No. 5
`
`'
`
`f
`
`1085
`
`Ch
`.Ph
`. B '.
`[23(iiniiossji330(19liii)
`
`,‘
`7
`,
`UPC 547. 92.93
`
`Solubilization of Steroids by Multiple co-Solvent Systems
`
`YIE W. 'CHIEN” and HOWARD J. LAMBERT
`
`Biopkarmacentios Section Product Development Department Searle Laboratories”
`
`(Received September 27, 1974)
`
`The aqueous solubility of steroids was exponentially increased following the addition
`of a co-solvent. A theoretical equation was derived to describe the relationship between
`the amount of drug solubilized and the volume fraction of co-solvent incorporated. Both
`single and multiple co-solvent systems fellowed the derived relationship.
`
`Introduction
`
`In the practice of pharmaceutical formulation, the challenge of improving the solubility
`of many poorly soluble drugs, leg, steroids, in solution dosage forms is common. To solve
`such problems, scientists often incorporate one or more co—solvents with distilled water to
`overcome the poor aqueous solubility.
`.'
`Ethanol, propylene glycol, and several members of the polyethylene glycol polymer
`series, e’.g., polyethylene glycol 400, represent the limited number of co—solVents that are both
`useful and generally acceptable in the formulation of aqueous liquids”.
`In addition, Spiegal
`and Noseworthy3) in their review of nonaqueous solvents used in parenteral products, also
`suggested a number of co-solvents, such as 2,2-dimethyl—l,3-dioxolane—4—methanol (Solketal),
`dimethylacetamide, glycerol formal, glycoiurol,
`'N—(fl—hydroxyethyl)-lactamide, and ethyl
`lactate.
`
`There have been several reports dealing with the systematic investigation of drug solubility
`and solvent compositionfia‘l‘” Observations to date indicate that the solubility of many
`drugs and druglike substances in binary! aqueous systems is enhanced exponentially by the
`addition of a c0-solvent.4’
`
`Past experiences in our laboratories with solution dosage formulation indicated that the
`semilogarithmic relationship of drug solubility to c0~solvent composition was followed not
`only in binary aqueous systems containing a single co-solvent, but also in the systems
`containing 2 or more different co-solvents.
`In this paper, the authors will report their obser-
`vations on the dependency of steroid solubility on the concentration of single, binary, and/or
`ternary co~solvents, and will present a theoretical analysis on the semilogarithmic relationship
`of drug solubility to co—solvent concentration.
`
`Experimental
`
`_ MaterialeC—9376, $04640, SC—11800, and SC-25152 (Fig. 1) (Searle Laboratories, Skokie, Illinois),
`propylene glycol, polyethylene glycol 400, 2,2-dimethyl-1,3-dioxolane—4—methanol, and N,N-dimethylacet—
`amide (Matheson Coleman Bell (30., Norwood, Ohio) and 95 % ethanol for medical use were utilized as obtained.
`
`.
`1) Location: Skokie, Illinois 60076, U.S.A.
`2) L. Lachman, ILA. Lieberman, and IL. Kanig, "The Theory and Practice of Industrial Pharmacy,”
`Lea & Febiger, Philadelphia, 1970, Chapter 15.
`'
`i
`3) A.J. Spiegel and M.M. Noseworthy, ]. Phomn. 805., 52,! 917 (1963).
`4) S.H. Yalkowsky, G.L. Flynn, and G.L. Amidon, J. Pharm. 5613, 61, 983 (1972).
`5) A.N. Parnta and S.A. Irani, J. Pharm. SOL, 54, 1334 (1965),
`V
`' 6) W.G. Gorman and G.D.'Hall, J. Pharm. Sell, 53, 1017 (196.4).
`7) KS. Lin, J. Anschel, and CJ. Swartz, Bull. Parenteral Drug Anson, 25, 40 (1971).
`
`NIT—Electronic Library Service
`
`InnoPharma Exhibit 1098.0001
`
`

`

`
`
`1086 Vol. 23 (197,5)
`
`gRrenteral quality distilled water was prepared and used for the preparation of 0»~SQ%~co-solvent—water
`combinations. Absolute methanol (spectroquality, ].T. Baker Chemical, Cleveland, Ohio) was"employed
`for the dilution of filtered drug solutions to an appropriate concentration for spectrophotometric measurement.
`
`0
`
`Hg-Co‘é
`
`l
`1
`i
`'
`"
`_
`SC 9376(pamemng)
`
`.
`
`‘
`
`8.071108%
`f
`"j
`$945640":
`-
`I;
`‘iethynodiol'.diacetate)
`U
`(‘nor‘ethindronfl
`.
`V
`2:515? ,
`,
`Fig. 1. Molecular Structures of the Steroids investigated
`
`i
`
`
`Determination of Drug Solubility
`The methodology reported earlier” was adopted here with minor
`modificationsto avoid the possible hydrolysis of SC-9376 and SC-25152 after‘48-hour equilibration at 37°.
`Excess drug solid was equilibrated with 10 ml of a co—solvent—Water combination at 37° for 2 hours with
`constant ‘sha'king.
`v Aftercooling to room temperature, the over-saturated drug solution Was quickly [filtered
`through a filter holder (millipore) containing a glass fiber membrane (Reeve’Angel). The filtered ‘dru'g
`solution was diluted 10 to 4000 times With methanol to an appropriate drug concentration and then read
`spectrophotometrically. Themagnitude'of the absorbance at the A max was used to calculate the amount
`of a drug solubilized in a given solvent system. The filtered solution" of ethynodiol diacetate (SC-11,800)
`was subjected to acidic hydrolysis” before dilution.
`v
`‘
`'
`
`Results and Discussions “
`
`I
`
`In nonideal‘solutions,‘9> the solubility of a drug“(Cw) in pure Water system is defined by I,
`
`gASw
`2.303RT
`
`10g cw = —
`
`(Tm—r) + log y“,
`
`"
`
`(1)
`
`Also, the solubilities of this drug (CA, CB, or CX) in pure systems of co-solvent A, B, or
`may be defined in the same way:
`I
`
`‘
`
`1
`
`0g
`
`ASA
`c =—-——T —T
`1
`'
`2.303RT,( m H 0g“
`. A53
`= —————T —T
`1
`25303RT( m
`)+ Ogys
`Asx
`w
`z
`= ——— T —:r
`1
`2.303RT( m
`)\+ 037’):
`
`10% CB
`
`,
`
`10% CX
`
`i
`
`7
`
`.
`
`~
`
`V
`
`‘
`
`I
`
`_
`
`'
`
`'
`
`_
`
`*
`
`2
`()
`3
`()
`
`4
`( )
`
`where ASW, ASA, ASE, and ASx are the entropies of thedrug species in the pure solvent systems
`and W, 32A, 3213, and 32,; are the corresponding activity coefficients; Tm is the drug melting point;
`. and T is the temperature of the system investigated.
`,
`it
`»
`In a nonideal multiple solvent system containing fA fraction of co—solvent A, fB fraction
`5of co-solvent B, fX fraction of co-solvent X, and [1—(fA —|—fB +fx)] fraction of water, the apparent
`solubility (CA, B, X) of the same drug species may be expressed as
`log CA,’B,X = fw 10g Cw
`fA 10g CA
`CB + fx
`CX
`f.
`= 10% Cw + fAGOg CA—Iog Cw.) .+. fB,(10g Cfi—logcw),
`
`,
`
`_
`
`#
`
`
`
`.
`8) Y.W. Chien, H.J. Lambert, and D.E. ‘Granti’ji 131mm; sax} 6,3, 365 (197.4).
`9) A.N. Martin, “Physical 'Pharmacy,"~Lea, & Febiger; Philadelphia',*19§0,x Chapter. 14.
`
`,
`
`NII—Electronic Library Service
`
`InnoPharma Exhibit 1098.0002
`
`

`

`No. ,5
`
`~
`
`-
`
`1087'
`
`since
`
`
`
`
`
`
`
`r
`
`‘
`(ASw—ASA) + log—31::—
`165; CA —- log CW =
`(Asa—ASE) + log if
`log CB —— 10g CW =
`(Ase—ASX) + log g:—
`log Cx — 10g 0.. =
`Substituting Eqs. (6), (7), and (8) into (Si-b) gives
`(9)
`10g CA,B,x =10g CW) + SAfAV'i' ssz + exfx
`where the slopes (EA, 3],, and ex) of the semilogarithmic relationship between the drug solubility
`(CA, B, X) and the composition of co-solvent (fA, f3, and fx) are defined by:
`
`'71».
`(meT)
`.
`:
`1 ——
`A w—AS
`>,:A)+ 0g VW
`2.303RT( S
`
`(TnamT) A
`Sw—A
`2.3031zr( ,
`
`(Tm'"T) A
`2.303RT( S
`
`(6) ‘
`(7)":
`at
`
`(10)
`
`(1D
`
`,
`
`(12)
`
`=
`
`8A
`
`=
`
`=
`
`8”
`
`ax
`
`7’3
`’
`l —
`3‘9ng
`A
`‘
`7X
`w~ s
`1
`.—
`x)+ 0g yw
`
`‘
`v
`
`-
`
`:
`
`.
`
`_
`
`Eqs. (10) to (12) point out that the magnitudes of the slopes (3A, 83, and ex) are determined by
`the difference in entropies, MSW—ASA), (ASw—ASB), and (ASWQ—ASX), respectively,the ratio
`of the activity coefficients (yA/yw, 723/32,, and yX/yw and the difference between the melting
`point temperature (Tm) and the temperature of the system studied (T).
`Eq. (9) indicates that the solubility of a drug species in a given multiple co—solvent system
`(CA, B, x,) is exponentially related to the volume fractions of co-solvents (A, B, and X) added.
`In the case of a ternary system containing a fixed (fA) of co-solvent A and a varying fraction
`(fx) of co-solvent X, Eq. (9) may be simplified. Since fB=O, the apparent drug solubility
`(CA, x) in such a ternary system is:
`
`log CA,X = log CW + £AfA + exfx
`
`(13)
`
`When using a binary system, (only one co—solvent), Eq. (13) may be further simplified to
`log Cx = log Cw + exfx
`,
`‘
`i (14)
`An equation similar to Eq. (14) was reported previously by Yalkowsky, et all“ to describe
`the solubility of alkyl p—aminobenzoates in propylene glycol~water systems.
`The experimental evidence for the dependence of drug solubility upon co—solvent composi-
`tion in a binary system (Eq. 14) is demonstrated in Fig. 2, where the solubility of poorly soluble
`steroids, e.g., progestins (SC—11800 and 804640) as well as anti-mineralcorticoid drugs (SC-9376
`and 8025152), is exponentially enhanced as the volume fraction of polyethylene glycol 400
`co—solvent increases. The molecular structures of the four synthetic steroids investigated
`were shown earlier in Fig. 1.
`I
`i
`.
`,
`Eq. (14) was also followed in binary systemslcontaining either ethanol, propylene glycol,
`solketal (2,2-dimethyl—1,3-dioxolane—4—methanol), or‘ dimethylacetamide. For example, the
`solubilization of canrenone by ethanol, propylene glycol, and PEG 400 is illustrated in Fig. 3,
`A common ordinate intercept, which was equivalent to the actual aqueous solubility (Cw)
`of canrenone (8.1 X 10%) measured independently, was obtained. The relative efficiency
`on the solubilization of canrenone by these five co-solvents is tabulated in Table I in order
`of solubilizing efficiency. Ethanol is the most effective solubiliz‘er and polyethylene glycol
`400 is least effective.
`,
`'
`v
`I.
`'_
`,
`j_
`;«
`A
`;
`The validity of Eq,
`(13) in a ternary system containing a fixed (fA) volume fraction of
`co—solvent A and a varying fraction (fx) of co—solvent X is demonstratedby the datavin‘Fig..4=-.
`In this experiment, the incorporation 'ofua fixed concentration (19 or 28.5% v/v), of ethanol
`
`NIT—Electronic Library Service
`
`InnoPharma Exhibit 1098.0003
`
`

`

`1088
`
`Vol. 23 (1975)
`l N
`
`ca X ._. o
`
`4X10—2
`
`N l'
`
`t—I o:
`
`EEI
`D0
`
`’8
`
`5 10‘8
`g;
`g
`w
`o
`>,
`3% 10“
`~g
`__5
`U)
`
`10-5
`
`V
`
`1.0
`
`0.8
`0.6
`0.4
`» 0.2
`0
`‘Volume fraction of polyethylene
`glyCOI 400
`Solubihmtion of some Syflthe_
`Fig 2
`tic Steroids by the- Addition of Poly_
`eth lene Glycol
`y
`keys: . EEC-11800, 0 804640, I 809376, and
`D 5025152
`
`(moles/liter)
`
` 10"2-
`
`‘
`
`canrenone
`
`' 10‘9
`
`“a
`<>a
`a
`L5
`3
`c8
`
`.
`
`10f4
`‘ 5X10_5___._I_._L_____1_~J_____1
`0
`0.2
`0.4
`0.6
`0.8
`1.0
`Volume fraction of co-solvent
`
`’
`
`'
`
`.
`
`-
`
`Fig. 3. Semilogarithmic Relationship
`between the Concentration of Caan-
`none (moles/liter) solubilized and the
`Volume Fraction of a Single 00—
`Solvent
`.
`'
`F
`A common intercept at 8.1 X 10‘“M was observed:
`keys: 0 ethanol, 0 propylene glycol, and .
`polyethylene glycol 400
`
`TABLE I. The Solubilization of Canrenone by the Binary
`Systems Containing a Single co-Solvent and Water
`
`
`
` Solubilizers ex“)
`
`Ethanol
`Solketal
`-
`I Propylene glycol
`v
`' Dimethylacetarnide
`Polyethylene g1ycol‘400'
`
`.
`
`4.75
`3.97
`3.42
`3.21
`2.93
`
`a) axis the slope of the log OK as. fx profiles as defined in Eq. (14).
`
`substantially enhanced the magnitude of the intercept (from log CW to leg CW+£AfA) (compare
`Eqs; 14 with 13); but, the linear relationship of log CA, x to fx was still followed and the magni—
`tude of the slope (ex) stayed about the same (ex=8.34 to 3.43).
`‘ When a fixed concentration of a third co~solvent, e.g., solketal (20% v/v) or dimethyl—
`acetamide (20% v/V) f was incorporated into mixtures of ethanol—propylene glycol—water,
`the solubility of canrenone was enhanced. The linearity of log CA, B, X to the volume fraction
`0f propylene glycol (as expected from Eq.
`(9)) was still observed. Following the addition
`of 20% V/V of either dimethylacetamide Or solketal (Fig 5) the intercept (at zero concentration
`‘of propylene glycol) was significantly increased from log Cw +8AfA (Eq. 13) to log Cw—l—eAfA +83fB
`(Eq. 9). The effectof addition of a third solubilizer on the magnitude of both the intercept
`(log CW+aAfA+stB) and the slope (ex) of the log CA, B, X vs.fW profiles are found in Table II.
`Along with the increase in the intercept, the slopes were slightly minimized due to the addition
`Of) a third co-solve‘nt. This may be due to the change in the entropy and activity coeflicient
`‘of the resultant solution. The predictive value of Eq.
`(9')
`is demonstrated in Table III.
`
`NIT—Electronic Library Service
`
`InnoPharma Exhibit 1098.0004
`
`

`

`N0. 5‘
`
`1089
`
`6X10’a
`
`3X10—l
`‘
`
`
`
`Solubilityofcanrenone(moles/liter)
`
`J.
`
`H
`
`..o
`
`
`
`n
`
`l—l
`
`,9
`
`
`
`A,
`
`H,9.
`
`
`
`
`—5
`10—47
`1'.o
`.0.8
`0.6
`0.4
`0,2
`5X10 0
`Volume fraction of propylene glycol
`I
`Fig. 5. Semilogarithmic Relationship
`between the Concentration of Game-
`
`none (moles/liter) solubilized and the
`Volume Fraction of Propylene Glycol
`in a Three co-Solvent Aqueous System
`
`keys’ 0 Propylene 313m] alone, l: Plus 95%
`ethanol and 20% dimethylacetamide, I plus 19%
`ethanol and 20% dimethylacetamide, and . plus
`19% ethanol and 20% solketal
`
`i.
`
`‘
`
`N
`
`t. on
`
`5—. O
`
`J»
`
`0\
`
`
`
`
`
`Solubilityofcanrenone(moles/liter)
`
`"”
`4{/
`1°C
`5x10‘5_.
`
` 0
`
`Volume fraction of propylene glycol
`
`Fig. 4. Semilogarithmic Relationship
`between the Concentration (moles/
`liter) of Canrenone solubilized and the
`Volume Fraction of Propylene Glycol
`in a Two co-Solvent Aqueous System
`
`keys: 0 propylene glycol alone, 0 plus 19% v/v
`of ethanol, and . plus 28.5% v/v of ethanol
`
`TABLE II. Effect of the Addition of Third co-Solvent on the Intercept
`and Slope Values of log CA,3,X vs. fx Profiles (Eq. 9)
`co—Solvent combinations
`
`
`:fi: 1
`
`# 2
`
`# 3
`
`Intercept
`(MX103)
`
`
`Slope
`
`Propylene glycol
`Propylene glycol
`Propylene glycol
`Propylene glycol
`Propylene glycol
`
`—
`19.0% ethanol
`9.5% ethanol
`19.0% ethanol
`19.0% ethanol
`
`——
`—
`20% DMA
`20% DMA
`20% solketal
`
`3.42
`0.081
`3.43
`0.80
`3.00
`2.06
`2.39
`7.00
`
`13.80 2.08
`
`The product of the slope (6) (obtained from the solubilization of canrenone with individual
`co—solvents) and the volume fraction (f’5) used was employed to estimate the expected can—
`renone solubility in multi—co—solvent systems. The agreement of the observed solubility
`with that calculated is good. The small deviation of the observed solubility from the estimated
`value (low ratio, ca. 0.76) may be due to the slight decline in slope (Table II) observed after
`the incorporation of a third co-solvent.
`The use of multiple co-solvent combinations to enhance steroid solubility is an improve—
`ment over solubilization with a high volume fraction of a single co—solvent. For example,
`55% ethanol, or 77% propylene glycol, or 87% polyethylene glycol 400 was required in order
`to effectively solubilize 3.25X 10—2M of canrenone in aqueous solution. The use of high co-
`solvent concentrations may unfavorably affect the desired Viscosity, and the esthetic accepta-
`bility of the resultant formulations. On the other hand, this drug concentration (3.25 X lO—ZM)
`was achieved with a multiple co-solvent system containing either 9.5% ethanol—20% dimethyl-
`acetamide—étOo/0 propylene glycol or 19% ethanol—20% solketal—18% propylene glycol.
`In
`
`NIT—Electronic Library Service
`
`InnoPharma Exhibit 1098.0005
`
`

`

`1090
`
`Vol. 23 (1975-)
`
`TABLE III. Agreement betweenlCalculated and Observed Canrenone Solubility
`‘
`in Multiple co-Solvent Systems
`
`log}.
`“
`Multiple
`co—Solvent sYstems“)
`CA, 3, x”)
`‘
`Calculatedd)
`Observed
`
`
`Canrenone ‘isolubility ,‘(M X 102)
`
`Ratio”)
`
`‘
`
`1.368
`40%
`Propylene glycol
`0 . 451
`9. 5%
`Ethanol
`0.642
`20%
`DMAe)
`2.461
`SUI“ ,
`p
`- 1.368:
`Propylene glycol
`40%
`0 . 903
`Ethanol
`19%
`0.642
`DMAe>
`20%
`1 2.91.3
`SUM
`'
`1.368
`Propylene glycol
`40%
`' 0.903
`Ethanol
`19%
`O . 794
`Solketal
`20%
`
`V 3.065 11.61 9.61 1SUM 0.83
`
`
`2.89
`
`'
`
`3.25
`
`1.12
`
`8.18
`
`6.25
`
`0.76
`
`
`
`
`
`
`
`a) 9.8. of distilled water is added to make 100%
`[1)
`calculated from Table I by multiplying the slope (s) with volume fraction (f) for each cosolvent used
`c)
`ratio of observed solubility over calculated value
`'
`d)
`antilogaritbmic of log OMB”;
`e ) dimethylacetamide
`
`practice, any number of combinations may be blended depending on specific fermulation
`requirements and the physico—chemical properties of the drug. Finally, the use of a multiple
`co—solvent system may prevent the occurence of undesirable toxicity which may result from
`the use of a high volume fraction of a single co~solvent.
`
`Appreciation is extended to Ms. Dianne M. Jefferson for her
`Acknowledgements and Addresses
`technical assistance and to Ms. Susan M. Justi for manuscript preparation.
`
`NII—Electronic Library Service
`
`InnoPharma Exhibit 10980006
`
`