Pharmaceutical Research, Vol. 11, No. 10, 1994 (cid:9)
`
`Report
`
`The Effect of Bulking Agent on the
`Solid-State Stability of Freeze-Dried
`Methylprednisolone Sodium Succinate
`
`Beth D. Herman,1 Brent D. Sinclair,2
`Nathaniel Milton,2 and Steven L. Nail23
`
`Received February 11, 1994; accepted May 11, 1994
`
`The rate of hydrolysis of methylprednisolone sodium succinate in
`the freeze dried solid state at 40°C was determined in the presence
`of two common bulking agents - mannitol and lactose - at two dif-
`ferent ratios of drug to excipient. Residual moisture levels were less
`than 1% in all samples tested, with no significant difference in re-
`sidual moisture among different formulations. Rate of hydrolysis
`was significantly higher in mannitol-containing formulations versus
`lactose-containing formulations, and the rate of hydrolysis increases
`with increasing ratio of mannitol to drug. Thermal analysis and x-ray
`diffraction data are consistent with a composition-dependent rate of
`crystallization of mannitol in the formulation and its subsequent
`effect on distribution of water in the freeze-dried matrix. Increased
`water in the microenvironment of the drug decreases the glass tran-
`sition temperature of the amorphous phase, resulting in an increased
`rate of reaction. The physical state of lactose remained constant
`throughout the duration of the study, and the rate of hydrolysis was
`not significantly different from the control formulation containing no
`excipient. Thermal analysis and x-ray diffraction data are consistent
`with formation of a liquid crystal phase in freeze-concentrated so-
`lutions of methylprednisolone sodium succinate containing no ex-
`cipient.
`
`KEY WORDS: thermal analysis; x-ray diffraction; excipients; crys-
`tallization; lyophilization.
`
`INTRODUCTION
`
`Bulking agents are commonly used in formulation of
`freeze dried products in order to provide an inert, easily
`reconstituted matrix containing a low dose of active drug
`substance. The study reported here was prompted by the
`need to prepare material for a blinded clinical trial where
`three different doses of drug are presented such that the
`freeze dried cakes all look alike by using an appropriate
`amount of a suitable bulking agent.
`Methylprednisolone sodium succinate is a soluble pro-
`drug of methylprednisolone used as an injectable corticoste-
`roid, where solubilization is achieved through the use of the
`ionizable hemisuccinate moiety. The prodrug is unstable in
`aqueous solution, the principal products being hydrolysis to
`methylprednisolone and acyl migration to form the isomeric
`17-ester (1). Therefore, methylprednisolone sodium succi-
`nate is marketed as a freeze dried powder (Solu-Medrole).
`
`Drug Delivery Research and Development, The Upjohn Company,
`7000 Portage Road, Kalamazoo, Michigan 49001.
`2 Department of Industrial and Physical Pharmacy, School of Phar-
`macy, Purdue University, West Lafayette, Indiana 47907.
`3 To whom correspondence should be addressed.
`
`The purpose of this study is to examine the effects of
`two of the most commonly used bulking agents in freeze
`dried injectable formulations- mannitol and lactose - on the
`stability of methylprednisolone sodium succinate as a freeze
`dried solid, with particular attention to the effect of excipient
`on the physical form of the freeze dried solid and the location
`of water within the solid matrix.
`
`C — CH2 — 0 — C — (CH2)2 — C — Oe Na0
`
`11
`
`OH
`
`CH
`
`HO
`CH
`
`00
`
`MATERIALS AND METHODS
`
`Materials
`
`Methylprednisolone hemisuccinate USP was provided
`by The Upjohn Company, Kalamazoo, MI. Mannitol and
`lactose were USP grade and were either provided by The
`Upjohn Company or purchased from Sigma Chemical Co.
`Phosphate buffer salts (Fisher Scientific) were analytical
`grade. Inorganic salts for water vapor adsorption experi-
`ments were reagent grade materials.
`
`Methods
`
`Drug with no excipient present (250 mg per vial) was
`used as a control for stability studies. Two strengths of drug
`were examined with each bulking agent - 40 mg and 125 mg
`- where the total amount of bulking agent plus drug was the
`same as the control. The formulation was prepared by sus-
`pending methylprednisolone 21-hemisuccinate in 0.08 M
`phosphate buffer, pH 7.5, containing the appropriate quan-
`tity of either mannitol or lactose. Conversion to the sodium
`salt was carried out by slow addition of 10% sodium hydrox-
`ide until essentially all solids were dissolved. The final pH
`was 7.5 - 7.7. Solutions were sterile filtered and 3.74 ml was
`filled into 20 ml tubing glass vials. Freeze drying was carried
`out by placing vials directly on the freeze dryer shelf, freez-
`ing at -50°C for 4-6 hrs, followed by primary drying at a shelf
`temperature of 10°C and a chamber pressure of 100 microns
`Hg for 24 hrs. Secondary drying was carried out for approx-
`imately 24 hrs at a shelf temperature of 30°C and a chamber
`pressure of 100 microns Hg. Vials were stoppered under full
`vacuum. Residual moisture was measured by Karl Fisher
`titration (Model 701, Metrohm, Inc., Herisau, Switzerland)
`where the freeze dried material was quickly transferred to
`the titration vessel containing anhydrous methanol. Coulo-
`metric end point detection was used.
`Stability of the freeze dried solid was measured at 25°C
`and 40°C using a reverse phase HPLC assay (1). A 4 p.m
`Nova-Pak® C-18 column (Waters, Inc., Milford, MA) was
`used. The mobile phase consisted of 33% acetonitrile in wa-
`ter buffered at pH 5.2 - 5.4 with 0.05M acetate buffer. A fixed
`wavelength UV detector at 254 nm and digital data station
`were used to quantitate the parent compound and degrada-
`
`1467 (cid:9)
`
`0724-8741/94/1000-1467$07.00/0 © 1994 Plenum Publishing Corporation
`
`Mylan Ex 1054, Page 1
`
`

`
`1468 (cid:9)
`
`Herman, Sinclair, Milton, and Nail
`
`10
`
`9
`
`a
`
`7
`
`6
`
`4
`
`3
`
`2
`
`Free Methylprednieolone (%)
`
`0
`0 (cid:9)
`
`1 (cid:9)
`
`2 (cid:9)
`
`5
`
`6 (cid:9)
`
`7
`
`4
`3 (cid:9)
`Time (Months)
`Figure 1 - Hydrolysis of methylprednisolone sodium succinate at
`40°C for 250 mg control (solid circles), 125 mg drug in mannitol
`(open triangles), 40 mg drug in mannitol (closed triangles), and 40 mg
`drug in lactose (open circles).
`
`difference in residual moisture between the samples contain-
`ing mannitol and those containing lactose. There was a small
`but significant increase in residual moisture of all formula-
`tions during six months storage at 40°C, probably caused by
`water vapor transfer from the rubber stopper to the freeze
`dried cake. The average increase was 0.29% with a range of
`0.05 - 0.45% and there was no significant difference between
`the formulations studied with respect to increase in residual
`moisture level.
`
`Thermal Analysis
`
`DSC thermograms of the formulations prior to freeze
`drying are shown in Figures 2-4. Note that endotherms are
`represented by upward deflections on all thermograms.
`Thermograms of methylprednisolone sodium succinate in
`phosphate buffer with no bulking agent present (Figure 2) are
`shown for three different heating rates after cooling at a rate
`of 20°C/min. Three separate thermal events occur prior to
`the melting endotherm of ice, and the thermograms recorded
`at different heating rates illustrate the effect of heating rate
`on both sensitivity and resolution. The endotherm on the
`leading edge of the ice melting endotherm is resolved at the
`slowest heating rate, but is not resolved at the highest heat-
`ing rate. Two exotherms are observed prior to the en-
`dotherms, with best sensitivity in detecting the exotherms
`observed at higher heating rates.
`The thermogram in Figure 2 is not consistent with low-
`temperature thermal behavior of solutions in which the sol-
`ute a) crystallizes readily upon freezing, b) remains amor-
`phous upon freezing, or c) initially forms a metastable amor-
`
`tion products. USP reference standards of methylpredniso-
`lone hemisuccinate and methylprednisolone were used as
`external standards. Percent methylprednisolone, the ester
`hyrolysis product, was used as the primary measure of
`chemical stability in the solid state.
`Both the freeze dried solids and solutions prior to freeze
`drying were examined by thermal analysis using a Perkin-
`Elmer Series 7 DSC with a mechanical cooling accessory
`and computer data station (2). Helium (40 ml/min) was used
`as the purge gas. Solution samples were prepared by placing
`approximately 20 µ1 of solution in an aluminum sample pan
`and sealing the sample by crimping an aluminum cover in
`place. Samples were frozen at a controlled rate to about
`-50°C, and the thermograms were recorded during warming
`at a controlled rate to just above 0°C. A nitrogen-purged
`glove box was placed over the sample compartment to pre-
`vent artifacts due to moisture condensation. Solid samples
`were prepared by packing about 10 mg of freeze dried pow-
`der into aluminum sample pans. For measuring the effect of
`water vapor sorption on glass transition temperatures in the
`solid state, DSC pans containing the solid sample were
`placed in dessicators at different relative humidities and
`equilibrated for about 48 hours, at which time the samples
`were quickly removed from the dessicator and sealed. Ther-
`mograms were recorded in the range of 20-120°C.
`Water vapor sorption isotherms were measured on
`freeze dried powders by placing vials (with lyostoppers in
`the partially inserted position) in dessicators containing var-
`ious saturated salt solutions. The following salts were used -
`sodium hydroxide, lithium chloride, potassium acetate, po-
`tassium carbonate, and sodium bromide. Samples were re-
`moved at two hour intervals during the first 12 hours, fol-
`lowed by sampling twice a day for five days. Vials were
`sealed upon removal from the dessicator, and moisture con-
`tent was determined by Karl Fisher titration. The water con-
`tent of the samples was found to reach approximate equilib-
`rium after about 48 hours at room temperature.
`The physical form of the freeze dried powders was stud-
`ied by x-ray powder diffraction. A Siemens Krystalloflex®
`diffractometer was used, with CuK, radiation at a voltage of
`40 kV and a current of 20 mA. Powder specimens were pre-
`pared by gently breaking up the freeze dried cakes and plac-
`ing in an aluminum powder mount. Samples were scanned
`from 2° to 40° at 0.1° per second.
`
`RESULTS AND DISCUSSION
`
`Effect of Bulking Agent on Stability
`
`The stability at 40°C of the control formulation is shown
`in Figure 1 along with that of formulations containing 40 and
`125 mg of drug in mannitol and 40 mg of drug in lactose.
`Each data point represents the average of duplicate measure-
`ments from different vials. Rate of hydrolysis of the drug
`with mannitol as the bulking agent is markedly faster than
`when lactose is used, and the rate of hydrolysis increases as
`the ratio of mannitol to drug increases. The rate of hydroly-
`sis of the 40 mg strength in lactose is not significantly differ-
`ent from that of the control.
`Residual moisture content of all formulations were in
`the range of 0.3 to 0.9% percent, and there was no significant
`
`Mylan Ex 1054, Page 2
`
`(cid:9)
`(cid:9)
`

`
`The Solid-State Stability of Freeze-Dried Methylprednisolone Sodium Succinate (cid:9)
`
`1469
`
`-60 (cid:9)
`
`-40 (cid:9)
`
`-20 (cid:9)
`-SO (cid:9)
`-10
`Temperature (C)
`
`0
`
`10
`
`Figure 2 - DSC thermograms of methylprednisolone sodium succinate in 0.08 M
`phosphate buffer. Cooling rate was 20°C/min and heating rates were at 10°C (A),
`5°C (B), and 2°C (C) per minute.
`
`phous phase which then crystallizes with subsequent
`warming (3,4). No glass transition is observed in Figure 2, as
`would be expected for an amorphous solute. The freeze
`dried drug does undergo a glass transition, however (see
`Figure 5). The endotherm on the leading edge of the ice
`melting endotherm suggests eutectic melting; however, the
`x-ray diffractogram of the freeze dried drug alone (see Figure
`7) is not consistent with well-defined crystallinity.
`Thermograms of the lactose formulations (Figure 3)
`show a glass transition (Tg') with a midpoint of about -27°C.
`This is consistent with the behavior of a solution where the
`solute does not crystallize when the system is frozen. The
`relative concentrations of lactose and drug have no signifi-
`
`cant effect on Tg' for the formulations studied here. This
`result is in reasonable agreement with a previously reported
`value of about -29°C for Tg' of lactose (2).
`Thermograms of the mannitol formulations (Figure 4)
`indicate that the physical state of the material depends on the
`ratio of drug to mannitol. For the formulation containing 40
`mg of drug, a crystallization exotherm is observed in the
`temperature range expected for mannitol. The formulation
`containing 125 mg of drug in mannitol does not crystallize
`under the conditions used for thermal analysis. Instead, a
`glass transition is observed, with a mid-point temperature of
`about -36°C. This is lower than the glass transition temper-
`ature of mannitol alone (about -28°C). This may be explained
`
`Heat Flow (m11)
`
`-60 (cid:9)
`
`-00 (cid:9)
`
`-40 (cid:9)
`
`-30 (cid:9)
`-20
`Temperature (C)
`Figure 3 DSC thermograms of formulations containing 40 mg (A) and 125 mg (B)
`of drug in lactose. Inset shows glass transition region.
`
`-10
`
`0
`
`10
`
`Mylan Ex 1054, Page 3
`
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`

`
`1470
`
`Herman, Sinclair, Milton, and Nail
`
`B
`
`Heat Flow (mW)
`
`z
`
`B
`
`A
`
`—50 (cid:9)
`
`—50 (cid:9)
`
`—40 (cid:9)
`
`—20 (cid:9)
`—90 (cid:9)
`Temperature (C)
`Figure 4 - DSC thermograms of formulations containing 40 mg (A) and 125 mg (B)
`of drug in mannitol.
`
`—10 (cid:9)
`
`0 (cid:9)
`
`10
`
`by a plasticizing effect of the drug itself or, more likely,
`increased unfrozen water in the freeze concentrate caused
`by the drug, the buffer, or both.
`Thermal analysis of freeze dried drug (no excipient
`present) shows a glass transition which is well above 100°C
`at low residual moisture levels. As shown in Figure 5, the
`glass transition temperature decreases continuously as the
`residual moisture level increases. This is consistent with the
`behavior of amorphous solids where water acts as a plasti-
`cizer, as previously reported by Zografi and coworkers (5,6).
`
`X-Ray Powder Diffraction
`X-ray powder diffractograms of a mannitol-containing
`formulation (125 mg drug) are shown in Figure 6 initially and
`at 2 and 6 months. The data show that the freeze dried solid
`is substantially amorphous initially, and that crystallinity de-
`velops during storage despite residual water levels on the
`order of 1.0 to 1.5 percent. The formulations containing lac-
`tose as the excipient were amorphous initially, and remained
`so for the duration of this study (data not shown). The for-
`mulation containing 40 mg of drug in mannitol showed sub-
`
`120
`
`100
`
`00
`
`0
`
`40
`2
`
`4
`
`a
`
`10
`
`12
`
`11 Water
`Figure 5 - Glass transition of freeze dried methylprednisolone so-
`dium succinate as a function of water content.
`
`stantial crystallinity initially after freeze drying, and the de-
`gree of crystallinity increased somewhat during the six
`month storage interval (data not shown). Samples of freeze
`dried mannitol alone were prepared for use as a standard to
`estimate the degree of crystallinity in stability samples; how-
`ever, both mannitol alone and mannitol in the formulation
`appear to be mixtures of polymorphs (7), and the composi-
`tion of the mixtures is not the same for the two materials.
`Further work is under way to better characterize the crystal
`forms of mannitol which are formed during freezing and dry-
`ing under a variety of conditions.
`The x-ray diffraction pattern of drug with no bulking
`agent present is shown in Figure 7. A single, sharp peak was
`observed in all vials tested at about 4° 20, and a broad, dif-
`fuse peak is centered at around 18° 20. This type of diffrac-
`togram is consistent with diffractograms reported for liquid
`crystals (8), where the reflection at the low angle arises from
`the spacing between layers, and the larger angle reflection
`arises from order within the individual layers. Additional
`polarized light microscopy experiments have confirmed the
`formation of liquid crystals by methylprednisolone sodium
`succinate. A lyotropic mesomorphic phase might also ex-
`plain the endotherm observed in Figure 2 prior to the main
`ice melting endotherm. Casillas and coworkers (9) observed
`two endotherms prior to the main ice melting endotherm in
`the thermogram of aqueous solutions of dioctyl sodium sul-
`fosuccinate (Aerosol OT). These additional endotherms
`were attributed to water in two different states of association
`with the solute in the mesophase.
`Given that the surface activity of the drug is well doc-
`umented (10), formation of a liquid crystalline phase from
`concentration of a micellar system would not be unexpected.
`However, further work is needed in order to better under-
`stand the pharmaceutical significance of the liquid crystal-
`line state within freeze dried solids. While the solid state
`stability of methylprednisolone sodium succinate in the ab-
`sence of any excipient is not a problem at sufficiently low
`water content, the general significance of liquid crystal
`phase formation with respect to stability could be related to
`
`Mylan Ex 1054, Page 4
`
`

`
`The Solid-State Stability of Freeze-Dried Methylprednisolone Sodium Succinate (cid:9)
`
`1471
`
`5 (cid:9)
`
`10 (cid:9)
`
`15 (cid:9)
`
`20 (cid:9)
`25
`Angle (2 Theta)
`Figure 6 - X-ray powder diffractograms of formulations containing 125 mg of drug in
`mannitol initially after freeze drying (A), and after 2 months (B) and 6 months (C)
`storage at 40°C.
`
`30
`
`35
`
`40
`
`preferred orientation between molecules in the liquid crystal
`phase, with subsequent effects on reaction rates through in-
`termolecular catalysis. As a side note, it was observed dur-
`ing x-ray diffraction analysis of stability samples that some
`vials of drug containing no excipient show evidence of crys-
`tallization during storage. Further work is needed to better
`understand the phase transitions involved; in particular,
`whether a crystalline solid is formed from the dehydrated
`liquid crystal phase, or whether the liquid crystalline phase
`must first convert to an amorphous phase prior to crystalli-
`zation of the solid.
`
`Water Vapor Sorption Isotherms
`
`Water vapor adsorption isotherms were measured for
`freeze dried, single component systems consisting of man-
`nitol, lactose, and methylprednisolone sodium succinate in
`order to compare water uptake characteristics of the indi-
`vidual formulation components. Moisture content versus
`time at relative humidities of 11, 21, 32, and 43 percent was
`determined. The approximate time for a reasonably constant
`moisture level to be established was 48 hours at all relative
`humidities examined. The plateau value of moisture was
`
`>.
`4-,
`in
`C
`a)
`-I ,
`C
`
`5 (cid:9)
`
`10
`
`20 (cid:9)
`25
`15 (cid:9)
`Angle (Two Theta)
`
`30
`
`35
`
`40
`
`Figure 7 - X-ray diffraction of freeze-dried methylprednisolone sodium succinate with no
`excipient present.
`
`Mylan Ex 1054, Page 5
`
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`

`
`1472 (cid:9)
`
`Herman, Sinclair, Milton, and Nail
`
`10
`
`a
`
`4
`
`2
`
`•
`
`0
`
`0
`10 (cid:9)
`
`15 (cid:9)
`
`20 (cid:9)
`
`90
`25 (cid:9)
`Relative Humidity
`Figure 8 - Water vapor adsorption isotherms of freeze dried mannitol (open tri-
`angles), lactose (open circles), and drug (closed circles).
`
`35
`
`40
`
`45
`
`plotted against relative humidity to construct the water va-
`por sorption isotherms shown in Figure 8. Lactose (amor-
`phous) and methylprednisolone sodium succinate (dried liq-
`uid crystal) are similar in water vapor sorption characteris-
`tics. Mannitol, however, adsorbs very little moisture for the
`relative humidity range examined. This is consistent with
`crystalline material, where the only surface available for wa-
`ter vapor sorption is the surface of crystals.
`
`Suggested Role of Excipient in Solid State Stability
`
`The data reported here are consistent with the work of
`Zografi and coworkers which support the idea that the effect
`of water on critical attributes of amorphous drugs is deter-
`mined not so much by how much water is present, but by
`where the water is located. The effect on critical product
`attributes of concentration of water within amorphous re-
`gions becomes more important as the fraction of amorphous
`material decreases (11). In formulations containing mannitol
`as the excipient, the mannitol crystallizes either during the
`freeze dry process or with time during storage, depending on
`the ratio of mannitol to drug. In either case, there is not a
`significant amount of water associated with crystalline man-
`nitol, as shown by the comparative water vapor sorption
`isotherms. For the formulation containing 40 mg of drug in
`mannitol, the effective water content of the amorphous
`phase would be expected to be more than five times the
`value measured by Karl Fisher titration, assuming that the
`mannitol is essentially all crystalline, since the mannitol to
`drug ratio is approximately 5.2:1. Referring to Figure 5, the
`increased water level would be expected to decrease T5' by
`20° to 30°C. Amplification of the water content of the amor-
`phous phase could occur by any or all of three mechanisms:
`1) While the initial levels of residual moisture are not signif-
`icantly different for lactose and mannitol formulations, the
`amount of water "seen" by the drug cannot be the same
`given the dramatically different water vapor adsorption iso-
`therms for the two materials. 2) Crystallization of mannitol
`with time would result in a redistribution of water in the
`
`freeze dried matrix and an increase in the amount of water in
`the drug microenvironment. 3) Water added to the freeze
`dried solid by transfer from the stopper would locate in the
`amorphous phase. Subsequent lowering of the glass transi-
`tion temperature of the amorphous phase would be expected
`to accelerate the rates of chemical reactions leading to in-
`stability of the drug substance.
`In contrast to the mannitol formulations, those contain-
`ing lactose are markedly more stable because lactose re-
`mains amorphous, resulting in a more uniform distribution of
`water in the freeze dried matrix. This is supported by the
`water vapor adsorption isotherm data, showing similar affin-
`ity of both drug and lactose for water. The similar water
`vapor adsorption isotherms support the idea that lactose
`could act, in part, as an internal dessicant by competing for
`the available water.
`
`Practical Implications
`
`Mannitol is often the first choice as a bulking agent
`among formulation development scientists, since it is inex-
`pensive, non-toxic, and can be freeze dried at relatively high
`temperatures with full retention of the desireable properties
`of a freeze dried product. Lactose, on the other hand, re-
`quires more conservative drying conditions because it re-
`mains amorphous on freezing and has a relatively low T5'
`temperature of about -30°C. However, it is important to be
`aware of potential adverse effects on stability because of the
`effect of mannitol crystallization on distribution of water in
`the freeze dried cake. Stability protocols for mannitol-based
`formulations should reflect the potential for adverse effects
`of mannitol on stability of drugs which are amorphous as
`freeze-dried solids and, in particular, for protein formula-
`tions containing mannitol (12). Particular attention should be
`given to monitoring moisture content of stability samples in
`mannitol-based formulations, since water vapor transfer
`from the stopper to the product can be very significant, par-
`ticularly when the moisture is concentrated in amorphous
`regions of the lyophilized cake.
`
`Mylan Ex 1054, Page 6
`
`(cid:9)
`(cid:9)
`(cid:9)
`

`
`The Solid-State Stability of Freeze-Dried Methylprednisolone Sodium Succinate (cid:9)
`
`1473
`
`Looking at only the DSC thermogram of the drug solu-
`tion at low temperature, a development scientist would be
`led to believe that the drug crystallizes during freezing or
`after the frozen solution is heated during initiation of drying.
`This drug is certainly not crystalline in the traditional sense
`after freeze drying. This points to the importance of charac-
`terizing the dried solid by x-ray diffraction initially after
`freeze drying and monitoring the physical state of the solid
`during stability studies.
`
`ACKNOWLEDGMENTS
`
`The authors gratefully acknowledge the support pro-
`vided by the Pharmaceutical Manufacturers Association
`Foundation for an Undergraduate Research Fellowship (B.
`Sinclair), Merck, Sharpe, & Dohme for an Academic Devel-
`opment Grant (S.L. Nail), Edwards High Vacuum, Ton-
`awanda, NY (freeze drying equipment grant), and NIH Bi-
`ological Research Support Grant #RR-05586-24.
`The technical assistance provided by May Mui and Lisa
`Nail is also greatly appreciated.
`
`REFERENCES
`
`1. B.D. Anderson and V. Taphouse. Initial rate studies of hydro-
`lysis and acyl migration in methylprednisolone 21-hemi-
`succinate and 17-hemisuccinate. J. Pharm. Sci., 70: 181-6
`(1981).
`2. L.M. Her and S.L. Nail. Measurement of glass transitions in
`
`frozen solutions by differential scanning calorimetry. Pharm.
`Res., 11: 54-59 (1994).
`3. S.L. Nail and L.A. Gatlin. Principles and practice of freeze
`drying. In K.E. Avis, H.A. Lieberman, an L. Lachman, Phar-
`maceutical Dosage Forms: Parenteral Medications, Vol. 2,
`Marcel Dekker, Inc., 1993, pp. 173-9.
`4. L.A. Gatlin and P.P. DeLuca. A study of phase transitions in
`frozen antibiotic solutions by differential scanning calorimetry.
`J Parenteral Drug Assoc., 34: 398-408 (1980).
`5. G. Zografi. States of water associated with solids. Drug Dev.
`Ind. Pharm., 14: 1905-26 (1988).
`6. C. A. Oksanen and G. Zografi. The relationship between glass
`transition temperature and water vapor adsorption of poly(vi-
`nylpyrrolidone). Pharm. Res., 7: 654-7 (1990).
`7. Organic and Organometallic Phases, Powder Diffraction File,
`International Center for Diffraction Data, Swarthmore, PA 1992,
`p. 643.
`8. G.H. Browne, J.W. Doane, and V.D. Neff, A Review of the
`Structure and Properties of Liquid Crystals. CRC Press, Cleve-
`land, 1971, page 18.
`9. N. Casillas, J.E. Puig, R. Olayo, T.J. Hart, and E.I. Franses.
`State of water and surfactant in lyotrophic liquid crystals. Lang-
`muir, 5: 384-9 (1989).
`10. B.D. Anderson, R.A. Conradi, and K. Johnson. Influence of
`premicellar and micellar association on the reactivity of meth-
`ylprednisolone 21- hemiesters in aqueous solution. J. Pharm.
`Sci., 72: 325-31 (1983).
`11. C. Ahlneck and G. Zografi. The molecular basis of moisture
`effects on the physical and chemical stability of drugs in the
`solid state. Int. J. Pharm. 62: 87-95 (1990).
`12. K. Izutsu, S. Yoshoika, and T. Terao. Effect of mannitol crys-
`tallinity on the stabilization of enzymes during freeze drying.
`Chem. Pharm. Bull. 42: 5-8 (1994).
`
`Mylan Ex 1054, Page 7