Pharmaceutics
`
`The Science of Dosage Form Design
`
`Edited by
`Michael E. Aulton BPharm PhD FAAPS MAPharmS
`Professor of Pharmaceutical Technology,
`School of Pharmacy,
`De Montfort University,
`Leicester, UK
`
`SECOND EDITION
`
`/,&\ CHURCHILL
`~~ LIVINGSTONE
`
`:u:
`
`EDINBURGH LONDON NEW YORK OXFORD PHILADELPHIA ST LOUIS SYDNEY TORONTO 2002
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 1 of 66
`
`

`

`CHURCHILL LIVINGSTONE
`An imprint of Elsevier Science Limited
`
`© Harcourt Publishers Limited 2002
`© Elsevier Science Limited 2002. All rights reserved.
`
`No part of this publication may be reproduced, scored in a
`retrieval system, or transmitted in any form or by any means,
`electronic, mechanical, photocopying, recording or otherwise,
`without either the prior permission of the publishers or a
`licence permitting restricted copying in the United Kingdom
`issued by the Copyright Licensing Agency, 90 Tottenham
`Court Road, Lo.ndon WIT 4LP. Permissions may be sought
`directly from Els:evier's Health Sciences Rights Department
`in Philadelphia, USA: phone: (+ I) 215 238 7869, fax: (+!)
`2 I 5 238 2239, e-mail: healtl1permissions@elsevier.com. You
`may also complete your request on-line via the Elsevier
`Science homepage (http://www.elsevier.com), by selecting
`'Customer Support' and then ' Obtaining Permissions'.
`
`First published I 988
`Second Edition ,1002
`Reprinted 2002 twice
`
`Standard edition ISBN O 443 05517 3
`
`International Student Edition ISBN O 443 05550 5
`Reprinted 2002 twice
`
`British Library Cataloguing in Publication Data
`A catalogue record for this book is available from the British
`Library
`
`Library of Cong,ress Cataloging in Publication Data
`A catalog record for this book is available from the Library of
`Congress
`
`Note
`Medical knowledge is constantly changing. As new
`information becomes available, changes in treatment,
`procedures, equipment and the use of drugs become
`necessary. The editor, contributors and the publishers have
`taken care to ensure that the information given in this text is
`accurate and up to date. However, readers are strongly
`advised to confirm. that the information, especially with
`regard to drug usage, complies with the latest legislation and
`standards of practice.
`
`,,;1w1 ~ournals and mu_ltimedia
`
`p your source for books,
`
`on t:he health sciences
`www.elsevierhec1lth.com
`
`Printed in Spain
`
`The
`put,lishe(s
`policy rs to use
`paper manulactured
`from suslalnable forests
`
`I
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 2 of 66
`
`

`

`Contents
`
`What is 'Pharmaceutics'? xiii
`
`1. The design of dosage forms
`Peter York
`
`PART ONE
`Scientific principles of dosage form
`design 13
`2. Dissolution and solubility 15
`M ichael A ulton
`
`3. Properties of solutions 33
`Michael Au/ton
`
`4. Rheology 41
`Ch,·is M arriou
`
`5. Surface and interfacial phenomena 59
`J ohn Fell
`6. Disperse systems 70
`David A uwood
`
`7. Kinetics and product stability 101
`J ohn Pugh
`
`8. Pharmaceutical preformulation 113
`James Wells
`
`PART TWO
`Particle science and powder
`technology 139
`9. Solid-state properties 141
`Graham Buckton
`
`10. Particle-size analysis 152
`John Staniforth
`
`11 . Particle-size reduction 166
`J ohn Sianiforth
`
`12. Particle-size separation 174
`John Siam/ orth
`
`13. Mixing 181
`Andrew Twitchell
`
`14. Powder flow 197
`J ohn Staniforth
`
`PART THREE
`Biopharmaceutical principles of drug
`delivery 211
`15. Introduction to biopharmaceutics 213
`M arianne A shford
`
`16. The gastrointestinal tract - physiology and
`drug absorption 217
`Marianne Ashford
`
`17. Bioavailability - physicochemical and
`dosage form factors 234
`M arianne Ashford
`
`18. Assessment of biopharmaceutical
`properties 253
`M arianne Ashford
`
`19. Dosage regimens 275
`Swart Proudfoot, (updated by John Colleu)
`20. Modified-release peroral dosage form 289
`John Co/leu, Chris M oreton
`
`PART FOUR
`Dosage form design and
`manufacture 307
`21. Solutions 309
`Michael Bil/any
`
`22. Clarification 323
`Andrew Twitchell
`
`23. Suspensions and emulsions 334
`M ichael Bil/any
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 3 of 66
`
`

`

`CONTENTS
`
`24. Powders and granules 360
`M alcolm Summers
`
`25. Granulation 364
`Malcolm Summers, Michael A ulton
`
`26. Drying 379
`Michael Aulwn
`
`27. Tablets and compaction 397
`Goran Alderbom
`
`28. Coating of tablets and
`multiparticufates 441
`J ohn Hogan
`
`29. Hard gelatin capsules 449
`Brian J ones
`
`30. Soft gelatin capsules 461
`Keith Hutchison, Josephine Ferdinando
`
`31. Pulmonary drug delivery 473
`Kevin Taylor
`
`32. Nasal drug delivery 489
`Peter Tay lor
`
`33. Transdermal drug delivery 499
`Bn·an Bany
`
`34. Rectal and vaginal drug delivery 534
`J osef Tukker
`
`35. Delivery of pharmaceutical proteins 544
`Daan Crommelin, Ewoud van Winden
`Albert M ekking
`
`36. Packs and packaging 554
`Dixie Dean
`
`37. Pharmaceutical plant design 571
`Michael Aulton, Andrew Twitchell
`
`38. Heat transfer and the properties and use
`of steam 586
`Andrew Twitchell
`
`PART FIVE
`Pharmaceutical microbiology 597
`39. Fundamentals of microbiology 599
`Geoff Hanlon
`
`40. Pharmaceutical applications of
`microbiological techniques 623
`Norman Hodges
`
`41 . The action of physical and chemical agents
`on microorganisms 643
`Geoff H anlon, Norman Hodges
`
`42. Microbiological contamination and
`preservation of pharmaceutical
`products 658
`Malcolm Parker, Norman Hodges
`Index 669
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 4 of 66
`
`

`

`8P
`
`harmaceutical preformulation: the
`physicochemical properties of drug substances
`
`James Wells
`
`CHAPTER CONTENTS
`
`The concept of preformulation 114
`
`Spectroscopy 114
`
`Solubility 115
`Aqueous solubility 115
`Intrinsic solubility (C0) 115
`pKj, from solubility data 116
`Salts 116
`Solvents 118
`Partition coefficient (K\) 119
`Solvent solubility 119
`Methodology and structure activity
`prediction 119
`Choice of non-aqueous solvent (oil) 119
`Structure-activity relationships 120
`Dissolution 122
`Intrinsic dissolution rate 122
`Measurement of intrinsic dissolution rate 123
`Common ion effect 123
`
`Melting point 124
`Techniques 124
`Capillary melting 124
`Hot-stage microscopy 124
`Differential scanning calorimetry and thermal
`analysis 124
`Polymorphism 124
`Pseudopolymorphism (solvates) 125
`True polymorphism 126
`Crystal purity 126
`Solubility 126
`
`Assay development 127
`UV spectroscopy 128
`Molecular weight 128
`pKa 128
`Thin-layer chromatography 128
`High-performance liquid chromatography
`(HPLC) 128
`
`Normal-phase HPLC 129
`Reverse-phase HPLC 129
`
`129
`
`Drug and product stability
`Temperature 130
`Order of reaction 130
`Hydrolysis 130
`The influence of pH 130
`Solvolysis 131
`Oxidation 131
`Chelating agents 131
`Photolysis 131
`Solid-state stability 132
`Hygroscopicity 132
`Stability assessment 132
`
`Microscopy 132
`Crystal morphology 133
`Particle size analysis 133
`
`Powder flow properties
`Bulk density 133
`Angle of repose 134
`
`133
`
`134
`
`Compression properties
`Plastic material 134
`Fragmentation 135
`Elastic material 135
`Punch filming (sticking) 136
`
`Excipient compatibility 136
`Method 136
`Interpretation 136
`
`Conclusions 138
`
`References 138
`
`Bibliography 138
`
`113
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 5 of 66
`
`

`

`SCIENTIFIC PRINCIPLES OF DOSAGE FORM DESIGN
`
`THE CONCEPT OF PREFORMULATION
`
`Table 8.1 Frequency distribution of dosage form
`types manufactured in the UK
`
`Frequency (%)
`
`46
`
`16
`
`15
`
`13
`
`3 3 2 1 1
`
`Dosage form
`
`Tablets
`
`Liquid oral
`Capsules
`
`Injections
`Suppositories and pessaries
`Topicals
`
`Eye preparations
`
`Aerosols (inhalation)
`
`Others
`
`Independent of this pharmaceutical profiling
`(Table 8.2), analysts will generate data (Table 8.3)
`to confirm structure and purity, and this should be
`used to complement and confirm pharmaceutical
`data. Their greater training and knowledge in
`analysis will assist in the identification of suitable
`stability-indicating assays by high-performance
`liquid chromatography (HPLC).
`
`SPECTROSCOPY
`
`The first step in preformulation is to establish a
`simple analytical method. Most drugs absorb light in
`the ultraviolet wavelengths (190-390 run) as they are
`
`Almost all new drugs are marketed as tablets, cap-
`sules or both (Table 8.1). Although only a few are
`marketed as an injection (25% of those marketed as
`tablets) the intravenous route is always required
`during early toxicity, metabolic, bioavailability and
`clinical studies to provide a precise drug and dose
`deposition. Other dosage forms may be required
`(Table 8.1) but these are drug specific and depend to
`a large extent on the successful development of
`tablets, capsules and injections.
`Prior to the development of these three major
`dosage forms, it is essential that certain fundamen-
`tal physical and chemical properties of the drug
`molecule and other derived properties of the drug
`powder are determined. This information dictates
`many of the subsequent events and approaches in
`formulation development. This first learning phase is
`known as prefor mutation.
`A recommended list of the information required
`in preformulation is shown in Table 8.2. This is
`assembled, recognizing the relative importance and
`probable existence of only limited quantities of new
`bulk drug (mg rather than g). Investigators must be
`pragmatic and generate data of immediate rele-
`vance, especially if the likely dosage forms are
`known.
`Two fundamental properties are mandatory for a
`new compound:
`1. Intrinsic solubility (C0),
`2. Dissociation constant (pKJ.
`
`Table 8.2 Preformulation drug characterization
`
`Test
`
`Method/function/characterization
`
`Spectroscopy
`Solubility
`aqueous
`pKa
`salts
`solvents
`partition coeff K^
`dissolution
`Melting point
`Assay development
`Stability (in solution and solid state)
`Microscopy
`Powder flow
`bulk density
`angle of repose
`Compression properties
`Excipient compatibility
`
`114
`
`Simple UV assay
`Phase solubility, purity
`Intrinsic solubility, pH effects
`Solubility control, salt formation
`Solubility, hygroscopicity, stability
`Vehicles, extraction
`Lipophilicity, structure activity
`Biopharmacy
`DSC - polymorphism, hydrates, solvates
`UV, TLC, HPLC
`Thermal, hydrolysis, oxidation, photolysis, metal ions, pH.
`Morphology, particle size
`Tablet and capsule formulation
`
`Tablet and capsule formation
`Excipient choice
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 6 of 66
`
`

`

`Table 8.3 Analytical preformulation
`
`Attribute Test
`
`Identity Nuclear magnetic resonance (NMR)
`Infra red spectroscopy (IR)
`Ultraviolet spectroscopy (UV)
`Thin-layer chromatography (TLC)
`Differential scanning calorimetry (DSC)
`Optical rotation, where applicable
`
`Purity
`
`Assay
`
`Quality
`
`Moisture (water and solvents)
`Inorganic elements
`Heavy metals
`Organic impurities
`Differential scanning calorimetry (DSC)
`
`Titration
`Ultraviolet spectroscopy (UV)
`High-performance liquid chromatography (HPLC)
`
`Appearance
`Odour
`Solution colour
`pH of slurry (saturated solution)
`Melting point
`
`generally aromatic and contain double bonds. The
`acidic or basic nature of the molecule can be predicted
`from functional groups (Perrin et al 1981). Using the
`UV spectrum of the drug, it is possible to choose an
`analytical wavelength (often Amax) suitable to quantify
`the amount of drug in a particular solution. Excitation
`of the molecule in solution causes a loss in light
`energy, and the net change from the intensity of the
`incident light (/0) and the transmitted light (7) can be
`measured. The amount of light absorbed by a solution
`of drug is proportional to the concentration (C) and
`the path length of the solution (/) through which the
`light has passed. Equation 8.1 is the Beer-Lambert
`law, where e is the molar extinction coefficient.
`
`In pharmacy it is usual to use the specific absorp-
`tion coefficient E\%
`cm (£]), where the pathlength
`is 1 cm and the solution concentration is 1% w/v
`(10 mg mLr1), as doses of drugs and concentrations
`are generally in unit weights rather than molarity
`(E\ = 10e/MW).
`
`SOLUBILITY
`
`Aqueous solubility
`The availability of a drug is always limited and the
`preformulation scientist may only have 50 mg. As the
`
`PHARMACEUTICAL PREFORMULATION
`
`compound is new the quality is invariably poor, so
`that a large number of impurities may be present and
`often the first crystals come down as a metastable
`polymorph. Accordingly, as a minimum, the solubil-
`ity and pKa must be determined. Solubility dictates
`the ease with which formulations for oral gavage and
`intravenous
`injection studies
`in animals are
`obtained. The pKa allows the informed use of pH to
`maintain solubility and to choose salts required to
`achieve good bioavailability from the solid state
`(Chapter 9) and improve stability (Chapter 7) and
`powder properties (Chapter 13 and 14).
`Kaplan (1972) suggested that unless a compound
`has an aqueous solubility in excess of 1% (10 mg
`mLr1) over the pH range 1-7 at 37°C, potential
`bioabsorption problems may occur. If the intrinsic
`dissolution rate was greater than 1 mg cnr2 mkr1
`then absorption was unimpeded. Dissolution rates
`less than 0.1 mg cm~2 mkr1 were likely to give dis-
`solution rate-limited absorption. This tenfold differ-
`ence in dissolution rate translates to a lower limit for
`solubility of 1 mg mL'1. Under sink conditions, dis-
`solution rate and solubilities are proportional.
`A solubility of less than 1 mg mL"1 indicates the
`need for a salt, particularly if the drug will be
`formulated as a tablet or capsule. In the range
`1-10 mg mLr1 serious consideration should be given
`to salt formation. When the solubility of the drug
`cannot be manipulated in this way (neutral mole-
`cules, glycosides, steroids, alcohols, or where the pKa
`is less than 3 for a base or greater than 10 for an
`acid) then liquid filling in soft or hard gelatin cap-
`sules may be necessary.
`
`Intrinsic solubility (C0)
`to
`An increase in solubility in acid compared
`aqueous solubility suggests a weak base, and an
`increase in alkali a weak acid. In both cases a disso-
`ciation constant (pKJ can be measured and salts
`should form. An increase in acidic and alkaline solu-
`bility suggests either amphoteric or zwitterion
`behaviour. In this case there will be two pKas, one
`acidic and one basic. No change in solubility sug-
`gests a non-ionizable neutral molecule with no mea-
`surable pKa, and solubility manipulation will require
`either solvents or complexation.
`When the purity of the drug sample can be
`assured, the solubility value obtained in acid for a
`weak acid or alkali for a weak base can be assumed
`to be the intrinsic solubility (C0), ie. the fundemen-
`tal solubility when completely unionized. The solu-
`bility
`should
`ideally be measured
`at
`two
`temperatures:
`
`115
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 7 of 66
`
`

`

`SCIENTIFIC PRINCIPLES OF DOSAGE FORM DESIGN
`
`1. 4°C to ensure physical stability and extend
`short-term storage and chemical stability until
`more definitive data are available. The maximum
`density of water occurs at 4°C.This leads to a
`minimum aqueous solubility.
`2. 37°C to support biopharmaceutical evaluation.
`
`However, as absolute purity is often in doubt it is
`more accurate to determine this crucial solubility by
`the use of a phase-solubility diagram (Fig. 8.1). The
`data are obtained from a series of experiments in
`which the ratio of the amount of drug to the amount
`of dissolving solvent is varied.
`Any deviation from the horizontal is indicative of
`impurities, which a higher drug loading and its
`inherent impurities either promotes or suppresses
`solubility. In cases where the observed result changes
`with the amount of solvent, the line is extrapolated to
`zero phase ratio, where solubility will be independent
`of solvent level and the true intrinsic solubility of the
`drug. The United States Pharmacopoeia uses this
`method to estimate the purity of mecamylamine
`hydrochloride.
`
`pKa from solubility data
`Seventy-five per cent of all drugs are weak bases;
`20% are weak acids and only 5% are non-ionic,
`amphoteric or alcohols. It is therefore appropriate to
`consider the Henderson-Hasselbalch equations for
`weak bases and acids.
`
`For weak bases:
`
`and for weak acids:
`
`Equations 8.2 and 8.3 can be used:
`1. to determine pKa by following changes in
`solubility
`2. to predict solubility at any pH, provided that the
`intrinsic solubility (C0) and pKa are known
`3. to facilitate the selection of suitable salt-forming
`compounds and predict the solubility and pH
`properties of the salts.
`
`Albert and Serjeant (1984) give a detailed account of
`how to obtain precise pKa values by potentiometry,
`spectroscopy and conductivity.
`
`Salts
`
`A major improvement in solubility can be achieved
`by forming a salt. Acceptable pharmaceutical salt
`counter-ions are shown in Table 8.4. As an example,
`the consequence of changing chlordiazepoxide to
`various salt forms is shown in Table 8.5.
`In some cases, salts prepared from strong acids or
`bases are freely soluble but very hygroscopic. This
`does lead to instability in tablet or capsule formula-
`tions, as some drug will dissolve in its own adsorbed
`films of moisture. It is often better to use a weaker
`acid or base to form the salt, provided any solubility
`requirements are met. A less soluble salt will gener-
`
`Fig. 8.1 Effect of drug: solvent ratio on solubility when the drug is impure. Assuming the compound is a base and the estimate of its
`solubility in 0.1 M NaOH was 1 mg mLr1, four solutions of 3 ml_ should be set up containing 3, 6, 12 and 24 mg of drug. These give the
`phase ratios shown here. 3 ml is the smallest volume that can be manipulated for either centrifugation or filtration and dilution of UV
`analysis. The vials should be agitated continuously overnight and then the concentration in solution determined.
`
`116
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 8 of 66
`
`

`

`PHARMACEUTICAL PREFORMULATION
`
`Table 8.4 Potential pharmaceutical salts
`
`Basic drugs
`
`Anion
`
`Hydrochloride
`
`Sulphate
`
`Mesylate
`
`Maleate
`
`Phosphate
`
`Salicylate
`
`Tartrate
`
`Lactate
`
`Citrate
`
`Succinate
`
`Acetate
`
`Others
`
`pKa
`
`-6.10
`
`-3.00, +1.96
`
`-1.20
`
`1.92,6.23
`
`2.15, 7.20, 12.38
`
`3.00
`
`3.00
`
`3.10
`
`3.13, 4.76, 6.40
`
`4.21,5.64
`
`4.76
`
`Acidic drugs
`
`% Usage
`
`Cation
`
`Potassium
`
`Sodium
`
`Lithium
`
`Calcium
`
`Magnesium
`
`Diethanolamine
`
`Zinc
`
`Choline
`
`Aluminium
`
`43.0
`
`7.5
`
`2.0
`
`3.0
`
`3.2
`
`0.9
`
`3.5
`
`0.8
`
`3.0
`
`0.4
`
`1.3
`
`31.4
`
`Others
`
`PKa
`
`16.00
`
`14.77
`
`13.82
`
`12.90
`
`11.42
`
`9.65
`
`8.96
`
`8.90
`
`5.00
`
`% Usage
`
`10.8
`
`62.0
`
`1.6
`
`10.5
`
`1.3
`
`1.0
`
`3.0
`
`0.3
`
`0.7
`
`8.8
`
`Table 8.5 Theoretical solubility and pH of salts of
`chlordiazepoxide
`
`pKg
`
`SaltpH
`
`Solubility (mg mL-1)
`
`Salt
`
`Base
`
`4.80 8.30
`
`Hydrochloride
`
`-6.10
`
`2.53
`
`Maleate
`
`Tartrate
`
`Benzoate
`Acetate b
`
`1.92 3.36
`
`3.00 3.90
`
`4.20 4.50
`
`4.76 4.78
`
`2.0
`<165a
`57.1
`
`17.9
`
`6.0
`
`4.1
`
`a Maximum solubility of chlordiazepoxide hydrochloride,
`achieved at pH 2.89, is governed by crystal lattice energy
`and common ions.
`b Chlordiazepoxide acetate may not form; pKa of acetate
`too high and too close to that of drug ion.
`
`ally be less hygroscopic and form less acidic or basic
`solutions (Table 8.5). Injections should ideally lie in
`the pH range 3-9 to prevent vessel or tissue damage
`and pain at the injection site. Oral syrups should not
`be too acidic, to enhance palatability. Packaging may
`also be susceptible: undue alkalinity will attack glass,
`and hydrochloride salts should not be used in
`aerosol cans as a propellant-acid
`reaction will
`corrode the canister.
`From Table 8.5, not only does the intrinsic pH of
`the base solution fall significantly if salt forms are
`produced but, as a consequence,
`the solubility
`
`increases exponentially (Eqns 8.2 and 8.3). This has
`important implications in vivo. A weak base with an
`intrinsic solubility greater than 1 mg mL"1 will be
`freely soluble in the gastrointestinal tract, especially
`in the stomach. However, it is usually better to for-
`mulate with a salt, as it will control the pH of the dif-
`fusion layer (the saturated solution immediately
`adjacent to the dissolving surface, known as the pH
`microenviroment). For example, although chlor-
`diazepoxide base (Cs = 2 mg mL^1 at pHsat 8.3)
`meets the requirements for in vivo 'solubility'
`(Kaplan, 1972); commercial capsules contain chlor-
`diazepoxide hydrochloride (Cs = 1 65 mg roLr1 at
`pHsat 2.53).
`A weak base will have a high dissolution rate in the
`stomach, but as it moves down the gastrointestinal
`tract the pH rises and dissolution rate falls.
`Conversely, a weak acid has minimal dissolution in
`the stomach but becomes more soluble and dissolu-
`tion rate increases down the gut. Paradoxically, as
`dissolution rate increases so absorption falls because
`the drug is ionized.
`The dissolution rate of a particular salt is usually
`much greater than that of the parent drug. Sodium
`and potassium salts of weak acids dissolve much
`more rapidly than do the parent acids, and some
`comparative data are shown in Table 8.6. On the
`basis of bulk pH these salts would be expected to
`have
`lower dissolution rates
`in the stomach.
`However, the pH of the diffusion layer (found by
`measuring the pH of a saturated bulk solution) is
`
`117
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 9 of 66
`
`

`

`SCIENTIFIC PRINCIPLES OF DOSAGE FORM DESIGN
`
`higher than that of gastric fluid (which is approxi-
`mately 1.5) because of its buffering action. The pH
`is the saturated unbuffered aqueous solution (calcu-
`lated pH in Table 8.6) and the dissolution rate is gov-
`erned by this pH and not the bulk medium pH.
`In the intestine the salt does not depress the pH,
`unlike the acid which is neutralized, and the diffusion
`layer pH is again raised to promote dissolution.
`Providing that the acid forming the salt is strong, the
`pH of the solution adjacent to the dissolving surface
`will be that of the salt, whereas for the dissolving free
`base it will be the pH of the bulk dissolving medium.
`With weak bases, their salts dissolve rapidly in the
`stomach but there is no absorption, as the drug is
`ionized and absorption is delayed until the intestine.
`Any undissolved drug, as salt, rapidly dissolves, as the
`higher diffusion layer pH compensates for the higher
`bulk pH, which would be extremely unfavourable to
`the free base. Data for chlordiazepoxide are shown in
`Table 8.5. The maleate salt has a predicted solubility
`of 57 mg mLr1 but, more importantly, reduces the
`pH by 5 units. By controlling diffusion layer pH the
`dissolution rate can increase manyfold, indepen-
`dently of its position in the gastrointestinal tract. This
`is particularly important in the development of con-
`trolled-release products.
`Different salts of a drug rarely change pharmacol-
`ogy, but only physical properties. This statement has
`been qualified to acknowledge that salts do affect the
`intensity of response. However, the salt form does
`change the physiochemical properties of the drug.
`Changes in dissolution rate and solubility affect the
`rate and extent of absorption (bioavailability), and
`changes on hygroscopicity and stability influence
`formulation.
`Consequently each new drug candidate has to be
`examined to choose the most suitable salt, because
`
`each potential salt will behave differently and require
`separate preformulation screening. The regulatory
`authorities also treat each salt as a different chemical
`entity, particularly in the context of toxicity testing.
`
`Solvents
`It is generally necessary to formulate an injection
`even if there is no intention to market. The first-
`choice solvent is obviously water. However, although
`the drug may be freely soluble, it may be unstable in
`aqueous solution. Chlordiazepoxide HC1 is such an
`example. Accordingly, water-miscible solvents are
`used:
`
`1. in formulations to improve solubility or stability
`2. in analysis to facilitate extraction and separation
`(e.g. chromatography).
`Oils are used in emulsions, topicals (creams and
`ointments), intramuscular injections and liquid-fill
`oral preparations (soft and hard gelatin capsules)
`when aqueous pH and solvent solubility and stabil-
`ity are unattainable Table 8.7 shows a range of sol-
`vents to fulfil these needs.
`Aqueous methanol is widely used in HPLC and is
`the standard solvent in sample extraction during
`analysis and stability testing. It is often made acidic
`or alkaline to increase solvent power and ensure con-
`sistent ionic conditions for UV analysis. Other phar-
`maceutical solvents are available but are generally
`only required in special cases. The most acceptable
`non-aqueous solvents pharmaceutically are glycerol,
`propylene glycol and ethanol. Generally for a
`lipophilic drug (i.e. a partition coefficient (log P >1),
`solubility doubles through this series.
`Where bulk is limited and the aqueous solubility is
`inadequate, it is better to measure the solubility in
`
`Table 8.6 Dissolution rates of weak acids and their sodium salts
`
`Drug
`
`pKa
`
`pH (at Cs)
`
`Dissolution rate (mg cm-2 min-1) x 102
`
`Dissolution media
`
`0.1 MHCI(pH 1.5)
`
`Phosphate (pH 6.8)
`
`Salicylic acid
`Sodium salicylate
`
`Benzoic acid
`Sodium benzoate
`
`Sulphathiazole
`
`Sodium sulphathiazole
`
`118
`
`3.0
`
`4.2
`
`7.3
`
`2.40
`
`8.78
`
`2.88
`
`9.35
`
`4.97
`
`10.75
`
`1.7
`
`1870
`
`2.1
`980
`
`<0.1
`550
`
`27
`
`2500
`
`14
`
`1770
`
`0.5
`810
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 10 of 66
`
`

`

`Table 8.7 Recommended solvents for preformulation screening
`
`PHARMACEUTICAL PREFORMULATION
`
`Solvent
`
`Water
`
`Methanol
`0.1 MHCI(pH 1.1)
`
`0.1 MNaOH(pH 13.1)
`Buffer (pH 6-7)
`
`Ethanol
`
`Propylene glycol
`
`Glycerol
`
`PEG 300 or 400
`
`Dielectric constant (e)
`
`Solubility parameter (8)
`
`Application
`
`80
`
`32
`
`24
`
`32
`
`43
`
`35
`
`24.4
`
`14.7
`
`12.7
`
`12.6
`
`16.5
`
`All
`
`Extraction, separation
`Dissolution (gastric), basic extraction
`
`Acidic extraction
`
`Dissolution (intestinal)
`Formulation
`
`approaches do not allow easy estimates for the
`behaviour of crystalline solids. For a wide range of
`drugs it is possible to relate solvent solubility and the
`partition coefficient (log K^ = log P). Yalkowsky and
`Roseman (1981) derived the following expression
`for 48 drugs in propylene glycol:
`
`Equation 8.4 can be applied more generally by intro-
`ducing a factor </> to account for the relative solvent
`power of pharmaceutical solvents (see Table 8.8 for
`examples).
`For a wide range of solvents Eqn 8.4 now
`becomes:
`
`Methodology and structure activity prediction
`Choice of non-aqueous solvent (oil) The oilrwater
`partition (.K^) is a measure of the relative lipophilic-
`ity (oil-loving) nature of a compound, usually in the
`unionized state (HA or B), between an aqueous
`phase and an immiscible lipophilic solvent or oil.
`
`Table 8,8 Solvent power (<f») of some pharmaceutical
`solvents
`
`Solvent
`
`Relative solvent power (4>)
`
`0.5
`
`1 1 2 4
`
`Glycerol
`Propylene glycol
`
`PEG 300 or 400
`
`Ethanol
`
`DNA, DMF
`
`119
`
`aqueous solvent mixtures rather than in a pure
`organic solvent. Whereas solubilities at other levels
`and their mixtures can be predicted, the solubility in
`pure solvent is often inconsistent because of cosol-
`vent effects. Furthermore, formulations rarely use
`pure non-aqueous solvent, particularly injections.
`For example, ethanol should only be used up to 10%
`in an injection to prevent haemolysis and pain at the
`injection site, and include isotonic salts.
`
`Partition coefficient (K°)
`Partition coefficient (the solvent:water quotient of
`drug distribution) has a number of applications
`which are relevant to preformulation:
`1. Solubility: both aqueous and in mixed solvents
`2. Drug absorption in vivo: applied to a
`homologous series for structure activity
`relationships (SAR)
`3. Partition chromatography: choice of column
`(HPLC) or plate (TLC) and choice of mobile
`phase (eluant).
`
`Solvent solubility
`The relative polarities of solvents can be scaled using
`dielectric constant (e), solubility parameter (8),
`interfacial (y) and hydrophilic-lipophilic balance
`(HLB). The best solvent in any given application is
`one whose polarity matches that of the solute; an
`ideal, fully compatible solution exists when Ssolvent =
`Solute- This can be ascertained by determining solu-
`bility maxima, using a substituent contribution
`approach or the dielectric requirement of the system.
`The most useful scale of polarity for a solute is K%,
`(oil:water partition coefficient), as
`the other
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1061, p. 11 of 66
`
`

`

`SCIENTIFIC PRINCIPLES OF DOSAGE FORM DESIGN
`
`Many partition solvents have been used. The largest
`database has been generated using n-octanol. The
`solubility parameter of octanol (8 = 10.24) lies
`midway in the range for drugs (8-12), although
`some non-polar (8 < 7) and polar drugs (8 > 13) are
`encountered. This allows measurable results between
`equal volumes of oil and aqueous phases.
`In the shake flask method the drug, dissolved in
`one of the phases, is shaken with the other partition-
`ing solvent for 30 minutes, allowed to stand for
`5 minutes, and then the majority of the lower
`aqueous phase (density of octanol = 0.8258 g mL L)
`is run off and centrifuged for 60 minutes at
`2000 rpm. The aqueous phase is assayed before (2G)
`and after partitioning (Cw) [the aqueous concentra-
`tion] to give K^ = CSC - CW)/(CJ.
`If the transfer of solute to the oil phase is small,
`ZiCw is small, and any analytical error increases
`error in the estimate of K°,. Indeed, to encourage
`greater aqueous loss (>ACW~) a considerably more
`polar solvent, w-butanol, has been used. Where the
`partition coefficient is high, it is usual to reduce the
`ratio of the oil phase from 1:1 to 1:4 or 1:9 in order
`to increase the aqueous concentration (Cw) to a
`measurable level.
`For a 1:9 oil:water ratio K* = (10 £C - CW)/CW).
`The partition of a polar solute between an inert
`non-polar hydrocarbon e.g. hexane, and water is quite
`different from that of hydrogen bonding solvents such
`as octanol. The behaviour of the weak acid phenol
`(pKa = 1 0) and weak base nicotine (pKa = 3.1) is
`worthy of note. For phenol, K°aan°l = 29.5, whereas
`^hexane = Q.ll. The acidic solvent chloroform sup-
`presses partition (K° = 2.239), whereas ethyl acetate
`and diethyl ether are more polar. The basic behaviour
`of the solvents give higher K%, values. With solvents
`capable of both hydrogen donation and acceptance
`(octanol, nitrobenzene and oleyl alcohol), K%, is inter-
`mediate. For nicotine the behaviour is reversed, and
`the hydrogen donor (acidic) solvent chloroform parti-
`tions most strongly K%, - 77.63), even though the
`neutral solvent nitrobenzene, which is marginally
`more lipophilic (logP= 1.87 against 1.96 for chloro-
`form), gives similar values for both phenol and nico-
`tine. Clearly both solute and solvent characteristics
`are important.
`In general, polar solvents are advocated to corre-
`late biological activity with physicochemical proper-
`ties. Solvents less polar than octanol, measured by
`water solvency, have been termed hyperdiscriminat-
`ing, whereas more polar solvents such as butanols
`and pentanols, are hypodiscriminating. This concept
`refers to the discriminating power of a partitioning
`solvent within a homologous series. With w-butanol
`
`120
`
`the values of log P tend to be close, whereas with
`heptane and other inert hydrocarbons the differences
`in solute lipophilicities are exaggerated. w-Octanol
`generally gives a range consistent with other physico-
`chemical properties when compared to drug absorp-
`tion in the GI tract. Hyperdiscriminating solvents
`reflect more closely the
`transport across
`the
`blood-brain barrier, whereas hypodiscriminating sol-
`vents give values consistent with buccal absorption
`(Fig. 8.2). In rationalizing the effects of different par-
`titioning solvents, a good correlation was found to
`exist between the solvent water content at saturation
`and solvent lipophilicity.
`Certainly it is imperative to standardize on
`methodology, especially for the solvent. Where solu-
`bility constraints allow, this should be w-octanol,
`especially as the existing data bank is extensive.
`Structure-activity relationships Since the pioneer-
`ing work of Meyer and Overton numerous studies
`on correlating molecular structure and biological
`activity have been reported. These structure-activity
`relationships (SAR) can rationalize drug activity
`and, particularly in modern medicinal chemistry,
`facilitate a scientific approach to the design of more
`effective, elegant structural analogues.
`The application of SAR depends on a sound
`knowledge of the physicochemical properties of each
`new drug candidate in a therapeutic class, and
`preformulation is an essential information source.
`It is assumed in SAR that:
`
`This relationship holds for all polar and semipolar sol-
`vents, but with non-polar solvents (hexane to iso-
`octane) correlations are poor, and this seems to be
`related to water content. Given the importance of
`water, it is imperative that the octanol is saturated
`with the aqueous phase and the aqueous phase with
`octanol prior to any determination, otherwise the par-
`ti