This material may be protected by Copyright law (Title 17 U.S. Code)
`
`1
`
`International Journal of Pharmaceutics, 33 (1986) 201-217
`Elsevier
`
`UP 01121
`
`201
`
`Salt selection for basic drugs
`
`Philip L. Gould
`Pharmaceutical Research and Development Department, Pfizer Central Research, Sandwich, Kent (UK)
`
`(Received 24 March 1986)
`(Accepted 30 May 1986)
`
`Key words: Salt form selection — Pharmaceutical salts
`
`
`Summary
`
`An attempt has been made using a Kepner-Tregoe decision analysis approach to provide rationale to salt selection for basic drugs.
`The selection objectives are reviewed in terms of -the ‘essential’ (MUSTS) and ‘desirable’ (WANTS) issues. The desired characteristics
`of the salt form, given sufficient strength and toxicological suitability of the conjugate acid, are then discussed on the basis of the
`various pivotal physicochemical properties; melting point,,aqueous solubility and dissolution rate, stability and _hydrophobicity.
`Several trends are established which can then assist the decision of which range of salt forms to evaluate to overcome a particular
`problem with a basic drug. It is concluded that it is important to view the choice of salt form for development as a compromise, with
`particular focus on the correctly weighted desires to obtain the best balanced choice.
`
`Introduction
`
`Salt formation provides a means of altering the
`physicochemical and resultant biological char-
`acteristics of a drug without modifying its chem-
`ical structure. The importance of choosing the
`‘correct’ salt form of a drug is well outlined in a
`published review (Berge et al., 1977) but, although
`salt form can have a dramatic influence on the
`
`overall properties of a drug, the selection of the
`salt form that exhibits the desired combination
`
`of properties remains a difficult semi-empirical
`choice.
`
`In making the selection of a range of potential
`salts, a chemical process group considers issues on
`the basis of yield, rate and quality of the crystalli-
`sation as well as cost and availability of the con-
`
`jugate acid. The formulation and analytical groups
`are, on the other hand, concerned with the hygro-
`scopicity, stability, solubility and processability
`profile of the salt form, while the drug metabolism
`group is concerned with the pharmacokinetic
`aspects and the safety evaluation group on the
`toxicological effects of chronic and acute dosing
`of the drug and its conjugate acid. Thus, a clear
`comprornise of properties for the salt form is
`required, but the difficulty remains of assessing
`which salt forms are best to screen for a particular
`drug candidate.
`A
`Little,
`if any,
`literature has been devoted (to
`discussing the compromise of properties for salt
`form selection. This review addresses the problem
`of salt form selection for basic drugs.
`'
`
`‘ Correspondence: P.L. Gould, Pharmaceutical Group, Product
`Research and Development Laboratories, Cyanarnid of Great
`Britain Limited, Gosport, Hants, U.K.
`
`Approach to the salt selection process
`
`Walking and Appino (1973) have used the
`Kepner-Tregoe (KT)
`techniques
`(Kepner and
`
`0378-5173/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
`
`Mylan Exhibit 1028, Page 1
`
`Mylan Exhibit 1028, Page 1
`
`

`
`TABLE 1
`
`FDA-APPROVED COMMERCIALLY MARKETED SALTS
`
`Anion
`Acetate
`
`Percent “
`1.26 -
`
`Anion
`Iodide
`
`Percent “
`2.02
`
`Benzenesulfonate
`Benzoate
`.
`Bicarbonate
`Bitartrate
`Bromide
`Calcium edetate
`
`‘
`
`0.25
`0.51
`0.13
`0.63
`4.68
`0.25 -
`
`Isothionate i
`Lactate
`Lactobionate
`Malate
`Maleate
`Mandelate
`
`0.88
`0.76
`0.13
`0.13
`3.03
`0.38
`
`Camsylate b ,
`Carbonate
`Chloride
`Citrate
`Dihydrochloride
`Edetate
`‘
`Edisylate °
`Estolate d
`-
`I
`Esylate °
`Fumarate
`
`A“
`
`;
`
`‘
`
`‘
`
`0.25
`0.38
`4.17
`_ 3.03
`0.51
`' 0.25
`0.38
`0.13" 7
`
`0.13
`0.25
`
`i
`
`"V
`
`'
`
`0.18
`Gluceptate I
`~ 0.51
`Gluconate
`0.25
`Glutamate
`Glycollylarsnilate 5 0.13
`Hexylresorcinate
`0.13
`Hydrabamine h
`0.25
`Hydrobromide
`1.90
`Hydrochloride
`42.98
`Hyd.roxynaph-
`thoate
`
`2.02
`0.76
`0.38
`0.88
`0.13
`0.25
`0.64
`1.01‘
`
`0.25
`3.16
`
`Mesylate
`Methylbrornide
`Methylnitrate
`Methylsulfate
`Mucate '
`Napsylate
`. Nitrate
`Pa.mo’a'te" ‘
`(Embonate)
`Pantothenate
`Phosphate/
`' diphosphate
`. Polygalacturonate 0.13
`V Salicylate
`0.88
`Stearate '
`0.25
`Subacetate
`0.38
`Succinate
`0.38
`Sulfate
`7.46
`Tannate
`0.88
`Tartrate
`3.54
`
`'7
`
`_
`
`'
`
`Teoclate J‘
`Triethiodide
`
`0.13
`0.13
`
`0.25
`
`'
`
`
`
`Cation .
`
`Percent “
`
`Cation
`
`Percent 3
`
`Metallic:
`Organic:
`0.66
`Aluminium _
`0.66
`Benzathinek
`10.49
`Calcium _
`‘0.33
`Chloroprocaine‘
`1.64
`Lithium
`0.33
`Choline
`1.31
`Magnesium
`0.98
`Diethanolamine
`10.82
`Potassium
`0.66
`Ethylenedianune
`61.97
`,
`Sodium
`..
`2.29
`Megluminel
`
`
`
`0.66 ZincProcaine 2.95
`
`“ Percent is based on total number of anionic or cationic salts
`
`in use through 1974. b Camphorsulfonate. °1,2-Etha.nedisu.l-
`fonate.
`’d Laurylsulfate.
`"‘ Ethanesulfonate.
`I Glucoheptonate.
`5 p-Glycollamidophenylarsonate.
`h N,N’-Di(dehydroabietyl)
`et_hy1enedia.mine.A i 231-Iydroxyethanesulfonate.
`J 8-Chlorot.h_eo—
`phyllinate. 1‘ N,N’-Dibenzylethylenediamine. 1N-Met.hylgluca-
`mine.
`I
`I
`'
`
`’
`
`Reproduced from .Berge ‘et al. (1977) with permission of the
`copyright owner (.7. Phgrm. Scz'.).
`’
`
`Mylan Exhibit 1028, Page 2
`
`202
`
`Tregoe, 1976) of decision analysis and potential
`problem analysis to aid the selection of a salt
`form. Although their application is more exem-
`plary of the KT method rather than of the specific
`application, the rational process decision analysis
`approach which ‘defines essential and desirable
`attributes as.‘MUSTS’ and ‘WANTS’,
`respec-
`tively, provides a route to initially address the
`problem of,salt form selection.
`
`“G0”/ “N0-G0” issues
`The major “GO”/“NO-GO” (MUSTS) issue
`for salt selection of an ionizable drug is the con-
`sideration of the relative basicity of the drug and
`the relative strength of the conjugate acid. Clearly
`to form a salt the pK,, of the conjugate acid has to I _
`be less than or equal to the pKa of the basic centre
`of the drug.
`.
`A
`.
`.
`Thus the potentialrange of salts of drugs ‘con-
`taining for example triazoyl bases (I; pKa ~ 2) is
`restricted to strong acids (mineral and sulphonic,
`but excluding the carboxylic), whereas imidazole
`bases (II; pKa 6-7) are far less restricted and the
`greatest scope for salt formation occurs for the
`aliphatic tertiary ‘amines (III; pKa 9—10).
`
`‘
`I\}~{\I
`( )[pK.21
`N
`H
`(1)
`
`CH,
`I
`bi \ I
`[pK.61tcH3—N tpK.9i
`N
`I
`CH,
`H
`(H)
`(III) ‘
`
`’
`
`0
`
`The relative acid/base strength of the resultant
`salts also dictates their stability to disproportiona-
`tion, since all salts will be acid and therefore
`potentially reactive towards basic formulation ad-
`ditives..
`__ _ T
`The other essential selection issue for a salt
`
`form is the relative toxicity of the conjugate an-
`ion; some salts clearly fall into a desirable cate-
`gory, some acceptable but
`less desirable ,(both
`“GO”) and some undesirable (“NO GO”). A ta-
`ble of salts used in pharmaceutical products
`marketed in the U.S. up to 1974 is given in Table
`1. It would seem sensible that any acid relating to
`normal metabolism, or present in food and drink
`can be regarded as a suitable candidate for prepar-
`ing salts. Clearly anions that cause irritancy to the
`
`Mylan Exhibit 1028, Page 2
`
`

`
`GI tract should be avoided for some types of drug,
`e.g. anti-inflammatories, laxative surfactant anions
`for anti-secretory drugs and conjugate anions with
`intrinsic toxicity, e.g. oxalatef
`
`Properties desired of the salt form (WANTS)
`Therdesires or ‘WANTS’ of a salt form are
`
`dictated by the nature of the required dosage
`forms,_ their process and desired biological perfor-
`mance. Thus, it is somewhat difficult to provide a
`complete overall specification of ‘WANTS’ for a
`series of salt forms, but
`ideally the bulk salt
`should be completely chemically stable, non—hy-
`groscopic, not cause processing problems, and dis-
`solve quickly from solid‘-’dos‘age forms.
`Because of simple availability
`physiological
`reasons,-the monoproticihydrochlorides have been
`by far the most frequent (f4'40%)‘choice of the
`available anionic salt-forming: species. Thus, there.
`is clear precedent, and an overwhelming argument
`on many grounds to immediately progress to the
`hydrochloride salt -and evaluate other forms only
`if problems with the hydrochloride emerge.
`
`Prepare the hydrochloride; pros andcons
`Kramer and Flynn (1972) suggest that the‘solu-
`bility of an amine hydrochloride generally sets the
`maximum obtainable» concentration _for a given
`amine.
`:
`'
`Many reports (Miyazal<_i et al., 1980,1981) have
`shown that hydrochloride-"shalt formation does not"
`necessarily enhance the solubility "of poorly solu-
`ble basic drugs and result in improved bioavaila-
`bility. This finding is based on the common ion
`effect of chloride on the solubilityproduct equi-
`libriumz
`1
`-
`
`. —
`— Ksp
`BH+C1(s) .= BH;“q + Claq
`
`Hydrochloride salts therefore, have the potential
`to exhibit a reduced dissolution rate in gastric
`fluid because of the abundance of chloride ion
`
`(0.1—O.15 M). Indeed, the Setschenowsalting-out
`constants (k) for chloride are greatest for drugs of
`very low solubility (Fig. 1), and can decrease the
`dissolution rate of the salt to below that of the
`. free base form (Migazaki et al., 1980), which shows
`
`.
`
`' 203
`
`
`
`0.4
`
`0.8
`
`1.2
`
`0.4
`1.6
`LOG S, mglml
`
`0.8
`
`1.2
`
`1.6
`
`Fig. 1. Relationship between solubility in water and saltmlg-out
`constant -at 25°C (left)
`and 37°C (right). Key:
`‘A =
`phenazopyridine; B = cyproheptadine; C = bromhexirie; D =
`trihexyphenidyl; E = isoxsuprine; F = chlortetracycline; G =
`rnethacycline; H = papaverine; and I = demeclocyline.
`
`Adapted from Miyazaki et al. (1981). Reproduced with permis-
`sion of the copyright owner (J. Pharm. Sci.).
`»
`
`that a precipitous drop in drug solubility occurs as
`the free Cl‘ level is increased.
`
`An example of a basic‘ drug showing a strong
`chloride-ion dependence is prazosin.
`1
`‘
`9
`
`N
`
`—N'
`
`N—CO
`
`0-
`
`CH3O
`
`\
`
`cH;o /
`
`Ksp = 2.2X10"6 M @ 30°C
`
`Solubility/mg.ml‘1 @ 30°C A
`
`Base
`Hydrochloride _
`0.1 M HCl ~
`. water water
`0.037 »
`1.40
`0.0083
`
`Chloride, as well as other inorganic anions have
`the potential to form insoluble complex salts with
`weak bases (Dittert at al., 1964), which are then
`potentially less bioavailable than the _free base
`form. The formation of these complex salts is
`controlled by their stability constant Kc.
`
`‘
`Kc
`’
`-J.‘
`‘
`Drug(S) ‘——* Drug(aq) + xH+ : Drug - H: (aq)
`
`.
`
`Mylan Exhibit 1028,’ Page 3
`
`Mylan Exhibit 1028, Page 3
`
`

`
`
`
`204
`
`Evaluation of KC for triamterene (Tr) yields val-
`ues of x=O.5 for. chloride, suggesting that one
`proton solubilizes two molecules of the drug’, i.e.
`the complex is Tr2H+Cl.
`With hydrochloride salts there is frequently an
`‘overkill’ on acid strength, which leads to a very
`low pH for-an aqueous solution (Nudelman et al.,
`1974) of the salt. This can then limit the utility of
`hydrochloride salts in "certain parenteral dosage
`forms, or lead to packaging incompatibilities with
`pharmaceutical metal containers (aerosols).
`Other problems frequently arise as the result of
`the polar nature of hydrochloride salts. Their high
`hydrophilic nature, favouring wettability probably
`as a result of
`the polar ionized groups being
`exposed on the crystal -surfaces,
`leads to. water
`vapour sorption (hygroscopicity) which on occa-
`sions, (may be excessive. This can result
`in
`processing difficulties_(e.g. powder flow) and re-
`, duce the stability of a hydrolytically unstable drug.
`—.This latter effect is exacerbated by the resulting
`very low pH of the loosely bound moisture.
`These-problems can be particularly acute with
`dihydrochlorides.(or disulphates). Also,
`the dif-
`ference in the strength" of the basic centres in
`dihydrochloride salts can lead to a gradual loss of
`one of the hydrochloride moieties by release of
`hydrogen chloride gas (Lin et al., 1972) at elevated
`temperatures or under
`reduced pressure (i.e.
`freeze-drying). Also,
`their extreme polar nature
`results in excessive hydroscopicity (Boatman and
`. Johnson, 1981). eventually leading to deliques-
`cence. .
`.
`.
`
`Thus, progression of a hydrochloride salt should
`be a first move, but if the problems with that salt
`form arises due to some of the reasons outlined,
`then the real selection issue for a salt
`form
`
`emerges-——what trends are available for guidance?
`
`‘ The pivotal issues for salt selection
`
`Each drug and its associated range of dosage
`forms _ will present different salt
`form require-
`ments, and it is perhaps best to discusssalt selec-
`tion further byloutlining some of the trends ‘in salt
`properties that may facilitate selection.
`
`The pivot of melting point
`
`A change in the development of a compound
`from the free base to a salt may be promoted by a
`need to moderate the kinetics and extent of drug
`absorption, or to modify drug processing. Unfor-
`tunately these desires may be mutually exclusive,
`as the balance between these properties is fre-
`quently pivoted around the melting point of the
`salt form. For example, an increase in melting
`point is usually accompanied by a reduction in
`salt solubility (the ideal solubility of a drug in all
`solvents decreases by an order of magnitude on an
`increase of 100°C in its melting point), whereas
`high melting crystalline salts are potentially easier
`toprocess...
`_
`..
`.
`The increase or decrease in_ melting point of a
`series of salts is usually dependent on the control-
`ling effect of crystallinity from the conjugate an-
`ion. This isgexemplified by considering an experi-
`mental drug candidate (UK47880) which has a
`basic pKa of T 8, and therefore gives access to a
`wide variety of salt forms:
`'
`
`
`
`‘Sf:
`
`UK-47880
`
`melting point, 74°C
`
`Salts prepared from planar, high melting aromatic
`sulphonic or hydroxycarboxylic acids yielded
`crystalline salts of correspondingly high melting
`point
`(see Table 2), whereas flexible aliphatic
`strong acids such as citric and dodecyl benzene
`sulphonic yielded oils. Thus,
`the comparative
`planar symmetry of the conjugate acid appears
`important for the maintenance of high crystal
`lattice forces. This is shown by the melting point
`of the conjugate acid being highly correlated with
`. the melting point of the‘resultant salt form (Fig.
`2). Therefore the highly crystalline salts are in this
`case best suited to reducing drug solubility.
`Alternatively it should also be feasible to build
`up crystal lattice forces of drugs with good hydro-
`gen bonding potential, by considering syrnmetry
`and hydrogen bonding potential of the conjugate
`acid. One salt
`series of
`interest-
`is
`that
`for
`
`Mylan Exhibit 1028, Page 4
`
`Mylan Exhibit 1028, Page 4
`
`

`
`TABLE 2
`
`MELTING POINT OF SALTS 01: EXPERIMENTAL COM-
`POUND (UK47880) AND THE CORRESPONDING =coN—
`IUGATE ACID
`
`Melting point (°C) Legend
`Salt
`Conjugate
`F13‘ 2
`acid
`
`
`.
`
`74
`
`235
`
`280
`
`170
`156
`
`223
`
`190
`158
`
`220
`
`4
`
`_ A
`G
`
`D
`C
`_
`E
`
`UK 47880’; free base
`pamoate (embonate)
`4—hydroxynaphthalene-
`1-sulphonate
`-
`Salicylate 1
`3—hydroxynaphthalene-
`2'-caxboxylate
`.
`2—hydroxynaphthalene—1
`B
`120
`145
`—ca.rboxylate
`'
`“_
`'
`F
`225
`2,34
`anthraquinone-3-sulphonate
`—
`‘—
`-
`20
`dodecylbenzene sulphonate
`—
`~ 20
`.
`113'
`mesylate
`'
`T
`
`
`
`. ‘ 20 153citrate -
`
`205
`
`3GQQ
`
`60
`
`80
`
`300
`200
`100‘
`MELTING POINT AClDl°C
`
`
`
`
`
`
`
`MELTINGPOINTSALT/t“c
`
`Fig. 2. Plot of melting point of UK47880 salts vs melting point
`of conjugate acids. Legend given in Table 2.
`
`A
`
`7
`
`N
`H
`
`OH
`I
`
`C
`
`' CF3 »
`
`
`
`
`
`
`
`epinephrine
`
`Ho—cHcH2NHcH;
`
`I
`OH
`
`OH __;'._
`
`Salt form
`
`epinephrine
`ta.r’r_rate
`
`maleate
`malate
`fumarate
`
`'
`
`Melting
`point,(°C)
`
`157 ‘
`149
`
`182 -_
`170 A
`103
`
`_
`
`p
`
`hydrogen bonding acids such
`where small
`as malonic andvmaleic gave higher melting salts,
`whereas the larger bitartrate and presumably.sym-
`metrically unfavoured fumarate. gave salts of lower
`melting point.
`‘
`
`Salt form
`
`m.p; salt m.p. acid
`
`solubility
`
`.
`\
`
`, M
`Free base
`HCl
`.
`DL-lactate
`L—lactate
`2-hydroxyethane
`sulphonate
`Mesylate
`. Sulphate
`
`/
`
`_
`
`<°c)
`215
`331
`172
`192
`
`. 251
`290
`270 -
`
`i
`
`<°C)
`
`17
`5'3
`
`.
`
`'
`
`‘
`
`(mg/ml)
`.7.5
`13‘
`, 1850
`925»
`
`'
`
`620
`300
`20
`
`.
`
`The relationship between aqueous’ solubility
`",(SW) and melting point is shown diagrammatically —
`.
`in Fig.» 3, where log1SW is correlated overa range
`. of salts with the inverse of the melting point.
`Interestingly with this compound, the solubility of
`the hydrochloride salt-in water is only approxi-
`mately twice that of the free base, whereas the low
`«melting DL-lactate provides a 200-fold advantage
`over the free base in terms of solubility, which is "a
`result in part of the reduced lattice energy.
`
`.
`»
`Melting point and aqueousis0lub'z"Zz'ty~ V
`The trends in melting point“ (m.p.) andaqueous
`solubility alluded to above are exemplified in the-
`salts of a high melting antimalarial drug (Aghar-
`kar et al., 1976).
`.
`'
`‘
`
`1Melting point and chemical stability '
`The stability of .organic compounds in the solid
`state-is intimately related to the strength of the
`‘crystal lattice. Since the forces between molecules
`: in a crystal are small compared with the energy.
`
`Mylan Exhibit 1028, Page 5
`
`
`
`Mylan Exhibit 1028, Page 5
`
`

`
`
`
`206
`
`essentially controls the formation and extent of
`eutectic melts.
`
`As an additional aspectto the strength of crystal
`forces, the balance of the amorphous to crystalline
`nature in solid salts» can dramatically affect their
`stability. This is exemplified by the sodium salts
`of ethacrynic acid (Yarwood et al., 1983).
`
`V
`
`V
`
`m.p. (9 C) ‘
`' % remaining after
`9 days @ 60° C
`
`Sodium ethacrynate
`
`Crystalline
`
`’ Amorphous
`
`200
`
`100
`
`'
`'
`
`_
`
`92
`
`These results are consistent with the concept of an
`amorphous material being a highly viscous con-
`centrated solution and show that
`the stronger
`crystalline. lattice forces result
`in superior solid
`state stability.
`''
`A
`
`Melting point and formulation processing
`Salt formation is frequently employed to raise
`the melting point (and crystallinity) of the drug
`species being processed. However, published work
`concerning this type of manipulation is somewhat
`sparse.
`,
`A
`
`The melting point of _drug salts can dra.mati-
`cally affect their physical storage. Drugs (or salts)
`with low melting points generally exhibit plastic
`deformation (Jones, 1979) and thus during storage
`the stress exerted by the bulk mass on the asperity
`points o_f interparticulate contact can lead to the
`formation of localized welds leading to bulk ag-
`-gregation. Also, if the sublimation temperature is
`low (e.g. ibuprofen, m.p. 76°C), intraparticulate
`voids can be bridged by sublimed drug again
`leading to aggregation. Thus on storage, the bulk
`drug salt will begin to cake and aggregate, thereby
`altering significantly its flow, compression and
`long-term dissolution properties.
`Melting point also has a crucial role in drug
`processing, in particular comminution and tablet-
`ing. Since low melting compounds tend to be
`plastic, rather thanbrittle, they comminute poorly,
`and frictional heating causes melting and deposi-
`tion of the drug on the screens and pins of the mill _
`causing it
`to ‘blind’. For production ‘of
`fine
`pharmaceutical powders this aspect is crucial to
`judging the correct level of filler to allow efficient
`
`Mylan Exhibit 1028, Page 6
`
`I
`
`
`
`SOLUBILITYIgm!“
`
`P I
`
`
`
`0.01 m_imm__lm_____1m_1__
`1.5
`1.3
`.
`2.0
`2.2
`
`[1/MELTING POINT) x
`
`1a3r<"
`
`< Fig. 3."Plo‘t of aqueous solubility vs inverse of absolute melting
`point for a series of salts of a hydrophobic antimalarial drug.
`Data taken from Agharkar et al. (1976).
`'
`A
`
`required to break chemical bonds, liquefaction of
`the solid (and an increased frequency of molecular
`collisions) occurs before degradation begins. Thus
`the melting point of a compound can be an‘ im-
`8 portant factor in determining stability.
`.
`) Degradation of solid drugs, when it is observed,
`usually occurs in the surface film phase and is
`accompariiecl bylthe formation of a liquid phase at
`temperatures below the normal meltingpoint of
`the solid. Using this so-called ‘liquid layer’ ap-
`proach, Guillory and Higuchi (1962) investigated
`the stability of esters of vitamin A employing the
`following equation to determine the relationship
`between degradation rate and melting point.
`
`log K=
`
`AAH[1}_AH
`
`where Tm = normal melting point; Td = depressed
`storage temp.‘ = storage temperature; K = degrad-
`ation rate constant; AH = heat of fusion.
`Thus, for a series of related compounds subject
`to a storage temperature Td, the logarithm of the
`degradation rate constant is inversely related to
`the absolute melting point of
`the compounds.
`Although this approach may be somewhat simplis-
`tic it may have utility as a method of assessing the
`bulk stability of ~non—hygroscopic salt forms.
`The melting point of a salt form also has some
`influence on its relative compatibility with drug
`combinations (Hirsch et'al., 1978) or formulation
`, excipients (Li Wan Po and ‘Mroso, 1984) since it
`
`Mesylate
`Edisylate
`L‘ Lactate
`DL-Lactate
`
`‘
`
`
`
`Mylan Exhibit 1028, Page 6
`
`

`
`manufacture using a cost-effective feed rate.
`Salt melting point can also have important
`implications for particle bonding on compression
`for tableting. Since bonding on compression oc-
`curs by point welding at the deformed or frag-
`mented particle surfaces, then at a fixed tempera-
`ture and pressure, a lower melting species would
`be expected to improve bonding. However,
`the
`pressure on the powder (and the eutectics formed
`with the other excipients) suppress the melting
`point further. The Skotnicky equation defines the
`fall in melting point (Tm) with the pressure on the
`solid (PS)
`A
`
`dTm
`dPs'
`
`~VST
`Hf
`
`where AHf = heat of fusion; V5 = volume of solid;
`T= temperature, and therefore as well as those
`salts which are intrinsically low melting, salts of
`different values _of AHf would be expected to have
`different abilities to cold weld in the compression
`process. If we compare, for example, the melting
`points and heats of fusion of the salts of an
`experimental drug candidate:
`
`.
`
`.
`
`—
`
`Salt_
`Hydrochloride
`Mesylate
`Tartrate
`Citrate
`Phosphate
`Acetate
`
`.
`
`'
`
`Tm (°c)
`280
`135
`213
`180 s
`2'50 1
`180
`
`,
`
`‘
`
`AHf (kJ-mol‘1)
`56.5
`V »20.5'
`63.6
`27.2
`136.5 -
`- 167.9
`
`‘
`
`-I
`
`r
`
`the data suggest that the low melting point and
`low AHf for the mesylate_ salt would make it the
`most suitable candidate, on bonding "grounds, for
`a direct compression tablet. Since the melting
`points of compounds are reduced under pressure,
`the solubility of salt forms would be expected to
`increase with increasing pressure. This can p_oten.-
`tially cause the formation of solutions of the salts
`in the film of absorbed moisture on the surface of
`the drug (and excipient) particles which then may
`have an effect on drug bonding (Parrot, 1982) or
`cause‘ the drug to adhere to the punches on com-
`pression (Wells and Davison, -1985).
`-
`
`Conclusion
`
`_
`
`The consideration of melting point. is a key
`
`207
`
`parameter in assessing the ‘viability’ of certain
`salt forms. In general, an increase in melting point,
`usually by maximizing or encouraging crystal sym-
`metry, leads to reduced solubility in all solvents,
`but generally improved stability, particularly if
`salt formation results in a crystalline solid, and
`easier formulation processing. For a specific salt
`form for parenteral use, i.e. where solubility and
`resultant pH is a major issue, a low melting point
`salt produced using a soluble fairly weak acid (see
`next section) probably made in situ is likely to be
`preferable.
`
`The pivot of drug solubility
`
`A There are various solubility issues that can de-
`cide the viability of a particular salt form and it is
`perhaps worth addressing these separately to iden-
`tify trends that may aid salt selection.
`
`Aqueous solubility per se
`As indicated earlier, the solubility of a drug can
`enhanced dramatically by salt;
`formation
`be
`(Agharkar et al., 1976). This enhancement may
`arise from‘ a reduction in melting point, or from
`improved water-drug interactions. A good exam-
`ple of this is with the salts of chlorhexidine (Senior,
`1973), where increased water solubility" was not’
`‘only produced by a lowering of melting point, but
`by increasing‘ the hydroxylation of the conjugate
`acid.
`'
`'
`'
`
`Chlorhexidine
`
`Cl@NH— pNHpNH(cH2)6NHfiNHfiNH@c1
`
`|
`NH NH
`
`NH NH
`
`— Salt 9
`
`Stnlcture
`
`above
`_
`.
`base
`dihydrochloride HCl
`
`_
`
`_
`
`_
`
`p
`
`,
`
`di-2hydroxy-
`
`naphthoate-
`
`co2H
`
`|
`
`.OH —
`
`_
`
`~
`
`I
`
`Melting Solubility
`point
`% W/v
`(° C)
`@ 20 ° C
`
`134
`261
`
`0,008
`(0.06
`
`—
`
`V 0.014
`
`diacetate .
`dilactate
`digluconate
`
`_
`
`154
`CH3CO2H
`—
`CH3CHOHCO2H
`HO2C(CHOH)4CO2H low
`
`1.8
`1.0
`70‘
`
`Mylan Exhibit 1028, Page 7
`
`Mylan Exhibit 1028, Page 7
`
`

`
`
`
`208
`
`-
`
`.
`
`The above data exemplify the importance of
`considering the hydrop_hilic nature of the con-
`jugate anion, as well as its role in controlling
`crystallinity, when considering the potential solu-
`bility of salts.
`.
`Reduced aqueous» solubility may occasionally
`be a crucial development factor for a drug, e.g. for
`an organoleptically acceptable or chemically sta-
`ble suspension. Such systems demand salts of low
`solubility, but recent experience with a series of
`purposely designed insoluble salts of an experi-
`mental drug candidate also highlighted the need to
`consider the solubility and pKa of the conjugate
`anion.
`
`BH4 X15, =BH;;q, +X;q
`B - free base of
`salt
`1LK,, K, 5lL+H*'
`experimental drug
`B<s> *‘—‘ B<aq>
`HX<aq> *‘—‘ H799
`H+
`
`-111C aoove 1OI11C eq111l.lDI‘1a snow that even sparing
`solubility of the salt means that the level of the
`conjugate anion in solution will depend markedly
`“ on the pH of the fluid. Consideration of the above
`for a pamoate salt, which has pKa’s of the parent
`acid of 2.5 and 3.1 and virtually insoluble un-
`ionized form, indicates that solutions of pH 5-6
`will drive the equilibria to the right, with full
`precipitation of the free acid HX(S) and liberation
`of a full component in solution of the ionized base
`(BH+). However, if we consider the hydroxynaph-
`thalene sulphonic acid (pKa = 0.11) then this sys-
`tem provides ‘insolubility’pover a much wider pH
`range and is therefore far more tolerant to fluctua-
`tions in the fluid pH.
`The ‘above aspect is important when consider- '
`ing the potential use of
`‘insoluble’
`salts (e.g.
`pamoate)
`to control
`the absorption of a drug
`candidate..For example,
`the in vitro dissolution
`rates of the dimaleate and pamoate salts of a drug
`candidate were compared in simulated gastric and
`-intestinal fluid. The dissolution rates were essen-
`tially identical
`in the former fluid, with rapid
`deposition of the pamoic acid and liberation of
`the free base, whereas in the latter the pamoate
`salt exhibited a much slower dissolution rate than
`the maleate. Therefore ‘control’ on the drug ab-
`sorption (and toxicity) may then depend -on the
`duration of gastric residence and the pH of the
`gastric contents. Thus aspects such as food vs the
`
`fasted state are also important. In fact in this case,
`the bioavailability in the dog of the two salt forms
`when dosed orally from ‘a standard capsule formu-
`lation were of the same order; 24% for the pamoate
`and 39% for the maleate.
`'
`
`is the v- dissolution rate of a drug
`Usually it
`which is of major importance to the formulation
`and as a rule a salt exhibits a higher dissolution
`rate than the base at an equal pH, even though
`theyhave the same equilibrium solubility. This
`latter effect, which is exemplified by theophylline
`salts (Nelson, 1957), is due to the salt effectively
`acting as its own buffer to alter the pH of the
`diffusion boundary layer,
`thereby increasing the
`apparent solubility of the parent drug in that
`layer. Thus, administration of basic drugs as their
`salt forms (e.g. tetracycline hydrochloride) ensures
`that stomach emptying rather than in vivo dissolu-
`tion will be" the rate-limiting factor in their absorp-
`tion. It is also possible that increased drug absorp-
`tion may occur with salts due to their effect on the
`surface tension ofithegastrointesftinal fluids (Berge
`et al., 1977).
`
`Salt solubility and pH of salt solutions
`__ Enhancement of the aqueous solubility of a
`drug by salt formation can occur due to_ dif-
`ferences in the pH of the saturated salt solutions.
`A soluble acid salt of a weakly basic drug will
`cause the pH to drop as the salt is added to the
`solution. This pH drop will, in turn cause more
`drug to dissolve, and this process will continue
`until the pH of maximum solubility is reached (see
`Fig. 4). The equilibrium solubility(ies) are then
`given by:
`
`s = s+(1 + 10PH~PK«)
`
`i.e. when the ionized form is
`for pH=pHmax,
`solubility limiting and
`
`s = s,(1 + 10PK«-PH)
`
`for pH=pHmax, where the unionized form is
`solubility limiting and pHmx is given by the solu-
`tion of the equality of pH for the above two
`
`Mylan Exhibit 1028, Page 8
`
`Mylan Exhibit 1028, Page 8
`
`

`
`SOLUBILITYmglml
`
`Fig. 4. Solubility of A in water at ambient temperature (~
`23° C) as a function of pH. -All data are in mg/ml calculated in
`terms of free base equivalent. The lines drawn through the data
`are theoretical and were calculated using 0.067 mg/ml as the
`free base solubility, 11.5 mg/ml as the hydrochloride solubility
`and 8.85 as the pK,,. Data by both gravimetric (I) and GLC
`(O) procedures were in good agreement.
`
`Adapted from Kramer and Flynn (1972), with permission of
`the copyright owner (J. Pharm. Sci.).
`
`equations where
`
`pHmax = pKa + log
`
`+
`
`1
`
`and implies that both freebase and salt form can
`exist
`simultaneously in equilibrium with the
`saturated solution.
`,
`
`large pH shifts on dissolution of salts
`Thus,
`suggests that a large amountof conjugate acid is
`dissociating and therefore, a relatively high solu~
`bility is- then obtained. If we consider physiologi-
`cal pH, a low pKa fora conjugate acid of high
`aqueous solubility, would appear to give the best
`change of obtaining the lowest pHmax and the
`highest aqueous salt solubility. For example, the
`solubilities of a series of salts of a drug candidate 0
`and the‘ pH of the saturated solutions were as
`follows:
`
`Type
`
`Salt
`
`Conjugate acid
`
`209
`
`pHm,,,,
`
`Solu-
`, Solu-
`bility
`bility
`(mg/
`.
`.
`(mg/
`
`ml)
`ml)
`
`pKa
`
`Hydro-
`chloride
`Mesylate
`Tartrate
`Citrate
`Phos—
`
`2.71
`2.57
`4.21
`3.30
`~
`
`35.9
`51.2
`0.49
`2.16
`
`— 6.1
`— 1.2
`3.03 _
`3.13
`
`1470 ‘
`2400 ,
`
`2.15
`10.31
`5.31
`phate
`
`
`
`5.29 8.04Acetate 4.76
`
`indicating that the salts of stronger acids (HCl,
`methanesulphonic) produce the lowest slurry pH
`and the highest salt solubility. In this case the
`solubility and resultant low pH of the hydrochlo-
`ride is suppressed by the common ion effect. The
`solubility of salts such as the lactate (pKa of
`conjugate acid is 3.85, with infinite solubility) may
`offer significant advantage over for example the
`acetate, tartrate and citrate.
`'
`-
`The pH of. a salt solution can‘ be a deciding
`factor in the selection of a salt for aparenteral
`dosage form. Ideally to avoid pain on injection the
`pH of i.v. parenterals should be between pH 3 and .
`9, and so highly acidic salts such as the hydrochlo-
`ride and mesylate are probably best replaced by
`an acetate salt.
`.
`'
`
`Kramer and Flynn (1972) have shown that for
`a series of hydrochloride salts, by making analysis
`of the differential heat of solutions of the ionized.
`
`that the temperature de-
`and unionized species,
`pendencies of the solubilities of the hydrochloride
`salts were considerably lower than those of the
`corresponding free base form. This may have im-
`portant implications for solution dosage form des-
`ign and storage conditions.
`
`‘A
`Salt solubility and salt stability
`As well as the relationships between salt melt-
`ing point and stabilityraised earlier, it is -also clear
`that
`low solubility and low hygroscopicity can
`contribute significantly to the stability of a salt
`form. The former aspect is obviously important in
`developing a stable aqueous suspension formula-
`tion of a hydrolytically unstable water soluble
`<;
`
`Mylan Exhibit 1028, Page 9
`
`Mylan Exhibit 1028, Page 9
`
`

`
`210
`
`drug; e.g. penicillin-benzithine.
`For salts of weak bases, the moisture associated
`with the bulk can be very acidic (the salt will
`
`buffer the available moisture), and can potentially
`cause severe hydrolytic degradation of the parent
`drug. Classic examples of this phenomenon are
`thiamine salts (Yamarnoto et al., 1956, 1957) where
`the stability is related to their hygroscopicity,’
`aqueous solubility and the resulting pH.
`,
`Thus, to improve drug stability by salt forma-
`tion,
`it is clearly not only important to control
`hygroscopicity, but also to consider carefully the
`strength‘ of the conjugate acid used to form the
`salt. This is particularly important for compacted
`dosage forms where salt. and excipient share the
`available moisture, particiularly when the majority
`of the available moisture comes from the excipient
`rather than the drug. Thus, assessment of salt
`stability in compressed and —non-compressed sys-
`tems is «an important activity inpreformulation
`studies. However,
`in selection terms
`salts of
`mineral acids will produce a lower pH, and higher
`solubility in the available moistureand therefore
`produce a more hostile stability environment than
`that from a sulphonate or carboxylate type salt. It
`is also