles of the spheres were determined. The Surelease® coated spheres retarded the release of the dye. Slower release was observed for the 10% coating level than for the 5% coating level (t5o% at approximately 6 hours and 4 hours respectively). Release of dye from the HPMC coated spheres was very fast (tso% at approximately 1 hour). Flat-faced tablets were made of uncoated spheres, spheres coated with the HPMC layer, spheres coated with the HPMC layer and 5% Surelease® and spheres coated with the HPMC layer and 10% Surelease®. The tablets were evaluated for tensile strength, work of compression, porosity, peak offset time, diametric strain and dissolution. Tensile strength of the tablets increased with increasing peak pressure and the values for the different formulations were comparable at low peak pressures. However, the tensile strength of the tablets made from spheres coated with 5% ethylcellulose was slightly lower at high peak pressures. A similar pattern was observed for the work of compression versus peak pressure plots except that the works of compression for Surelease® coated spheres were slightly lower at higher peak pressures. Similar decreases in porosity with increases in peak pressure were observed for all formulations. No discernible patterns could be seen in the Young's moduli for any of the formulations. However, differences in the diametric strain were observed for the different tablets. Tablets made from uncoated and HPMC coated spheres had similar values for diametric strain. However, the diametric strain values increased with the addition of an ethylcellulose coating. A larger value was 25 observed for the 10% ethylcellulose coating than for the 5% coating level. The researchers suggested that these changes were due to the flexibility of the plasticized ethylcellulose which accommodates greater deformation of the compact prior to failure. 2.5. Tabletting Excipients 2.5.1. Diluents The major criterion for the selection of diluent materials was their compression behaviour. When compressed, materials can undergo deformation (both elastic and plastic) and brittle fracture. Most materials exhibit a combination of these two reactions. Three materials were selected: a primarily plastic material, a primarily brittle material and a combination material. As compression-induced damage to the ethylcellulose membrane was inevitable, these three types of materials were selected to examine how their respective compression behaviours affected the extent of membrane damage. Secondary considerations included compressibility and disintegrant properties. Diluents had to be inherently compressible so that tablets could be prepared by direct compression. Microcrystalline cellulose (MCC) is cellulose from which the amorphous portions are removed by mineral acids and which is then washed and spray dried. It is water insoluble. Microcrystalline cellulose has a variety of uses in tabletting including as a diluent and as a disintegrant. This material is directly compressible and can produce very strong tablets that disintegrate rapidly when in contact with aqueous media (FMC Corp., 1984). It tends to undergo plastic deformation when subjected to typical compression forces. The grade used in this study had a particle size of approximately 90 urn and a moisture content of < 5% (FMC Corp., 1984). 26 Pregelatinised starch is processed from corn starch such that the final product is flowable and directly compressible. This is achieved by rupturing all or part of the starch granules via chemical and mechanical processes (Wade and Weller, 1994). Starch 1500®, a commercially available pregelatinised starch product, contains 80% unmodified corn starch (Colorcon Inc., 1993). This material possesses intrinsic disintegrant properties and gels when in contact with water (Ferrrari et al., 1997, Van der Voort Maarschalk et al., 1997 and Hudson et al., 2000). It is mostly water insoluble although a small amount of water soluble components may be present. Pregelatinised starch undergoes both brittle fracture and plastic deformation when subjected to typical compression pressures. A variety of brittle diluent materials were examined. These included dicalcium phosphate dihydrate, compressible cane sugar, ground sugar spheres and co-spray-dried sorbitol-mannitol mixture (CSD). All of these materials required a disintegrant to produce satisfactory tablet disintegration. Dicalcium phosphate dihydrate was examined as the primary brittle diluent for the study. It is an alkaline material. Dicalcium phosphate is practically insoluble in water. The major incompatibilities listed for this material occur with tetracycline antibiotics and indomethacin. Also, it should not be used with active ingredients which are sensitive to a pH of 7.3 or more (Wade and Weller, 1994). Use of this material was eventually halted when it was found to interfere with the assay of the model drug. Compressible cane sugar is refined from cane sugar such that it is directly compressible. It is prepared by co-crystallisation and the final product contains 97% sucrose and 3% maltodextrin. This material is incompatible with dilute acids and alkaline earth hydroxides (Wade and Weller, 1994). This eventually became the preferred brittle 27 diluent because of the lower level of disintegrant required (as compared to that of other acceptable, brittle diluent materials). Sugar spheres, NF (uncoated and ground) were examined as possible diluents. As a significant portion of the ethylcellulose coated spheres consist of the substrate materials, it was believed that use of pristine and ground spheres would have good adhesive properties with the coated spheres and therefore, strong tablets would be produced. It was also believed that a smaller particle size may better protect the ethylcellulose membrane from compression-induced damage. However, high levels of disintegrant were required to produce satisfactory disintegration of the tablets. As a result, this material was discarded. Sorbitol-Mannitol CSD is a relatively new material and was examined as a possible brittle, diluent for the disintegrating tablet formulations. It is a co-spray dried (agglomerated) material consisting of approximately 88% sorbitol and 12% mannitol. The product is free flowing, directly compressible and produces hard tablets (EM Industries, 1999). Sorbitol-Mannitol CSD was eventually not selected as a higher disintegrant level was required (compared to that of other materials). 2.5.2. Disintegrants A disintegrant was necessary when a brittle diluent was used. Three materials were examined: croscarmellose sodium, corn starch and crospovidone. Preference was given to the disintegrant that produced adequate disintegration of the tablets at the lowest possible disintegrant level, was compatible with the other materials present and was compressible. 28 Croscarmellose sodium was examined as a possible disintegrant for the disintegrating tablet formulations. It is a cross-linked polymer of carboxymethylcellulose sodium. It is insoluble in water but swells to 4-8 times its original volume when placed in contact with water (Wade and Weller, 1994). Corn starch was examined as a possible disintegrant for the disintegrating tablet formulations. It is insoluble in cold water but swells by about 5-10% in water @ 37°C. It has poor flowability and does not compress easily (Wade and Weller, 1994). Crospovidone is a cross-linked, homopolymer of 7V-vinyl-2-pyrrolidone. It was used as a tablet disintegrant. It is practically insoluble in water and is compatible with most organic and inorganic pharmaceutical ingredients (Wade and Weller, 1994). Crospovidone was eventually selected as the disintegrant for the formulations as it provided adequate disintegration at relatively low levels and did not interfere with the assay of the model drug. 2.5.3. Lubricant Stearic acid was used as a lubricant in all of the disintegrating tablet formulations. It is processed from edible fats and oils and is practically insoluble in water. It is incompatible with most metal hydroxides and may be incompatible with oxidising agents (Wade and Weller, 1994). A lubricant material was necessary to improve ejection of the tablet from the die cavity. When used at low levels, stearic acid does not interfere with tablet disintegration. 29 2.6. Drug Release from Spheres 2.6.1. Dissolution Testing Dissolution is a compendial method for determining the in vitro release of a drug from a dosage form (USP 24, 1999). It is designed to simulate, under laboratory conditions, the in vivo release of drug from the dosage form. The dosage unit is placed in a known amount of dissolution fluid which is agitated and maintained at 37°C. Samples of dissolution fluid are removed at pre-determined time intervals and analysed for drug concentration. In the traditional dissolution test, the dosage unit can be either contained in a wire mesh basket which is rotated to provide the necessary agitation of the fluid (USP Apparatus I) or could move freely in the test fluid while agitation is provided by a rotating paddle (USP Apparatus II). Additional apparati have recently been included in the USP. The speed of the paddle or basket is kept constant throughout the test. The dissolution fluid may be water or another buffered aqueous system. Representative digestive enzymes, co-solvents or surfactants are added in specific cases. The choice of the dissolution fluid will depend on the nature of the experiment and what aspects of the drug release are being investigated. For example, if the general release characteristics of the dosage form in an aqueous medium are of concern, water may be a suitable dissolution fluid. If the release requires the presence of acid and or digestive enzymes, a buffered acid solution or addition of some digestive enzymes may be needed. The accuracy of the simulation is not very high since factors such as food content, variable extent of agitation, changes in pH due to stomach content and absorption are not included. Generally, researchers will attempt to find a correlation between some in vitro release parameter and 30 in vivo release. This correlation becomes useful in inexpensively determining the quality and consistency of any production batch. In research, dissolution provides an inexpensive method to study the general release of drug from a solid dosage form. Other dissolution testing apparati include the reciprocating cylinder (Apparatus 3) and the flow-through cell (Apparatus 4). In the reciprocating cylinder method, the sample under test is placed into a hollow glass vessel which has been equipped with mesh screens on both ends. The cylinder is then oscillated vertically within a flat-bottomed glass tube containing dissolution fluid. Dissolution samples are collected from the glass tube (USP, 2000). This method has the advantage of allowing for changes in the pH of the dissolution medium by physically moving the reciprocating cylinder from a tube containing one pH medium to a different tube containing fluid at a different pH. In the flow-through cell method, the sample is contained in a stationary, vertically oriented, conical-based cylindrical cell equipped with a filter on the (distal) cylindrical end. Dissolution fluid is pumped from a large reservoir, into the conical end of the cell, out the cylindrical end and back into the reservoir (USP, 1999). This apparatus is useful for low solubility drugs where a high volume of dissolution fluid is required. 2.6.2. Models for Drug Release Drug release during dissolution testing of modified-release formulations has been studied by various researchers (O'Connor and Schwartz, 1993, Hutchings and Sakr, 1994 and Neau et al., 1999). Though the mechanism of drug release was not scrutinised in this project, some of the various models of drug release are presented as background information. The application of any of the models below to the project under investigation 31 (could not be accomplished due to lack of additional information and) was beyond the scope of the current study. It is believed that the mechanism of drug release from the membrane coated beads is diffusion of the dissolution fluid through the membrane into the beads, dissolution of the water soluble components within the bead and then diffusion of the Resulting solution through the membrane into the bulk dissolution fluid (Moore, 1989). Another mechanism for drug release was proposed by Bodmeier amd Paeratakul (1994). Based on their studies of mechanical strength of various cellulosic and acrylic polymer films and transport of water soluble drug through cast films, they postulated that osmotic pressure arising from the dissolution fluid entering into the sphere was sufficient to create microchannels in the polymeric film resulting in microruptures through which the drug was released. Release from matrix tablets may occur through (but is not limited to) drug diffusion, polymer relaxation, tablet erosion or a combination of these (Neau at al., 1999). The two more common models of drug release are presented below. 2.6.2.1. Membrane Diffusion The drug release profile can be mathematically modelled to determine the mode of drug release from the solid dosage form. For example, simple diffusion of drug from a coated (water insoluble membrane) sphere follows Fick's 1st law of diffusion as applied to movement across a membrane (Equation 6) (Hogan, 1995). dM kA, x = y \Ccore ~ Cbu,k ) Equation 6 where dM/dt = the amount of drug in the bulk at time t 32 k diffusion rate constant A surface area of coating h thickness of the coating core concentration of drug in the core of the sphere and concentration of drug in the bulk dissolution fluid. Water will first pass through the coating and will dissolve the drug layer. The dissolved drug will then cross the membrane (or coating) into the stagnant drug solution layer because of a concentration gradient and possibly because of osmotic pressure which builds up inside the sphere. The concentration profile of the stagnant layer ranges from the drug saturation concentration at the surface of the sphere (i.e. at the membrane) to the concentration of the bulk fluid. The thickness of the stagnant layer is affected by the amount of agitation the fluid undergoes (i.e. the amount of mixing). Bulk fluid Stagnant drug solution layer Modrfied-release membrane Drug layer Figure 6: Schematic of Sphere in Dissolution Fluid 33 The bulk fluid should act as a sink i.e. the volume of the bulk fluid should be sufficient such that the concentration in the bulk is always significantly less than saturation, typically less than 10% of saturation. A perfect sink would maintain a constant concentration gradient at all times. However, in reality, as the concentration of drug inside the sphere approaches the concentration in the bulk, an equilibrium situation develops and the driving force is reduced to zero. For practical purposes, the sink should be sufficient to provide enough of a concentration gradient so that the steady-state release of the vast majority of the drug can be studied. Steady-state conditions are achieved when the difference between the concentrations in the sphere (i.e. saturated drug solution) and bulk fluid (i.e. sink condition) remains constant. As the concentration gradient is the only variable in Equation 6, the predicted amount of drug leaving the sphere per unit time will be constant (Abdou, 1989). According to the Figure 6, the solubulized drug must pass through a series of layers (i.e. the membrane and the stagnant solution layer). Each of these layers will provide some resistance to the flow of dissolved drug from the inside of the sphere to the sink (i.e. bulk fluid). The flow of drug for the above model can alternately be expressed in terms of resistance to drug movement (Martin et al., 1993). 2.6.2.2. Matrix Diffusion Diffusion of drug from a homogeneous, polymer matrix can be described by the Higuchi equation (Equation 7) (Martin et al., 1993). Equation 7 where 0 = amount of drug depleted from the matrix per unit area 34 / = time D = diffusion coefficient of drug in the matrix A = total amount of drug in unit volume of matrix Cm - the solubility of the drug in the polymeric matrix Typically, dissolution fluid dissolves the drug at the surface of the unit. As the dissolved drug leaves the unit, pores or channels are created through which more water enters the unit to dissolve drug which then can exit the unit through the channels. The process then continues until all available drug is depleted from the unit. Drug contained in the polymer itself will add to the complexity of drug release. The release can be affected by several factors such as tortuosity and channel pore size. 2.6.3. Factors Influencing Drug Release from Spheres Drug release from spheres can be influenced by a variety of factors. Such factors which have been studied include the formulation of the sphere (e.g. method of preparation, ratio of drug to excipient), the presence of a coating (Batra et al., 1994, Ho et al., 1996) and nature of plasticiser used in the film coating (Frohoff-Hiilsmann et al., 1999). Drug release from uncompressed and compressed spheres can be affected by many factors including sphere diameter, presence of a coating and the coating level and inclusion of excipients both within the sphere and/or external to the spheres (Beckert et al., 1996, Bodmeier, 1997, Mathir et al., 1997, Altaf et al., 1999, Miller et al., 1999, Tang et al., 2000). Uncoated, sustained-release spheres can be formulated as matrices which act to retard drug release. Chandy and Sharma compared drug release from microspheres of a 35 drug dispersed in a polysaccharide (chitosan) matrix with the drug release from microgranules prepared by wet granulation of the drug, polysaccharide and other excipients (Chandy and Sharma, 1992). Microspheres were prepared using a solvent evaporation method. The spheres and granules were approximately- the same size. The dissolution profiles of each of the formulations in acidic (pH 2.0) and neutral (pH 7.4) media were assessed. Drug release from the microspheres was much slower and near zero-order as compared to the faster, first-order release of the microgranules. Also, more drug was released in the lower pH medium than in the neutral one. The researchers suggested that the differences in the release mechanism were due to the effects of the excipients on the polysaccharide in the microgranules. In the microspheres, the chitosan behaves like a hydrogel. Initially, there is swelling of the outside of the sphere followed by outward diffusion of the drug then subsequent penetration of the dissolution medium into and diffusion of drug out of the core of the sphere. In the microgranules, one of the excipients interacts with the chitosan allowing it to swell much faster and hence, release drug much faster. The greater release of drug in the acidic medium may be due to the ability of the chitosan to form a salt (chitosan hydrochloride). Kaewvichit and Tucker incorporated a protein compound into a fatty acid matrix. Protein and stearic acid powder were mixed and compressed into cylindrical pellets (Kaewvichit and Tucker, 1994). The particle sizes of the protein and the stearic acid, the percentage of protein in the sphere and the compression force to make the sphere were varied. Results of this study indicated that the protein was released through an interconnecting network of pores formed in the matrix by the dissolution and diffusion of the drug (protein) and also possibly through void spaces within the sphere. Such a 36 mechanism would explain the initial burst of drug release as the surface particles dissolved and the subsequent extended period of slower release. Thus, at low concentrations of drug in the sphere or when using small stearic acid particles, less bulk released would be observed because the pore network created was not sufficiently interconnected to release drug from deep within the pellet. Compression force did not significantly affect release over the range of compression forces used. Sometimes, the release from a matrix formulated drug delivery system can be further retarded by the application of a polymer coating or membrane. Batra et al. studied drug release from uncoated and coated matrix type pellets (Batra et al., 1994). Variables evaluated included the ratio of drug to matrix-forming excipient and the type of hydrophobic polymer coating used. Ferrous sulphate was formulated with gum arabica to produce pellets with sustained-release characteristics by compressing the drug and powdered gum together into a small disc. They found that increases in the gum arabica content in the formulation resulted in longer release periods approaching a first-order model (i.e. release of drug was dependent on the concentration of drug within the dosage form). The researchers suggested that drug release from the matrix was controlled by the amount of water held by the gum arabica which demonstrated gel-forming properties in water. They suggested that the held or bound water increased the density of the gel and affected the diffusional behaviour of the drug. Addition of a polymer coating onto the pellet further decreased the amount released per unit time. In addition to diffusing through the matrix, the drug had to diffuse through a barrier (i.e. the coating). The rate of diffusion through the membrane appeared to depend on the ability of water to diffuse through the polymer into the pellet and the permeability of the drug through the polymer. 37 Frohoff-Hiilsmann, Schmitz and Lippold investigated the drug release mechanism of drug pellets coated with ethylcellulose membranes containing different plasticisers and a water soluble pore former, hydroxypropylmethylcellulose (Frohoff-Hiilsmann et al., 1999). Drug release rate increased with increasing pore former content and was-also affected by the nature of the plasticiser itself. The researchers suggested that, pores were formed in the membrane due leaching of the pore former. The variations in the drug release profiles were attributed to the differences in the properties of the plasticisers. After migration of the pore former from the film, a decrease in film permeabilty was observed for the films containing the less water soluble plasticiser but not for those films containing the water soluble plasticisers. The researchers found that the more water soluble plasticisers leached from the film into the bulk solution. They suggested that the migration of the plasticiser resulted in a more rigid film with a higher glass transition temperature. The less water soluble plasticiser produced films which were able to swell and reduce the size of the pores created by the migration of the pore former from the film. 2.6.3.1. Sphere Diameter The effect of sphere size on the release of drug from uncompressed and compressed spheres has been investigated. Li and others evaluated the effect of sphere size on drug release from sustained-release spheres (Li et. al., 1995). Nonpareil spheres were coated with a slurry of polymer (pharmaceutical glaze or shellac) and drug (diltiazem hydrochloride) and then subsequently coated with a ethylcellulose-dibutyl sebecate dispersion to form a rate-controlling membrane. Mesh sizes of the nonpareil seeds were varied between 25-30 and 40-50. In addition, drug loading was increased for some of the 38 spheres (from 60% drug content to 75% drug content) by applying the polymer-drug slurry in a two step process. Some of the dried, drug loaded spheres were subjected to further drug loading by the same process before application of the rate-controlling membrane. The researchers found that a greater drug load per unit volume of sphere could be achieved by using the 35-40 mesh size sphere rather than the smaller 40-50 mesh size sphere. The larger spheres also had a smaller surface area to volume ratio than the smaller spheres. These two factors (i.e. amount of drug loading and surface area) appeared to affect the rate of drug release since the larger spheres with a higher drug load were observed to release the drug more slowly. Chiao and Price also found that drug release from larger microspheres was slower than from smaller microspheres (Chiao and Price, 1994). In that study, microspheres were prepared by an emulsion-solvent evaporation technique and contained the model drug, propranolol hydrochloride, dispersed within a polymer matrix. Tablets consisted of microspheres, microcrystalline cellulose (diluent), carboxymethyl starch (disintegrant) and magnesium stearate (lubricant). The same trend was observed for tabletted microspheres of various diameters. However, the rate of drug release was much faster for the tablets than that for the uncompressed spheres. The percentage change in the release rate constant increased with increasing sphere diameter over the range of sphere sizes evaluated. 2.6.3.2. Coating As stated earlier, matrix type spheres can be coated with a functional polymer membrane or a drug dispersion and/or rate-controlling membrane can be applied to an 39 inert sphere. A wide variety of pharmaceutical coatings and methods of applying such coatings are available. Rate-controlling membranes were traditionally applied via solutions in organic solvents. More recently, the preferred method of use for such materials is in the form of latex or pseudolatex dispersions. The rate of drug release is controlled by many factors including the membrane permeability, thickness and the concentration gradient (Martin et al., 1993). The actual mechanism(s) of drug release through ethylcellulose based films have been investigated by various researchers (Frohoff-Hiilsmann et al., 1999 and Lippold et al., 1999). Rekhi, Porter and Jambhekar studied the release kinetics and mechanism involved in the transport of an ionizable drug through two kinds of commercially available polymeric films (Rekhi et al., 1995). They examined the effects of"drug' loading and membrane thickness on drug release. The model drug, propranolol hydrochloride, was combined with hydroxypropylmethycellulose (film-former) and polyethylene glycol (plasticizer) in a 60.40 mixture of alcohol and distilled water. This solution was then applied to sugar spheres (18-20 mesh size) at levels between 8 and 16% drug (dry weight per 500 mg of the core). The rate-controlling membranes, Aquacoaf® and Surelease® (both pseudolatex dispersions of ethylcellulose plasticized with dibutyl sebacate) were applied at levels between 5 and 10%. Rate of drug release was reported to be controlled by the rate of diffusion of drug solution through the membrane. Constant release rates were observed for different membrane thicknesses up until 70-80% of drug was released. It was suggested that beyond this point, the concentration of the drug solution inside the sphere fell below saturation and hence, the release rate began to decline (as predicted by Fick's law where the rate of drug release is affected by the concentration difference on 40 either side of the membrane). Osmotic pressure differences between the inside of the sphere and the surrounding fluid appeared to have an effect as well. Drug release rates decreased with increases in the osmotic pressure of the dissolution fluid (i.e. a reduced osmotic pressure gradient between the two sides of the membrane). In fact, no drug release was observed when the osmotic pressure of the dissolution medium was equal to or greater than that of the core leading to the conclusion that the release was diffusional but was also modulated by osmotic pressure. Membrane thickness also appeared to be a factor in drug release. It was observed that the rate of release decreased with increasing membrane thickness. Drug loading onto the spheres did not affect the release rate for spheres coated with Surelease®. However, increased release rates with increased drug loading were observed for spheres coated with Aquacoat®. This increase may have been attributable to increased migration of the drug into the membrane during application of the final coating thus encouraging a more porous membrane. The pH of the Surelease® and Aquacoat® dispersions may have affected the degree of migration. The model drug would be ionized and very soluble in the slightly acidic Aquacoat® dispersion (as it is applied) while the drug would be largely unionized when it came in contact with the Surelease® dispersion which is alkaline in nature. Similar observations about decreasing drug release rates with increasing membrane thickness were seen by other researchers (Li et al., 1995; Batra et al., 1994; and Maganti and Ce l ik , 1994). The effectiveness of the rate-controlling membrane can be influenced by many other factors including the choice of plasticizer (if needed), curing time and the pH of the dissolution medium into which the sphere will be immersed. Hutchings and Sakr investigated the influence of pH and plasticizers on drug release from ethylcellulose 41 pseudolatex (Aquacoat ) coated spheres (Hutchings and Sakr, 1994). A plasticizer is used to lower the glass transition temperature of the ethylcellulose thereby enabling coalescence of the pseudolatex particles as the solvent (water) is driven off. The ability of the latex particles to coalesce is dependent on the type and amount of plasticizer used. Since, the amount of plasticizer will also affect the drug release rate, consideration must be made in selecting the type and amount of plasticizer. Previous studies had indicated that pH of the dissolution medium affected release rates from Aquacoat® coated systems. The researchers evaluated six different plasticizers for Aquacoat® at three different coating levels. Sugar spheres (14-18 mesh size) were coated with a dispersion containing propranolol hydrochloride, the model drug. The Aquacoat® membrane was applied after the drug layer was dry. The drug release profiles in acidic (dilute HC1) and in neutral (phosphate buffer) media were determined. The release profile data were then fitted to two mathematical models so that comparison of several parameters affecting plasticizer efficiency could be made. It was observed that the choice of plasticizer, amount of plasticizer and pH of the dissolution medium affected drug release. For example, increases in the amount of four of the six plasticizers resulted in decreases in the release rate while increases beyond a certain point in the amount of the other two plasticizers led to no changes in the release rate. Also, the rate of drug release for one plasticizer was similar to that of other plasticizers when the spheres were tested in an acidic medium but was significantly different when a neutral medium was used. Release can also be affected by the manner in which the film was prepared. Wesseling and Bodmeier investigated several factors influencing drug release using beads coated with ethylcellulose films prepared from an aqueous dispersion and an organic 42 solution (Wesseling and Bodmeier, 1999). This study used drug layered, nonpareil beads onto which the ethylcellulose coating, plasticised Aquacoat® or an ethanolic solution of ethycellulose (plasticiser and stabilisers), was applied. Drug release in media of different pH and release before and after curing of the spheres was examined. Release from spheres coated using the organic solution was unaffected by curing or by the pH of the dissolution medium while these factors did affect drug release from spheres coated with the aqueous dispersion. The differences in behaviour of the drug release were attributed to the method of film-formation from the two different coating formulations. The distribution of additives (such as sodium lauryl sulphate) in the organic solution is most likely more homogeneous while in the dispersion, such additives may exist at the interface between the polymer and the water. Consequently, the additives in the films produced from the solution may be distributed more evenly throughout the film as compared to additives in films produced from the dispersion. As a result, curing of the ethylcellulose films and pH of the medium may more easily affect the film properties and hence, drug release. L i and others evaluated sustained-release spheres prepared by coating a slurry of polymer (pharmaceutical glaze) and drug (diltiazem hydrochloride) onto nonpareil seeds and then subsequently coating them with a ethylcellulose-dibutyl sebecate dispersion to form a rate-controlling membrane (Li et al., 1995). Drug release in dissolution media of different pH were evaluated. It was found that curing of the membrane for 24 hours was required for adequate coalescence. Additional drying time (20 minutes) in the coating column produced inadequate films. The spheres showed a slight increase in release rate when the pH of the dissolution medium was changed from 1.2 to 7.4. This was explained by the presence of shellac in the drug layer. Shellac solubility is pH dependent, displaying 43 enteric properties thus retarding drug release at low pH. When the pH was raised, the solubility of the shellac increased resulting in a slightly faster drug release rate. The release of drug appeared to be diffusion controlled and followed first-order release kinetics. Maganti and Celik studied the dissolution behaviour of compressed, coated spheres in water (Maganti and Celik, 1994). Spheres consisted of a core of model drug, microcrystalline cellulose and either lactose or dicalcium phosphate surrounded by an ethylcellulose membrane at coating levels of 10, 15 and 20%w/w. Increased coating level decreased the release rate of the drug. The lactose in one of the formulations increased the osmotic pressure and the resulting release of drug was faster than .that for the other formulation. Drug release from uncoated spheres was very fast by comparison to the coated spheres. Compaction of the coated spheres resulted in much faster release rates as compared to the uncompacted spheres regardless of the level of coating and approached the release profile of uncoated, uncompacted spheres. Approximately 95% of the drug was released from the compacts within 30 minutes. This significant change in the drug release profile suggests that damage to the film coating (i.e. formation of cracks) and fragmentation of the spheres had occurred due to the compression at low pressures and thus the spheres lost their sustained-release characteristics. 2.6.3.3. Compression with External Excipients Much attention has been paid to the effects of internal and external excipients to sustained-release spheres and tablets made from such spheres. Maganti and Celik compared the compression parameters of tablets made spheres and a variety of excipients 44 (Maganti and Celik, 1993). The spheres were made by extrusion-spheronization of the powder blend and were uncoated. All three of the formulations used contained drug and microcrystalline cellulose. In addition, one formulation contained lactose and another contained dicalcium phosphate. External excipients used included magnesium stearate, soy polysaccharide, microcrystalline cellulose and pregelatinized starch. The nature and amount of external additive resulted in increases in the pressure required to form a compact of the desired porosity. Compacts made with microcrystalline cellulose as an external additive produced compacts of higher tensile strength. This was attributed to the compressibility of the material used to form most of the sphere. Other excipients produced compacts with lower tensile strengths than compacts made from spheres alone. Beckert and others evaluated disintegrating tablets compressed from enteric coated spheres (Beckert et al., 1996). Soft spheres (produced by powder-layering sugar crystals) and hard spheres (commercially available sugar spheres) were coated with an aqueous dispersion containing the drug, a methacrylic acid copolymer, triethyl citrate (plasticizer) and talc (detackifier). Assay of the spheres revealed that 77 to 100 % of the theoretical amount of drug was loaded onto the spheres. The spheres were then coated with the enteric polymer. Coated spheres were combined with a disintegrant. (crospovidone), a lubricant (magnesium stearate) and a diluent (polyethylene glycol, microcrystalline cellulose, 75:25 mixture of lactose to cellulose or dicalcium phosphate). The blend was tabletted at five different compression force levels (between 5 and 25 kN). One of the formulations and one of the compression pressures were selected for the next part of the experiment in which the spheres were coated with one of nine enteric coating formulations (including several formulations and combinations of Eudragit® and two experimental 45 enteric polymers). The crushing strength of the spheres (i.e. the deformation and cracking behaviour under an applied force), tensile strength of the tablets, disintegration time of the tablets and dissolution rates were investigated. For the hard spheres, the applied force increased steadily until a maximum was reached and the spheres fractured. However, for the soft spheres, this trend was not observed; there was no distinct maximum as the applied force was increased. The researchers suggested that the soft spheres deformed slowly under the applied load. Coating did not significantly affect crushing strength. The enteric coated spheres released drug as required (i.e. less than 10% in 0.1 M HC1 and more than 80% within 45 minutes at pH 6.8). The data suggested that disintegration of the tablets was affected by the order of the mixing sequence and by the compression force applied to make the tablet. Some tablets formed matrices which required much more time to disintegrate (greater than 15 minutes). The tensile strength of the tablets was also influenced by the nature of the materials used in the formulation. For example, tablets containing up to 70% spheres and compressed at 1 OkN or more had a tensile strength of greater than 50 N . Liberation of the drug in 0.1M HC1 after two hours appeared to be independent of compression force for all tablet formulations. However, percentage of drug released in an acidic medium increased as percentage of spheres in the tablet increased. At 10% w/w sphere content in the tablets, marginal differences in the liberation of drug were observed between the formulations. At 90% w/w sphere content, the liberation of drug in the acidic medium was high for all tablet formulations. Between 30 and 70% however, the type of diluent appeared to affect the degree of liberation. Tablets containing the polyethylene glycol had the least amount of drug release while those containing dicalcium phosphate had the greatest. Plastic deforming diluents appeared to 46 better retard liberation of the drug in an acidic medium than brittle diluents. It is possible that the excipient with the greater ability to plastically deform provided better cushioning for or protection from rupture of the film coating than the brittle excipient. Soft spheres released more drug in an acidic medium than the hard spheres. It was suggested that the hard spheres were better able to withstand the compression forces as compared to the soft spheres. Evaluation of the different enteric coatings suggested that the more flexible films could follow the deformation of the spheres during compression. Tablets made with spheres coated with a brittle film liberated more drug in an acidic medium than those prepared from spheres coated with a flexible film. Chiao and Price studied the effects of compression pressure on the physical properties and dissolution characteristics of microspheres (Chiao and Price, 1994). The microspheres consisted of propranolol hydrochloride dispersed within a polymer matrix. Each tablet formulation contained carboxymethyl starch, magnesium stearate and one of three direct compression grade diluents (microcrystalline cellulose, sugar and corn starch). Tablets were prepared by direct compression at five different compression pressures between 17.6 and 175.4 MPa. The tablets were evaluated for density, tensile strength, friability and disintegration and Heckel plots were examined to study the compaction behaviour. Heckel plots of the compaction of the three formulations used by Chiao and Price were non-linear suggesting that the consolidation of the compact may be a combination of brittle fracture and plastic deformation. The rupture or breakage of the microspheres was confirmed by comparison of scanning electron micrographs of the uncompressed and compressed formulations. Tensile strength appeared to vary linearly with the logarithm of compression pressure. The strongest tablets were prepared from the 47 microcrystalline cellulose containing formulation. Friability decreased with increasing compression pressure. The microcrystalline cellulose formulation at any given compression pressure produced the least friable tablet of all the three formulations. Fragmentation of the other two formulations was observed for compression pressures at 35.1 and 70.2 MPa. All of the microcrystalline cellulose formulations disintegrated within 2 minutes. Vilivalam and Adeyeye developed a tabletted microsphere formulation for diclofenac (Vilivalam and Adeyeye, 1994). The drug was dispersed in a wax matrix. The tablet formulation consisted of microspheres, diluent (microcrystalline cellulose) and disintegrant (sodium starch glycolate). Based on the data generated, the researchers concluded that compression pressure (within the range studied) did not affect the dissolution profile and suggested that this observation was due to the cushioning effect of the diluent. Drug release from the compressed tablets was marginally faster than that from uncompressed spheres. They concluded that though some damage to the microspheres may have occurred during tabletting, a sustained-release profile was maintained. Lopez-Rodriguez and others compared elastic recovery, tensile strength and dissolution rate for tablets made of acetylsalicylic acid (ASA) crystals, A S A spheres, A S A spheres coated with an acrylic resin and A S A acrylic resin-coated spheres combined with microcrystalline cellulose (Lopez-Rodriguez et al., 1993). The A S A spheres were made by granulating milled A S A crystals with an enteric polymer solution and pelletizing by a pan coating method. The observed elastic recovery was highest for the microcrystalline-coated sphere formulation (13.6%) and was lowest for the coated sphere formulation (6%). The A S A crystal and ASA uncoated sphere formulations exhibited a similar elastic 48 recovery value (11.3 and 11.9% respectively). The researchers noted that under compression, ASA undergoes elastic deformation. The tensile strength of the microcrystalline cellulose and coated sphere formulation was the greatest of the four formulations and this was most likely due to the nature of microcrystalline cellulose. The coated sphere formulation produced a non-disintegrating matrix tablet. Matrix tablets were also produced when coated spheres were combined with low levels of microcrystalline cellulose (5%). In such cases, the film coatings on the spheres came into contact and fused together during compression. Fillers can be used to prevent this fusion but there is always the danger of damaging the rate-controlling film. These matrix type formulations not only retained a sustained-release profile but the rate of release decreased from that of uncompressed coated spheres. Yao and others examined the effects of excipient particle size on compression-induced damage to ethylcellulose coated theophylline powder (Yao et al., 1997 and 1998). Theophylline powder, 53 - 105 um, was coated with an ethylcellulose or an ethylcellulose/hydroxypropylcellulose film from an organic solvent solution. The coated powder was tabletted with microcrystalline cellulose (Avicel®) and chitosan of different sizes and the tablets were subjected to dissolution testing. Avicel® was found to be ineffective against compression-induced damage to the membrane regardless of the size as demonstrated by the faster release of drug from coated powder and Avicel® formulations. Tablets containing chitosan (6 um) demonstrated similar release to the uncompressed, coated powder (control) (Yao et al., 1997). Subsequently, the researchers examined other additives (specifically Avicel®, cornstarch, lactose, mannitol, croscarmellose and low-substituted hydroxypropylcellulose) of various sizes. Comparison of drug release from all 49 the different formulations suggested a maximum particle size of 20 um regardless of additive (Yao et al., 1998). 2.7. Data Evaluation 2.7.1. Data Modelling Often, it is desirable to fit dissolution data to a mathematical model. This may be done to normalize certain parameters so that comparisons of different formulations can be made. Hutchings and Sakr fit dissolution data from 20 formulations to natural logarithmic functions relating percent of drug released to time (Hutchings and Sakr, 1994). The value of these models was that the parameters determined (i.e. release rate and time of fastest release) were meaningful and could be used as a basis of comparison between the different formulations. Such models can be employed when the actual release mechanism is not known. Sometimes, dissolution data can be fit to a mathematical model which describes a type of drug release. O'Connor and Schwartz studied the pattern of release and factors affecting the release from inert matrix spheres prepared by extrusion-spheronization (O'Connor and Schwartz, 1993). Various sphere formulations were prepared by extrusion spheronization using several model drugs (at various levels), microcrystalline cellulose and lactose. It was postulated that the microcrystalline cellulose in the spheres formed a continuous matrix within which the drug particles were dispersed. The set of experiments was designed to evaluate the effects of drug loading in the sphere, porosity and aqueous solubility of the model compounds on the rate of release of drug from the sphere. These parameters are part of the Higuchi square root of time equation .which describes the 50 amount of drug released from an inert matrix as dependent on the square root of time. This equation is applied to both spherical and planar systems. Mathematical modelling of the release rates of the model drugs suggested that drug was released from an inert matrix. In a later study, Zimm, O'Connor and Schwartz modelled the same release data from spheres (made by extrusion-spheronization) using the Higuchi square root of time equation and the Higuchi cubic equation and studied the differences in the behaviour of the equations with sphere size (Zimm et al., 1996). (The Higuchi cubic equation is applied specifically to spherical inert matrix systems.) The initial findings of the research suggested that both equations can be used to describe the release of drug from a spherical system but the Higuchi cubic equation was more sensitive to changes in the size of the sphere. 2.7.2. Similarity Factor Comparison of dissolution profiles was accomplished via the use of a fit factor, the similarity factor, f2. The similarity factor was first introduced by Moore and Flanner (Moore and Flanner, 1996) as a method to compare the degree of similarity between two dissolution profiles. The equation appears as Equation 8. This equation is the logarithmic transformation of the average sum of squares of the difference between the two profiles. A weighting factor is also included for cases where a particular part of the curve may of greater interest. For example, if a certain amount of drug must be released by 45 minutes, the weighting factor may be higher for the averages at t = 45 min. In this study, the weighting factor was set to unity since no such acceptance criteria had been accepted. 51 h = 5 0 x lo 100 g « Equation 8 1 + 1 n where f2 = the similarity factor n = number of sampling points w, = weighting factor (set to unity) Rt = average % released for the reference batch at time / T, = average % released for the test batch at time t This equation allows one to comment on the extent of similarity of two curves that may graphically appear to be the same. A 10% difference between two profiles results in a similarity factor of 50, a 5% difference yields 65 and a 2% difference yields 83 (Moore and Flanner, 1996). Generally, an acceptance limit of 50 or greater (corresponding to a 10% or less difference) is sufficient to declare two dissolution profiles similar (US Dept. of Health and Human Services, 1997). There are some limitations associated with the use of this fit factor (US Dept. of Health and Human Services, 1997). Firstly, the sampling timepoints for both the reference and test profiles must be the same. Secondly, only one measurement beyond 85% dissolution may be included in the calculation as after this point, the two curves will naturally converge. Finally, the percent coefficients of variation for averages at early timepoints cannot exceed 20% and those for all averages at subsequent timepoints cannot exceed 10%. This last limitation arises from the use of averages without accounting for variability in the data. Also, data must be generated under the same test conditions. 52 The similarity factor has been adopted by the United States Food and Drug Administration (US FDA) as an industry guideline for determining sufficient similarity between a reference and test batch for scale-up testing and post-approval changes for immediate release products (SUPAC IR) (US Dept. of Health and Human Services, 1997). It is appropriate for use in comparison of model-independent drug release profiles. Use of the similarity factor has been explored by various researchers. Liu and others raised some statistical concerns (Liu et al., 1997). The similarity factor is a random, sample statistic for which a statistical hypothesis (with regards to population parameters) cannot be formulated. As a consequence of this, evaluations of false positive and negative rates of decisions cannot be determined. Further, the distribution of f2 is complex to the extent that expected variance and therefore, confidence interval for the mean cannot be determined. Also, the similarity factor is insensitive to curve shape and unequal spacing between sampling timepoints. The similarity factor has been subjected to simulation testing and comparisons to other statistical dissolution evaluation methods (Ju and Liaw, 1997). Ju and Liaw examined the properties of the similarity factor and other methods to profiles generated using linear and quadratic models. They concluded that the similarity factor has a lower probability of declaring dissimilarity that other comparison methods based on their simulation studies. They also raised concerns regarding the lack of Type I and II errors (i.e. false positive and negatives), the inability of the method to compare profiles where different sampling timepoints were used and the inclusion of data after the asymptote was reached (as such data will increase the similarity factor). 53 Shah and others also conducted testing of the similarity factor using dissolution results of one reference and four test batches to address some of these statistical concerns (Shah et al., 1998). Among their conclusions, they found that the similarity factor was sensitive to the number of sampling timepoints included after the dissolution plateau was achieved and that the estimated similarity factor (f2 , the calculated value from Equation 1 and based on the dissolution results) was a conservative estimate of f2. They suggest the use of a "bootstrap" approach to simulate the confidence interval for the estimated similarity factor. Despite its statistical limitations, the similarity factor has been used by researchers as a means of evaluating different formulations against a control (Edddington et al., 1998: and Huang et al., 1998). As long as the conditions specified in the SUPAC-IR are followed, the similarity factor is used to determine similarity between the dissolution profiles of a reference batch and a test batch. 54 3. Objectives and Specific Aims The objectives of the study are listed below. The primary objectives of the study were: • To design a drug-loaded, coated sphere where the rate of drug release is constant regardless of whether the spheres are encapsulated or compressed into disintegrating or non-disintegrating tablets. • To investigate the influence of sphere diameter and the mechanical properties of spheres and excipients on drug release rates from the compressed, coated spheres. Specifically, the study was designed: • To evaluate the influence of sphere diameter on drug release from encapsulated and compressed spheres. • To evaluate excipients which could be used to form either disintegrating or non-disintegrating tablets and to determine the rate of drug release from such tablets. • To evaluate the influence of coating level and mechanical properties of excipients on the rate of drug release from disintegrating and non-disintegrating tablets. 55 4. Experimental 4.1. Materials All materials were used as received. Materials were of compendial grade unless otherwise stated. 4.1.1. Model Drug Chlorpheniramine maleate USP, Novopharm Ltd. (Toronto, Ont.) 4.1.2. Substrate Sugar spheres, NF: Nu-Pareil® PG, Crompton and Knowles (Mahwah, NJ); 14-18, 20-25 and 30-35 mesh sizes Sugar spheres, NF: Nu-Core®, Crompton and Knowles (Mahwah, NJ); 45-60 mesh size 4.1.3. Coating Materials Povidone USP (PVP): Kollidon® 30, BASF (Parsippany, NJ) Alcohol (House Std): Alcorub , Stanley Pharmaceuticals (North Vancouver, BC) Hydroxypropylmethylcellulose (HPMC), USP: Methocel , E-5 Premium, Dow Chemical Company (Midland, Michigan) Talc, USP: Talc, Cyprus Industrial Minerals Co. (Alpine, AL) Ethylcellulose, NF Surelease aqueous dispersion, Colorcon Ltd. (West Point, PA) 56 4.1.4. Tablet and Capsule Materials Microcrystalline cellulose (MCC), NF: Avicel®, PH 102, F M C Corporation (Philadelphia, PA) Compressible sugar, NF: Di-Pac®, Domino Specialty Ingredients (Baltimore, MD) Sorbitol-mannitol CSD: E M Industries Inc. (Toronto, Ontario) Pregelatinised starch, NF: Starch 1500®, Colorcon Ltd. (West Point, PA) Dibasic calcium phosphate, USP: Calstar® Type CS-400, FMC Corporation (Philadelphia, PA) Crospovidone, NF: Polyplasdone® X L , GAF Corporation (Wayne, NJ) Croscarmellose sodium, NF: Ac-Di-Sol®, F M C Corporation (Philadelphia, PA) Starch, NF: Pure-Dent B700, Grain Processing Corporation (Muscatine, IA) Stearic acid, NF: Triple Pressed, Witco Corporation (Willowdale, Ontario) Capsules: Hard gelatin capsule shells, Capsugel (Greenwood, SC); size 0 4.2. Equipment 4.2.1. Coating Equipment Fluid bed column: Aeromatic A G Strea-1 coating column, Niro Inc. (Columbia, MD) Column insert: Leading Edge Counters (Richmond, British Columbia) Pump: H.R. Flow Inducer, Watson-Marlow Ltd. (Cornwall, England) 57 Mixer: Dyna-Mix, Fisher Scientific Canada (Toronto, Ontario) Compressor: 5.0/25 Gallon, Single Cylinder air compressor, Sears (Toronto Ontario) A schematic of the column appears in Figure 4. The column was equipped with a grounded, stainless steel insert which dissipated any static charge built up from the motion of the small spheres. This insert had narrow, vertical viewing slits to observe the motion of the spheres during coating. Any openings between the insert and the walls of the column are sealed with an inert tape to prevent any trapping of material. Top screen Wire mesh Column wall Metal insert Drying air Atomised coating solution Wire mesh Perforated base plate Figure 7: Schematic of Coating Column 4.2.2. Balances Low masses: Mettler A E 163 (Hightstown, NJ) Top-loading: Sartorius models: BP6100, B310P and H120 (Mississauga, Ontario) 58 4.2.3. Dissolution Equipment Dissolution: Vanderkamp 600 six-spindle dissolution tester Autosampler: Vanderkamp model 10 fraction collector Vanderkamp 10400 motorised drive bar Vanderkamp V K IPS 12 pump Vanderkamp EDS-10 programmable sequencer All dissolution equipment manufactured by Van-Kel Industries Inc. (Edison, NJ) 4.2.4. UV Spectrophotometer Systems HP System: Hewlett Packard 8452A diode array spectrophotometer (equipped with a deuterium lamp) (Mississauga, Ontario) HP89532A UV-vis software (rev. A.00.00) Ocean Optics system: Ocean Optics Inc. S2000 fibre optic spectrometer (Dunedin, FL) Ocean Optics PX-1 xenon flash lamp, (Dunedin, FL) Ocean Optics C U V sample holder, (Dunedin, FL) Ocean Optics Inc. Basic Acquisition Software (1.5.07) 4.2.5. Compression Equipment Tablet press: Manesty Betapress, Manesty Machines Ltd. (Liverpool, England) Punch tooling: Vi \ round, flat-faced, IPT tooling, Thomas Engineering (Hoffman Estates, IL) Hardness tester: CT40, Engineering Systems (Nottingham, UK) 59 Micrometer: model CD-6"B, Mitutoyo Corp. (Japan) The Manesty Betapress was fitted with strain gauges on the upper compression roll pin to measure force during compression. Details of the press instrumentation have been reported by Mitchell and others (Oates and Mitchell, 1989, Oates and Mitchell, 1990 and Dwivedi et al., 1991). Press deformations as a function of force have been determined and used in the calculation of porosity and work where exact positions of the punch faces must be known. The output of the strain gauges is transmitted to a computer where the signal is converted to digital readings as a function of time. Parameters calculated from these data include "in-die" porosity, work of compression and peak offset time. Tablet dimensions are determined manually following ejection of the tablet from the die. Die fill is manual and compression pressure is controlled by the mass of material present in the die cavity. Control of actual mass of material in the die cavity is not exact. For example, a dense but finely ground material may be difficult to quantitatively transfer to the die and hence the mass actually compressed would be less than the mass measured out. Tablets are stored in polystyrene vials for at least 24 hours after compression. Dimensions are measured again and then the tablets are crushed in the hardness tester. Tablets are placed between two platens and a load is applied by mechanically pulling down the upper platen. The CT-40 hardness tester has the advantage of constant rate of movement of the upper platen and hence, uniform application of the load. The upper platen is fitted with a linear variable-differential transformer (LVDT) to measure the movement of the platen and deformation of the hardness tester as a function of the applied force had been determined. During the test, both force and positions of the platens are transmitted to the computer. Parameters calculated using these data include tensile 60 strength (calculated from the force at which tablet fracture occurs), work (calculated from the force against time profile) and diametric strain (calculated from the change in tablet diameter at the point of failure). 4.2.6. Miscellaneous Equipment Gas pycnometer: Quantachrome MVP-1 Multipycnometer (Boyton Beach, FL) Blender: Erweka AR400 cube blender (Milford, CT) Homogeniser: Polytron PT10-35 homogeniser (Kinematica A G , Switzerland) Disintegration apparatus: Erweka GmbH ZT2 disintegration apparatus (Frankfurt, Germany) 4.3. Methods 4.3.1. Drug Layering: Method Development and Final Method Initially, drug layering was attempted via continuous application of an aqueous solution of 10% chlorpheniramine maleate, 3% talc and 10% w/v H P M C using the parameters listed in Table 1 (initial setting). A batch size of-424 g of uncoated spheres Table 1: Coating Parameters (Initial and After Method Development) Parameter Initial Setting Final Setting Air volume > 130 m3/hr - 100 m3/hr Inlet air temperature ~70°C ~60°C Outlet air temp, (dial) ~40°C ~40°C Outlet air temp, (thermometer) ~ 60°C ~50°C Temperature setting 70°C 55°C Pump setting 0.3 0.25 Application rate ~ 4-5 g/min -3-4 g/min Atomising pressure 1.0 bar 0.4 bar was used. At the end of the coating, spheres were allowed to dry in an oven set at approximately 40°C for at least 12 hours. The spheres were then sieved through 61 appropriate size meshes (determined based on the size of uncoated spheres used). A viable sphere was defined as an individual sphere containing model drug. Yield was calculated as the weight percentage of spheres in each mesh fraction compared to the total amount of spheres recovered and sieved. The initial coating solution and method were found to be unsuitable. Method development for the production of drug-layered spheres is documented in Sections 5.1 and 6.1. Uncoated spheres of 30-35 mesh size were used in this activity. Several modifications to the coating solution and application method were attempted and evaluated based on the yield of viable spheres (mesh fraction 25-35). The final method used to prepare the drug layered spheres utilised a 20% excess of a 6% and 0.5% w/v hydro-alcoholic solution of model drug and povidone respectively. The solution was applied at 5 minute intervals with 5 minutes of in-column drying in between coating intervals. Column parameters are detailed in Table 1 (final setting). Spheres were allowed to dry overnight (or longer) at ~ 40°C before application of the control release membrane. A typical product formula appears in Table 2. Table 2: Typical Formulation for Preparation of CPM Layered Spheres and 10% Surelease Coated Spheres Material Applied Quantity Target Quantity Uncoated spheres 424 g 424 g Chlorpheniramine maleate 30 g 24 g Povidone 2.5 g 2 g Alcohol 375 mL 300 mL Water qs 500 mL -100 mL C P M layered spheres 300 g 300 g Surelease (15% solids) 250 mL 30 g 62 Drug layered spheres prepared for the size study were released for further use once the assay for drug content demonstrated that the appropriate amount of model drug had been loaded onto the spheres. In some cases, additional drug layering was necessary. Two batches of drug layered spheres were prepared for each size ancTcGinbined into one batch using a cube blender. The combined batches were released for ethylcellulose coating after drug content was determined. Drug layered spheres prepared for the coating level study were not assayed for drug content until after application of the ethylcellulose coating as final release of these spheres was based mainly on the dissolution release profiles. 4.3.2. Application of Ethylcellulose Membrane Ethylcellulose (Surelease®) coating was applied to the drug layered spheres using the column parameters detailed in Table 1 (final setting). A 20% excess of coating dispersion was applied to account for expected losses during the process. The dispersion was diluted to a 15% solids content and was gently agitated with a propeller mixer during coating to ensure even distribution of the solids. Coating was applied continuously until the reservoir was empty. The spheres were then stored overnight at ~ 40°C to remove any remaining moisture. Ethylcellulose coated spheres were released for, use following assay for drug content. For spheres used in the coating study, dissolution testing was also required. The coating level was determined by the % weight of Surelease® applied to the drug layered spheres. For example, a batch of 10% Surelease® coated spheres would result from coating 30 g of Surelease® solids (36.0 g including excess) onto the surface of 300 g of 63 drug layered spheres. The actual thickness of the Surelease® coating was not determined directly. 4.3.3. Preparation of Various Dosage Units 4.3.3.1. Capsules The required amount of spheres was measured out and transferred quantitatively to the empty capsule shell. Each capsule had a drug content of 24 mg of model drug. The target mass of spheres was determined based on the calculated drug loading. 4.3.3.2. Non-disintegrating tablets Approximately 570 mg of spheres were measured out and transferred to the tablet die cavity. Compression pressure is proportional to the mass of material in the die and this amount of material was sufficient to produce a compression pressure of about 100 MPa for each tablet. The compression profile was acquired and stored for later reference. 4.3.3.3. Disintegrating tablets The individual tablet components were measured out and combined in the same vessel where they were manually mixed. The mixture was transferred to the die cavity and compressed. Masses were selected such that a compression pressure of approximately 100 MPa was used to form the tablet. The compression profile was acquired and stored for later reference if needed. For dissolution samples used to develop the formulations, a 20 g blend of the necessary components was prepared. Mixing of the blend was accomplished by gentle tumbling of the container. As separation of the blend occurred quickly due to the differences in particle sizes of the individual components, the blend was 64 re-mixed immediately prior to material removal. Blend material was measured out and transferred to the die cavity where it was compressed at approximately 100 MPa. 4.3.4. Dissolution Testing Dissolution parameters used for dosage units containing Surelease® coated spheres are listed in Table 3. USP Apparatus II was used. Six samples were tested during each dissolution run. The sampling probes were fitted with in-line, 35 um filters. Solvent replacement after sampling was not done as the automated system was not equipped to do this. Losses due to sample removal were accounted for in the calculation of drug released for the dissolution profiles. Since the presence of Starch 1500® in the some samples complicated the UV assay of drug content, minor modifications to the method were made when this material was used. To minimise the effects of the interference, the initial medium volume was reduced to 500 mL thus producing a more concentrated sample. Any additional absorption due to the water soluble components of Starch 1500® in the sample was negligible. Table 3: Dissolution Parameters Paddle speed 75 rpm Temperature 37°C Medium glass distilled water Initial volume 900 mL * Sampling time 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 16 & 24 hours Sample volume approximately 10 mL * * Note: modified for Starch 1500® samples 65 4.3.5. Assay of Model Drug Content 4.3.5.1. UV Assay Method The amount of model drug in the dissolution samples was quantified by U V spectrophotometry. The maximum absorbance at a wavelength of-261 nm was recorded for all samples. As water was chosen as the dissolution fluid, glass distilled water served as the blank. Drug content in dissolution samples was determined directly from the sample solutions removed during dissolution testing. Drug content of the spheres was determined by dissolving a known mass of ground (Surelease® coated) spheres or (drug layered) spheres in water. A portion of the sample was filtered through a 0.45 um filter and then analysed by U V spectrophotometery for C P M content. Drug content of spheres was reported as mg C P M per 450 mg of spheres. Drug content for the drug layered spheres used in the size study was based on the results of 10 independent determinations of 100 mg of spheres diluted to 100 mL with distilled water. Drug content for Surelease® coated spheres was based on 4 independent determinations (500 mg of ground spheres diluted to 500 mL). In addition, a solution containing a known amount of chlorpheniramine maleate (standard solution) was run with each set of dissolution samples as a check of the U V system (system suitability). If the absorbance agreed with the expected value (calculated from the calibration curve) and the U V absorbance spectrum agreed with that of the reference spectrum for C P M , the system was accepted for use. The standard solution was assigned a shelf life of 2 months at ambient light and temperature. 66 The primary equipment was the HP system. Data were collected between 190 and 82Q nm. The secondary equipment was the Ocean Optics system which employs a xenon pulse flash lamp light source. The acquisition range on this system was much less than that of the other system, between 220 and 400 nm. These cut-offs were a limitation of the light source. This range limitation was acceptable since the quantifying peak of the model drug absorbed within this range. 4.3.5.2. Validation of Assay Method Linearity of the Beer's law plot was established on both U V spectrophotometers by analysing several solutions containing various, known amounts of model drug. Linearity was re-determined on several occasions after the initial determination as part of investigations into possible sources of errors. Interference solutions were prepared for the soluble components of the EC coated spheres (i.e. uncoated spheres, HPMC, PVP) and soluble excipient (i.e. Di-Pac) at the concentrations expected in the dissolution testing. Also, when necessary, verification of correct operation of the Ocean Optics U V spectrophotometer was confirmed against that of the HP U V spectrophotometer via comparison of the data generated by both systems for the same solutions. 4.3.6. Investigation of Dissolution Related Errors 4.3.6.1. Solvent Evaporation Dissolution testing of early samples produced unusually high results; greater than 100 % drug released values were observed. A number of possible causes were examined including sampling method (i.e. manual or automatic), effects of stirring and heating of solutions and operation of supporting equipment such as the balance. To investigate 67 sampling method, known amounts of C P M were added to dissolution vessels containing known amounts of distilled water, samples were obtained manually and via the autosampler and the solutions analysed for the model drug. The effects of heating and stirring were examined by adding a known amount of model drug to volume of distilled water at room temperature, removal and analysis of a sample of the solution, transfer of the solution to a dissolution vessel where it was heated to 37°C and agitated by a paddle at 75 rpm and finally removal (after 3 hours) and analysis of another sample of the solution. Verification of operation of supporting equipment (such as balances) was accomplished through re-calibrations and checks using external standards (e.g. calibration weights). The rate of evaporative loss was determined by loss of mass for both covered and uncovered dissolution vessels. The mass of dissolution fluid was determined at various timepoints in a 24-hour period. The values were then converted to volumes using the density of water at the fluid temperature (measured at each timepoint). Solvent losses due to evaporation were accounted for in the calculation of drug released for the dissolution profiles. 4.3.6.2. Apparent Drug Gains and Losses Due to Insoluble Materials Apparent losses of drug from disintegrating tablet formulations containing croscarmellose sodium were conducted by adding known amounts of model drug and various external excipients and combinations of excipients to known volumes of distilled water. The water was maintained at 37°C and agitated at 75 rpm. The Avicel® formulation was used for this investigation. Samples of fluid were removed from each vessel at 2 and 3 hours and analysed for model drug. 68 The three initial formulations containing crospovidone were tested for apparent drug losses and gains by subjecting two sample tablets of each formulation (each containing a known mass of EC coated spheres) to dissolution testing. Samples were removed at 2 and 3 hours and assayed for model drug. In addition, the insoluble remnants of the dicalcium phosphate tablets were filtered from the dissolution fluid and treated with methanol to extract any drug adsorbed to the surface of the water insoluble particles. The extracts were filtered and analysed for model drug. Standard and interference solutions in methanol were also prepared and analysed to more accurately interpret the methanol extract data. Interference due to water soluble components of Starch 1500® was examined by preparing solutions of Starch 1500® in distilled water at concentrations between 0.27 and 0.33 mg/mL (corresponding respectively to masses of 243 and 297 mg of Starch 1500® in 900 mL of dissolution fluid) and determining the U V spectrum and absorbance at 262 nm of each solution. Solutions were filtered through a 0.45 um filter prior to analysis. 4.3.7. Formulation Development 4.3.7.1. Initial Formulations (25% Excipient Level) Spheres were compressed with various excipients in the development of a disintegrating tablet. The initial formulations contained 75% spheres (10% EC coated, C P M layered, 20-25 mesh size spheres), 24% diluent and 1% stearic acid (as a lubricant). Four possible diluent materials were examined: microcrystalline cellulose (Avicel®), dicalcium phosphate, pregelatinised starch (Starch 1500®) and ground, uncoated sugar 69 spheres. Tablets were prepared from a formulation blend and subjected to dissolution testing. 4.3.7.2. Formulations Containing 50% Excipients Incorporation of a disintegrant prompted examination of a suitable disintegrant level. The disintegrant and level initially selected were croscarmellose sodium and 3% respectively. The formulation composition was changed to 50% (10% EC coated, C P M layered, 30-35 mesh size) spheres, 46% diluent, 3% disintegrant and 1% stearic acid powder. The change in mesh size was made because of the larger amount of 30-35 mesh size spheres available. A comparison of the release profiles of tablets consisting of 75% (10% Surelease® coated, 20-25 or 30-35 mesh size) spheres, 24% Avicel® and 1% stearic acid demonstrated like drug release (see Figure A2-1 in Appendix 2). Tablets were prepared from a formulation blend. Suitability of the disintegrant level was determined for the dicalcium phosphate and ground spheres formulations. Tablets of these two formulations were subjected to dissolution testing conditions and observed for adequate break up of the tablet. A disintegrant level of 3% was examined for the dicalcium phosphate formulation while levels of 3 and 5% were examined for the ground spheres formulation. Tablets of acceptable formulations were subjected to dissolution testing. 4.3.7.3. Evaluation of an Alternate Disintegrant As a disintegrant was necessary for the dicalcium phosphate formulation, a test formulation containing that diluent was prepared to test the suitability of an alternate disintegrant, crospovidone. Possible interference of the new disintegrant with the model drug was examined using drug recovery. Known amounts of model drug and 70 crospovidone were added to a known volume of dissolution fluid (@ 37°C and 75 rpm). Samples of fluid were removed from each vessel at 2 and 3 hours and analysed for model drug. Adequate tablet break-up during dissolution was determined visually during dissolution testing. The test formulation consisted of 50% (10% Surelease® coated, 30-35 mesh size) spheres, 46% dicalcium phosphate, 3% crospovidone and 1% stearic acid. Tablets were prepared from a formulation blend and subjected to dissolution testing. Additional tablet samples were prepared individually and also subjected to dissolution testing. 4.3.7.4. Evaluation of Alternate Brittle Diluents Each of the alternate brittle diluents (uncoated spheres, sorbitol-mannitol CSD and compressible cane sugar, Di-Pac) was examined for possible interference with the model drug. Known quantities of diluent were dissolved in distilled water and the U V spectra determined. The lack of interference for uncoated spheres was determined previously in the validation studies of the U V assay of drug. Ground, uncoated spheres were re-examined as a possible diluent due to the use of a different disintegrant. Disintegration testing was conducted as per the USP using water at 37°C. For ground, uncoated spheres, six samples of tablets at each disintegrant level were tested. Tablets were prepared from formulation blends. For the remaining diluent materials, one sample at each disintegrant level was prepared individually and subjected to disintegration testing. 71 4.3.7.5. Comparison of Variability in Dissolution Profiles Between Tablet Preparations (Individual vs. Blend Sampling) This experiment examined the variability produced when tablets prepared individually and via a formulation blend were subjected to dissolution testing. This variability was observed previously for tablets containing dicalcium phosphate. Three tablets were prepared individually and three via a formulation blend and subjected to dissolution testing. The data were averaged for each preparation method. The formulation consisted of 50% (10% EC coated, C P M layered, 30-35 mesh size) spheres, 46% sorbitol-mannitol CSD, 3% crospovidone and 1% stearic acid powder. 4.3.7.6. Final Formulations The final composition of the three formulations are listed in Table 4. In all formulations, stearic acid was included as a lubricant at 1%. Crospovidone was higher for the Di-Pac® formulation at 7%. Half of the tablet mass consisted of Surelease® coated, 30-35 mesh size spheres. The compositions of these formulations were kept consistent for all Surelease® coating levels so that appropriate comparisons could be made. Diluent made up the remainder of the tablet mass. All tablets used in the remainder of the study and prepared for dissolution testing were prepared individually. Table 4: Final formulations for the disintegrating tablet Formulation 1 Formulation 2 Formulation 3 Surelease® coated, 30-35 spheres (50%) Avicel® (46%) Di-Pac® (42%) Starch 1500® (46%) Crospovidone (3%) Crospovidone (7%) Crospovidone (3%) Stearic acid (1%) 72 4.3.8. Compression Studies 4.3.8.1. True Density Determination The true density of the compressed materials was needed for calculation of various parameters in the compression studies. A Quantachrome Multipycnometer (helium gas pycnometer) equipped with the large sample cell was used. About 10 g of sample were ground to a fine powder using a mortar and pestle. The documented experimental value was the average of three individual determinations (of the same sample). True density was determined for all materials used in the compression studies. 4.3.8.2. Tablet Compression Compressed sphere samples were prepared from spheres only. Where necessary, lubricant was added to the die cavity using a 5% solution of stearic acid in chloroform. For formulations, samples were taken from tablet blends as the exact sphere content of the tablets (i.e. drug content) was not critical. Samples were measured individually and manually filled into the die cavity of the Manesty Betapress. Round, flat-faced, half-inch tooling was used. The compression event was recorded and stored electronically for subsequent analysis. The mass and physical dimensions (diameter and thickness) of the ejected tablet as well as the peak compression pressure were recorded. Tablets were kept individually.in plastic vials for at least 24 hours before being tested for hardness. 4.3.8.3. Hardness Testing Tablets were broken radially using the CT40 hardness tester. The data for each tablet was combined with the corresponding data from the compression event and used to calculate the various compression parameters. 73 4.3.8.4. Compression Data Evaluation The main objective of the compression work was to investigate any possible relationships for porosity, work of compression, peak offset time and tensile strength with respect to compression pressure. Compression pressure was selected as the independent variable as it could be accurately measured though it could not be controlled precisely. This part of the study was initially designed to seek out general comparisons over a range of compression pressures for the different materials to examine any possible prediction of dissolution behaviour based on compression properties. It was thought that a general approach might be most advantageous as this would identify the more relevant parameters and provide focus for any subsequent compression work in which meaningful comparisons between materials could be made. To cover a range of compression forces, the mass of the tablets spanned 450 to 685 mg. As possible trends in compression parameters against pressure were being investigated, one sample was taken at each compression pressure. This approach efficiently permitted an overall view of the compression behaviour over the range of pressures. As this investigation was performed as a probe in the overall study, statistical analysis was not required and hence was not completed. However, some statistical analysis can be performed if needed (Er, 2000). Statistical comparisons between materials based on variability at any particular compression pressure (e.g. A N O V A ) could not be made as replicate determinations were not done. However, it is possible to apply linear and non-linear regression to the data to generate 95% confidence intervals. Subsequent comparison of overlap of the confidence intervals can be used to assess the general differences in the regression lines (Er, 2000). 74 5. Results Data pertaining to method and formulation development in support of the main project will be presented first. This includes information regarding coating of the spheres, assay for model drug content, investigations concerning dissolution irregularities and disintegrating tablet formulation development. The results of the primary study are then presented mostly in the form of dissolution profiles and comparisons of similarity factors. The similarity factors have been tabulated and the time point of the last data points used in the similarity factor calculation is included. 5.1. Coating Method Development 5.1.1. Drug Layer Coating Solution Formulation Development Use of the initial coating solution formulation led to large masses of fused spheres. Alterations to the coating solution and application were attempted in an effort to produce viable spheres. The changes to the coating solution and application are listed in Table 5. Table 5: Drug Layer Solution Formulation Development Coating Solution # Components (% of total) Details of Coating Procedure Observations DLS-1 - C P M (10%) • HPMC (10%) • Talc (3.5%) • Distilled water (76.5%) • Initial coating parameters used. • Continuous coating application. " Spheres agglomerated into large masses. • Yield of viable spheres could not be calculated. DLS-1 • C P M (10%) • HPMC (10%) • Talc (3.5%) • Distilled water (76.5%) • Initial coating parameters used. • Alternate coating application and drying intervals of 10 min. • Sphere agglomeration observed (smaller masses). • Yield of viable spheres was 1.6%) of total material recovered. 75 Table 5 continued DLS-2 • HPMC (10%) • Talc (3.5%) • Distilled water (86.5%) • Initial coating parameters used. • Alternate coating application and drying intervals of 10 min. • Some sphere agglomeration observed (-10% of batch). • Yield of viable spheres was 90% of total material recovered. DLS-3 • HPMC (10%) • Talc (3.5%) • NaCl (4.5%) • Distilled water (82%) • Initial coating parameters used. • Alternate coating application and drying intervals of 10 min. • Some sphere agglomeration observed (-15% of batch). • Yield of viable spheres was 85% of total material recovered. DLS-4 • HPMC (10%) • Talc (3.5%) • 45:55/EtOH: H 2 0 (86.5%) • Initial coating parameters used. • Alternate coating application and drying intervals of 10 min. • Much sphere agglomeration observed (81%) mostly as groups of three spheres. • Yield of viable spheres was 19% of total material recovered. DLS-5 • C P M (6%) • PVP (0.5%) • 75:25/EtOH: H 2 0 (93.5%) • Initial coating parameters used. • Alternate coating application and drying intervals of 10 min. • Little sphere agglomeration observed (-0.9% of batch). • Yield of viable spheres was 95% of total material recovered, • Drug content of spheres was -30 % lower than expected . (theoretically). Assay of the drug-layered spheres demonstrated a typical sub-potency of 20% from the target application of 24 mg CPM/450 mg (drug-layered) spheres. Applications of excesses of 20 and 40% coating solution yielded similar coating efficiencies. Also, some variability of drug loading on the spheres was observed although all'parameters were kept constant. 5.1.2. Ethylcellulose Application Development A batch of 15% EC coated, C P M layered, 30-35 mesh size spheres was prepared and subjected to dissolution testing. Comparison of the profile to that of the 10% EC 76 coated, drug layered spheres of the same mesh size produced an estimated similarity factor for the encapsulated spheres of 61 (using data points up to and including 12 hours). Variation was required for the application of the 30% ethylcellulose coating as the yield of useable spheres was low. The yields for each mesh fraction for both methods are listed in Table 6. The desired fraction, 25-40 mesh, increased by about 24% when the 2-part application was used instead of the single application. Table 6: Comparison of Yields for 30% Ethylcellulose Application Mesh Size Process Comments 1-Step 2-Step >25 36.0 % 12.1% Fraction contained mostly clusters of 2 or more spheres 25-40 63.9% 87.8% Fraction consisted of mostly single spheres <40 0.1% 0.1% Fraction contained broken spheres Figure 8 compares the dissolution profiles for the capsule and tablets prepared from the two batches of 30% EC coated, C P M layered, 30-35 mesh size spheres. The estimated similarity factors for the two batches were 89, 74 and 76 for the capsules, non-disintegrating tablets and disintegrating tablets respectively. These values were calculated with all data points up to and including the 24, 12 and 24 hours timepoints respectively. These results indicate a difference of <5% between the profiles for each dosage form. 77 Figure 8: Dissolution Profiles for 30% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Coated by 1 and 2 Step Coating Applications 78 5.2, Validation of UV Assay of Model Drug A typical U V spectrum of chlorpheniramine maleate in distilled water appears in Figure 9. The Beer's Law plot for chlorpheniramine maleate was found to be linear. A typical linearity plot appears in Figure 10. This plot used data generated using the HP U V system: Table 7 details the linearity equations and the concentration ranges for the two U V spectrophotometers used. 3 o-a b o.o-| , , —) 220 242 2«3 235 307 326 350 wavelength (nm) Figure 9: UV Spectrum of CPM in Water Table 7: CPM Linearity Data for HP and Ocean Optics UV Spectrophotometers U V System Beer's Law Lower C P M Upper C P M Correlation Equation Concentration Concentration Coefficient ,r2 (mg/mL) (mg/mL) (# of observations) HP A^14.26c +0.002 0.0001 0.055 0.99994 ( ID Ocean A=14.26c + 0.06 0.003 0.055 0.9994 Optics (6) The U V spectrum of the soluble sphere components and external excipients were also examined. The materials examined were uncoated spheres, HPMC, PVP and Di-Pac. The absorbance values for each component at 262 nm (i.e. in the region of the quantifying peak of the C P M spectrum) were indistinguishable from the baseline. 79 L j I I I I I I I I I I 1 1 1 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Concentration (mg/mL) Figure 10: Typical Beer's Law Plot of Chlorpheniramine Maleate At times, extraneous, interference peaks were observed at about 261 nm when using the Ocean Optics U V system. Analysis of the same samples on the HP U V system demonstrated no such peaks in the C P M spectrum. These peaks were present in the standard solution as well as samples of distilled water. Pve-scanning of the blank usually resulted in removal of these extraneous peaks. Acceptable performance of the Ocean Optics system against that of the HP system was confirmed on several occasions when samples tested on both systems yielded the same results. 80 5.3. Investigations of Dissolution Related Errors 5.3.1. Solvent Evaporation This investigation explored possible reasons for unexplained increases in drug content after heating in the dissolution vessels. The operation of supporting equipment such as the balances was found to be normal. No appreciable differences were found in the drug assayed from dissolution samples taken manually and via the autosampler (< 1% difference). A check of the linearity of the Beer's law plot of C P M demonstrated agreement of the linearity equation with previous determinations. Comparison of drug i I I I i I I I I I I I 1 I L 0 2 4 6 8 10 12 14 16 18 20 22 24 T ime (hours) Figure 11: Dissolution Solvent Losses Due to Evaporation Over a 24-Hour Period 81 content from a solution before and after heating and stirring resulted in a slight increase (3%) in the apparent amount of drug recovered. Loss of solvent (water) over a 24-hour period was then examined. The results appear in Figure 11. The experimental data is plotted as individual points while the lines represent the calculated, theoretical trend. The rates of evaporative loss of solvent were calculated to be -4.92 mL/hr (r2 = 0.9999, 6 observations) and -1.01 mL/hr (r2 = 0.9997, 6 observations) for the uncovered and covered vessels respectively. 5.3.2. Apparent Drug Gains and Losses Due to Insoluble Materials Systematic drug loss was observed with the use of two external excipients, croscarmellose sodium and dicalcium phosphate. The results for the recovery of drug from combinations of drug and excipients (Avicel® formulation) are listed in Table 8. All samples containing croscarmellose have a lower than expected recovery of drug at three hours (<70%). Table 8: Drug Recovered from Dissolution Vessels Containing Various Combinations of Excipients and Disintegrating Tablet Formulations External Components Drug Recovered (%) t= 2 hours t= 3 hours None 101.3 102.2 Avicel® 99.9 99.8 croscarmellose 68.4 69.3 Avicel® and croscarmellose 65.7 67.2 Avicel®, croscarmellose and stearic acid 66.1 67.9 Avicel®, uncoated spheres, croscarmellose and 66.5 68.8 stearic acid (as a tablet) Avicel®, EC-coated spheres, crospovidone and 96.8 97.3 stearic acid (as a tablet) Dicalcium phosphate, EC-coated spheres, 87.4 89.0 crospovidone and stearic acid (as a tablet) Starch 1500®, EC-coated spheres, 101.9 108.3 crospovidone and stearic acid (as a tablet) 82 Table 8 also lists the recovery of drug from the three early disintegrating tablet formulations. There was an apparent loss in amount of drug for the dicalcium phosphate formulation. Additionally, model drug was detected in the methanol solution of the insoluble materials present after the tablets containing dicalcium phosphate had disintegrated (-0.8 mg). Interference solutions in methanol demonstrated no interference from any of the components in the region of the C P M peak. There was an apparent gain of drug for the Starch 1500® formulation. Some absorbance was observed for the water soluble components of Starch 1500®. Filtered, aqueous solutions of Starch 1500® at various concentrations absorbed weakly at 200 and 235 nm though the absorbance did not appear to vary with concentration (see Figure 12). Absorbance at 261 nm was above the baseline (average of - 0.02 absorbance units) resulting from the shoulder of the peak at 235 nm. 0.03 - + j= 0 .02 cz CM t o CM -+ + TO 0) O c + 1 a < 0.01 -+ i , i i i i i > i i i o .ou 240 250 2 6 0 270 280 290 300 Mass of Starch 1500 in 900 mL of Distilled Water (mg) Figure 12: Absorbance of Water-Soluble Components of Starch 1500® at 262 nm 83 5.4. Formulation Development 5.4.1. Initial Formulations (25% Excipient Level) The Avicel® and Starch 1500® tablet formulations disintegrated completely during the dissolution testing while the dicalcium phosphate and ground, uncoated spheres formulations remained intact. The dissolution profiles for all formulations appear in Figure 13. The spread of the data is somewhat large. Blend separation during tablet preparation was observed for all formulations especially for the dicalcium phosphate, Starch 1500® and ground spheres formulations. 120 80 h a: _ 0 _ Avicel formulation Dicalc ium p h o ^ h a t e formulation - A — Starch 1500 formulations Ground spheresformulat ion mean ± SD, n=6 10 12 14 16 Time (hours) 18 20 22 24 Figure 13: Drug Release Profiles for Tablets Containing 10%-EC Coated, CPM Layered, 20-25 Mesh Size Spheres (75%), Diluent (24%) and Stearic Acid (1%) 84 5.4.2. Formulations Containing 50% Excipients The dicalcium phosphate formulation fragmented into small pieces within 1 minute after contact with the dissolution fluid. The ground spheres formulation containing 3% croscarmellose required 22 minutes to break into smaller fragments. An increase to 5% croscarmellose had no observable effect on the disintegration. The dissolution profiles for the three formulations containing croscarmellose are depicted in Figure 14. A large variability was observed for the data point for each profile. Also, low drug recovery was observed for all formulations. "O » ro CD or 110 100 90 60 50 40 30 20 10 -#— Avicel formulation Dicalcium phosphate formulation - A - - Starch 1500 formulation mean ± SD, n=6 _ ! I I l _ J I I ! _ 10 12 14 Time (hours) 16 18 20 22 24 Figure 14: Drug Release Profiles for Tablets Containing 10%-EC Coated, CPM Layered, 30-35 Mesh Size Spheres (50%), Diluent (46%), Croscarmellose Sodium (3%) and Stearic Acid (1%) 85 5.4.3. Evaluation of Alternate Disintegrant Replacement of croscarmellose sodium with crospovidone did not adversely affect disintegration of the tablets. Interference studies of the model drug and crospovidone resulted in complete recovery of model drug (100.5%). Dissolution data for the tablets prepared from the blend demonstrated a wide variability as previously observed and an apparent loss of drug. Variability was smaller for the dissolution results of individually prepared tablets but loss of model drug was still observed (85% of drug was recovered). 5.4.4. Evaluation of Alternate Brittle Diluents U V spectra of aqueous solutions of alternate diluent materials demonstrated no absorbance above the baseline in the region of the quantifying C P M peak (-261 nm). Table 9: Results of Disintegration Testing for Alternate Brittle Diluent Diluent Crospovidone Level Time Observations Ground, uncoated spheres 3% 15 min. Tablets fractured into several pieces. 5% 15 min. Tablets fractured into smaller pieces. 6% 7.5 min. Tablets fractured into smaller pieces. 8% 7.5 min. Tablets fractured into smaller pieces. Uncoated spheres 3% 15 min. Tablet mainly intact. 5% 15 min. Tablet mainly intact. 7% 10 min. Tablet fractured into smaller pieces. 9% 10 min. Tablet fractured into smaller pieces. Sorbitol-mannitol CSD 5% 10 min. Tablet mainly intact. 7% 10 min. Tablet mainly intact. 9% 10 min. Tablet mainly intact. 11% 10 min. Tablet fractured into smaller pieces. 13% 10 min. Tablet fractured into smaller pieces. 15% 10 min. Tablet fractured into smaller pieces. Di-Pac 3% 10 min. Tablet partly intact. 5% 10 min. Tablet fractured into smaller pieces. 7% 10 min. Tablet fractured into many, very small pieces. 9% 5 min. Tablet fractured into very small pieces. 11% 5 min. Tablet fractured into very small pieces. 13% 5 min. Tablet fractured into very small pieces. 86 Disintegration testing results for the alternate, brittle, diluent materials appear in Table 9. Small fragments or pieces of tablets consisted of several spheres fused together. Complete disintegration (i.e. no material retained on the screen) was not achieved for all samples. 5.4.4. Comparison of Variability in Dissolution Profiles Between Tablet Preparations (Individual vs. Blend Sampling) Figure 15 demonstrates the difference in variability between dissolution profiles of tablets prepared individually and those sampled from a formulation blend. The observed (A) 130.0 r 97.5 65.0 h 32.5 T3 CD CO CD _CD CD 01 (B) 0.0 130.0 97.5 65.0 32.5 Individual Tablet Preparation (mean ± SD, n=3) - j i i i i i i i _ 10 12 14 16 0.0 S H E -31-Blend Tablet Preparation (mean ± SD, n=3) 4 6 8 Time (hours) 10 12 14 16 Figure 15: Dissolution Profiles of 10% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Compressed with Sorbitol-Mannitol CSD and Other Excipients: (A) Tablets Prepared Individually and (B) Tablets Prepared from a Formulation Blend 87 variability in dissolution profiles for tablets prepared individually was much smaller than those prepared from the formulation blend. The relative standard deviations for the averages generated from the individually prepared tablets and those from the formulation blend ranged from 0.2-4% and 7-12% respectively. Also, the quantity of drug released from the individually prepared tablets is closer to the theoretical value of 100% released. 5.5. Effects of Substrate Size on Drug Release from Encapsulated and Compressed Spheres The drug release profiles for the encapsulated spheres coated at 10%> Surelease® appear in Figure 16. Each data point is an average of 6 individual units unless otherwise stated. Error bars indicate ± 1 standard deviation. The similarity factor estimates for profile similarity between the dissolution profiles for the controls are listed in Table 10. Values greater than 50 indicate a difference of less than 10%> between profiles. The capsule shells dissolved within 10 minutes of contact with the dissolution medium. Compression of the spheres resulted in fusion of the ethylcellulose membranes on each sphere to produce an ethylcellulose network which visually resembled a matrix. These structure remained intact after dissolution testing. The in vitro profiles appear in Figure 17. The estimated similarity factors comparing the different profiles for the compressed spheres appear in Table 11. Comparisons of drug release profiles of the compressed spheres with those of the corresponding controls appear in Figure 18. The estimated similarity factors for drug release from the encapsulated (control) and compressed spheres are listed in Table 12. 88 Time (hours) Note: for 45-60 mesh fraction, n=5 for t=16 and 24 hours Figure 16: Drug Release Profiles for Encapsulated, 10% EC Coated, CPM Layered Spheres of Various Substrate Size (Controls) Table 10: Estimated Similarity Factors for Encapsulated, 10% EC Coated, CPM Layered Spheres of Various Substrate Size (Controls) Estimated Similarity Factor (final time point)* Substrate Mesh Size 14-18 20-25 30-35 20-25 27(12hr) 30-35 25 (12 hr) 60(12hr) 45-60 NC NC NC Note: NC = not calculated * Values > 50 indicate differences between profiles of < 10%. 89 110 100 90 80 $ 7 0 03 I 60 cn 50 40 30 20 10 0 14-18 mesh fraction 20-25 mesh fraction 30-35 mesh fraction - - - o - - - 45-60 mesh fraction average ± SD, n=6 unless otherwise stated 10 12 14 16 18 20 22 24 Time (hours) Note: for 14-18 mesh fraction, n=5 fort=4 hours Figure 17: Drug Release Profiles for Compressed, 10% EC Coated, CPM Layered Spheres of Various Substrate Size Table 11 : Estimated Similarity Factors for Compressed, 10% EC Coated, CPM Layered Spheres of Various Substrate Size Substrate Mesh Size Estimated Similarity Factor (time point) 14-18 20-25 30-35 20-25 55 (6 hr) 30-35 43 (6 hr) 63 (8 hr) 45-60 49 (6 hr) 77 (8,hr) 75 (8 hr) 90 2 « .2 E 1 a i-a. •a ey /—s I. 4* S N a. ™ •3 ^ — o m u £ oin oin .5 oj hr hr 0 / 3 S (N O mated tor (ti r-IT) 72 (1 z w fa trate Size trate Size 00 in IT) o trate Size -—i CM m c/j _C i i O o 1 in <N in T3 <U 3 o o a 5.6. Effects of Compression with Excipients The in vitro release profiles of disintegrating tablets consisting of 10% Surelease coated, 30-35 mesh size spheres and excipients appears in Figure 19. The profiles for the encapsulated and compressed spheres appear for comparison. Al l drug was released quickly, within 2 hours. Similarity factors could not be calculated as the drug release was too rapid. 100 80 h •o CD CD 60 _CD CD 40 20 - 9 — Avicel formulation DiPac formulation Starch 1500 formulation ---0--- Encapsulated spheres (no excipient) — Compressed spheres (no excipient). average ± SD, n=6 unless otherwise stated _i i i_ 6 8 10 12 14 16 18 20 22 24 Time (hours) Note: For Starch 1500 formulation at t=2,3,8 and 12 hours, n=5 and at t=24 hours, n=4. Encapsulated and compressed spheres profiles shown for comparison. Figure 19: Drug Release Profiles for 10% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Tabletted with Various Excipients 92 5.7. Effects of Ethylcellulose Coating Level The drug release profiles of encapsulated, 30-35 mesh size spheres coated at the three different levels of Surelease® are shown in Figure 20. As expected, the apparent rate of release is decreased with increasing coating level. The estimated similarity factors appear in Table 13. Each of these profiles differs from the adjacent profile by more than 10%. As dissimilarity in release profile was an acceptance criteria for the batches, these results are not surprising. The drug release profiles of the compressed spheres coated at various Surelease® levels are illustrated in Figure 21 and the estimated similarity factors are listed in Table 14. The effect of decreased apparent release rate with increasing coating level is not as pronounced. The profiles for the 10% and 20% Surelease® coated spheres differed by A more than 10% (f2 of 43 at 12 hours) with a decrease in the apparent rate of drug release at the higher coating level. However, there is a striking similarity between the 20% and 30% Surelease® coated batches (f2 of 88 at 24 hours indicating a difference of <2%). The drug release profiles for the encapsulated and compressed spheres at each coating level is illustrated in Figure 22 and the estimated similarity factors for the encapsulated and compressed spheres for each coating level are listed in Table 15. The profiles for the encapsulated and compressed, 20% Surelease® coated, 30-35 mesh size spheres were similar as were the profiles for the encapsulated and compressed, 30% Surelease® coated, 30-35 mesh size spheres. Also, as the % drug release for the compressed spheres was plotted with respect to the square root of time to consider a possible matrix type release (Figure 23). The regression coefficients for each profile is listed in Table 16. 93 Figure 20: Drug Release Profiles for Encapsulated, CPM Layered, 30-35 Mesh Size Spheres at Various EC Coating Levels Table 13: Estimated Similarity Factors for Encapsulated, CPM Layered 30-35 Mesh Size Spheres at Various EC Coating Levels Coating Level Estimated Similarity Factor (time point) 10% 20% 20% 38 (12 hr) 30% 27(12hr) 44 (12 hr) 94 100 ® @ JR 80 b Released O b Released O i f' 40 T M f t ' 20 i — © — 10% E C coated spheres • • • A - • • 20% E C coated spheres • - • ! ! - • - 30% E C coated spheres average ± SD, n=6 0 1 . 1 . 1 , 1 , 1 , 1 , 1 , 1 0 2 4 6 8 10 12 14 Time (hours) 16 18 20 22 24 Figure 21: Drug Release Profiles of Compressed, CPM Layered 30-35 Mesh Size Spheres at Various EC Coating Levels Table 14: Estimated Similarity Factors for Compressed, CPM Layered 30-35 Mesh Size Spheres at Various EC Coating Levels Coating Level Estimated Similarity Factor (time point) 10% 20% 20% 43 (12 hr) 30% 42 (12 hr) 88 (24 hr) 95 110 100 90 80 70 d) CO 60 0) CD EC 50 40 30 -20 10 0 10% EC coated, 14-18 mesh spheres 10% EC coated, 20-25 mesh spheres 10% EC coated, 30-35 mesh spheres 10% EC coated, 45-60 mesh spheres 20% EC coated, 30-35 mesh spheres 30% EC coated, 30-35 mesh spheres average ± SD, n=6 unless otherwise stated Square Root of Time (hours**0.5) Note: for 10% EC coated, 14-18 mesh spheres att=4 hours, n=5 Figure 23: Drug Release for Compressed, EC Coated, CPM Layered Spheres - % Released vs. Square Root of Time Plot Table 16: Regression Coefficients for % Drug Released Versus Square Root of Time for Compressed EC Coated, CPM Layered Spheres Description r 2 10% Surelease® coated, C P M layered, 14-18 mesh size spheres 0.958 10% Surelease® coated, C P M layered, 20-25 mesh size spheres 0.986 10% Surelease® coated, C P M layered, 30-35 mesh size spheres 0.991 10% Surelease® coated, C P M layered, 45-60 mesh size spheres 0.984 20% Surelease® coated, C P M layered, 30-35 mesh size spheres 0.993 30% Surelease® coated, C P M layered, 30-35 mesh size spheres 0.999 Note: Coefficient calculated using 7 data points (t=0.5 to t=6 hours) for all spheres except 14-18 mesh size where 6 data points (t=l to t=6 hours) were used. 97 5.8. Effects of Ethylcellulose Coating Level and Compression with Excipients The drug release profiles of 20% and 30%> Surelease® coated, 30-35 mesh size spheres tabletted with various excipients are depicted in Figures 24 and 25 respectively. The profiles for the encapsulated and compressed spheres at the same coating level are included in each figure for reference. The estimated similarity factors are listed in Tables 17 and 18. The drug release profiles of the formulations plotted by diluent materials are shown in Figure 26. The estimated similarity factors appear in Table 19. In each case, increases in coating level led to decreases in apparent rate of release. In all cases, the profiles for each coating level differed by more than 10% for each diluent examined. 98 Time (hours) Note: For Starch 1500 formulation att=16hours, n=5 Figure 24: Dissolution Profiles for Disintegrating Tablet Formulations Containing 20% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Table 17: Estimated Similarity Factors for Disintegrating Tablet Formulations Containing 20% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Diluent Estimated Similarity Factor (time point) Avicel DiPac Starch 1500 None: Encapsulated spheres NC NC 23 (6 hr) None: Compressed spheres 13 (1.5 hr) 16 (1.5 hr) 27 (6 hr) DiPac 50 (1.5 hr) Starch 1500 23 (1.5 hr) 28 (1.5 hr) 99 Note: For Starch 1500 formulation at t=12 hours, n=4 and at t=16 hours, n=5 Figure 25: Dissolution Profiles for Disintegrating Tablet Formulations Containing 30% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Table 18: Estimated Similarity Factors of Disintegrating Tablet Formulations Containing 30% EC Coated, CPM Layered, 30-35 Mesh Size Spheres Diluent Estimated Similarity Factor (time point) Avicel DiPac Starch 1500 None: Encapsulated spheres 18(6 hr) 24 (6 hr) 34 (12 hr) None: Compressed spheres 23 (6 hr) 26 (4 hr) ' 51 (12 hr) DiPac 50 (6 hr) Starch 1500 27 (8 hr) 31 (6 hr) 100 Avicel Formulations DiPac Formulations 100 80 60 40 T3 0) 20 to 03 ® 0 CD or: 100 80 60 40 20 5 15 25 Starch 1500 Formulations* 5 15 Time (hours) 25 100 80 60 40 20 0 15 25 — • — 10% EC coated spheres • 20% EC coated spheres • 30% EC coated spheres average ± SD, n=6 unless otherwise stated *Note: For 10% EC coated spheres at t= 2,3,8 and 12 hours, n=5 and at t=24 hours, n=4 For 20% EC coated spheres at t=16 hours, n=5 For 30% EC coated spheres at t=12 hours, n=4 and at t=16 hours, n=5 Figure 26: Dissolution Profiles for Disintegrating Tablet Formulations Table 19: Estimated Similarity Factors for Each Formulation for the Various EC Coating Levels Diluent Estimated Similarity Factor (time point) Avicel DiPac Starch 1500 Coating level 10% 20% 10% 20% 10% 20% 20% 38 (1.5 hr) 29 (1.5 hr) 22(1.5hr) 30% 22 (1.5 hr) 37 (2 hr) 15 (1.5 hr) 30 (2 hr) 12 (1.5 hr) 31 (6 hr) 101 5.9. Preliminary Compression Studies The results of most of the preliminary compression work appear in Appendix 1. Plots of porosity, peak offset time, work of compression and tensile strength against peak compression pressure are shown for the materials and formulations for comparison of substrate size, coating level and excipient formulations. The plot of tensile strength against peak pressure for the different coating levels appears in Figure 27. This data suggests a change in tensile strength behaviour at the higher peak pressures for the Surelease® coated spheres. 2.6 ra f> 1 6 c CD i_ 00 c 0.6 10% Surelease coated 20% Surelease coated 30% Surelease coated Uncoated Drug layered © v •of4 v © o 20 40 60 80 100 120 140 160 180 200 220 240 260 Peak Pressure (MPa) Figure 27: Tensile Strength of Tablets Compressed from 30-35 Mesh Size Spheres Coated with Various Coating Materials and EC Coating Levels 102 5.10. Comparison of Profiles for Proposed System The drug release profiles for the most promising formulations, the Starch 1500® formulations were compared to those from encapsulated and compressed spheres at each coating level. The estimated similarity factors for these profiles appear in Table 20. The dissolution profiles for the proposed system appears in Figure 28. 1 1 0 1 0 0 9 0 8 0 7 0 h 6 0 5 0 4 0 3 0 2 0 1 0 0 T i m e ( h o u r ) • E n c a p s u l a t e d , 1 0 % E C C o a t e d , 3 0 - 3 5 M e s h S i z e S p h e r e s a . . . . C o m p r e s s e d , 1 0 % E C C o a t e d , 3 0 - 3 5 M e s h S i z e S p h e r e s B 3 0 % E C C o a t e d , 3 0 - 3 5 M e s h S i z e S p h e r e s T a b l e t t e d w i t h S t a r c h 1 5 0 0 " e r a g e ± S D , n = 6 u n l e s s o t h e r w i s e s t a t e d • N o t e : F o r S t a r c h 1 5 0 0 f o r m u l a t i o n , a t t = 1 2 h o u r s , n = 4 a n d a t t = 1 6 h o u r s , n = 5 Figure 28: Dissolution Profiles for Dosage Forms in the Proposed System Table 20: Estimated Similarity Factors for Starch 1500 Formulations Compared with Encapsulated and Compressed, EC Coated, CPM Layered, 30-35 Mesh Size Spheres Sphere coating .level for Starch 1500® formulations Estimated Similarity Factor (Timepoint) 10% Coated Spheres 20% Coated Spheres 30% Spheres Encapsulated Compressed Encapsulated Compressed Encapsulated Compressed 10% NC 17 (1.5 hr) NC 13 (1.5 hr) NC 12 (1.5 hr) 20% 41 (6 hr) 40 (6 hr) 23 (6 hr) 27 (6 hr) 17 (6 hr) 26 (6 hr) 30% 51 (12 hr) 54 (12 hr) 54 (12 hr) 52 (12 hr) 34 (12 hr) 51 (12 hr) NC = not calculated due to insufficient number of data points 103 6. Discussion 6.1. Coating Method Development 6.1.1. Drug Layer Coating Solution Formulation Development The main problem encountered in the preparation of drug-layered spheres was agglomeration or fusion of two or more spheres into an inseparable mass. A change in the application procedure (i.e. alternate coating application and drying intervals) was tried but did not prevent nor sufficiently reduce sphere agglomeration. However, as smaller fused masses of spheres were observed, all subsequent applications were carried out with alternate coating application and drying intervals. It was suspected that the model drug at the concentration used in the coating soluition (DLS-1) behaved as a plasticiser for the HPMC and lowered the glass transition temperature below any practical column drying temperature. To test this, coating of spheres was attempted using coating solution without the model drug (DLS-2) while keeping all other coating parameters constant. The yield of individual beads increased to about 90%. However, the remaining 10%> were agglomerate of 2 or more spheres. These results suggested that the combination of the model drug and film-forming polymer employed in the coating solution led to the agglomeration problem. It was believed at the time that modifications to the coating.solution might have produced a viable product. Two modifications without model drug were tested: 1) use of a coating solution containing sodium chloride as suggested by Yuasa, Nakano and Kanaya (1997) and 2) use of a more volatile solvent. Use of the 5% sodium chloride coating solution (DLS-3) resulted in no significant increase to the yield of viable spheres. For the 104 alternate option, use of a hydroalcoholic vehicle (2:l/water:alcohol) containing FfPMC and talc resulted in an even lower yield of viable spheres (19%). However, use of the more volatile solvent resulted in the majority of agglomerates consisting of 3 spheres and was suspected to be useful in future attempts of drug layer application' If was concluded that examination of another film-forming polymer might be a better course of action than further modification of the HPMC coating solution. A non-cellulosic film-forming polymer, polyvinylpyrrolidone or povidone, was examined. The coating solution consisted of 6% w/v of chlorpheniramine maleate and 0.5% povidone in a 75:25/alcohol:water mixture. Alternate 10 minute intervals of coating application and in-column drying was used. This application-drying cycle was continued as initial agglomeration of spheres was observed in the column during coating. Most of the final product consisted of individual spheres (about 95%). This coating method appeared very promising and was further refined such that the column parameters were optimised to increase the yield and efficiency. Also, the coating application/in-column drying interval was reduced to 5 minutes to prevent clogging of the atomiser nozzle resulting from overheating. The final method produced only individual spheres (no agglomerates were observed). In response to the losses of coating material during the coating process, the coating efficiencies when using excesses of 20 and 40% coating solution were examined. As the amounts of drug assayed were similar for the two excess amounts, an excess of 20% coating solution was adopted (for both drug layering and ethylcellulose coating) in all subsequent coating applications. As variability and losses inherent with this coating 105 equipment were inevitable (and observed), assay of drug content for each batch was necessary. 6.1.2. Ethylcellulose Application Development The estimated similarity factor comparing the drug release profiles of the 10% and 15% EC coated, C P M layered, 30-35 mesh size spheres indicated a difference of less than 10%. As a result, the 15% EC coated batch was discarded as it failed to meet the release criteria for spheres to be used in the coating level study. The failure of the 15%> EC coated spheres to meet the dissolution release criteria may have been a limitation of the variability in the coating equipment. Therefore, it was decided to increase the coating level by 10% increments as the coating equipment could not guarantee dissimilarity in the drug release profiles of spheres coated with 5% coating level differences. Continuous application of coating solution onto the 30% Surelease® coated batch resulted in many instances of 2 spheres or more being strongly agglomerated together. The batch was useable after the removal of the agglomerates by screening and functioned as the primary batch. The yield of viable spheres was increased when the Surelease® coating was applied in two parts over a period of two days. Comparison of the dissolution profiles for each type of dosage form (i.e. capsule, non-disintegrating tablet and one of the disintegrating tablets) demonstrated similar drug release between the two batches of spheres for all three dosage forms. Based on these results, any subsequent 30% ethylcellulose coating of spheres was conducted using the two-step application. 106 6.2. Validation of UV Assay of Model Drug The U V spectrum of the model drug agreed with that from the literature (Eckhart and- McCorkle, 1978). Linearity of the Beer's Law plot of absorbance against concentration was established for both U V systems down to model drug concentrations of 1.1 x 10'4 and 2.8 x 10"3 mg/mL for the HP and Ocean Optics systems respectively. The Beer's law equations also agreed with previous determinations conducted prior to this project. Also, no significant absorbance was observed at 261 nm for any of the soluble components of the Surelease® coated spheres nor the soluble tablet excipients on either U V system. Based on this information, the U V assay was accepted for use in the study. The extraneous peaks observed on the Ocean Optics U V system were thought to result from irregularities in the absorbance of the reference (i.e. blank) spectrum and were another limitation of the xenon flash lamp light source. These peaks were present intermittently and if present, the U V system was not used to analyse the test solutions. However, when these extraneous peaks were absent, the system operated well as confirmed by comparison of data generated on both U V systems. Therefore, a system suitability test was adopted for use of the Ocean Optics system. The system would be acceptable for use if the spectrum generated for a C P M solution agreed with the documented spectrum (i.e. Figure 9). Use of the Ocean Optics system was eventually discontinued as the increasing amount of time required to achieve system suitability became impractical. 107 6.3. Investigations of Dissolution Related Errors 6.3.1. Solvent Evaporation Quantitative results demonstrated that the supporting equipment was functioning normally and that sampling method had no appreciable effect on the calculated recovery values. A small increase in amount of drug recovered was observed over a 3 hour period whenever heating of samples was involved. This led to an investigation of possible significant losses due to evaporation of the medium (Tang and Schwartz, 1998). Rates of evaporative loss over 24 hours were found to be approximately 5 and 1 mL/hr from uncovered and covered vessels respectively. Both rates were linear with regression coefficients of > 0.999. Fluid loss due to evaporation became more significant for samples taken later in the dissolution run and hence it was important to account for this loss in determining the drug release profile. 6.3.2. Apparent Drug Gains and Losses Due to Insoluble Materials A 30% loss of model drug was observed for all vessels containing croscarmellose sodium. It is possible that an ionic attraction existed between the positively charged, aliphatic nitrogen on the chlorpheniramine molecule and the negatively charged croscarmellose. As the croscarmellose is insoluble, the adsorbed C P M would be effectively removed from solution and hence would not be detected in the dissolution samples. As no reproducible stoichiometric relationship was apparent, the data could not be salvaged. Alternate, non-ionic disintegrants were then evaluated. A 10% loss of drug was observed for the disintegrating tablets containing dicalcium phosphate. It is again possible that an ionic attraction exitsed between the 108 aliphatic nitrogen and negatively charged surfaces of the insoluble dicalcium phosphate. As previously observed, the adsorbed C P M would be effectively removed from solution. Presence of C P M in the methanolic solutions of the insoluble remnants of the dicalcium phosphate tablets supported this suspicion. An alternate brittle diluent was then investigated. The apparent gain in drug observed for the Starch 1500® tablets appeared to be due to U V interference from water soluble components of the diluent. The absorbance appeared to be constant over the range of concentrations studied and seemed to suggest saturation of the soluble components. Therefore, alterations were made to the dissolution method to counter this effect when Starch 1500® was used. The initial dissolution fluid volume and sample volumes were reduced to 500 mL and 5 mL respectively. The reduction of the initial fluid volume increases the concentration of model drug and reduces the interference due to the Starch 1500® to negligible amounts. 6.4. Formulation Development 6.4.1. Initial Formulations (25% Excipient Level) Blend separation problems appeared to contribute to a noticeably large spread of the profiles between the six individual determinations. In addition, the exact mass of Surelease® coated spheres was not precisely known, only estimated based on blend composition. As this data was used primarily to develop the formulations, the trends in the general release profiles were important and the variability was not of great concern during this stage of the development activity. The rapid disintegration of the tablets containing Avicel® and Starch 1500® most likely resulted from the inherent disintegrant 109 properties of these two diluent materials. Dicalcium phosphate and ground spheres do not have such disintegrant properties and this may have contributed to the failure these tablets to disintegrate during dissolution. Drug release from the dicalcium phosphate and ground spheres formulations did not quite reach 100% released. This observation was thought to be a result of using a blend to prepare the tablets. As the formulation exercise required disintegrating tablets, disintegrant was needed for the brittle diluent materials. For consistency, disintegrant was included in all subsequent formulations. Drug release from the Avicel® and Starch 1500® formulations was relatively rapid with most drug released at 2 and 6 hours respectively. Of these two formulations, the Starch 1500® formulation appeared to exhibit a slower release of drug. However, an estimated similarity factor (f 2) could not be calculated due to the rapid release of drug. Also, the rapid drug release suggested damage to the ethylcellulose membrane during tabletting. Compression-induced damage is suspected to occur whenever the spheres are tabletted regardless of the presence of excipients. In the absence of external excipients, the ethylcellulose film on each sphere was able to fuse with that on neighbouring spheres thus forming the matrix-like network. In the disintegrating tablets, the diluent acted as a physical barrier preventing the intimate contact needed for formation of the network. When the tablet disintegrates, the damage to the EC coating permits the rapid release of drug. Compression-induced damage to coating may have occurred for the dicalcium phosphate and ground spheres formulations but the failure of these tablets to disintegrate may have masked the damage. Water penetration into the tablet may have been slower 110 thus observed drug release would be slower. Formation of an ethylcellulose network may have occurred but such a possibility is not supported by these data. 6.4.2. Formulations Containing 50% Excipients Based on the results for the initial formulations (25% excipient level), the formulation composition was changed. Adequate break-up of the tablets containing dicalcium phosphate was observed. Inclusion of croscarmellose sodium in the ground spheres formulation did not provide satisfactory tablet break-up at either level examined. As a result, ground, uncoated spheres were discarded as a possible diluent material. The large variability of the data was attributed to the preparation of tablets from formulation blends. The low recovery of model drug (approximately 60-90%) appeared too large to be due solely to blend preparation. This prompted an investigation and revealed that the sources of the problem were the disintegrant and the use of dicalcium phosphate as a diluent material (see Section 6.3.2). The formulations were re-designed to include a non-ionic disintegrant. A replacement for the brittle diluent was also investigated. 6.4.3. Evaluation of an Alternate Disintegrant Crospovidone was examined as a replacement for croscarmellose sodium. Complete model drug recovery was demonstrated for this material indicating no interaction between the two materials with regards to the drug assay. In addition, relatively low disintegrant levels (i.e. 3%) were maintained. Complications developed in the in vitro evaluation of the suitability of crospovidone due to the use of dicalcium phosphate (see Section 6.3.2) and the use of blend sampling in tablet preparation. Once 111 these complications were corrected, the suitability of crospovidone was established and used in all formulations. Crospovidone was included in the Avicel® and Starch 1500® formulations at a low level (3%) for consistency even though a disintegrant was not necessary. 6.4.4. Evaluation of Alternate Brittle Diluents The U V spectra of the aqueous solutions of the alternate diluent materials demonstrated no interference with the U V assay of the model drug. Complete disintegration (i.e. no material retained on the mesh) may not have been achieved due to similarity in size of the mesh and the EC coated spheres. As a result, adequate disintegration was achieved if the retained material was very small (i.e. pieces consisted of several spheres fused together). Samples of the formulations containing ground, uncoated spheres were unable to demonstrate adequate disintegration. Formulations containing uncoated spheres required a high level of disintegrant to achieve fracture into smaller pieces. The sorbitol-mannitol CSD formulations required a high level of disintegrant to produce reasonable tablet break-up. These three materials were discarded as possible replacements for the brittle diluent because of the lack of adequate disintegration demonstrated or the high level of crospovidone required. Sufficient disintegration was achieved for the Di-Pac formulations at disintegrant levels of 7% and higher. Di-Pac was found suitable to replace dicalcium phosphate because of the lack of interference with the assay of model drug and the adequate disintegration observed within 10 minutes for formulations containing a moderate amount of disintegrant. In an effort to keep the disintegrant level as low as practically possible, 7% was selected. 112 6.4.5. Comparison of Variability in Dissolution Profiles Between Tablet Preparations (Individual vs. Blend Sampling) Variability in the dissolution results was reduced noticeably When tablets were prepared individually rather than from a formulation blend. This result agrees qualitatively with previous comparisons (see Section 5.4.3). This reduction in variability is intuitive as the amount of drug is measured more accurately when tablets are prepared individually instead of being estimated as is the case when a formulation blend is used. Based on these and previous results, all disintegrating tablets subjected to dissolution testing were prepared individually. 6.5. Effects of Substrate Size on Drug Release from Encapsulated and Compressed Spheres Release from the 14-18 mesh spheres was slower than that from the other sizes while release from 45-60 mesh spheres was rapid (Figure 16). Release from the 20-25 and 30-35 mesh spheres was similar, i.e. within 10% (Table 10). The differences in apparent release rates may result of from differences in surface area. Coating solids were added to give a certain overall increase in sample weight, i.e. 10%, so the amount used for the membrane was constant for all sizes of spheres. However, the available surface area for each size was different for each unit mass of spheres. The surface area of the 45-60 mesh spheres, the smallest spheres size, is much greater than that of the other sizes. It is possible that the amount of Surelease® material applied could not provide effective enough coverage of this large a surface. Consequently, soluble materials (such as drug) may have been released rapidly from inadequately coated areas of the spheres. Conversely, the 14-113 18 mesh size spheres had the smallest available surface area and this may have resulted in a thicker coating which in turn led to a slower release. Slower release through thicker membranes has been observed by other researchers (Sadaghi et al.,J2000) as well as slower release from larger spheres with the same theoretical coating (Ragnarsson and Johansson, 1988 and Lee at al., 1996). The reason for the similarity in drug release for the 20-25 and 30-35 mesh fractions is not immediately apparent and may be coincidental. Though the available surface areas for the two batches are not the same, the difference between them is smaller than the differences between the available surface areas for the other sizes. Additionally, variability inherent in the coating process may also have contributed to the observed similarity. Compression of 10% ethylcellulose coated spheres altered the drug release profiles for the 14-18 and 45-60 mesh sizes (Figures 17). The release from the 14-18 mesh size spheres is faster (complete release by 12 hr instead of >24 hr) and that from the 45-60 mesh size spheres is much slower (24 hr instead of 1 hr). The release profiles from the compressed, 20-25, 30-35 and 45-60 mesh sizes were similar (Table 11). Drug release profiles for the 14-18 and 20-25 mesh size spheres were similar. Of the four sizes of substrate, faster apparent drug release from tablets made with the 14-18 mesh size may be a result of larger pores or channels throughout the ethylcellulose network which forms as the soluble components, largely sucrose, dissolve. These larger pores result from use of the larger spheres and may have allowed faster penetration of water into the tablet and release of dissolved drug. The larger pore size may also permit more direct (i.e. less meandering) access into the tablet. For the 45-60 114 mesh spheres, the formation of the matrix provided a degree of release control not seen in the encapsulated spheres. The formation of the non-disintegrating network and alteration of the drug release profile have been observed by other researchers (Lopez-Rodriguez et al., .1993, Sarisuta and Pupreuk, 1994 and Shlieout and Zessin, 1996). Also, the percent of drug release is somewhat linear with respect to the square root of time (r2 > 0.9) until about 6 hours for all the compressed spheres (Figure 23). With slight curvature observed in the plots, the relationship is largely consistent with the model of drug diffusion from an insoluble matrix which is visually observed following complete drug release. Further work is needed before such a possibility can be entertained. However, the mechanism of drug release from non-disintegrating tablets of this sort has not been of great interest in the literature most likely because matrix-like tablets can be prepared by other methods (e.g. melts and granulations). Another reason for the lack of interest is that formation of non-disintegrating tablets is a by-product of the research; the primary aim in compressing such pellets is to create a dosage form which will disintegrate into controlled-release pellets after administration (Bodmeier, 1997). Interestingly, non-disintegrating matrices are also produced when granules prepared using an ethylcellulose dispersion as the granulating fluid are tabletted (Klinger et al., 1990 and Ghaly and Ruiz, 1996). In the case of such granules, the-dispersion solids exist both on the surface of and within the granules while in the case of coated, drug layered spheres, the dispersion solids exist only on the surface. However, during compression, fracture of both granules and spheres may render this difference meaningless and a produce similar dosage form. Ruiz and Ghaly found that drug release from Surelease® granulated granules was proportional to the square root of time (Ruiz and 115 Ghaly, 1997). More work would be required to determine if comparisons between tablets prepared from such granules and from coated spheres can be made and at which Surelease® levels is drug release equivalent. Similarities in drug release profiles were not found for the encapsulated and compressed 14-18 nor 45-60 mesh size spheres (Figure 18 and Table 12). Possible reasons for this change have already been discussed. However, more work is necessary to examine and confirm any reasons for this change. Use of these two substrate sizes was discontinued as they failed to meet the overall objective of the project. Drug release from the 20-25 and 30-35 mesh size spheres were similar regardless of encapsulation or compression. Any suspected changes in the manner in which drug is released was not suggested solely by the in vitro release profiles of these two substrate sizes. If there were changes in the manner of drug release between the encapsulated and compressed dosage forms, the similarity in release profile may be coincidental. Alternatively, the similarity may be a result of the amount of coating present, the substrate sizes or a combination of these two factors. Based on the release studies, both the 20-25 or the 30-35 mesh size spheres could be used in the capsule and non-disintegrating tablet dosage forms. However, the 30-35 mesh size was chosen because of the larger amount available and because of the closer similarity between the encapsulated and compressed release profiles. 6.6. Effects of Compression with Excipients Tabletting with excipients resulted in increases in the apparent rate of drug release (Figure 19). Again, the rapid release of drug may be indicative of damage to the 116 ethylcellulose film. The release of drug from the Avicel and Di-Pac formulations appears to be similar. However, an estimated similarity factor could not be calculated due to the rapid release. There appeared to be a slight delay in release from the Starch 1500® formulation but again, an estimated similarity factor could not be generated to support such a claim. Any possible observable differences resulting from the selection of diluent cannot be concluded based on this data only. These results are consistent with other researchers (Torrado and Augsburger, 1994). The presence of diluent appeared to prevent re-coalescence of the ethylcellulose into a matrix. However, the diluent material was not able to protect the ethylcellulose coating from compression-induced damage. The possible reasons for this are many. The particle size of the materials may be a factor. Some success was observed with ethylcellulose coated particles and diluent of much smaller particle size, 42-75 urn for the coated powder and <20 um for the diluent materials (Yao et al., 1997 and 1998). The researchers suggested that the smaller particles would be more likely to deform rather than fragment and provide better cushioning of the coated powder. In addition, at the early stages of compression, the smaller particles would more easily be rearranged under the compression pressure such that they could better protect the membrane. The particle sizes of the materials used to prepare the disintegrating tablets of the system under investigation were much larger (>500 um for the ethylcellulose coated spheres and > 80 um for the diluent materials). Such relatively large particles may not be able to be rearranged effectively enough to prevent compression-induced damage to the membrane nor might they be more likely to deform rather than fragment. 117 Another factor may be the flexibility of the ethylcellulose film itself. Surelease films prepared under similar conditions of this study are relatively weak and brittle (Parikh et al., 1993 and Bodmeier and Paeratakul, 1994). As a result, the film may be more likely to tear during compression due to shear forces encountered with the additives and fracture of the substrate. Both particle size and mechanical properties of the film are limitations of the multiparticulate system under study. Consequently, it may not be possible to adequately protect the membrane from compression-induced damage within the current confines of the system under investigation. 6.7. Effects of Ethylcellulose Coating Level Increases in coating level led to progressive decreases in apparent rate of drug release (Figure 20 and Table 13). This agrees with the observations of others (Rekhi et al., 1995, Lee et al., 1996, Miller at al., 1999 and Opota et al., 1999) and was a requirement for acceptance of the spheres for use. However, this effect in release is not as pronounced for the compressed spheres (Figure 21 and Table 14). There is a decrease in the apparent drug release rate from the 10% to 20% ethylcellulose coated spheres (f2 = 43 at 12 hours) but no such significant decrease was observed from the 20% to 30% ethylcellulose coated spheres (f2 = 88 at 24 hours). These data suggest a practical or effective maximum coating level above which further coating level increases do not further decrease the apparent rate of drug release from the compressed spheres. Decreases in apparent drug release rate with increasing ethylcellulose content have previously been observed for tablets prepared from ethylcellulose coated spheres (Rekhi et al., 1995; Opota et al., 1999 and Sadeghi et al., 2000) and from Surelease® granulated 118 granules (Klinger et a l , 1990, Ghaly and Ruiz, 1996 and Ruiz and Ghaly, 1997). However, ethylcellulose dispersion solids contents of <20% for spheres or granules have not been examined and a practical maximum ethylcellulose level has not been observed. As a result, comparison of a practical maximum level cannot be made. The reasons for this effective maximum coating level are not readily apparent. It is possible that substrate size and amount of ethylcellulose present in the tablet may be factors. As seen from Table 12, the best linear correlation for the 10% Surelease® coated, drug-layered spheres is observed for the 30-35 mesh fraction. The correlation increases with decreasing size of spheres for those spheres with complete membrane coverage. This is not surprising considering that sphere size will affect the size of the matrix pores. The linear correlation also improves with increases in Surelease® coating level. It is possible that the amount of ethylcellulose present affects the quality or continuity of the matrix, i.e a matrix made of compressed 10% ethylcellulose coated spheres may not be as coherent as one made from 20% ethylcellulose coated spheres. At the 20% coating level, there may be adequate matrix coherency such that additional Surelease® material does not produce significant changes in matrix structure and drug release. Another possible reason for the practical maximum coating level may be due to the compressional properties (such as degree of plasticity) of the ethylcellulose material itself. At the higher coating levels, the amount of ethylcellulose film materials comprises a larger proportion of the tablet and may more greatly affect the compression behaviour of the overall "powder bed." The increased amount of Surelease® material may be better able to absorb the compression energy through elastic and plastic deformation and reduce the 119 amount of fracture experienced by the substrate itself. However, work using the spray dried Surelease® material is necessary before such a possibility can be evaluated. The drug release profiles for the encapsulated and compressed spheres were similar at each ethylcellulose coating level (Figure 22 and Table 15). However, the degree of similarity was not as high at the higher coating levels compared with that at the initial coating level (i.e. 10%). This reduction in the degree of similarity may further support the possibility of a maximum practical coating level for the compressed spheres. 6.8. Effects of Ethylcellulose Coating Level and Compression with Excipients At each coating level, the profiles for the Avicel® and Di-Pac® formulations were similar (Figures 19, 24 and 25 and Tables 17 and 18). The slower release from spheres tabletted with Starch 1500® first suggested at the 10% Surelease® coating level is more apparent at these higher coating levels. At the 20% and 30% coating levels, drug release from the Avicel® and Di-Pac® formulations differed from that of the Starch 1500® formulation. In all cases, the diluent was unable to adequately protect the membrane from compression-induced damage as evidenced by the faster release of drug from the disintegrating tablets compared to that of the encapsulated spheres. The extent of damage to the membrane appeared to be the same for the plastically deforming material (i.e. Avicel®) and the brittle material (i.e. Di-Pac®). Other researchers have observed differences in the protective abilities of various diluents (Torrado and Augsburger, 1994, Lehmann et al., 1994, Torrado-Santiago et al., 1995, Khan and Zhu, 1998 and El-Mahdi and Deasy, 2000). The lack of difference may be related to the particle size of the materials. As discussed earlier, the sizes of the substrate and diluent 120 powder may be too large to be effectively rearranged under the compression forces such that the membrane is protected and may be more likely experience fracture than smaller particles. This size effect may predominate and as a result, may mask any subtle differences between the effects of the two different types of diluents. The slower release from the Starch 1500® formulations may be a result of the material's ability to gel when in contact with water (Ferrari et al., 1997; Van der Voort Maarschalk et al., 1997 and Hudson et al., 2000). It is possible that during compression, any tears to the Surelease® membrane are filled with Starch 1500®. When the tablet is placed in water, the Starch 1500® in the tears expands and wedges itself into the tear. The gel-like material then acts as a barrier to retard drug release. For all diluent materials examined, increases in coating level led to decreases in apparent rate of release. In all cases, the profiles for each coating level differed by more than 10% for each diluent examined (Figure 26 and Table 19). From this data, it can be concluded that coating level does affect the apparent rate of drug release when the spheres are tabletted with an excipient to form a disintegrating tablet. This decrease in release with increasing coating level may be due to the compression properties of the Surelease® material itself and/or the amount of material present in the compact. As the film thickness increases, the strength of the overall film increases and the greater the amount of compression force required to produce the same extent of stress and shear damage. Since the compression force was constant for all the formulations, the extent of damage to the film membranes may have decreased with increasing coating level. 121 6.9. Preliminary Compression Studies There was no apparent correlation seen between the compression parameters examined and the drug release patterns from the actual dosage forms. The close proximity of the data points for the various different materials (e.g. in the plot of porosity vs. peak pressure for the substrate size study, Figure Al-1) suggested no difference for the materials examined (e.g. spheres of different mesh size). No apparent differences were observed for substrate size, coating levels and compression with excipients for porosity and work of compression. The results of peak offset time with respect to peak pressure were inconclusive. The scatter of the data points for all of the materials suggests that this parameter may not be sufficiently sensitive to evaluate differences in the materials used. However, some provocative possibilities for future compression work may be seen in some plots. In plots of tensile strength against peak pressure at the various coating levels, there appeared to be a change in the trend for the ethylcellulose coated spheres as compared to the data for the uncoated and drug layered spheres (see Figure 27). Between 20 and 100 MPa compression pressure, there appears to be no difference for any of the materials. However, at compression pressures greater than 100 MPa, there appears to be a drop in tensile strength for the ethylcellulose coated spheres. Typically, the greater the force used in the compression, the greater the tensile strength of the tablet except at high compression pressures. These results are in agreement with those of Maganti and Celik (1994). Tensile strength is an indication of the strength of cohesive and adhesive forces between the particles and is often affected by compression pressure (Parrott, 1990). The energy used to form the tablet and keep it together is reflected in the tensile strength of the tablet. This change in the trend suggested from the probe work may indicate that the 122 tablets compressed from ethylcellulose coated spheres have a different or additional bonding mechanism other than adhesion and cohesion (i.e. the compression-induced formation of the ethylcellulose network). However, if the tablet is formed by re-coalescence of the EC material, only the amount of energy needed to induce that re-coalescence would be used and the rest of the energy of the compression event would be transferred into heat, elastic recovery, etc. Hence, the tensile strength of the tablet will be the same regardless of increasing coating level once an effective maximum EC level is reached. Such a possibility may also be supported by the similarity of dissolution profiles for the compressed spheres at the 20 and 30% EC coating levels. These results may suggest an effective maximum EC component level in the tablets for formation of a non-disintegrating, matrix-like tablet. Any increases in the EC component level may not affect the compression nor dissolution properties of the tablet. Such an effective EC component level may be seen in high EC content tablets prepared by wet granulation with ethylcellulose dispersions. 6.10. Proposed System The original aim of the study was to formulate a coated sphere which could be used in the three dosage forms, the capsule, the non-disintegrating tablet and the disintegrating tablet. Unfortunately, at all coating levels, the profile generated by the most promising formulation, Surelease® coated, 30-35 mesh size spheres compressed with Starch 1500®, was different from those of the corresponding encapsulated spheres (control) and non-disintegrating tablets. Increasing coating level for these formulations progressively reduced the apparent drug release rate to the point that drug release from the 30% Surelease® coated spheres formulated with Starch 1500® was similar to those of 123 the capsules and non-disintegrating tablets made of 10% Surelease coated spheres (Figure 28 and Table 20). This led to the proposal of the final system which is listed in Table 21. Table 21: Proposed System Capsule 10% ethylcellulose coated, 30-35 mesh size spheres Non-disintegrating tablet 10%o ethylcellulose coated, 30-35 mesh size spheres Disintegrating tablet 30%) ethylcellulose coated, 30-35 mesh size spheres (50%) tabletted with Starch 1500 ® (46%), crospovidone (3%) and stearic acid powder (1%>) Though matches were found for the Starch 1500® formulation containing 30% EC coated spheres with the 10% and 20% encapsulated and compressed spheres, the 10% coating level is proposed for practical reasons. Drug layered spheres would be coated with 10% Surelease® for the capsule and non-disintegrating tablet dosage forms. The required amount of spheres would be filled into hard gelatin capsule shells. The non-disintegrating tablet would be formed using a compression force of. 100 MPa and Vi" diameter punch tooling. The same spheres would then be further coated to a 30% Surelease® coating level for the disintegrating tablet formulation. The difference in drug content between the 10% and 30% EC coated spheres would require a slight increase in the weight of spheres included in each dosage unit. The disintegrating tablet formulation would contain 50% spheres, 46% Starch 1500®, 3% crospovidone and 1% stearic acid and would also be compressed at 100 MPa. The in vitro drug release profiles for the three dosage forms of the above system were similar and released about 90% of the model drug by 12 hours. The degree to which 124 drug release has been modified by compression is clear when drug release from encapsulated 14-18 and 45-60 mesh spheres is compared to that from the corresponding compressed spheres. Compression with excipients and the choice of those excipients also modifies drug release appreciably as observed from the release profiles of encapsulated spheres and spheres compressed with different diluent materials at all coating levels. These stark differences may be attributable to changes in the drug release mechanism or changes in the contribution of any particular release mechanism to the observed drug release profile. Identification and optimisation of common or interrelated factors for each dosage form may help design variations of the proposed system. '' Factors specific to any particular drug release mechanism would need to be carefully examined. For example, drug release from a matrix tablet is normalised by surface area (See Equation 7). Consequently, reduction of the surface area of the tablet by changes in tablet shape or dimensions may prolong the release of drug if desired. The effects on drug release and similarity to release from the other dosage forms of the proposed system would need to be evaluated. It is possible that changes in such factors may further complicate the proposed system. Disintegrant was used in tablets containing Starch 1500® for consistency with tablets prepared using other diluents. The necessity of the disintegrant in the Starch 1500® tablet formulation was not evaluated but based on the inherent disintegrant properties of this diluent and its already significant level in the formulation, the Starch 1500® level might be increased to 49% and the crospovidone eliminated from the formulation. Stearic acid should remain in the formulation as a lubricant is necessary in the tabletting process. As it is unlikely that the presence of the crospovidone retarded drug release (as this was not 125 observed in the other two disintegrating formulations at the similar coating levels), elimination of that component should not have any major impact on the drug release. Such a conclusion can be confirmed by dissolution testing of the crospovidone-absent Starch 1500® formulation. 126 7. Summary and Conclusions 7.1. Method and Formulation Development A method was developed to successfully apply the model drug, chlorpheniramine maleate, onto sugar spheres, NF. This method avoided the problem of excessive tack which results in the adhesion of spheres to form large agglomerates. Efficiency of the drug layering process was improved by alternating application and in-column drying intervals. Increased yield of spheres coated with higher levels of ethylcellulose was produced when EC coating was executed as two separate applications. Three disintegrating tablet formulations containing EC coated spheres were developed to study the effects of sphere compression on drug release when diluent materials with various mechanical properties were used. Finally, a method for detection of the model drug was developed and validated. 7.2. Effects of Substrate Size on Encapsulated and Compressed Spheres For encapsulated spheres, decreases in apparent drug release rate were observed with increasing substrate size. The similar in vitro drug release profiles observed for 10% EC coated, C P M layered spheres of 20-25 and 30-35 mesh sizes was most likely coincidental. Drug release from the 14-18 mesh size spheres was relatively slower suggesting a thicker EC coating while that from the 45-60 mesh size spheres was relatively fast suggesting insufficient or incomplete coating on the individual spheres. Compression of EC coated spheres resulted in a matrix-like structure which remained intact after release of drug during dissolution testing. Similar in vitro drug 127 release profiles were observed for 10% EC coated, compressed spheres of 20-25, 30-35 and 45-60 mesh sizes. Compression altered the in vitro release of drug from the 14-18 and 45-60 mesh size spheres. Drug release was respectively faster and slower for the two sizes. For the smaller sphere size, the formation of the matrix-like structure, appeared to compensate for insufficient coating material. For all of the compressed spheres, drug release appeared to be somewhat linear with respect to the square root of time suggesting a change in mechanism of drug release. Similar in vitro drug release profiles were observed for 10% EC coated, encapsulated and compressed spheres of 20-25 and of 30-35 mesh sizes. The reason for this similarity is not apparent based on the current data. Preliminary compression studies demonstrated no apparent correlation between the tabletting parameters examined and final drug release. 7.3. Effects of Compression of 10% EC Coated 30-35 Mesh Spheres with Excipients None of the selected excipients were able to protect the EC membrane from compression-induced damage. In all cases, in vitro drug release was much faster for the spheres tabletted with excipients as compared to those from the encapsulated and compressed spheres. The main reasons for this are believed to be the particle sizes of the substrate and diluent materials and the strength of the film itself. It is suspected that such film damage occurs regardless of the presence of any excipients. In the absence of excipients, the EC coating on the individual spheres fuse to form the matrix-like structure thus "healing" the compression-induced tears. When 128 excipients are used, they block the necessary contact between the EC membranes of the individual spheres thus preventing the formation of the non-disintegrating structure. 7.4. Effects of Ethylcellulose Coating Level Increases in coating level led to progressive decreases in in vitro drug release from encapsulated spheres. This result is expected as increases in membrane thickness often results in decreases in drug release. Similar in vitro drug release from encapsulated and compressed spheres were observed for all coating levels. For compressed spheres, increases in coating level initially led to decreases in in vitro drug release. However, similar in vitro drug release from spheres coated at 20 and 30% EC coating levels were observed. This similarity suggests a practical or effective maximum coating level above which further increases in coating material does not strongly affect drug release rate. This possibility appears to be supported by the preliminary tensile strength data. The tensile strength for the EC coated spheres suggests that tablets made of the EC coated spheres may have a different compaction mechanism than uncoated spheres. 7.5. Effects of Ethylcellulose Coating Level and Compression with Excipients Increases in coating level led to progressive decreases in in vitro drug release for all diluent materials examined. In all cases, in vitro drug release from the spheres tabletted with excipients was faster than those for the encapsulated and compressed spheres at all coating levels indicating some damage to the rate-controlling, ethylcellulose membrane. Drug release profiles for formulations containing M C C and compressible sugar were similar at the 20 and 30% EC coating levels. The nature of the diluent, i.e. its tendency to 129 deform or fracture, did not appear to have an effect on the extent of damage. This lack of effect may be related to the particle sizes of the materials used. The effect of size may be more dominant and mask any effects due to the nature of the diluent. Drug release from tablets containing pregelatinised starch was slower than release from tablets compressed with the other two diluent materials at the 20 and 30% coating levels. It is suspected that the any compression-induced tears in the membrane are filled with the diluent material. Upon contact with water, the pregelatinised starch gels and retards drug release more than the other two diluent materials. Success of the compressed formulations was not predicted by similarity of compression properties examined in the preliminary compression studies. 7.6. Proposed System Similar in vitro dissolution profiles were observed for the encapsulated and compressed, 10% EC coated, 30-35 mesh size spheres and for the 30% EC coated, 30-35 mesh size spheres tabletted with pregelatinised starch. The proposed system therefore consists of the former spheres (i.e. 10% EC coated) for the encapsulated and non-disintegrating tablet dosage forms and the latter spheres (i.e. 30% EC coated) tabletted with Starch 1500® for the disintegrating tablet dosage form. 130 8. Future Work There are several possible avenues for future work. 1) The suggested trends regarding compression discussed above (such as mechanism of tablet formation) can be confirmed with repeated experimentation and then further investigated. Also, the compression behaviour of any modified formulations should also be investigated. 2) The proposed system could be further developed to determine and evaluate the extent of influence of any possible limitations (e.g. tablet size of the non-disintegrating tablet). A major undertaking would be to investigate the drug release mechanisms for each dosage form. The information resulting from such a study may prove useful in uncovering the limitations of the system. 3) Dissolution studies could be expanded to examine the robustness of the proposed system. Such a study would include effects of process scale-up then potentially, comparisons of release behaviour for less soluble drugs from the proposed system. In addition, small modifications to the system such as a larger range of substrate sizes. Also, the methods used to prepare the components of the proposed system could be optimised (e.g. the drug layering process). 4) The effects of amount of ethylcellulose coating material present in the dosage unit could be further investigated through drug release from an ethylcellulose dispersion (wet) granulation formulation (instead of a substrate coated system). This could examine the influence of amount of coating material (which could be more effectively controlled) and also, the location of the ethylcellulose material (i.e. dispersed within each granule as opposed to on the surface). Furthermore, any possible relationship 131 between a critical ratio of ethylcellulose material to substrate material and a maximum rate of drug release may be investigated. Other possible applications for Surelease® coating dispersion could be investigated. Such a study could include alternate formulations (such as an osmotic pump system). 132 9. References Abdou, H M Dissolution, bioavailablity and bioequivalence. Mack Printing Company, Easton, PA, 1989 Altaf, SA, SW Hoag and JW Ayres. Bead compacts II: evaluation of rapidly disintegrating, nonsegregating, compressed bead formulations. Drug Dev. Ind. Pharm. 25(5): 635-642 (1999). Ansel, HC, N G Popovich and L V Allen Jr. Pharmaceutical dosage forms and drug delivery systems. Williams and Wilkins, Media, PA, 1995. Arwidsson, H , O Hjelstuen, D Ingason and C Graffner. Properties of ethylcellulose films for extended release 111: influence of process factors when using aqueous dispersions. Acta Pharm. Nord. 3(4): 223-228 (1991). Bansal, P, S Vasireddy, F Plakogiannis and D Parikh. Effect of compression on the release properties of polymer coated niacin granules. ./ Controlled Release 27: 157-163 (1993). Bates, TR and M Gibaldi. Gastrointestinal absorption of drugs in Current concepts in pharmaceutical sciences: biopharmaceutics. J Swarbrick (ed.), Lea and Febiger, Philadelphia, 1970. Batra, V, A Bhowmick, B K Behera and AR Ray. Sustained release of ferrous sulfate from polymer-coated gum arabica pellets../. Pharm. Sci. 83(5):632-635 (1994). 133 Beckert, TE, K Lehmann and PC Schmidt. Compression of enteric coated pellets to disintegrating tablets. Int. J. Pharm. 143:13-23 (1996). Bodmeier, R. Tableting of coated pellets. Eur. J. Pharm. Biopharm. 43:1-8 (1997). Bodmeier, R and O Paeratakul. Mechanical properties of dry and wet cellulosic and acrylic films prepared from aqueous colloidal polymer dispersions used in the coating of solid dosage forms. Pharm. Res. 11(6): 882-888 (1994). Chandy, T and CP Sharma. Chitosan beads and granules for oral sustained delivery of nifedipine: in vitro studies. Biomaterials 13(13): 949-952 (1992). Chiao, CSL and JC Price. Effect of compression pressure on physical properties and dissolution characteristics of disintegrating tablets of propranolol microspheres. J. Microencapsulation 11 (2): 161 -170 (1994). Colorcon Inc. Starch 1500® technical data sheets, PT-8 (1-93). Colprcon Inc., West Point, PA, 1993. Crompton and Knowles Corporation. Pharmaceutical ingredients. Crompton and Knowles Corporation, Mahwah, NJ, 1995. Dabbagh, M A , JL Ford, M H Rubinstein and JE Hogan. Effects of polymer particle size, compaction pressure and hydrophillic polymers on drug release from matrices containing ethylcellulose. Int. J. Pharm. 140: 85-95 (1996). 134 Dwivedi, SK, RJ Oates and A G Mitchell. Peak offset times as an indication of stress relaxation during tableting on a rotary tablet press. ./. Pharm. Pharmacol. 43: 673-678 (1991). Dwivedi, SK, RJ Oates and A G Mitchell. Estimation of elastic recovery, work of decompression and Young's modulus using a rotary tablet press. J. Pharm. Pharmacol. 44:459-466 (1992). Eckhart, CG and T McCorkle. Chlorpheniramine maleate in Analytical profiles of drug substances, volume 1, K. Florey (ed.). Academic Press Inc., New York, 1978. Eddington, ND, P Marroum, R Uppoor, A Hussain and L Augsburger. Development and internal validation of an in-vitro-in vivo correlation for a hydrophilic metoprolol tartrate extended release tablet formulation. Pharm. Res. 15(3): 466-473 (1998). El-Mahdi, I'M and PB Deasy. Tableting of coated ketoprofen pellets. ./. Microencapsulation 17(2): 133-144 (2000). E M Industries Inc. Sorbitol-manriitol CSD technical information and product data sheets. E M Industries Inc., 1999. Er, LS. Design and methods of analysis for manufactured tablets data. Statistical Consulting and Research Laboratory (SCARL File# 00-002-003), Dept. of Statistics, University of British Columbia, Vancouver, BC, May 4, 2000. Fenner, RT Mechanics of Solids. Blackwell Scientific Publications, Oxford, UK, 1989. 135 Ferrari, F, S. Rossi, A Martini, L Muggetti, R De Ponti and C Caramella. Technological induction of mucoadhesive properties on waxy starches by grinding. Eur. J. Pharm. Sci. 5: 277-285 (1997). FMC Corporation, Food and Pharmaceutical Products Division. Avicel® problem solver and reference manual. FMC Corporation, 1984. Frohoff-Hiilsmann, M A , A Schmitz and BC Lippold. Aqueous ethyl cellulose dispersions containing plasticizers of different water solubility and hydroxypropyl methylcellulose as coating material for diffusion pellets I: Drug release rates from coated pellets. Int. J. Pharm. Ill: 69-82 (1999). Ghebre-Sellassie, 1 Pellets: a general overview in Pharmaceutical Pelletization Technology. I. Ghebre-Sellassie (ed), Marcel Dekker Inc., New York, 1989. Ghaly, ES and NR Ruiz. Compressibility characteristics of matrices prepared with ethylcellulose aqueous dispersion. Drug Dev. Ind. Pharm. 22(2): 91-95 (1996). Gilligan, CA and A L W Po. Factors affecting drug release from a pellet system coated with an aqueous colloidal dispersion. Int. J. Pharm. 73: 51-68 (1991). Guo, J-H. Aging processes in pharmaceutical polymers. Pharm. Sci. Tech. Today 2(12):478-483 (1999). Guo, J-H, RE Robertson and GL Amidon. Influence of physical aging on mechanical properties of polymer free films: the prediction of long-term aging effects on the water 136 permeability and dissolution rate of polymer film-coated tablets. Pharm. Res. 8(12): 1500-1504 (1991). Guo, J-H, RE Robertson and GL Amidon. An investigation into the mechanical and transport properties of aqueous latex films: a new hypothesis for the film-forming mechanism of aqueous dispersion system. Pharm. Res. 10(3): 405-410 (1993). Ho, H-O, H-L Su, T Tsai and M-T Sheu. The preparation and characterization of solid dispersions on pellets using a fluidised-bed system. Int. J. Pharm. 139: 223-229 (1996). Hogan, JE Film-coating materials and their properties in Pharmaceutical coating technology, G Cole (ed), Taylor and Francis Inc., Bristol, PA, 1995. Huang, Y, R-C Hwang, GK Berlin and AD Moore. Formulation development of controlled release tablets using bio-dis and similarity factor. Proceed. Int'l Symp. Conttrol. Rel. Bioact. Mater. 25: 796-797 (1998). Hudson, H M , CR Daubert and E A Foegeding. Rheological and physical properties of derivitized whey protein isolate powders. J. Agr. Food Chem. 48(8): 3112-3119 (2000). Hutchings, DE and A Sakr. Influence of pH and plasticizers on drug release from ethylcellulose pseudolatex coated pellets. ./. Pharm. Sci. 83(10): 1386-1390 (October 1994). Hyppola, R, I Husson and F Sundholm. Evaluation of physical properties of plasticized ethylcellulose films cast from ethanol solution, part I. Int. J. Pharm. 133: 161-170 (1996). 137 Ju, H L and S-J Liaw. On the assessment of similarity of drug dissolution profiles - a simulation study. Drug Info. J. 3 1: 1273-1289 (1997). Kaewvichit, S and 1G Tucker. The release of macromolecules from fatty acid matrices: complete factorial study of factors affecting release. J. Pharm. Pharmacol. 46:708-713 (1994). Khan, G M and JB Zhu. Ibuprofen release kinetics from controlled-release tablets granulated with aqueous polymeric dispersion of ethylcellulose II: influence of several parameters and excipients. ./ Controlled Release 56: 127-134 (1998). Klinger, GH, ES Ghali, SC Porter and JB Schwartz. Formulation of controlled release matrices by granulation with a polymer dispersion. Drug Dev. Ind. Pharm. 16(9): 1473-1490 (1990). Lee, B-J, K A Parrot, JW Ayres and RL Sack. Development and characterization of an oral controlled-release delivery system for melatonin. Drug Dev. Ind. Pharm. 22(3): 269-274 (1996). Lehmann, K, H-U Petereit and D Dreher. Fast disintegrating controlled release tablets from coated particles. Drugs Made in Germany 37(2): 53-60 (1994). Li , LC and Y - H Tu. Latex film matrix systems with a concentration gradient for controlled drug delivery. Drug Dev. Ind. Pharm. 17(15): 2041-2054 (1991). 138 Li , SP, RG Felt, LC Di Paolo, M - Y Huang and RO Williams. Development and in vitro-in vivo evaluation of a multiparticulate sustained release formulation of diltiazem. Pharm. Res. 12(9): 1338-1342 (1995). Lippold, BC, W Gunder and B H Lippold. Drug release from diffusion pellets coated with the aqueous ethyl cellulose dispersion Aquacoat® ECD-30 and 20% dibutyl sebacate as plasticizer: partition mechanism and pore diffusion. Eur. J. Pharm Biopharm. 47: 27-32 (1999). Liu, J-P, M-C Ma and S-C Chow. Statistical evaluation of similarity factor f2 as a criterion for assessment of similarity between dissolution profiles. Drug Info. J. 31: 1255-1271 (1997). Lopez-Rodriguez, FJ, JJ Torrado, C Escamilla, R Cadorniga and L L Augsburger. Compression behaviour of acetylsalicylic acid pellets. Drug Dev. .Ind. Pharm. 19(2): 1369-1377 (1993). Maganti, L and M Celik. Compaction studies on pellets: I. uncoated pellets. Int. J. Pharm. 95: 29-42 (1993). Maganti, L and M Celik. Compaction studies on pellets: II. coated pellets. Int. J. Pharm. 103: 55-67 (1994). Martin, A, P Bustamante and AHC Chun. Physical pharmacy: physical chemical principles in the pharmaceutical sciences. Lea and Febiger, Philadelphia, 1993. 139 Mathir, Z M , C M Dangor, T Govender and DJ Chetty. In vitro characterization of a controlled-release chlorpheniramine maleate delivery system prepared by the air-suspension technique. J. Microencapsulation 14(6): 743-751 (1997). The Merck index, an encyclopedia of chemicals, drugs and biologicals, 12 th ed. S Budavari (ed), Merck Research Laboratories Division of Merck and Co. Inc., Whitehouse, NJ, 1996. Miller, RA Personal communication, November 1996. Miller, RA, E M K Leung and RJ Oates. The compression of spheres coated with an aqueous ethylcellulose dispersion. Drug Dev. Ind. Pharm. 25(4): 503-511 (1999). Moore, JW and H H Flanner. Mathematical comparison of dissolution profiles. Pharm. Tech. 20(6): 64-74 (1996). Moore, KL Physicochemical properties of Opadry®, Coaler ic® and Surelease® in Aqueous polymeric coatings for pharmaceutical dosage forms. JW McGinty (ed.) Marcel Dekker Inc., New York, 1989. Neau, SH, M A Howard, JS Claudius and DR Howard. The effect of the aqueous solubility of xanthine derivatives on the release mechanism from ethycellulose matrix tablets. Int. J. Pharm. 179: 97-105 (1999). Oates, RJ and RA Miller. The mechanical properties of tablets compressed from sugar spheres NF, sucrose and corn starch. Presented at AAPS 1996. 140 Oates, RJ and A G Mitchell. Calculation of punch displacement and work of powder compaction on a rotary tablet press. ./. Pharm. Pharmacol. 41: 517-523 (1989). Oates, RJ and A G Mitchell. Comparison of calculated and experimentally determined punch displacement on a rotary tablet press using both the Manesty and IPT punches. ./. Pharm. Pharmacol. 42: 388-396 (1990). O'Connor, RE and JB Schwartz. Extrusion and spheronization technology in Pharmaceutical Pelletization Technology. I Ghebre-Sellassie (ed.), Marcel Dekker Inc., New York, 1989. O'Connor, RE and TB Schwartz. Drug release mechanism from a microcrystalline cellulose pellet system. Pharm. Res. 10(3): 356-361 (1993). Opota, DO, G Joachim, G Kalantzis, P Piccerelle, JP Reynier and J Joachim. Controlled-release behavior of diphenhydramine hydrochloride loaded neutral microgranules and coated using ethylcellulose water dispersion. Drug Dev. Ind. Pharm. 25(1): 81-87 (1999). Ozturk, A G , SS Ozturk, BO Palsson, T A Wheatley and JB Dressman. Mechanism of release from pellets coated with an ethylcellulose-based film. J. Controlled. Release 14: 203-213 (1990). Palmieri, GF and P Wehrle Evaluation of ethylcellulose-coated pellets optimized using the approach of Taguchi. Drug Dev. Ind. Pharm. 23(11): 1069-1077 (1997). Parikh, NH, SC Porter and BD Rohera. Tensile properties of free films cast from aqueous ethylcellulose dispersions. Pharm. Res. 10(6): 810-518 (1993) 141 Parrott, EL. Compression in Pharmaceutical dosage forms: tablets, volume 2. H A Lieberman, L Lachman and JB Schwartz (eds.) Marcel Dekker Inc., New York, N Y , 1990. Pernarowski, M , W Woo and RO Searl. Continuous flow apparatus for the determination of the dissolution characteristics of tablets and capsules. J Pharm Sci. 57(8): 1419-1421 (1968). Pipel, N, CI Igwilo and S Malamataris. Effects of molecular coatings on the compression and tableting of some pharmaceutical powders. Int. ./. Pharm. 68: 157-166 (1991). Polli, JE, GS Rekhi, LL Augsburger and VP Shah. Methods to compare dissolution profiles and a rationale for wide dissolution specifications for metoprolol tartrate tablets. J. Pharm. Sci. 86(6): 690-700 (1997). Ragnarsson, G and MO Johansson. Coated drug cores in multiple unit preparations influence of particle size. Drug Dev. Ind. Pharm. 14(15-17): 2285-2297 (1988) Rekhi, GS, SC Porter and SS Jambhekar. Factors affecting the release of propranolol hydrochloride from beads coated with aqueous polymeric dispersions. Drug Dev. Ind. Pharm. 21(6): 709-729 (1995). Ruiz, NR and ES Ghaly. Mechanisms of drug release from matrices prepared with aqueous dispersion of ethylcellulose. Drug Dev. Ind. Pharm. 23(1): 113-1 17 (1997). 142 Sadeghi F, JL Ford, MFf Rubinstein and A R Rajabi-Siahboomi. Comparative study of drug release from pellets coated with HPMC or Surelease®. Drug Dev. Ind. Pharm. 26(6): 651-660 (2000). Sarisuta, N and K Punpreuk. In vitro properties of film-coated diltiazem hydrochloride pellets compressed into tablets. J. Controlled Release 31: 215-222 (1994). Schmidt, C and R Bodmeier. Incorporation of polymeric nanoparticles into solid dosage forms. ./. Controlled Release 57: 115-125 (1999). Shah, VP, Y Tsong, P Sathe and J-P Liu. In vitro dissolution profile comparison -statistics and analysis of the similarity factor, f2. Pharm. Res. 15(6): 889-896 (1998). Sheen, P-C, PJ Sabol, GJ Alcorn and K M Feld. Aqueous film coating studies of sustained release nicotinic acid pellets: an in-vitro evaluation. Drug Dev. Ind. Pharm. 18(8): 851-860 (1992). Shlieout, G and G Zessin. Investigation of ethycellulose as a matrix former and a new method to regard and evaluate the compaction data. Drug Dev. Ind. Pharm. 22(4): 313-319 (1996). Steuernagel, CR Latex emulsions for controlled drug delivery in Aqueous polymeric coatings for pharmaceutical dosage forms. JW McGinty (ed.) Marcel Dekker Inc., New York, 1989 Sun, Y - M , W-F Huang and C-C Chang. Spray-coated and solution-cast ethylcellulose pseudolatex membranes. J. Membrane Sci. 157: 159-170 (1999). 143 Tang, L-J and JB Schwartz. Introduction of dissolution error as a result of different openings in vessel covers. Pharm. Dev. Tech. 3(2): 261-267 (1998). Tang, L, JB Schwartz, SC Porter, RL Schnaare and RJ Wigent. Drug release from film-coated chlorpheniramine maleate nonpareil beads: effect of water soluble polymer, coating level and soluble core material. Pharm. Dev. Tech. 5(3): 383-390 (2000). Tang, L, RJ Wigent and JB Schwartz. Drug release from film-coated chlorpheniramine maleate nonpareil beads: water influx and development of a new drug release model. Pharm. Dev. Tech. 4(4): 481-490 (1999). Tang, Y and K Gan. Statistical evaluation of in vitro dissolution of different brands of ciproflaoxacin hydrochloride tablets and capsules. Drug Dev. Ind. Pharm. 24(6): 549-552 (1998) . Tho, I, P Schultz, T Waaler and P Kleinebudde. Mechanical properties of dry and wet free polymer films. Proceed. Int'l Symp. Conttrol. Rel. Bioact. Mater. 26: 549-552 (1999) . Torrado, JJ and L L Augsburger. Effect of different excipients on the tableting of coated particles. Int. ./. of Pharm. 106: 149-155 (1994). Torrado-Santiago, JJ Torrado and R Cadorniga. Effect of different excipients on release characteristics of acetylsalicylic acid from compressed pellets. Pharwazie 50: 476-478 (1995). 144 Tsai, T, Y-P San, H-0 Ho, J-S Wu and M-T Sheu. Film-forming polymer-granulated excipients as the matrix materials for controlled release dosage forms. J. of Controlled Release 51: 289-299 (1998). US Department of Health and Human Services, Food and Drug Administration, Centre for Drug Evaluation and Research. Guidance for industry: dissolution testing of immediate release solid oral dosage forms. August 1997. United States Pharmacopeia 24, <711> US Pharmacopeial Convention Inc., Rockville, MD. 1999. Upadrashta, SM, PR Katikaneni, G A Hileman, SH Neau and CE Rowlings. Compressibility and compactibility properties of ethylcellulose. Int. J. of Pharm. 112: 173-179 (1994). Valleri, M and P Mura. Factors influencing the release of aminophylline from tableted ethylcellulose microcapsules. Drug Dev. Ind. Pharm. 21(18): 2139-2146 (1995). Vanderhoff, JW, MS El-Aasser and J Ugelstad. Polymer emulsification process. United States Patent 4,177,177 December 1979. Van der Voort Maarschalk, K, H Vromans, W Groenendijk, GK Bolhuis and CF Lerk. Effect of water on deformation and bonding of pregelatinized starch compacts. Eur. J. Pharm. Biopharm. 44: 253-260 (1997). Van Wylan, GJ and RE Sonntag. Fundamentals of classical thermodynamics, SI version 2. John Wiley and Sons, New York, 1978. 145 Velasco, M V , JL Ford and AR Rajabi-Siahboomi. Use of dynamic differential scanning calorimetry in identifying glass transition temperatures of ethylcellulose and Surelease® films. Pharm. Pharmacol. Commun. 5: 259-263 (1999). Vervaet, C, L Baert and IP Remon. Extrusion-spheronization: a literature review. Int. J. of Pharm. 116: 131-146 (1995) Vilivalam, VD and C M Adeyeye. Development and evaluation of controlled-release diclofenac microspheres and tabletted microspheres. ./ of Microencapsulation 11(4): 455-470 (1994). Wade, A and PJ Weller (ed.). Handbook of pharmaceutical excipients. 2 n d edition. American Pharmaceutical Association, Washington D C , 1994. Wang, C-C, G Zhang, N H Shah, M H Infeld, A W Malick and JW McGJnty. Mechanical properties of single pellets containing acrylic polymers. Pharm. Dev. and Tech. 1(2): 213-222 (1996). Watt, PR. Tablet machine instrumentation in pharmaceutics: principles and practice. Ellis Horwood Ltd. Chichester, England, 1988. Wesseling, M , F Kuppler and R Bodmeier. Tackiness of acrylic and cellulosic polymer films used in coating of solid dosage forms. Eur. J. Pharm. and Biopharm. 47: 73-78 (1999). 146 Wheatley, TA and CR Steuernagel. Latex emulsion for controlled drug delivery in Aqueous polymeric coatings for pharmaceutical dosage forms. JW McGinty (ed.) Marcel Dekker Inc., New York, 1997. Yao, T, M Yamada, H Yamahara and M Yoshida. Tableting of coated particles I: small particle size chitosan as an agent protecting coating membrane from mechanical damage of compression force. Chem. Pharm. Bull. 45(9): 1510-1514 (1997). Yao, T, M Yamada, H Yamahara and M Yoshida. Tableting of coated particles II: influence of particle size of pharmaceutical additives on protection of coating membrane from mechanical damage during compression process. Chem. Pharm. Bull. 46(5): 826-830 (1998). Yuasa, H, T Nakano and Y Kanaya. Suppression of agglomeration in fluidized bed coating 1: suppression of agglomeration by adding NaCI. Int. J. of Pharm. 158: 195-201 (1997). Zimm, ICR, JB Schwartz and RE O'Connor. Drug release from a multiparticulate pellet system. Pharm. Dev. and Techn. 1(1): 37-42 (1996). 147 Appendix 1: Compression Study Data A1.1. Substrate Size Study 30 27 24 21 18 3 ~ 15 £ 12 CO O £" 9 9 a 80 180 280 Peak Pressure (MPa) <u 4 Q) CO o ro Q. * H * • 14-18 20-25 a 30-35 • 45-60 80 180 280 Peak Pressure (MPa) Figure A l - 1: Porosity and Peak Offset Time for 10% EC Coated Spheres of Various Mesh Size 148 30 .2 ; tn oo <D L-Q. E o O o -XL 10 80 180 280 Peak Pressure (MPa) 2 5 2.0 TO Q. O) c <D CO OJ CO c h-1 5 1.0 0.5 0 0 • 14-18 A 20-25 a 30-35 45-60 80 180 280 Peak Pressure (MPa) Figure A l - 2: Work of Compression and Tensile Strength for 10% E C Coated Spheres of Various Mesh Sizes 149 A1.2. Coating Level Study 30 0 V • o V ° 20 - • o • 10% Surelease coated & 20% Surelease coated ;ity (%) o * , v o B H 30% Surelease coated o Uncoated v Drug layered V) 2 10 W ° V / \ o a v A V a 0 i B B I . I . I V • 89 B & 9 H 0 20 40 60 80 100 120 140 160 180 200 220 240 260 Peak Pressure (MPa) Figure A l - 3: Porosity for 30-35 Mesh Size Spheres Coated with Various Coating Materials and EC Coating Levels 150 8 7 ffi & A 6 CD E _ • 10% Surelease coated A 20% Surelease coated P 5 0 V 9 30% Surelease coated "53 CO S 4 O V Uncoated Drug layered Peak M CO o s o ° A A o as V a A V 1 H • A B y a • 8 59 0 i 1 , 1 . 1 . 1 . 1 , 1 , 1 1 i , i , i , i , i 0 20 40 60 80 100 120 140 160 180 200 220 240 260 Peak Pressure (MPa) Figure A l - 4: Peak Offset Time for 30-35 Mesh Size Spheres Coated with Various Coating Materials and EC Coating Levels 151 24 22 -O) 20 C 18 O '(/> "> 16 0) Q. E 14 o O 12 O !=; 10 i o v R 10% Surelease coated 20% Surelease coated 30% Surelease coated Uncoated Drug layered 20 40 60 80 100 120 140 160 Peak Pressure (MPa) 180 200 220 240 260 Figure A l - 5: Work of Compression for 30-35 Mesh Size Spheres Coated with Various Coating Materials and EC Coating Levels 152 A1.3. Formulation and Coating Level Study Major Materials Formulations 35 25 0_ 5 h a a Avicel a Di-Pac a 88 Starch 1500 a 10% EC Coated Spheres a 30 80 130 180 230 30 20 10 Avicel formulation Di-Pac formulation Starch 1500 formulation • A a 40 90 140 190 240 Peak Pressure (MPa) Figure A l - 6: Porosity for Diluent Materials and Formulations Containing 10% EC Coated, 30-35 Mesh Size Spheres 153 Major Materials Formulations 10 9 10 01 E CD D O 5 CO Q_ 4 3 2 1 0 «• a ft B Avicel 9 • Avicel formulation & Di-Pac Di-Pac formulation I S Starch 1500 8 H Starch 1500 formulation •» 10% EC Coated Spheres 30 80 130 180 230 40 90 140 190 240 Peak Pressure (MPa) Figure A l - 7: Peak Offset Time for Diluent Materials and Formulations Containing 10% EC Coated, 30-35 Mesh Size Spheres 154 Major Materials 30 c o '(/) (/) 0) Q . 20 E o o o a Avicel A Di-Pac J S Starch 1500 «> 10% EC Coated Spheres A 9 A 10 %4 A <JB _ J L -30 80 130 180 230 Formulations 30 20 10 Avicel formulation Di-Pac formulation Starch 1500 formulation • a A A • A ^ 40 90 140 190 240 Peak Pressure (MPa) Figure A l - 8: Work of Compression for Diluent Materials and Formulations Containing 10% EC Coated, 30-35 Mesh Size Spheres 155 io r CD Major Materials • Avicel Di-Pac 81 Starch 1500 10% EC Coated Spheres s • c CD CO 5 C/3 c CD h-s a B S « s 30 80 130 180 230 Formulations, 10 r Avicel formulation Di-Pac formulation Starch 1500 formulation a -88—H-ffl-40 90 140 190 240 Peak Pressure (MPa) Figure A l - 9: Tensile Strength for Diluent Materials and Formulations Containing 10% EC Coated, 30-35 Mesh Size Spheres 156 Major Materials Formulations 35 25 -i—• o 0_ • a Avicel 0 A Di-Pac • I S Starch 1500 • » 20% EC Coated Spheres a 35 25 15 Avicel formulation Di-Pac formulation Starch 1500 formulation A , A « A a a « 30 80 130 180 230 50 100 150 200 250 Peak Pressure (MPa) Figure A l - 10: Porosity for Diluent Materials and Formulations Containing 20% EC Coated, 30-35 Mesh Size Spheres 157 Major Materials Formulations CD E i -10 9 8 7 O 5 CD fX 4 3 2 1 0 a Avicel A Di-Pac H Starch 1500 «> 20% EC Coated Spheres 88 30 80 130 180 230 10 9 8 7 6 5 4 3 2 1 0 Avicel formulation Di-Pac formulation Starch 1500 formulation 50 100 150 200 250 Peak Pressure (MPa) Figure A l - 11: Peak Offset Time for Diluent Materials and Formulations Containing 20% EC Coated, 30-35 Mesh Size Spheres 158 Major Materials 30 • Avicel Di-Pac 8 Starch 1500 * 20% EC Coated Spheres c o 'co co CD a. 20 E o O o i _ o 10 30 80 130 180 230 Formulations 30 20 h 10 Avicel formulation Di-Pac formulation Starch 1500 formulation A B • a B A S • A 50 100 150 200 250 Peak Pressure (MPa) Figure A l - 12: Work of Compression for Diluent Materials and Formulations Containing 20% EC Coated, 30-35 Mesh Size Spheres 159 10 CD c CD CO _CD c/5 c CD Major Materials a Avicel Di-Pac Starch 1500 ® 20% EC Coated Spheres a e * « a 30 80 130 180 230 10 Formulations • Avicel formulation & Di-Pac formulation s. Starch 1500 formulation • a a I 1 — B _ 50 100 150 200 250 Peak Pressure (MPa) Figure A l - 13: Tensile Strength for Diluent Materials and Formulations Containing 20% EC Coated, 30-35 Mesh Size Spheres 160 Major Materials Formulations 35 25 o o D_ 15 a a Avicel e Di-Pac • ss Starch 1500 e 0 30% EC Coated Spheres 9 A » „ 30 80 130 180 230 35 25 15 • Avicel formulation A Di-Pac formulation a Starch 1500 formulation • * B 9. SS • • 50 100 150 200 Peak Pressure (MPa) Figure A l - 14: Porosity for Diluent Materials and Formulations Containing 30% EC Coated, 30-35 Mesh Size Spheres 161 Major Materials Formulations 10 a Avicel 9 a Avicel formulation & Di-Pac A. Di-Pac formulation ss Starch 1500 3 H Starch 1500 formulation © 30% EC Coated Spheres Peak Pressure (MPa) Figure A l - 15: Peak Offset Time for Diluent Materials and Formulations Containing 30% EC Coated, 30-35 Mesh Size Spheres 162 Major Materials Formulations 30 c o co co CD Q_ 20 E o O o o 10 a Avicel Di-Pac s Starch 1500 30% EC Coated Spheres A 30 80 130 180 230 30 20 10 • Avicel formulation A Di-Pac formulation a Starch 1500 formulation • A Q e A si 9 50 100 150 200 Peak Pressure (MPa) Figure A l - 16: Work of Compression for Diluent Materials and Formulations Containing 30% EC Coated, 30-35 Mesh Size Spheres 163 10 03 0_ Major Materials e Avicel A Di-Pac H Starch 1500 «> 20% EC Coated Spheres • • CD CO • i 5 CD ^ a 30 80 130 180 230 Formulations 10 r Avicel formulation Di-Pac formulation Starch 1500 formulation A A I B i B 50 100 150 200 Peak Pressure (MPa) Figure A l - 17: Tensile Strength for Diluent Materials and Formulations Containing 30% EC Coated, 30-35 Mesh Size Spheres 164 Avicel 50 40 30 20 10 0 -10 ' A l l * * « . H A 50 100 150 200 250 50 40 30 20 10 0 -10 Di-Pac 'SB A o 50 100 150 200 250 t o O t_ o 50 40 30 20 10 0 -10 Starch 1500 • Excipient A 10% Spheres Formulation a 20% Spheres Formulation * 30% Spheres Formulation 50 100 150 200 250 Peak Pressure (MPa) Figure A1- 18: Porosity for Disintegrating Tablet Formulations 165 Avicel 10 <D E a a — 0 50 100 150 200 250 CD o CO CD D_ Starch 1500 10 Di-Pac 10 0 50 100 150 200 250 * Excipient * 10% Spheres Formulation a 20% Spheres Formulation * 30% Spheres Formulation 0 50 100 150 200 250 Peak Pressure (MPa) Figure A l - 19: Peak Offset Time for Disintegrating Tablet Formulations 166 30 20 CD .2 1 0 00 oo 0) O Avicel 50 100 150 200 250 30 20 10 Di-Pac 50 100 150 200 250 O Starch 1500 30 20 10 Excipient 10% Spheres Formulation 20% Spheres Formulation 30% Spheres Formulation 50 100 150 200 250 Peak Pressure (MPa) Figure A I- 20: Work of Compression for Disintegrating Tablet Formulations 167 Avicel Di-Pac 9 • • 3 0 7 0 0 * 5 2 1 D_ J rength o , 0 /JSP3 rength o , 50 100 150 200 250 ( 3 50 100 150 200 250 CO 0) t75 Starch 1500 EC <D 0.7 • 0 • Excipient * 10% Spheres Formulation 0.5 0 H 20% Spheres Formulation 0.3 0 * 30% Spheres Formulation 0.1 Note: Ordinate scales are different 0 50 100 150 200 250 Peak Pressure (MPa) Figure Al - 2 1 : Tensile Strength for Disintegrating Tablet Formulations 168 Appendix 2: Formulation Development Figures 30 c/> CD 60 _ C U cu o r so - • — 20-25 mesh size spheres ••A--- 30-35 mesh size spheres 10 12 14 Time (hours) 18 20 22 24 Figure A2- 1: Drug Release for 10% Surelease Coated Spheres (75%) Compressed with Avicel (24%) and Stearic Acid (1%) 169 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090285/manifest

Comment

Related Items