Pharmaceutical Research, Vol. 19, No. 11, November 2002 (© 2002)
`
`Research Paper
`
`Hydrogen Bonding with Adsorbent
`during Storage Governs Drug
`Dissolution from
`Solid-Dispersion Granules
`
`Manish K. Gupta,1 Yin-Chao Tseng,2
`David Goldman,2 and Robin H. Bogner1, 3, 4
`
`Received July 14, 2002; accepted August 2, 2002
`
`Purpose. To investigate changes in drug dissolution on storage of
`ternary solid-dispersion granules containing poorly water-soluble
`drugs.
`Methods. Hot-melt granulation was used to prepare ternary solid-
`dispersion granules in which the drug was dispersed in a carrier and
`coated onto an adsorbent. Seven drugs including four carboxylic acid-
`containing drugs (BAY 12-9566, naproxen, ketoprofen, and indo-
`methacin), a hydroxyl-containing drug (testosterone), an amide-
`containing drug (phenacetin), and a drug with no proton-donating
`group (progesterone) were studied. Gelucire 50/13 and polyethylene
`glycol (PEG) 8000 were used as dispersion carriers whereas Neusilin
`US2 (magnesium aluminosilicate) was used as the surface adsorbent.
`Results. Two competing mechanisms have been proposed to explain
`the complex changes observed in drug dissolution upon storage of
`solid dispersion granules. Conversion of the crystalline drug to the
`amorphous hydrogen bonded (to Neusilin) state seems to increase
`dissolution, whereas, the phenomenon of Ostwald ripening can be
`used to explain the decrease in drug dissolution upon storage. The
`solubility of the drug in Gelucire is a crucial factor in determining the
`predominant mechanism by governing the flux toward the surface of
`Neusilin. The mobility for this phenomenon was provided by the
`existence of the eutectic mixture in the molten liquid state during
`storage.
`Conclusions. A competitive balance between hydrogen bonding of
`the drugs with Neusilin and Ostwald ripening determines drug disso-
`lution from solid-dispersion granules upon storage.
`
`KEY WORDS: solid dispersion; dissolution; hydrogen bonding;
`crystallinity; stability.
`
`INTRODUCTION
`
`There is renewed interest in the formulation of solid dis-
`persions as a promising approach to enhance the dissolution
`of poorly water-soluble drugs (1–4). However, the key limi-
`tations to their widespread commercial use include problems
`in processing solid dispersions into dosage forms and rever-
`sion of the amorphous drug to the lower energy crystalline
`state on storage (3,4). Reversion to the lower energy crystal-
`line state from the non-equilibrium higher energy amorphous
`state leads to a decrease in drug dissolution, thereby, defeat-
`
`1 School of Pharmacy, University of Connecticut, Storrs, Connecticut
`06269
`2 Pharmaceutical Technology, Pharmaceutical Division, Bayer Cor-
`poration, West Haven, Connecticut 06269
`3 Institute of Materials Science, University of Connecticut, Storrs,
`Connecticut 06269
`4 To whom correspondence should be addressed at University of
`Connecticut, U-2092, Storrs, Connecticut 06269. (e-mail:
`BOGNER@UCONNVM.UCONN.EDU)
`
`ing the very purpose of formulating solid dispersions. As dis-
`cussed in the reviews by Serajuddin and by Leuner and Dress-
`man (3,4), physical stability of the drugs in solid dispersions
`seems to be the rate-limiting step in using this approach to
`enhance dissolution and oral bioavailability of poorly water-
`soluble drugs.
`In previous work a combination of solid dispersion and
`surface adsorption was used to prepare ternary solid-
`dispersion granules, with enhanced drug dissolution and fea-
`sibility of compression into tablets (5). The dissolution of a
`poorly water-soluble drug, BAY 12-9566, was enhanced by
`dispersing it in Gelucire 50/13 and coating the dispersion on
`Neusilin US2 using hot-melt granulation (5). Interestingly it
`was found that drug dissolution from the granules was further
`enhanced on storage at 40°C/ 75% RH for 4 weeks. This
`result is in sharp contrast to the more general phenomenon of
`reduced dissolution on storage resulting from reversion to the
`crystalline state. An increase or a decrease in drug dissolution
`on storage of a formulation is not acceptable from a stability
`standpoint and mandates further investigation for the devel-
`opment of a stable robust formulation.
`In a follow-up study it was shown that hydrogen bonding
`between several drugs and the adsorbent, Neusilin was key to
`the further enhancement of drug dissolution on storage of
`granules (6). On storage, a decrease in drug crystallinity (from
`X-ray powder diffractometry) and a corresponding increase
`in the drug hydrogen-bonded to Neusilin (from Fourier trans-
`form infrared spectroscopy) were observed for both BAY
`12-9566 and naproxen granules but not for progesterone gran-
`ules (6).
`Under the storage conditions (40°C/ 75% RH) used pre-
`viously (5,6), relatively high mobility in the eutectic melt of
`the solid-dispersion was expected to allow nucleation and re-
`version of the drug to the crystalline state leading to de-
`creased drug dissolution upon storage. However, in the pres-
`ence of another amorphous phase, Neusilin, there was the
`potential for the drugs to diffuse toward the surface of the
`adsorbent. We proposed that Neusilin hydrogen bonded with
`drugs (such as BAY 12-9566 and naproxen), thereby prevent-
`ing their reversion from the amorphous to crystalline state
`(6). Neusilin (magnesium aluminosilicate) has silanol groups
`on its surface that make it a potential proton donor as well as
`an acceptor. The hydrogen bonding potential of silanols in the
`local environment on silica surfaces is well-documented (7,8).
`In a recent study the hydrogen-bonding interaction between
`silanol groups in colloidal silicon dioxide and indomethacin (a
`carboxylic acid containing drug) was revealed using 29Si and
`13C solid-state nuclear magnetic resonance studies (9). Thus,
`hydrogen bonding between the drugs under investigation
`(with proton accepting and/ or donating potential) and Neu-
`silin is possible. In contrast to a previous report that addresses
`the prevention of reversion of the amorphous state of a drug
`(indomethacin) due to hydrogen bonding with a carrier (poly-
`vinyl pyrrolidone) in a single phase system (10), we found
`evidence for formation of additional hydrogen bonds of the
`drug onto a phase separated amorphous interface on storage.
`This study was designed to test the generalizability of the
`proposed mechanism using seven drugs, two solid dispersion
`carriers, and an adsorbent. The drugs include four carboxylic
`acid containing drugs (BAY 12-9566, naproxen, ketoprofen,
`
`1663
`
`0724-8741/02/1100-1663/0 © 2002 Plenum Publishing Corporation
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`1664
`
`Gupta et al.
`
`and indomethacin), a hydroxyl containing drug (testoster-
`one), an amide containing drug (phenacetin), and a drug with
`no proton-donating group (progesterone) were studied.
`These poorly water-soluble drugs were chosen to have a
`range of functional groups having potential to hydrogen bond
`with Neusilin so as to investigate the generalizability of the
`proposed mechanism. Gelucire 50/13 (melting range: 47–
`53°C) and PEG 8000 (melting range: 60–63°C) have been
`used in solid dispersion formulations and make good choices
`for low melting solid dispersion carriers (11–14). Neusilin
`US2 has a high specific surface area (∼300 m2/g) and consists
`of amorphous microporous granules with potential for hydro-
`gen bonding (15). Neusilin US2 was, therefore, used as the
`adsorbent.
`The physical properties of solid dispersions can be stud-
`ied using techniques such as Fourier transform infrared spec-
`troscopy (FT-IR), X-ray powder diffraction (XPD), and
`modulated differential scanning calorimetry (MDSC) (16–
`20). Hydrogen bonding between drugs and excipients was
`investigated using FT-IR spectroscopy. XPD was used to ex-
`amine the crystallinity of drugs in the dispersion granules.
`MDSC was performed to determine the melting point of the
`eutectic mixture of the drug and dispersion carrier in the
`dispersion granules. Solubility of the drugs in the dispersion
`carrier was estimated gravimetrically in a temperature con-
`trolled UV spectrophotometer. Ternary solid-dispersion
`granules of the seven drugs were prepared by hot-melt granu-
`lation. Drug dissolution from the initial granules (upon for-
`mulation) was compared with that after storage at 40°C/ 75%
`RH for 2 and 4 weeks. FT-IR, XPD, MDSC, and solubility
`data were used to explore underlying mechanisms resulting in
`changes in the dissolution profiles upon storage. The gener-
`alizability of the previously proposed mechanism was thereby
`tested.
`
`MATERIALS AND METHODS
`
`Experimental
`
`All the seven drugs were poorly water-soluble, thermo-
`stable compounds with different hydrogen bonding potential.
`The four carboxylic acid-containing drugs include BAY 12-
`9566, naproxen, ketoprofen, and indomethacin. BAY 12-9566
`has a melting point of 110°C. Naproxen USP, ketoprofen
`USP, and indomethacin USP have melting points of 152°C,
`94°C, and 162°C, respectively, and were obtained from PCCA
`(Houston, TX). The amide-containing phenacetin (purified
`powder) and hydroxyl-containing testosterone USP, have
`melting points of 135°C and 155°C, respectively, and were
`obtained from PCCA (Houston, TX). Progesterone USP (␣-
`form), with no proton-donating group, has a melting point of
`130°C and was obtained from Sigma (St. Louis, MO). The
`structures of these seven drugs are shown as Fig. 1. Gelucire
`50/13, a polyglycolized glyceride, was used as received from
`Gattefosse (Westwood, NJ). Gelucire 50/13 is obtained from
`hydrogenated vegetable oils consisting of mono-, di-, and tri-
`glycerides and mono- and di-fatty acid esters of PEG 1500.
`Polyethylene glycol 8000 (PEG 8000, Carbowax, Sentry
`grade) was obtained from Union Carbide (Danbury, CT).
`Fuji Chemicals (Englewood, NJ) supplied magnesium alumi-
`nosilicate (Neusilin US2). A laboratory scale low shear granu-
`
`Fig. 1. Structures of the seven drugs.
`
`lator (Mini MGT, L. B. Bohle Incorporated, Bristol, PA),
`fitted with a heating jacket, was used to prepare granules.
`
`Formulation of Solid-Dispersion Granules
`
`Ternary solid-dispersion granules were prepared using
`hot melt granulation. The details of preparation were de-
`scribed earlier (5). Briefly, the drug was added into the mol-
`ten dispersion carrier and heated to obtain a clear molten
`mixture. Neusilin US2 was preheated to 80°C in the granula-
`tor with stirring at 300 rpm. The molten mixture was then
`added dropwise over a period of one minute to Neusilin with
`continued stirring. Hot melt granulation was performed at an
`increased stirring speed of 600 rpm for one more minute to
`obtain ternary solid-dispersion granules of each of the drugs,
`dispersion carrier and adsorbent in a ratio of 1:1:1. The ter-
`nary dispersion granules of the seven drugs were prepared
`with Gelucire 50/13 as the dispersion carrier and Neusilin as
`the adsorbent. Ternary dispersion granules of BAY 12-9566,
`naproxen, and progesterone were also prepared using PEG
`8000 in place of Gelucire 50/13. The granules were sieved
`through mesh # 18 BSS. The abbreviations used to describe
`the components of these solid-dispersion granules are in-
`cluded in parentheses: BAY 12-9566 (Bay), naproxen (Nap),
`ketoprofen (Ket), indomethacin (Ind), phenacetin (Phe), tes-
`tosterone (Tes), progesterone (Pro), Gelucire 50/13 (G), PEG
`8000 (P), and Neusilin US2 (N). For example ternary granules
`containing BAY 12-9566, Gelucire 50/13 and Neusilin US2
`would be represented as Bay/G/N.
`
`Preparation of Amorphous State of Drugs
`
`The amorphous state of each of the seven drugs was
`prepared by melting followed by quench cooling in liquid
`nitrogen. The amorphous nature of these drugs was ensured
`by absence of birefringence under cross-polarized light.
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Hydrogen Bonding with Adsorbent during Storage Governs Drug Dissolution
`
`1665
`
`Dissolution Testing
`
`Dissolution profiles of the drugs from their solid-
`dispersion granules were determined using a USP Type II
`apparatus at 50 rpm. The dissolution medium consisted of 0.1
`N HCl with sodium lauryl sulfate (SLS) for all the drugs other
`than phenacetin, where 0.05 M, pH 7.0 phosphate buffer was
`used. All the drugs were, thereby, maintained in the union-
`ized state in the dissolution medium. Dissolution profiles in
`the previous study were determined using 30-mg dose of
`BAY 12-9566 (5). This dose corresponds to 17% of saturation
`solubility at 37°C in 900 ml of 0.1 N HCl and 1% w/v SLS.
`Based on the equilibrium solubility values, 25 mg doses of
`naproxen, 70 mg of ketoprofen, 70 mg of indomethacin, 25 mg
`of testosterone, and 25 mg of progesterone required 0.25%,
`0.25%, 1%, 0.1%, and 0.1% w/v SLS, respectively in 0.1 N
`HCl to maintain similar sink conditions in the dissolution
`medium (∼17% of saturation solubility). For phenacetin, 175-
`mg dose was used in 0.05M, pH 7.0 phosphate buffer to main-
`tain similar sink conditions and the unionized state of the
`drug. Dissolution samples were analyzed for drug concentra-
`tion from the absorbance values at the respective wavelength
`of peak absorbance.
`
`Examination of Hydrogen Bonding
`
`FT-IR spectra were measured using a spectrometer
`(Model Magna IR 560, Nicolet Instrument Technologies Inc.,
`Madison, WI). KBr pellets of each drug alone as well as solid-
`dispersion granules were prepared. An average of 100 scans
`of each sample was collected at 4 cm−1 resolution over a
`wavenumber region of 4000–600 cm−1. FT-IR spectra of pure
`crystalline as well as pure amorphous state of drugs were
`measured for comparison with the spectra of the granules.
`
`Evaluation of Crystallinity
`
`Samples of each drug in its pure crystalline state and
`formulated as solid-dispersion granules were studied for crys-
`tallinity using a diffractometer (Model D5005, Bruker AXS
`Inc., Madison, WI) using CuK␣ radiation, a voltage of 40kV,
`and a current of 40mA. The scanning rate was 1.25°/min over
`a 2␪ range of 5–50° with a sampling interval of 0.02°.
`
`Storage of Solid-Dispersion Granules
`
`Solid-dispersion granules were stored at 40°C/ 75% RH
`for 4 weeks. Their dissolution profiles were determined after
`storage and compared to those of the initial granules. FT-IR
`spectra of the initial and stored granules were compared to
`examine changes in hydrogen bonding of drugs, if any. X-ray
`powder diffractograms of the initial and stored granules were
`compared to evaluate any changes in drug crystallinity.
`
`Determination of the Melting Point of the Eutectic Mixture
`
`The temperature for the onset of melting of the eutectic
`mixture in the solid dispersion granules was determined using
`a modulated DSC instrument (Model 2920, TA Instruments).
`The instrument was calibrated in the modulated mode using
`high purity indium with helium as the purge gas. The disper-
`sion granules (∼4 mg) were hermetically sealed in aluminum
`sample pans (Perkin–Elmer, Norwalk, CT, USA). A modu-
`lated temperature program with a period of 60 s, an amplitude
`
`of ±0.4°C, and an overall heating rate of 2°C/ min was used
`from −40°C to a temperature above the melting point (Tm +
`10°C) of the drug.
`
`Estimation of Solubility of Drugs in the Dispersion Carriers
`
`The solubilities of the seven drugs were determined in
`the dispersion carriers using a temperature controlled UV
`spectrophotometer (Model Cary 50 Bio, Varian Instruments).
`The dispersion carrier (2 g) was allowed to melt in the cuvette
`at 60°C, which is well above the melting point of the disper-
`sion carrier. Accurately weighed increments of the drug (5 mg
`at a time) were then added to the cuvette followed by the
`stirring of the sample using a magnetic stirrer. Upon addition
`of drug to the cuvette beyond its saturation solubility, the
`absorbance baseline in the visible region shifted up due to
`scattering of light from the undissolved particles of the drug.
`None of the drugs or the dispersion carriers was found to
`absorb in the visible region. These solubility determinations
`were performed in triplicate using Gelucire 50/13 and PEG
`1450 as the dispersion carriers at 60°C. PEG 1450 was used
`instead of the PEG 8000 (used in preparing the solid disper-
`sion granules) due to problems encountered during stirring
`resulting from the high viscosity of the PEG 8000 melt. It has
`been reported that the dielectric constant of the PEGs does
`not depend on their average molecular weight but is rather a
`function of the concentration of its subunit (CH2CH2O) (21).
`The solubility of the drugs, therefore, should not be signifi-
`cantly different in PEG 1450 than in PEG 8000.
`
`RESULTS
`
`Ternary solid-dispersion granules of all the seven drugs
`were prepared by hot melt granulation using Gelucire 50/13
`as the dispersion carrier and Neusilin US2 as the adsorbent.
`Dispersion granules of BAY 12-9566, naproxen, and proges-
`terone were also prepared using PEG 8000 in place of Gelu-
`cire 50/13 as the dispersion carrier. Dissolution of BAY 12-
`9566 was further enhanced from the granules upon storage at
`40°C/ 75% RH for 2 and 4 weeks (see Bay/G/N in Table I).
`Similar dissolution enhancement was also observed with
`
`Table I. Drug Dissolution from Ternary Solid-Dispersion Granules:
`Comparison of Initial Dissolution Values with Those after Storage
`
`Percentage drug dissolved after 30 minutes
`(standard deviation)
`
`Batch name
`
`Initial
`
`2 Wk/40°C/
`75%RH
`
`4 Wk/40°C/
`75%RH
`
`Bay/G/N
`Bay/P/N
`Nap/G/N
`Nap/P/N
`Ket/G/N
`Ind/G/N
`Phe/G/N
`Tes/G/N
`Pro/G/N
`Pro/P/N
`
`39.5 (1.7)
`39.7 (1.3)
`37.6 (1.6)
`44.2 (1.4)
`92.7 (1.2)
`67.9 (1.9)
`97.0 (1.5)
`56.0 (3.5)
`73.0 (1.2)
`45.3 (1.4)
`
`82.0* (2.2)
`64.3* (2.4)
`54.4* (1.5)
`41.0 (3.0)
`91.3 (0.6)
`66.3 (1.6)
`85.6* (1.3)
`51.0* (2.7)
`46.7* (0.6)
`45.5 (0.6)
`
`82.7* (1.1)
`69.8* (0.8)
`54.7* (1.3)
`43.0 (2.9)
`91.8 (1.0)
`63.5* (1.7)
`86.2* (0.9)
`43.2* (2.9)
`39.4* (1.7)
`42.9* (0.8)
`
`Student’s independent t-test was performed at an ␣-value of 0.05
`* indicates significant difference between the initial and the stored
`granules.
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`1666
`
`Gupta et al.
`
`naproxen (see Nap/G/N in Table I). Although ketoprofen
`granules (Ket/G/N) did not show any change in dissolution,
`indomethacin granules (Ind/G/N) showed a slight decrease in
`drug dissolution (see Ket/G/N and Ind/G/N in Table I). On
`the other hand, granules containing phenacetin, testosterone,
`and progesterone exhibited the more commonly reported de-
`crease in the enhanced dissolution on storage (see Phe/G/N,
`Tes/G/N, and Pro/G/N, in Table I). The percentage drug dis-
`solution (after 30 min) upon storage compared to the initial
`sample was calculated and is shown in Fig. 2. As shown in Fig.
`2, the magnitude of changes in drug dissolution from PEG-
`containing dispersion granules was significantly less than that
`observed with Gelucire-containing granules. In the following
`sections, drug dissolution data is discussed relative to corre-
`sponding changes in hydrogen bonding (FT-IR studies), crys-
`tallinity of drug (XPD studies), melting point of the eutectic
`(MDSC studies), and the solubility data.
`
`BAY 12-9566 Solid-Dispersion Granules
`
`As shown in Table I, dissolution of BAY 12-9566 was
`enhanced further from the ternary solid-dispersion granules
`(Bay/G/N and Bay/P/N) after storage at 40°C/ 75% RH.
`Upon storage at 40°C/ 75% RH for 4 weeks, dissolution after
`30 min was found to be 210% of the initial dissolution from
`granules, Bay/G/N (Fig. 2).
`The hydrogen bonding interaction between BAY 12-
`9566 and Neusilin was explained in detail in a previous pub-
`lication (6). BAY 12-9566, like many other carboxylic acid-
`containing drugs, exists as a dimer stabilized via intermolecu-
`lar hydrogen bonding. In the crystalline state the character-
`istic dimer peak of the carboxylic acid group is at 1695 cm−1,
`whereas in the amorphous state it is at 1707 cm−1 (6). As
`shown in Fig. 3, the dimer peak at 1695 cm−1 was only appar-
`ent as a shoulder in the spectrum for the initial granules (Bay/
`G/N), indicating a disruption of dimer hydrogen bonds. Fur-
`ther, upon storage at 40°C/ 75% RH for 4 weeks, the dimer
`peak disappeared and only one peak was observed at 1687
`cm−1 (Fig. 3), indicating complete absence of any dimers. In
`the crystalline state, intramolecular hydrogen bonding is
`
`probable between oxygen of the benzoyl carbonyl and hydro-
`gen of the carboxyl group of BAY 12-9566, leading to a peak
`at 1681 cm−1. The blue shift for the benzoyl carbonyl stretch-
`ing peak from 1681 cm−1 to 1687 cm−1 on storage is an indi-
`cation of a tendency toward hydrogen bonding between BAY
`12-9566 and Neusilin compared to that for intramolecular hy-
`drogen bonding. Similar changes in FT-IR spectra for Bay/
`P/N granules indicate the presence of a further increase in the
`drug hydrogen-bonded to Neusilin upon storage in the pres-
`ence of PEG as well (6).
`XPD studies were performed to detect any changes in
`the crystallinity of BAY 12-9566, corresponding to the in-
`creased hydrogen bonding to Neusilin, indicated by FT-IR
`studies. The absence of any peaks in the diffractogram of
`Neusilin confirmed its amorphous nature (data not shown).
`The intensity of the most intense peak for crystalline BAY
`12-9566, at a 2␪ value of 20.8, was found to be dramatically
`reduced for the initial granules, Bay/G/N and Bay/P/N, indi-
`cating significant conversion to the amorphous state on for-
`mulation of solid-dispersion granules. Upon storage of these
`granules at 40°C/ 75% RH for 4 weeks, the intensity of this
`peak was found to decrease further (shown for granules, Bay/
`G/N, in Fig. 4). The increase in the amount of drug hydrogen-
`bonded to Neusilin is therefore accompanied by a corre-
`sponding decrease in the amount of crystalline BAY 12-9566
`upon storage of granules.
`However, the change in drug dissolution upon storage is
`greater for Gelucire-containing granules (Bay/G/N) than
`PEG-containing (Bay/P/N) granules. BAY 12-9566 showed
`good solubility in both Gelucire as well as PEG (Table II and
`Fig. 5). The onset of the melting point of the eutectic mixture
`of BAY 12-9566 and Gelucire 50/13 (30.2°C) was lower than
`that of the drug and PEG 8000 (37°C) (see Table II).
`
`Naproxen Solid-Dispersion Granules
`
`Like BAY 12-9566, naproxen is a carboxylic acid con-
`taining drug with the potential to accept as well as donate
`protons and hydrogen bond with Neusilin. Dissolution en-
`hancement of naproxen was observed upon formulation as
`
`Fig. 2. Comparison of drug dissolution (after 30 min) from initial and stored solid-dispersion
`granules using USP Type II apparatus at 50 rpm. Data are shown for drug dissolution (% of
`initial) from solid-dispersion granules after storage at 40°C/ 75% RH: Initial (black); 2 weeks
`(gray); and 4 weeks (squares).
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Hydrogen Bonding with Adsorbent during Storage Governs Drug Dissolution
`
`1667
`
`Fig. 3. FT-IR spectra for solid-dispersion granules before and after
`storage. Spectra are shown for granules: Bay/G/N, Nap/G/N, Ket/G/
`N, Ind/G/N, Phe/G/N, Tes/G/N, Pro/G/N.
`
`Fig. 4. X-ray diffractograms for solid-dispersion granules before and
`after storage. Spectra are shown for granules: Bay/G/N, Nap/G/N,
`Ket/G/N, Ind/G/N, Phe/G/N, Tes/G/N, Pro/G/N.
`
`granules, Nap/G/N and Nap/P/N. In contrast to the further
`increase in dissolution from Gelucire-containing granules
`(Nap/G/N), the PEG-containing granules (Nap/P/N) did not
`show any change in drug dissolution on storage at 40°C/ 75%
`RH for 4 weeks (see Nap/G/N and Nap/P/N in Table I and
`Fig. 2).
`As previously reported, FT-IR studies revealed the pres-
`ence of hydrogen bonding between naproxen and Neusilin in
`the granules, Nap/G/N and Nap/P/N (6). As observed with
`BAY 12-9566, naproxen granules, Nap/G/N, showed a reduc-
`tion in the dimer peak (at 1680 cm−1 in both crystal and
`amorphous state) upon formulation and a further reduction
`during storage (Fig. 3). Also the peak at 1605 cm−1 split and
`showed a new red shifted peak at 1593 cm−1 for the stored
`granules, indicating changes in the hydrogen bonding of drug
`
`upon storage. Naproxen in the granules was, therefore, found
`to hydrogen bond with Neusilin and the interaction increased
`during storage.
`In the X-ray diffractogram, the intensity of the most in-
`tense peak of naproxen, at a 2␪ value of 19.0 in the crystalline
`state, decreased significantly for the granules, Nap/G/N and
`Nap/P/N, indicating decreased crystallinity of naproxen upon
`formulation of granules. Drug crystallinity was found to de-
`crease further upon storage of granules, Nap/G/N and Nap/
`P/N at 40°C/ 75% RH for 4 weeks (shown for granules, Nap/
`G/N, in Fig. 4). Similar to BAY 12-9566, a decrease in drug
`crystallinity accompanied by an increase in the hydrogen-
`bonded drug leads to the further enhancement in drug disso-
`lution from naproxen granules upon storage.
`The solubility of naproxen in Gelucire (14.0% w/w) is
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`1668
`
`Gupta et al.
`
`Table II. Changes in State of Drug and in Drug Dissolution upon Storage of Ternary Solid-Dispersion Granules: Comparison of Initial
`Dissolution Values with Those after Storage
`
`Dispersion granules
`
`Change in drug
`dissolution
`
`Crystalline drug
`(XPD)
`
`Drug hydrogen-bonded
`to Neusilin
`(FT-IR)
`
`Onset of melting point of the
`eutectic (°C)
`(MDSC)
`
`Solubility of the drug
`in the dispersion
`carrier (% w/w)
`
`Changes upon storage of dispersion granules
`
`Bay/G/N
`Bay/P/N
`Nap/G/N
`Nap/P/N
`Ket/G/N
`Ind/G/N
`Phe/G/N
`Tes/G/N
`Pro/G/N
`Pro/P/N
`
`Increase
`Increase
`Increase
`No change
`No change
`Decrease
`Decrease
`Decrease
`Decrease
`Decrease
`
`Decreases
`Decreases
`Decreases
`Decreases
`No change
`No change
`No change
`No change
`Complete ␤ to ␣
`Incomplete ␤ to ␣
`
`Increases
`Increases
`Increases
`Increases
`No change
`No change
`No change
`No change
`No change
`No change
`
`30.2
`37.0
`30.5
`38.0
`16.8
`30.4
`34.2
`35.5
`33.4
`39.0
`
`33.8
`42.0
`14.0
`20.6
`39.3
`10.0
`6.2
`3.0
`7.4
`8.3
`
`less than that of BAY 12-9566 and ketoprofen, but is higher
`than the other four drugs used in this study (Fig. 5). Although
`the onset of melting of the eutectic mixture was 30.5°C in
`Nap/G/N, it was 38°C in the granules, Nap/P/N.
`
`Ketoprofen Solid-Dispersion Granules
`
`Dissolution of ketoprofen (a drug with a carboxylic acid
`group) was enhanced upon formulation as granules, Ket/G/N
`(data not shown). There was no change in drug dissolution
`from granules, Ket/G/N, upon storage for 4 weeks (see Table
`I and Fig. 2). Ketoprofen had the highest solubility in Gelu-
`cire (39.3% w/w) and showed the lowest onset of melting of
`the eutectic (16.8°C).
`The FT-IR spectra showed the acid dimer peak at 1697
`and 1706 cm−1 in the crystalline and the amorphous state of
`ketoprofen, respectively. Upon formulation as granules, Ket/
`G/N, the acid dimer peak disappeared, indicating breaking of
`the intermolecular hydrogen bonds and preferential hydrogen
`bonding between ketoprofen and Neusilin. As shown in Fig.
`3, no further change in the dimer peak was observed upon
`storage of granules, Ket/G/N. In contrast to BAY 12-956 and
`naproxen, the amount of the ketoprofen hydrogen-bonded to
`Neusilin remains the same after storage for 4 weeks.
`Upon formulation as granules, Ket/G/N, the X-ray dif-
`fractogram showed no peaks at all, indicating complete con-
`version to the amorphous state of the drug during initial pro-
`cessing of the granules (Fig. 4). No peaks were observed upon
`storage either (Fig. 4), indicating the absence of any reversion
`from the amorphous hydrogen bonded (to Neusilin) state to
`the crystalline state.
`
`Fig. 5. Solubility (%w/w) of the seven drugs in Gelucire 50/13 at 60°C.
`
`Indomethacin Solid-Dispersion Granules
`
`Indomethacin is another drug with a carboxylic acid
`group. Similar to BAY 12-9566, naproxen, and ketoprofen,
`this drug also showed enhanced dissolution upon formulation
`as granules, Ind/G/N (data not shown). In contrast to the
`other three carboxylic acid containing drugs, a slight decrease
`in drug dissolution was observed upon storage of granules,
`Ind/G/N, at 40°C/ 75% RH for 4 weeks (see Table I and Fig.
`2).
`
`In the FT-IR spectrum, indomethacin showed the acid
`dimer peak at 1717 cm−1 and 1710 cm−1 for the crystalline and
`the amorphous state, respectively. Unlike BAY 12-9566,
`naproxen, and ketoprofen, which showed a red shift (from
`amorphous to crystalline state) in the acid dimer peak, indi-
`cating greater hydrogen bonding in the crystalline state, in-
`domethacin does not seem to favor hydrogen bonding upon
`crystallization. Also, there was no reduction in the acid dimer
`peak at 1717 cm−1 for indomethacin granules, Ind/G/N, be-
`fore or after storage (Fig. 3), indicating the absence of any
`changes in the hydrogen-bonded state of indomethacin.
`From the XPD studies, it was found that drug was
`present at least in part in the crystalline state upon formula-
`tion as granules, Ind/G/N and the crystallinity did not change
`upon storage at 40°C/ 75% RH for 4 weeks (Fig. 4). The
`absence of any increase in drug crystallinity upon storage
`indicates the absence of any reversion from the amorphous to
`crystalline state. The onset temperature of the eutectic mix-
`ture was found to be 10°C lower than the storage temperature
`(see Table II). The solubility of indomethacin in Gelucire
`(10.0% w/w) was lower than the solubility of the other three
`carboxylic acid-containing drugs (Fig. 5).
`
`Phenacetin Solid-Dispersion Granules
`
`Phenacetin contains an amide group with a pKa of 2.2.
`Because the carboxylic acid-containing drugs were in the
`unionized state in 0.1 N HCl, it was decided to perform the
`dissolution studies for phenacetin in 0.05M phosphate buffer
`(pH 7) to keep the drug in an unionized state. Upon formu-
`lation as granules, Phe/G/N, drug dissolution was enhanced
`when compared to the drug alone (data not shown). How-
`ever, dissolution was found to decrease significantly upon
`storage at 40°C/ 75% RH for 4 weeks (see Table I and Fig. 2).
`
`Purdue 2028
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Hydrogen Bonding with Adsorbent during Storage Governs Drug Dissolution
`
`1669
`
`Both the FT-IR spectra (Fig. 3) as well as XPD diffrac-
`tograms (Fig. 4) did not show any evidence of changes in
`hydrogen bonding and drug crystallinity, respectively, upon
`storage at 40°C/ 75% RH for 4 weeks. Although the onset of
`melting of the eutectic was higher, the solubility of phenacetin
`in Gelucire was lower than that of indomethacin (Table II and
`Fig. 5).
`
`Testosterone Solid-Dispersion Granules
`
`Testosterone is a steroid compound containing a hy-
`droxyl group. Similar to phenacetin, after the initial dissolu-
`tion enhancement upon formulation as granules, Tes/G/N,
`dissolution was found to decrease significantly upon storage
`at 40°C/ 75% RH for 4 weeks (Table I and Fig. 2).
`Similar to phenacetin granules, both the FT-IR spectra
`(Fig. 3) as well as XPD diffractograms (Fig. 4) of testosterone
`granules, Tes/G/N, did not show any evidence of changes in
`hydrogen bonding and drug crystallinity, respectively, upon
`storage at 40°C/ 75% RH for 4 weeks. The onset of melting of
`the eutectic of testosterone was higher than that of the eu-
`tectic of phenacetin (see Table II). The solubility of testos-
`terone in Gelucire (3.0% w/w) was lowest among the seven
`drugs studied.
`
`Progesterone Solid-Dispersion Granules
`
`Progesterone does not have a proton-donating group but
`its carbonyl group can act as a proton acceptor. Progesterone
`granules, Pro/G/N and Pro/P/N, showed a decrease in disso-
`lution upon storage at 40°C/ 75% RH for 4 weeks (see Table
`I). The magnitude of the decrease in dissolution upon storage
`of granules, Pro/G/N, was greater (54% of initial) than that of
`phenacetin or testosterone granules (Fig. 2). However, the
`magnitude of the change in dissolution upon storage of PEG-
`containing granules (Pro/P/N) was much less (95% of initial)
`than Gelucire-containing granules (Pro/G/N) (Fig. 2).
`No changes in the FT-IR spectra of the granules, Pro/G/
`N, before or after storage, were found indicating the absence
`of any changes in the hydrogen-bonded state of the drug in
`
`granules (Fig. 3). Progesterone exists as the more stable (high
`melting point) ␣-form and the less stable (low melting point)
`␤-form (22). The characteristic diffraction peak for the
`␣-form seems at 2␪ value of 16.9, whereas that for the ␤-form
`is at 16.1 (as noted from the Joint Committee of Powder
`Diffraction Standards, JCPDS database). It was found that
`the drug converted from the ␣-form to the more soluble
`␤-form on formulation as granules, Pro/G/N (Fig. 4) and Pro/
`P/N. However, on storage the peak at 16.9 reappeared, indi-
`cating a reversion to the less soluble ␣-form (Fig. 4). The
`solubility of progesterone is similar in both Gelucire (7.4%
`w/w) and PEG (8.3% w/w). The onset of melting of the eu-
`tectic mixture in granules Pro/P/N was only 1°C below the
`storage temperature whereas for Pro/G/N it was 6.4°C below
`the storage temperature (see Table II).
`
`DISCUSSION
`
`Neusilin US2 was used