May 2002
`
`Chem. Pharm. Bull. 50(5) 567—570 (2002)
`
`567
`
`Characterization of Physical State of Mannitol after Freeze-Drying:
`Effect of Acetylsalicylic Acid as a Second Crystalline Cosolute
`
`Susana TORRADO and Santiago TORRADO*
`Department of Pharmaceutical Technology, School of Pharmacy, Complutense University; Madrid, Spain.
`Received June 18, 2001; accepted January 16, 2002
`
`Freeze-drying of mixed solutes is a preparative technique widely used in the pharmaceutical industry. The
`presence of an amorphous form or changes in the crystalline form can affect solid state stability. In this work,
`acetylsalicylic acid (AAS) was chosen as a model drug, and was mixed with mannitol, a commonly used bulking
`agent in formulation of tablets. Variations in the final freeze-dried crystalline forms were found after changing
`the ratios of the two co-solutes. Samples were analysed by powder X-ray diffractometry and differential scanning
`calorimetry. A major amorphous form and a minor crystalline dd-mannitol form were produced during the man-
`nitol freeze-drying process. The crystal form of mannitol in the two-component system depended on the
`AAS : mannitol ratio. The AAS was mostly crystalline, regardless of the amount of mannitol present. A major dd-
`mannitol and a minor amorphous form were obtained when AAS was present in a high percentage (75% w/w).
`When AAS percentages of 50 and 25% (w/w) were present during the drying process, the mannitol was found in
`a highly crystalline form.
`
`Key words
`
`acetylsalicylic acid; mannitol; frozen solution; crystallization
`
`The physical characteristics of frozen solutions have been
`studied mainly with aqueous single solutes.1) The complexi-
`ties of solutions used for pharmaceutical formulations re-
`quire information on the physical characteristics of multiple
`component solutions.2)
`When two different substances are mixed in a solution and
`then freeze-dried, some physical changes can affect the crys-
`tallographic characteristics of the final product.2,3) Pikal et al.
`have reported the mutual inhibition of crystallization of man-
`nitol and glycine during lyophilization.3) However, there are
`not many articles which study the effect of a crystalline ac-
`tive substance on the physical state of mannitol forms.
`Mannitol is an excipient broadly used in the freeze-drying
`process,4—8) and its crystalline and amorphous forms have
`been extensively studied.4,6—8) The freeze-drying process can
`produce a partially amorphous and partially crystalline mate-
`rial6); the crystallization of mannitol during this process can
`lead to different anhydrous polymorphs (a, b, and d). The
`existence of crystalline mannitol hydrate obtained during the
`freeze-drying process has been confirmed by Yu et al.7)
`As a result of these changes to the crystal form of manni-
`tol during freeze-drying there is a difference in the physical
`and chemical stability of the freeze-dried solid, which can in-
`duce adverse effects. Therefore, by studying the physical
`chemistry of freeze-dried mannitol containing formulations
`one can anticipate these changes and prevent adverse effects.
`During the freeze-drying process, samples are subject
`to different cycles of freeze and vacuum drying. These
`processes can lead to modifications of the initial crystalline
`form of the drug.4,5,8,9) The effects of freezing rate, tempera-
`ture and mannitol concentration may also induce crystallo-
`graphic changes producing a previously known mannitol
`form.6,7) Yu et al. have shown that mixtures of d (major) and
`b (minor) polymorphs were produced by freeze-drying a 4%
`w/v pure mannitol solution.7) Nevertheless, other authors
`when employing a fast freezing process favored the forma-
`tion of b form when 5% (w/v) mannitol solution was freeze-
`dried.6) The difficulties which arise when attending to control
`these variables (as temperature and cooling rate) during the
`
`freeze-drying process are shown by Yu et al.,7) who identify
`several factors that can contribute to the vial-to-vial variation
`in the amount and stability of the mannitol hydrate.
`In the current study, acetylsalicylic acid (AAS) is used as
`model crystalline drug. Possible changes in crystallinity of
`AAS can be easily detected because its crystallographic char-
`acteristics are well known.10,11)
`Therefore, the objectives of this study are to (1) Identify
`and measure the different crystalline mannitol forms ob-
`tained when a second crystallizing solute is present. (2) De-
`termine the effect of a similar to industrial scale freeze-dry-
`ing process conditions on the physical state of freeze-dried
`mannitol present as a single component.
`
`Experimental
`Materials AAS USP 23 (Merck, Darmstadt, Germany). b Mannitol NF
`(ICI Americas Inc., U.S.A.). d-mannitol was obtained using a procedure de-
`scribed by Kim et al.6) All other ingredients were of reagent grade or better
`(Merck).
`Methods. Powder Samples Physical Mixtures: Samples of AAS and
`b-mannitol (weight ratios 100 : 0, 75 : 25, 50 : 50, 25 : 75 and 0 : 100) were
`prepared and used as references for the different analytical methods.
`Recrystallized Formulations: AAS and mannitol powder mixtures were
`prepared at various weight ratios (100 : 0, 75 : 25, 50 : 50, 25 : 75 and 0 : 100).
`All these formulations were dissolved in water as the solvent. The freeze-
`drying process was carried out using a Liolabor 7 (Telstat, Inc., Madrid,
`Spain). The freeze-drying vials were of 20 ml capacity with 24.7 mm i.d.,
`different samples were prepared from a 5% w/v. The fill volume was
`2.0 ml/vial. A fast freezing step at 240 °C was employed for the different
`samples and approximately 2 h were required to reach 240 °C. An extended
`primary drying was carried out at a shelf temperature of less than 220 °C
`for 40 h, followed by secondary drying at a shelf temperature of 25 °C for
`6 h. The chamber pressure was set to 100mm of Hg throughout the drying
`process.
`Key Formulations Physical Mixtures: PM-100 : 0, PM-75 : 25, PM-
`50 : 50, PM-25 : 75 and PM-0 : 100.
`Recrystallized Formulations: R-100 : 0, R-75 : 25, R-50 : 50, R-25 : 75 and
`R-0 : 100.
`Differential Scanning Calorimetry Samples of 2—6 mg in covered
`aluminum pans were heated from 280 to 300 °C at the rate of 10 °C/min
`under nitrogen purge, with an empty, covered aluminum pan as the reference
`in a DSC Mettler TA 4000. Temperature was calibrated using indium as
`standard.
`The differential scanning calorimetry (DSC) technique may be used to
`
`* To whom correspondence should be addressed.
`
`e-mail: torrado1@farm.ucm.es
`
`© 2002 Pharmaceutical Society of Japan
`
`Mylan Ex 1050, Page 1
`
`

`

`568
`
`Vol. 50, No. 5
`
`confirm the existence of different crystalline forms of mannitol or to study
`the possible presence of amorphous phases of both substances.2,12,13)
`The crystallinities of AAS and mannitol in each freeze-dried formulation
`were obtained from the heat absorption at the melting temperatures and re-
`lated to the corresponding fully crystalline raw materials. A mathematical
`program (PeakFit®) based on deconvolution curves was used to calculate the
`enthalpy of fusion.
`Powder X-Ray Diffraction The structural characterization included
`conventional q–2q powder X-ray diffraction (Philips X’Pert-MPD) (CAI X-
`ray diffraction, School of Pharmacy, UCM) of all samples under study. Mea-
`surements were carried out with 2q 5—40°, using a step size of 0.04° (2q)
`and 1 s/step.
`Two peaks at 7.86° and 15.87° for 2q, corresponding to reflections 100
`and 002, respectively, were used to detect changes in the AAS crystallinity
`after the freeze-drying process.
`Several peaks were used to detect the crystalline forms of D-mannitol.
`
`Fig. 1. Powder X-Ray Diffraction Patterns
`a) AASStandard, b) R-100 : 0, c) R-75 : 25, d) R-50 : 50, e) R-25 : 75 and f) R-0 : 100.
`
`Results and Discussion
`Evaluation of the Crystalline Characteristics of Mono-
`component Formulations (AAS or Mannitol) after Freeze
`Drying Samples of formulations containing pure AAS ob-
`tained by freeze-drying were evaluated and compared to a
`reference AAS by powder X-ray diffractometry (Fig. 1). The
`relative intensity values did not exhibit any change when
`compared with the diffraction patterns of the AAS crystalline
`forms according to the ASCIT tables (monoclinic crystal,
`PDF number 12-850).
`Figure 2 shows the powder X-ray pattern of a standard b-
`mannitol and a freeze-dried mannitol sample as a single
`component (R-0 : 100) (a, b, and d polymorphs patterns are
`shown for comparison). According to this figure the standard
`mannitol corresponds to the b form while the freeze-dried
`mannitol corresponds to the d form. Variations in the inten-
`sity values can be attributed to the presence of a major amor-
`phous form of mannitol in the freeze-dried sample. The ab-
`sence of mannitol hydrate form was confirmed because its
`characteristic peaks at 16.5, 17.9, 25.7 and 27.0° (2q) were
`not present in the assayed samples.7) Most likely, this d man-
`nitol form is obtained when low concentrations of mannitol
`in solution (5% w/v) are employed. A similar result was ob-
`tained by Haikala et al., when a 5% mannitol solution was
`freeze-dried.4)
`The DSC results of b and d mannitol standard forms and a
`freeze-dried mannitol sample (formulation R-0 : 100) were
`compared, and the melting points were similar for all three
`samples (see Table 1). The enthalpy values of both mannitol
`forms were similar (DH–d mannitol5270.6; DH–b man-
`nitol5273.40 (J/g), therefore only one of these forms
`(b-mannitol) was added to the physical mixtures assayed
`through the deconvolution technique. However, there were
`great differences between the enthalpy values 25.70 J/g for
`mannitol freeze-dried formulation (R-0 : 100) and the two
`mannitol reference forms (DH5273.40, 270.60 J/g, see Table
`1). These results confirmed that a major amorphous form and
`a minor d form were obtained under our freeze-dried condi-
`tions. Similar mixtures of crystalline and amorphous manni-
`
`Fig. 2. Powder X-Ray Diffraction Patterns
`a) Standard mannitol and b) freeze-dried mannitol (R-0 : 100). The a, b, and d polymorphs patterns are shown for comparison.
`
`Mylan Ex 1050, Page 2
`
`

`

`May 2002
`
`569
`
`Table 1. Melting Points and Enthalpy of Fusion for AAS and Mannitol in Different Formulations
`
`Product
`
`AASStandard
`R-100 : 0
`R-75 : 25
`R-50 : 50
`R-25 : 75
`b-MannitolStandard
`d-MannitolStandard
`R-0 : 100
`
`Melting point (°C)
`
`AAS
`
`142.1
`139.3
`139.4
`137.8
`135.2
`—
`—
`—
`
`Mannitol
`
`—
`—
`142.6
`150.9
`154.4
`167.4
`166.1
`166.0
`
`DHTotal
`(J/g)
`
`198.00
`167.50
`203.03
`221.61
`235.49
`273.4
`270.6
`25.70
`
`DHAAS
`(J/g)
`
`198.00
`167.50
`193.93
`179.52
`168.18
`—
`—
`—
`
`DHMannitol
`(J/g)
`
`—
`—
`230.32
`263.70
`257.94
`273.4
`270.6
`25.7
`
`tol were obtained with different freeze-dried processes.14)
`Even though the shelf temperature was controlled at 240 °C,
`positive control of the sample temperature is uncertain due to
`lateral heat transfer from the chamber walls.7)
`Evaluation of the Crystalline Characteristics When
`Mixtures of AAS and Mannitol Were Used in the Same
`Formulations The effect of AAS on the crystallinity of
`freeze-dried mannitol was studied by powder X-ray diffrac-
`tometry and DSC.
`The powder X-ray diffraction of formulations AASstandard,
`R-100 : 0, R-75 : 25, R-50 : 50, R-25 : 75 and R-0 : 100 are
`shown in Fig. 1. The peak at 9.6 (2q) suggests the presence
`of a d-mannitol form in all the freeze-dried samples. Figure
`3 shows the DSC scans: (a) shows the melting points of the
`standard mannitol and freeze-dried mannitol, and (b) shows
`that the melting points of AAS and mannitol were signifi-
`cantly close in freeze-dried samples. A deconvolution pro-
`gram was used to evaluate the DSC results in the recrystal-
`lized mixtures of AAS and mannitol. Similar techniques have
`been used previously by Lee et al.15) and Lier et al.16) to
`study mixtures of binary crystalline polymers and determine
`the DSC melt endotherm of individual crystalline substances.
`The measured endotherms were related to the percentage
`crystallinity of each component.
`The enthalpy values of AAS and d-mannitol (R-0 : 100)
`were evaluated by the physical mixtures PM-100 : 0, PM-
`75 : 25, PM-50 : 50, PM-25 : 75 and PM-0 : 100. From a sta-
`tistical point of view, the best fit was obtained when the ex-
`perimental data were fitted to a linear model: measured en-
`dotherm5a1b3crystallinity percent. AAS presented a good
`determination coefficient (r250.9995) for values between
`0—5 mg of AAS. R2 statistics indicates that model as fitted
`explains 99.95% of the variability in measured endotherm.
`The intercept and slope values were 3.41 mJ and 191.0 J/g re-
`spectively. The slope value showed an AAS melt enthalpy
`was similar to that of the raw sample (DH5198.00 J/g).
`The mannitol determination coefficient was r250.9988.
`This relationship confirmed the utility of the deconvolution
`technique in the determination of crystallinity percentage for
`a mannitol range of 1—5 mg. The intercept and slope values
`were 2273.10 mJ and 284.71 J/g, respectively. The intercept
`(2273.1 mJ), confirmed the presence of an amount less than
`1 mg. The slope (284.73 J/g) was close to the raw material
`melt enthalpy (DH5273.40 J/g). These results are similar to
`those obtained by Haikala et al.4)
`The DSC scans of AAS reference and the freeze-dried
`sample showed the same pattern. The first endotherm (137—
`
`Fig. 3A. DSC Scans of a) Standard Mannitol and b) Freeze-Dried Manni-
`tol (R-0 : 100)
`
`Fig. 3B. DSC Scans of a) Freeze-Dried AAS (R-100 : 0), b) R-75 : 25, c)
`R-50 : 50 and d) R-25 : 75
`
`Mylan Ex 1050, Page 3
`
`

`

`570
`
`Vol. 50, No. 5
`
`(w/w) of mannitol is required in mixtures with different non-
`crystallizing cosolutes to detect the crystalline phase. Under
`our experimental conditions, the presence of a second crys-
`talline substance provides a heterogeneous nucleation site for
`the crystallization of d-mannitol. The presence of AAS in ra-
`tios of 50 : 50 or 25 : 75 (AAS : mannitol) induce mannitol
`enthalpy values which are similar to the d-mannitol reference
`form. Mannitol was found in a high crystalline form when 50
`or 25% (w/w) of a second, crystallizing solute such as AAS
`was present in the cake.
`
`Acknowledgments This work was supported by a FISS project n°
`99/0118.
`
`References
`1) Brittain H. G., J. Pharm. Sci., 4, 405—412 (1997).
`2) Takeuchi H., Yasuji T., Yamamoto H., Kawashima Y., Chem. Pharm.
`Bull., 48, 585—588 (2000).
`3) Pikal M. J., Dellerman K. M., Roy M. L., Riggin R. M., Pharm. Res.,
`8, 427—436 (1991).
`4) Haikala R., Eerola R., Tanninen V. P., Yliruusi, J., PDA J. Pharm. Sci.
`Technology, 51, 96—101 (1997).
`5) Trovão M. C., Cavaleiro M. V., Pedrosa J., Carbohydr. Res., 309,
`363—366 (1998).
`6) Kim A. I., Akers M. J., Nail S. L., J. Pharm. Sci., 87, 931—933
`(1998).
`7) Yu L., Milton N., Groleau E. G., Mishra D. S., Vansickle R. E., J.
`Pharm. Sci., 88, 196—198 (1999).
`8) Torrado G., Fraile S., Torrado S., Torrado S., Int. J. Pharmaceut., 166,
`55—63 (1998).
`9) Saleki-Gerhardt A., Stowell J. G., Byrn S. R., Zografi G., J. Pharm.
`Sci., 84, 318—323 (1995).
`10) Masaki N., Machida K., Kado H., Yokoyama K., Tohda T., Chem.
`Pharm. Bull., 39, 1899—1901 (1991).
`11) Goczo H., Szabo-Revesz P., Farkas B., Hasznos-Nezdei M., Serwanis
`S. F., Pintye-Hodi K., Kasa P., Eros I., Antal I., Marton S., Chem.
`Pharm. Bull., 48, 1877—1881 (2000).
`12) Carpenter J. F., Pikal M. J., Chang B. S., Randolph, T. W., Pharm. Res.,
`14, 969—975 (1997).
`13) Akkaramongkolporn P., Yonemochi E., Terada K., Chem. Pharm.
`Bull., 48, 231—234 (2000).
`14) Ward G. H., Schultz R. K., Pharm. Res., 12, 773—779 (1995).
`15) Lee J. C., Tazawa H., Ikehara T., Nishi T., Polymer J., 30, 327—339
`(1998).
`16) Lier A. S., Liau W. B., Chier W. Y., Macromolecules, 31, 6593—6599
`(1998).
`17) Herman B. D., Sinclair B. D., Milthon N., Nail S. L., Pharm. Res., 11,
`1467—1473 (1994).
`18) Hancock B., Zografi G., J. Pharm. Sci., 86, 1—12 (1997).
`
`Fig. 4. DSC Scan and Deconvoluted Graphics of AAS : Mannitol 75 : 25
`w/w Obtained by Freeze-Drying Process (R-75 : 25)
`Curve (a) AAS and (b) mannitol.
`
`143 °C) was due to the melting point and the second en-
`dotherm (181—183 °C) to the decomposition of this sub-
`stance.
`Figure 4 shows how this program resolves the DSC scan of
`formulation R-75 : 25. Both curves show
`the best fit
`(r250.995) for two different mathematical models. The first
`curve (a) is sigmoidal and the melting point is 139.4 °C,
`close to that observed (139.3 °C) for R-100 : 0. Curve (b) cor-
`responds to deconvoluted mannitol and shows a Gaussian
`distribution. The fit presents a lower d-mannitol melting
`point (142.6 °C). This may be explained by a plasticizing ef-
`fect of the amorphous mannitol form present in the cake.
`Herman et al.17) considered that these changes can have an
`effect on distribution of water within the freeze-dried matrix.
`However, further work is needed in order to have a better un-
`derstanding of the pharmaceutical significance of these
`changes within freeze-dried solids.
`The AAS and mannitol enthalpies of the different formula-
`tions after the deconvolution process are shown in Table 1.
`There is an increase in the mannitol enthalpy of R-75 : 25
`(230.32 J/g) when compared to R-0 : 100 (25.70 J/g). In sam-
`ple (R-75 : 25), a major crystalline mannitol form and a
`minor amorphous mannitol form were shown to coexist. The
`presence of one phase in another can act as a focal point for
`spontaneous phase transitions such as crystallization.18) Kim
`et al.6) have described how a threshold concentration of 30%
`
`Mylan Ex 1050, Page 4
`
`