The Physical State of Mannitol after Freeze-Drying: Effects of Mannitol
`Concentration, Freezing Rate, and a Noncrystallizing Cosolute
`
`ALEXANDRA I. KIM,t MICHAEL J. AKERS,1 AND STEVEN L. NAIL* ,t
`
`Contribution from Department of Industrial and Physical Pharmacy, Purdue University, West Lafayette, Indiana 47907, and
`Pharmaceutical Development, Eli Lilly & Company, Indianapolis, Indiana 46285.
`
`Received January 2, 1998. Accepted for publication May 11, 1998.
`
`Abstract I=1 The objectives of this study were to (1) measure the
`effects of freezing rate and mannitol concentration on the physical
`state of freeze-dried mannitol when mannitol is present as a single
`component, (2) determine the relative concentration threshold above
`which crystalline mannitol can be observed by X-ray powder diffraction
`in the freeze-dried solid when a variety of noncrystallizing solutes are
`included in the formulation, and (3) measure the glass transition
`temperature of amorphous mannitol and to determine the degree to
`which the glass transition temperature of freeze-dried solids consisting
`of mannitol and a disaccharide is predicted by the Gordon—Taylor
`equation. Both freezing rate and mannitol concentration influence
`the crystal form of mannitol in the freeze-dried solid when mannitol is
`present as a single component. Slow freezing of 10% (w/v) mannitol
`produces a mixture of the cc and 13 polymorphs, whereas fast freezing
`of the same solution produces the 6 form. Fast freezing of 5% (w/v)
`mannitol results primarily in the i3 form. The threshold concentration
`above which crystalline mannitol is detected in the freeze-dried solid
`by X-ray diffraction is consistently about 30% (w/w) when a second,
`noncrystallizing solute is present, regardless of the nature of the second
`component. The glass transition temperature of amorphous mannitol
`measured from the quench-cooled melt is approximately 13 1C.
`Accordingly, mannitol is an effective plasticizer of freeze-dried solids
`when the mannitol remains amorphous. Glass transition temperatures
`of mixtures of mannitol and the disaccharides sucrose, maltose,
`trehalose, and lactose are well predicted by the Gordon—Taylor
`equation with values of k in the range of 3 to 4.
`
`Introduction
`
`Mannitol is one of the most commonly used excipients
`in freeze-dried pharmaceutical products. One of the rea-
`sons for the widespread use of mannitol is its tendency to
`crystallize from frozen aqueous solutions and the high
`melting temperature of the mannitol/ice eutectic mixture
`(about —1.5 °C). This property promotes efficient freeze-
`drying and a physically stable, pharmaceutically elegant
`freeze-dried solid. However, there have been reports of
`adverse effects of mannitol on stability of drugs as freeze-
`dried solids. Herman et al. reported that the rate of
`hydrolysis of methylprednisolone sodium succinate in the
`freeze-dried solid state is significantly faster when man-
`nitol is used as the bulking agent versus an amorphous
`excipient such as lactose.' This instability of drug in the
`presence of mannitol was attributed at least in part to
`continued crystallization of mannitol from a system which
`is initially only partially crystalline. This can result in
`
`* Corresponding author. email: slnail@pharmacy.purdue.edu.
`phone: (765) 494-1401. fax:(765) 494-6545.
`Purdue University.
`Eli Lilly.
`
`"amplification" of water activity in amorphous regions
`where the drug is located, with subsequent adverse effects
`on stability.2 The physical state of mannitol during and
`after freeze-drying is particularly important in protein
`formulations where mannitol is present as a lyoprotectant.
`Izutsu et al., using three different model proteins, demon-
`strated that recovery of activity is inversely related to the
`degree of crystallinity of mannito1.3.4 In particular, an-
`nealing during freeze-drying—which promotes crystalliza-
`tion—was associated with marked loss of activity after
`freeze-drying of these model systems.
`There is a need for a better understanding of the physical
`chemistry of freeze-drying of mannitol-containing formula-
`tions in order to anticipate and avoid adverse effects of
`mannitol on physical and chemical stability of the freeze-
`dried solid. The purpose of this report is to identify
`formulation and processing factors which influence crystal-
`lization of mannitol when mannitol is present as both a
`single solute and in systems containing a second, noncrys-
`tallizing solute.
`
`Experimental Section
`
`Materials—The materials used in this study were reagent
`grade and were used as received. Mannitol, sucrose, and lactose
`were obtained from J. T. Baker, Inc. (Phillipsburg, NJ). Maltose,
`trehalose, dextran, and lysozyme were purchased from Sigma
`Chemical Co. (St. Louis, MO).
`Preparation and Characterization of Mannitol
`Polymorphs—The three known polymorphs of mannitol were
`prepared using a procedure described by Walter-Levy.5 Ten
`milliliter aliquots of mannitol solutions at concentrations of 0.4,
`0.8, and 1.2 M were placed in separate watch glasses and
`evaporated at room temperature. Upon evaporation, three distinct
`crystal forms were observed. One form, observed primarily at the
`edge of the watch glass, was opaque, looked like lichens, and grew
`vertically to about 7 mm in height. The X-ray diffractogram of
`this material was consistent with the reference diffractogram° for
`the a polymorph (see Figure 1). The second form was observed
`mostly in the center of the watch glass, and crystals were
`translucent with a parallelepiped shape about 6-8 mm long. The
`X-ray powder diffractogram of this form was consistent with the
`reference diffractogram of the /3 polymorph (Figure 1). The third
`form was also translucent, but in the shape of needles in a coarse
`spherulite morphology. The X-ray powder diffractogram of this
`material was consistent with the reference diffractogram of the 6
`form (Figure 1). In general, lower concentrations of mannitol in
`solution favored formation of the 6 form, while higher solution
`concentrations favored formation of the /3 form. The a polymorph
`was observed around the edges of the watch glass.
`Thermal Analysis—Thermal analysis was carried out using
`modulated DSC (Model 2920, TA Instruments, Newcastle, DE).
`Indium and mercury, with melting points of 156.6 °C and —38.83
`°C, respectively, were used for temperature calibration.
`Glass transition temperatures of freeze-dried powders were
`measured by modulated DSC. Samples of freeze-dried powders
`were equilibrated over phosphorus pentoxide for 3 days and
`prepared by forming a powder compact in a punch and die with
`
`© 1998, American Chemical Society and
`American Pharmaceutical Association
`
`50022-3549(98)00001-X CCC: $15.00 (cid:9)
`Published on Web 07/01/1998 (cid:9)
`
`Journal of Pharmaceutical Sciences / 931
`Vol. 87, No. 8, August 1998
`
`Mylan Ex 1048, Page 1
`
`(cid:9)
`(cid:9)
`

`

`.. I (cid:9)
`
`1
`
`111 (cid:9)
`
`I II
`
`110 mW
`
`exo#1
`
`7_
`
`C
`
`C
`
`-0 0
`
`5 10 15 20
`
`exo#2
`
`-20 (cid:9)
`
`0 (cid:9)
`
`B
`
`exo#1
`
`exo#2
`
`40 (cid:9)
`
`20 (cid:9)
`Temperature (°C)
`
`60 (cid:9)
`
`80 (cid:9)
`
`100
`
`* 4
`
`'d; NAjtkiAAritAldit
`
`ii
`
`,OLbut
`
`0 (cid:9)
`
`10 (cid:9)
`
`20 (cid:9)
`Angle (20)
`Figure 2—DSC thermogram of quench-cooled mannitol melt, showing the
`glass transition temperature and two exotherms (A) and X-ray powder
`diffractograms of quench-cooled mannitol melt after the first and second
`exotherm (B).
`
`30
`
`40
`
`Results and Discussion
`
`Studies on Amorphous Mannitol—Attempts to pre-
`pare amorphous mannitol as a single-component freeze-
`dried solid were unsuccessful. Despite freezing by dropwise
`addition of mannitol solution to liquid nitrogen and freeze-
`drying at the lowest attainable temperature, the resulting
`freeze-dried solid was crystalline by X-ray diffraction.
`Preparation of amorphous mannitol by quench-cooling
`of the melt was successful in producing amorphous man-
`nitol, however. The resulting thermogram is shown in
`Figure 2 (a). The thermogram consists of a glass transition
`at about 13 °C, followed by two crystallization exotherms.
`To better characterize the two exotherms, the DSC experi-
`ment was interrupted after each of the exotherms, and
`samples were held at 4 °C until X-ray powder diffracto-
`grams could be measured. As illustrated in Figure 2(b),
`the quench-cooled mannitol melt formed a mixure of the a
`and /3 polymorphs at the first exotherm, which then
`converted to the a polymorph at the second exotherm, as
`indicated by the disappearance of the designated peaks (*)
`at 14.0°, 23.4°, 24.7°, 29.5°, and 38.8° 20.
`The low glass transition temperature of amorphous
`mannitol may help to explain our inability to prepare
`amorphous mannitol as a single-component freeze-dried
`solid. Even if mannitol were amorphous following freeze-
`drying, holding the lyophile for even a brief period of time
`at room temperature would be expected to result in
`crystallization. Slow crystallization during freeze-drying
`cannot be ruled out, however. Even though the shelf
`temperature was controlled at —50 °C, positive control of
`the sample temperature is uncertain due to lateral heat
`
`0 (cid:9)
`
`10 (cid:9)
`
`20 (cid:9)
`
`30 (cid:9)
`
`40 (cid:9)
`
`0
`Angle (20)
`Figure 1—X-ray diffractograms of mannitol polymorphs prepared in this study
`(left panel) and corresponding reference diffractograms (right panel).
`
`10
`
`20
`
`30
`
`40
`
`an approximate inside diameter of 4 mm in a dry nitrogen-purged
`glovebox. A heating rate of 3 °C/min was used with modulation
`of +11°C every 60 s.
`Glass transition temperatures of quench-cooled mannitol/
`sucrose melts were measured by heating mixtures of mannitol/
`sucrose above the melting point, holding for 5 min, and then
`quenching in liquid nitrogen. DSC thermograms were recorded
`at a heating rate of 10 °C/min.
`X-ray Powder Diffraction —A Siemens Krystalloflex diffrac-
`tometer was used with Cu Ka radiation at a voltage of 40 kV and
`a current of 20 mA. Alignment was verified with a silicon standard
`using a reflection at 28.466° 20 before each measurement.
`Samples were prepared by placing powders on a low background
`aluminum powder mount and scanning from 2 to 40° 20 at a rate
`of 0.1° per second.
`Freeze-Drying—Freeze-drying experiments were carried out
`using an FTS Dura-Stop freeze-dryer (FTS Systems, Inc., Stone
`Ridge, NY). Two milliliters of solution was filled into 10 mL serum
`vials, and the vials were placed directly on the shelves of the
`freeze-dryer. Samples were typically frozen for 6 h at —45 °C.
`Primary drying was done at a shelf temperature of —25 °C and a
`chamber pressure of 100 mTorr for 48 h, followed by secondary
`drying at a shelf temperature of 25 °C and a chamber pressure of
`100 mTorr for 12 h. Vials were stoppered under vacuum.
`Two freezing rates were used to determine the effect of freezing
`rate on mannitol crystallization. Slow freezing was carried out
`by placing vials on the shelf of the freeze-dryer and ramping the
`shelf temperature at a rate of 0.2 °C/min from room temperature
`to —45 °C, followed by freeze-drying under the conditions described
`above. Fast freezing was done by placing vials in liquid nitrogen
`and transferring them to a precooled shelf at —45 °C.
`Measurement of Reconstitution Time —Reconstitution time
`of fast-frozen versus slow-frozen vials of freeze-dried mannitol was
`measured by injecting 2.0 mL of sterile water for injection into
`each vial of freeze-dried powder. The water was added along the
`side wall of the vial, and the vial was gently swirled. A blank
`was prepared by adding 2 mL of water to an empty 10 mL vial.
`Each sample was compared with the blank at 30 s intervals, and
`the reconstitution time was recorded as the first interval at which
`the sample and the blank were not distinguishable with respect
`to visual clarity. Five vials each of slow-frozen and fast-frozen
`freeze-dried solid were tested.
`Preparation of Amorphous Mannitol—Two methods were
`attempted for preparation of amorphous mannitol. Solutions of
`5% and 10% mannitol were added dropwise to liquid nitrogen in
`a Dewar flask. The frozen pellets were transferred to precooled
`freeze-dryer shelves at —50 °C and freeze-dried at —50 °C under
`full vacuum for 5 days. In the second method, a mannitol melt
`was quench-cooled by placing mannitol powder in an aluminum
`DSC pan, heating to 200 °C, and holding for 15 min. This sample
`was then quench-cooled in liquid nitrogen externally to the DSC.
`The sample compartment of the DSC was then cooled to —70 °C,
`and the quench-cooled melt was placed back in the instrument.
`The thermogram was then recorded at a heating rate of 10 °C per
`minute.
`
`932 / Journal of Pharmaceutical Sciences
`Vol. 87, No. 8, August 1998
`
`Mylan Ex 1048, Page 2
`
`(cid:9)
`

`

`fA,..AL,„AAjk_
`
`(n
`
`C
`
`0
`
`10
`
`20
`
`30
`
`40
`
`Angle (20)
`Figure 3—X-ray powder diffractograms of freeze-dried mannitol: (A) 5% and
`(B) 10%.
`
`10
`
`20
`
`30
`
`40
`
`Angle (20)
`
`11.J (cid:9)
`
`I
`
`0 (cid:9)
`
`I (cid:9)
`10 (cid:9)
`
`I (cid:9)
`20 (cid:9)
`Angle (28)
`Figure 5—X-ray powder diffractograms of freeze-dried mannitol/sucrose (A)
`and mannitol/lysozyme (B), showing threshold concentration below which
`mannitol remains amorphous.
`
`I (cid:9)
`30 (cid:9)
`
`I
`40
`
`0 (cid:9)
`
`10 (cid:9)
`
`20 (cid:9)
`
`30 (cid:9)
`
`40
`
`Angle (20)
`Figure 4—X-ray powder diffractograms of freeze-dried 10% mannitol frozen
`slowly (A) and fast (B).
`
`transfer from the chamber walls. The glass transition
`temperature of the freeze-concentrated amorphous phase
`is approximately —30 °C, and it is well recognized that
`considerable molecular mobility is present well below the
`glass transition temperature.
`Freeze-Drying of Mannitol as a Single Solute—X-
`ray powder diffractograms of mannitol freeze-dried from
`5% and 10% (w/v) solutions after fast freezing are shown
`in Figure 3. Freeze-drying was carried out as described
`above. The j9 polymorph was formed when 5% mannitol
`solution was freeze-dried, and the 6 polymorph was formed
`by freeze-drying of 10% solutions. This is in contrast to
`the behavior observed when mannitol is crystallized from
`aqueous solutions, where the 13 polymorph tends to form
`from more concentrated solutions.
`The rate of freezing of mannitol solutions also influences
`crystallization behavior. Figure 4 shows X-ray powder
`diffractograms of 10% mannitol solutions frozen slowly and
`rapidly. The slowly frozen solutions resulted in a mixture
`of a and /3 polymorphs, while rapidly frozen solutions
`produced primarily the 6 polymorph.
`Reconstitution time was significantly different between
`fast-frozen and slowly frozen freeze-dried solids. The
`average of five determinations of reconstitution time for
`fast-frozen and slowly frozen samples resulted in average
`reconstitution times of 36 s (SD = 13.4 s) and 78 s (SD =
`26.8 s), respectively. However, this cannot be attributed
`solely to differences in dissolution rates of mannitol poly-
`morphs, since fast freezing would be expected to result in
`
`a higher specific surface area of the freeze-dried solid,
`which would promote more rapid reconstitution.
`Mannitol Crystallization From a Two-Component
`System —To determine factors influencing crystallization
`of mannitol from a two-component system, it is necessary
`to identify the relative concentration threshold below which
`mannitol remains amorphous. Mannitol was freeze-dried
`with several noncrystallizing cosolutes, including sucrose,
`lactose, maltose, trehalose, dextran, and lysozyme in vari-
`ous ratios at a total solids concentration in the starting
`solution of 10% (w/w). It was observed that the relative
`concentration threshold above which crystalline mannitol
`is detected by X-ray diffraction is about 30% (w/w), and
`that this ratio is largely independent of the nature of the
`second solute. Figure 5 illustrates X-ray diffractograms
`of a freeze-dried mannitol/sucrose mixture (a) and a man-
`nitol/lysozyme mixture (b). These diffractograms illustrate
`the extremes of the difference in apparent degree of
`crystallinity between 30:70 and 40:60 ratios. Considering
`the wide range of molecular weights of these cosolutes, it
`appears that weight ratios are more important than mole
`ratios in determining the threshold concentration above
`which crystalline mannitol is observed by X-ray powder
`diffraction.
`Amorphous Mannitol as a Plasticizer of the Freeze-
`Dried Solid —The glass transition temperatures of two-
`component freeze-dried solids were studied by modulated
`DSC in the range of mannitol concentration below which
`crystalline mannitol can be detected by X-ray powder
`diffraction. The effect of mannitol as a plasticizer is clearly
`illustrated by Figure 6, where the glass transition de-
`creases markedly as the relative concentration of mannitol
`increases. The observation of only one glass transition is
`
`Journal of Pharmaceutical Sciences / 933
`Vol. 87, No. 8, August 1998
`
`Mylan Ex 1048, Page 3
`
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`

`

`A
`
`Table 1-Calculated and Measured Glass Transition Temperatures of
`MannitollDisaccharide Mixtures' Using Gordon-Taylor Equation with
`the Best-Fit Value of k
`
`LL
`
`D
`
`20 (cid:9)
`
`40 (cid:9)
`
`60 (cid:9)
`80
`Temperature (°C)
`Figure 6-Modulated DSC thermograms (reversing component) of freeze-
`dried lactose (A), mannitol/lactose (10:90) (B), mannitol/lactose (20:80) (C),
`and mannitol/lactose (30:70) (D), illustrating plasticizing effect of mannitol.
`
`100 (cid:9)
`
`120 (cid:9)
`
`140
`
`70
`
`60
`
`50
`
`9.- 40
`)--
`
`30
`
`20
`
`10
`
`0 0 (cid:9)
`
`•
`
`experimental
`
`G-T fitted (cid:9)
`
`_
`
`0.2 (cid:9)
`
`0.4 (cid:9)
`
`0.6
`
`0.8
`
`10
`
`Weight Fraction of Mannitol
`Figure 7-Glass transition temperature of quench-cooled mannitol/sucrose
`melt vs weight fraction of mannitol (0, experimental; -, Gordon-Taylor
`equation fit).
`
`consistent with a homogeneous amorphous phase. The
`composition dependence of the glass transition temperature
`of a binary mixture can be described by the Gordon-Taylor
`equation,7.8
`
`T g = (w ) T gi + kw 2T g2)/(w + kw 2) (cid:9)
`
`(1)
`
`where 7' is the glass transition temperature of the mixture,
`k is a constant, wi and w2 are weight fractions, and To
`and Tg2 are glass transition temperatures for component
`1 and 2, respectively. The Gordon-Taylor equation as-
`sumes ideal volume-mixing in a binary mixture, which
`means that a mixture is homogeneous and specific volume
`remains constant.? To determine the degree to which the
`Gordon-Taylor equation can be used to describe the
`composition dependence of the glass transition temperature
`of mannitol/disaccharide mixtures over a broad range of
`composition, T, data from quench-cooled mannitol/sucrose
`melts are plotted in Figure 7. Unlike freeze-dried manni-
`tol/disaccharide mixtures, quench-cooling produces an
`amorphous system over the entire range of compositions.
`Curve fitting using the Marquardt -Levenberg algorithm
`results in the illustrated curve, with a best-fit k value of
`3.1 (r2 = 0.98). Use of this k value for freeze-dried solids
`over the composition range for which mannitol remains
`amorphous is predictive of the glass transition temperature
`
`934 / Journal of Pharmaceutical Sciences
`Vol. 87, No. 8, August 1998
`
`lit.9
`T9 (1C)
`
`67b
`
`101c
`
`92b
`
`107'
`
`calcd
`(1C)
`T9
`
`measured T9 range
`(Ic)) (average)
`
`51.4
`42.3
`35.5
`
`78.2
`60.2
`47.9
`
`69.5
`53.9
`43.2
`
`77.1
`57.4
`44.9
`
`62.6-66.2 (64.1)
`51.6-55.1 (53.2)
`44.9-48.9 (46.5)
`36.0-40.5 (38.1)
`101.4-115.6 (107.1)
`78.6-85.2 (82.7)
`60.7-63.3 (62.4)
`45.8-51.0 (48.9)
`88.9-99.2 (94.5)
`63.5-68.6 (66.7)
`52.5-56.2 (54.4)
`41.7-46.7 (43.4)
`106.8-117.0 (112.5)
`78.3-83.6 (81.8)
`56.4-61.1 (58.8)
`43.0-49.3 (47.0)
`
`k
`
`3
`3
`3
`
`4
`4
`4
`
`4
`4
`4
`
`5
`5
`5
`
`sucrose
`m:S (10:90)
`(20:80)
`(30:70)
`lactose
`m:L (10:90)
`(20:80)
`(30:70)
`maltose
`m:M (10:90)
`(20:80)
`(30:70)
`trehalose
`m:T (10:90)
`(20:80)
`(30:70)
`
`a m: mannitol, S: sucrose, L: lactose, M: maltose, T: trehalose. b Midpoint
`value. c Onset value.
`
`of freeze-dried solids, as shown by the data in Table 1.
`Table 1 also lists the measured glass transition tempera-
`tures of disaccharides along with literature values.9 The
`best agreement between measured T s values and calculated
`values is obtained with values of k in the range of 3 to 5
`for all disaccharides listed in Table 1. These are in
`reasonable agreement with values of k reported by Roos
`and Karel8 for frozen solutions of sucrose, lactose, and
`maltose of 4.7, 7, and 6, respectively. The physical
`significance of k is uncertain.
`Practical Considerations-Given the wide use of
`mannitol as an excipient in freeze-dried products, phar-
`maceutical scientists should recognize that the physical
`state of mannitol in the freeze-dried solid is affected by both
`formulation and processing parameters. If mannitol is
`desired as a crystalline component of the formulation, then
`it is important to ensure that the relative concentration is
`high enough to result in a crystalline solid. Below the
`threshold concentration for crystallization, mannitol is an
`effective plasticizer of the lyophilized solid. This could have
`adverse effects on both physical and chemical stability of
`the product as a result of glass transition-associated
`mobility. In addition, the potential for changes in physical
`state of the solid due to different processing parameters
`such as freezing rate should be recognized when carrying
`out process validation studies intended to identify critical
`processing variables.
`
`References and Notes
`
`1. Herman, B. D.; Sinclair, B. D.; Milton, N.; Nail, S. L. The
`effect of bulking agent on the solid-state stability of freeze-
`dried methylprednisolone sodium succinate. Pharm. Res.
`1994, 11, 1467-1473.
`2. Ahlneck, C.; Zografi, G. The molecular basis of moisture
`effects on the physical and chemical stability of drugs in the
`solid state. Int. J. Pharm. 1990, 62, 87-95.
`3. Izutsu, K.; Yoshioka, S.; Terao, T. Decreased protein stabiliz-
`ing effects of cryoprotectants due to crystallization. Pharm.
`Res. 1993, 10, 1232-1237.
`4. Izutsu, K.; Yoshioka, S.; Terao, T. Effect of mannitol crystal-
`linity on the stabilization of enzymes during freeze-drying.
`Chem. Pharm. Bull. 1994, 42, 5-8.
`5. Walter-Levy, L. The Crystalline Varieties of n-mannitol (in
`French). C. R. Acad. Sci. Paris 1968, 267,1779-1782.
`6. Powder Diffraction File, Organic and Organometallic Phases
`
`Mylan Ex 1048, Page 4
`
`(cid:9)
`(cid:9)
`(cid:9)
`

`

`Search Manual; International Centre for Diffraction Data:
`Newtown Square, PA, 1995.
`7. Gordon, M.; Taylor, J. S. Ideal copolymers and the second-
`order transitions of synthetic rubbers. I. Noncrystalline
`copolymers. J. Appl. Chem. 1952, 2, 493-500.
`8. Roos, Y.; Karel, M., Nonequilibrium ice formation in carbo-
`hydrate solutions. Cryo-Lett. 1991, 12, 367-76.
`9. Roos, Y. Melting and glass transitions of low molecular
`weight carbohydrates. Carbohydr. Res. 1993, 238, 39-48.
`
`Acknowledgments
`
`This work was supported by the National Science Foundation
`Industry/University Cooperative Research Center in Pharmaceuti-
`cal Processing. We also gratefully acknowledge the assistance
`provided by Dr. Jan Fang of G. D. Searle for measurement of
`reconstitution times of freeze-dried powders.
`
`JS980001D
`
`Journal of Pharmaceutical Sciences / 935
`Vol. 87, No. 8, August 1998
`
`Mylan Ex 1048, Page 5
`
`