`
`PHYSICO-CHEMICAL CHARACTERIZATION OF
`HYDRATED AND ANHYDROUS CRYSTAL FORMS OF
`AMLODIPINE BESYLATE
`
`J. M. Rollinger* and A. Burger
`
`Institute of Pharmacy/Pharmacognosy, University of Innsbruck, Innrain 52, Josef-Moeller-Haus,
`A-6020 Innsbruck, Austria
`
`Abstract
`
`The antihypertensive drug substance amlodipine besylate crystallizes in two stable crystal forms, an
`anhydrate and a hitherto unknown monohydrate. Both forms have been characterized by thermal
`analysis, X-ray powder diffractometry, FTIR- and FT Raman spectroscopy. Moisture sorption- and
`desorption investigations reveal their unusual physical stability in a broad range of relative
`humidities. The monohydrate forms an isomorphic dehydrate upon dehydration, which was eluci-
`dated by variable temperature X-ray powder diffractometry. Physico-chemical properties as well as
`relative stabilities of the crystal forms are described and discussed based on a comprehensive analyt-
`ical identification, and enable an estimation of practical relevance for manufacturing of amlodipine
`besylate solid dosage forms.
`
`Keywords: amlodipine besylate, FTIR- and FT Raman spectroscopy, isomorphic dehydrate,
`monohydrate, pseudopolymorphism, thermal analysis, variable temperature X-ray
`powder diffractometry, water sorption
`
`Introduction
`
`Amlodipine besylate (rINNM), (–)-3-ethyl-5-methyl-2-(2-aminoethoxymethyl)-4-
`(2-chlorophenyl)-1,4-dihydro-6-methyl-3,5-pyridinedicarboxylate benzenesulphon-
`ate (Fig. 1), belongs to the group of dihydropyridine calcium-channel blockers. It is
`used as the racemate in the management of angina pectoris and hypertension, for ex-
`ample in Norvascfi, Istinfi or Amlorfi [1], and is one of the world(cid:146)s most widely pre-
`scribed cardiovascular drugs [2]. As reported earlier, the tendency of dihydro-
`pyridines to crystallize in more than one crystal form is extensively high [3(cid:150)7]. Both,
`polymorphism and pseudopolymorphism (solvated or hydrated forms) result in dif-
`ferent physical properties of the respective crystal forms. Hence, they affect analytics
`as well as crucial pharmaceutical properties, such as density, morphology [8], relative
`
`*
`
`Author for correspondence: Phone +43-512-507-5308; Fax: +43-512-507-2939;
`E-mail: judith.rollinger@uibk.ac.at
`
`1418(cid:150)2874/2002/ $ 5.00
`
`' 2002 AkadØmiai Kiad(cid:243), Budapest
`
`AkadØmiai Kiad(cid:243), Budapest
`
`Kluwer Academic Publishers, Dordrecht
`
`Merck Exhibit 2239, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`362
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`stability, dissolution rate, solubility [9], and finally the performance of a solid dosage
`form [10].
`The present study deals with the physico-chemical characterization of (–)-amlo-
`dipine besylate crystal forms, an anhydrous form (AnH), a monohydrate (MH) and its
`dehydrated isomorphic form (DeH). Beside their thermoanalytical characterization, a
`specific aim was to investigate their physical stability and structural features with wa-
`ter vapor sorption- and desorption studies, X-ray powder diffractometry at variable
`temperatures, as well as FTIR- and FT Raman spectroscopy.
`
`Fig. 1 Molecular structure of amlodipin besylate
`
`Experimental
`
`Materials and solvents
`
`Available amlodipine besylate (Solvias AG, Basel, Switzerland) consisted of pure
`AnH. This anhydrous crystal form can also be obtained by crystallization experi-
`ments using organic solvents, whereas the MH crystallizes from aqueous solutions.
`Large quantities of the MH are easily obtained by stirring an aqueous suspension of
`the AnH (magnetic stirrer, 900 rpm, 24 h at ambient conditions). All solvents and
`chemicals used for this study were of analytical grade.
`
`Thermal analysis
`
`Hot stage microscopy was performed with a Reichert-Thermovar polarizing micro-
`scope and a Kofler hot-stage (Reichert, Vienna, Austria).
`Differential scanning calorimetry (DSC) was carried out with a DSC-7 (Perkin
`Elmer, Norwalk, CT) using Pyris Software Ver. 2.0 for Windows. Sample masses for
`quantitative analysis were 1 to 3–0.0005 mg (Ultramicroscales UM3, Mettler,
`CH-Greifensee, Switzerland) weighed into perforated aluminum sample pans
`(25 m L). Nitrogen 5.0 (20 mL min(cid:150)1) was used as purge gas. The temperature axis was
`calibrated with benzophenone (m.p. 48.0(cid:176)C) and caffeine (m.p. 236.2(cid:176)C). Enthalpy
`calibration of the DSC signal was performed with indium 99.999% (Perkin Elmer,
`Norwalk, CT). The applied heating rate (HR) was 5 K min(cid:150)1.
`Thermogravimetry (TG) was carried out with a TGA-7 instrument (Perkin
`Elmer, Norwalk, CT) using 50 m L platinum sample pans and a nitrogen purge (nitro-
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`363
`
`gen 5.0, balance purge: 40 mL min(cid:150)1, sample purge: 20 mL min(cid:150)1). The sample mass
`was in the range of 1 to 5–0.0005 mg and the heating rate was 5 K min(cid:150)1. Mass cali-
`bration was performed with 100 mg calibration mass (Perkin Elmer), temperature
`calibration with alumel (magnetic transition temperature 163.0(cid:176)C) and nickel (mag-
`netic transition temperature 354.0(cid:176)C).
`
`X-ray powder diffraction (XRPD)
`
`XRPD patterns were obtained with a Siemens D-5000 X-ray diffractometer (Siemens
`AG, Karlsruhe, Germany) equipped with q /q -goniometer, a G(cid:246)bel mirror (Bruker
`AXS, Karlsruhe, Germany), a 0.15(cid:176) soller slit collimator, and a scintillation counter.
`The patterns were recorded at a tube voltage of 40 kV, and a tube current of 35 mA,
`applying a scan rate of 0.005(cid:176) 2q s(cid:150)1 in the angular range of 2 to 40(cid:176) 2q
`. For variable
`temperature X-ray powder diffraction the samples were stored in a low temperature
`camera (TTK Anton Paar KG, Kat.Nr.57478, Graz, Austria). A heating rate of
`10 K min(cid:150)1 was used to the desired temperature, which was maintained for the analy-
`sis period (from 2 to 40(cid:176), 0.020(cid:176) 2q s(cid:150)1, 31.7 min). The patterns were collected at 0%
`relative humidity (RH) using an air purge, dried over silicagel and phosphorus
`pentoxide before the sample chamber.
`
`FTIR spectra
`
`FTIR spectra were recorded with a Bruker IFS 25 FTIR-spectrometer (Bruker
`Analytische Messtechnik GmbH, Karlsruhe, Germany) connected with a Bruker
`FTIR-microscope (15x Cassegrain-objective and visible polarization). Samples were
`scanned as potassium bromide pellets (diameter 13 mm; 1 mg amlodipine besylate to
`270 mg potassium bromide; pressure 740 MPa) at an instrument resolution of 2 cm(cid:150)1
`in the spectral range from 4000 to 600 cm(cid:150)1 (50 interferograms per spectrum). For
`FTIR-microscopy, small samples were rolled on a zinc selenide window (13(cid:215)2 mm)
`and recorded at an instrument resolution of 4 cm(cid:150)1 (focus diameter 50 m m, 100
`interferograms per spectrum).
`
`FT-Raman spectra
`
`FT-Raman spectra were recorded with a Bruker RFS 100 FT-Raman spectrometer
`(Bruker Analytische Me(cid:223)technik GmbH, Karlsruhe, Germany) equipped with a di-
`ode-pumped Nd:YAG laser (1064 nm) as the excitation source and a liquid nitro-
`gen-cooled high-sensitivity detector. The powder samples were packed into small
`aluminum cups, and the spectra were recorded at an output power of 200 mW (64
`scans at 4 cm(cid:150)1 instrument resolution).
`
`Sorption kinetics
`
`The determination of moisture uptake was studied gravimetrically at 25(cid:176)C and 92%
`RH, using special hygrostates [11] and a below-weighing balance (Mettler semi-mi-
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`364
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`cro balance AT 261, Mettler Instruments AG, CH-Greifensee). The sample mass was
`about 250 to 300 mg. RH in the semi-micro hygrostates was adjusted with dried silica
`gel or phosphorus pentoxide (P2O5, RH 0%), and saturated salt solution of potassium
`nitrate (RH 92%) at 25(cid:176)C.
`
`Results
`
`Physico-chemical characterization of the crystal forms
`
`The most important physico-chemical parameters of the amlodipine besylate crystal
`forms are summarized in Table 1.
`
`Table 1 Physico-chemical data of amlodipine besylate crystal forms
`
`Crystal form
`
`Preparation by
`
`AnH
`
`DeH
`
`MH
`
`Crystallization
`from organic
`solvents
`
`Dehydration of
`MH at 0% RH,
`50 to 70(cid:176)C
`
`Crystallization from
`aqueous solution or
`suspension (AnH)
`
`M.p.:
`
`TM/(cid:176)C
`DSC onset/(cid:176)C
`DSC peak/(cid:176)C
`Enthalpy of fusion/kJ mol(cid:150)1 a
`Entropy of fusion/J mol(cid:150)1 K(cid:150)1 a
`
`Content of water/%
`Characteristic FTIR frequencies/cm(cid:150)1
`
`197.5(cid:150)200
`198
`200
`
`44.1–3.6
`
`93.7–7.6
`
`85.0(cid:150)100
`88
`93
`9.7 b
`26.9 b
`
`<0.1
`
`(cid:150)
`
`3301
`
`3156
`
`1698
`
`1676
`
`(cid:150)
`
`1616
`
`1494
`
`1445
`
`1433
`
`1366
`
`1303
`
`(cid:150)
`
`1265
`
`<0.1
`
`(cid:150)
`
`3305
`
`3063
`
`1690
`
`(cid:150)
`
`1650
`
`1608
`
`1484
`
`1445
`
`1436
`
`1347
`
`1306
`
`1287
`
`1260
`
`70(cid:150)100
`70(cid:150)100
`80(cid:150)110
`60.1–5.4 c
`
`(cid:150)
`
`2.9(cid:150) 3.1
`
`3550(cid:150)3350
`
`3311
`
`3061
`
`1689
`
`(cid:150)
`
`1648
`
`1607
`
`1485
`
`1446
`
`1436
`
`1348
`
`1308
`
`1288
`
`1261
`
`a–95% C.I.; bsingle value; centhalpy of fusion inclusive enthalpy of dehydration
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`365
`
`Thermal analysis
`AnH consists of colorless, small needles and prisms of about 50 to 100 m m length,
`showing a melting interval between 197.5 and 200(cid:176)C (Fig. 2a). Decomposition is vis-
`ible at temperatures >190(cid:176)C indicated by a brown coloration. The melt is thermally
`unstable. Thermogravimetrically, a distinct and continuous mass loss is observed be-
`yond the melting range. Due to the decomposition, the crystallization of polymorphic
`forms from the supercooled melt was not feasible.
`The micro crystalline aggregates of the MH dehydrate and melt simultaneously
`between 70 and 100(cid:176)C. The TG-curve shows a reproducible mass loss of 2.9 to 3.1%
`in this temperature interval (Fig. 2d), which corresponds to one molecule of water per
`molecule amlodipine besylate (theoretical value: 3.08%). Figure 2b shows the
`DSC-trace of a MH sample, containing crystal seeds of the AnH. The endothermic
`dehydration and melting process of the MH is directly followed by the exothermic
`crystallization process of the AnH (inhomogeneous melting). However, the com-
`bined process (dehydration and melting) proceeds homogeneously, if no crystal seeds
`of the AnH are present in the MH sample. Figure 2c shows the corresponding
`DSC-run of the MH. P1 represents the dehydration and fusion reaction. Between P1
`and P2 the substance is amorphous (liquid), which was confirmed by hot stage mi-
`croscopy, FTIR-microscopy, and XRPD. At about 140(cid:176)C an exothermic crystalliza-
`tion of the AnH (P2) and finally its melting at about 190(cid:176)C (P3) take place. The
`enthalpy of fusion and melting range of this AnH are significantly decreased (com-
`pare Fig. 2a and b). This behavior can be explained by a thermal instability of the
`amorphous state, which existed from about 90 to 140(cid:176)C (Fig. 2c, P1 to P2).
`
`Fig. 2 DSC-curve of the AnH (a); DSC-curve of the MH with inhomogeneous melt-
`ing (b); DSC-curve of the MH with homogeneous melting (c); TG-curve of the
`MH (d)
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`366
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`Fig. 3 Isothermal TG-curves of the MH at 40, 50, and 60(cid:176)C, respectively (a);
`DSC-curve of in situ prepared DeH (1 h, 60(cid:176)C) starting with the MH (b)
`
`In addition, isothermal TG investigations were carried out to gain an insight in
`the thermal stability of MH. At temperatures of 25, 30 , and 40(cid:176)C (0% RH) no signifi-
`cant loss of water was observed within 6 h (mass loss <0.2%). However, at 50(cid:176)C and
`60(cid:176)C the crystal water continuously escapes within 90 min and 17 min, respectively
`(Fig. 3a). The obtained products quickly reuptake moisture after opening the
`TG-oven. The process of re- and dehydration is reversible at temperatures below the
`melting range of this crystal form (<75(cid:176)C). Thus, the two processes of dehydration
`and melting could be separated. In analogy to the TG-measurements, MH was kept
`isothermally at 60(cid:176)C (under nitrogen purge) for one hour in an open DSC Alu-pan to
`release its crystal water without destroying the crystal structure. When heated with a
`rate of 5 K min(cid:150)1, last traces of water are removed below 75(cid:176)C (P1). The melting of
`the dehydrated material (DeH) starts at 88(cid:176)C (P2), followed by the crystallization of
`AnH at about 130(cid:176)C (P3) and its decreased melting at about 190(cid:176)C (P4). The
`enthalpy of fusion of DeH is 9.7 kJ mol(cid:150)1, which represents an extremely low value
`for an enthalpy of fusion probably due to an extensive loss in lattice energy while the
`water is released from the crystal lattice.
`
`Sorption and desorption measurements
`
`AnH was stored over 92% RH at 25(cid:176)C for almost 2 months without a significant wa-
`ter uptake (<0.1%, Fig. 4a). On the other hand, MH showed no release of water by
`storage over dried silica gel (0% RH, Fig. 4b), which is in accordance with isothermal
`TG investigations at 25(cid:176)C. Surprisingly, the dehydration occurs within 6 days under
`the same conditions, when phosphorus pentoxide is used as a drying agent (Fig. 4c).
`The rehydration of the previously formed DeH takes only a few minutes, when stored
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`367
`
`Fig. 4 Water sorption and desorption of amlodipine besylate crystal forms at 25(cid:176)C
`
`over 92% RH. This re- and dehydration process is reproducible as demonstrated in
`Fig. 4c. The crystal lattice is essentially maintained during loss or uptake of water,
`which was proved by FTIR-microscopy and X-ray powder diffractometry.
`
`Vibrational spectroscopy
`
`FTIR-spectra of AnH and MH were recorded with the potassium bromide method,
`whereas the spectrum of DeH could only be obtained by FTIR-microscopy. In order
`to dehydrate MH, the sample was purged for 1 h with nitrogen 5.0 at 65(cid:176)C before and
`during recording FTIR-micro-spectrum. As shown in Fig. 5, the spectra of MH and
`DeH are essentially the same, except for the water absorption bands in the
`O(cid:150)H-stretching region between 3550 and 3350 cm(cid:150)1 in the spectrum of MH.
`In Table 1 the most characteristic FTIR frequencies are listed. Small differences
`within the three forms are observed for the (cid:150)NH2 stretching vibration at about
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`368
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`Fig. 5 FTIR-spectra of amlodipine besylate crystal forms. AnH (potassium bromide
`method) (a); DeH (in situ prepared from MH, 1 h, 65(cid:176)C, 0% RH) on zink
`selenide window (b); MH (potassium bromide method) (c)
`
`Fig. 6 FT Raman-spectra of amlodipine besylate AnH and MH
`
`3300 cm(cid:150)1. However extensive differences exist between the spectra of the AnH and
`+ -stretching
`the other forms (MH and DeH), especially in the range of C(cid:150)H(cid:150) and NH 3
`vibrations between 2900 and 3200 cm(cid:150)1, in the C=O stretch of the carbonyl groups
`(1676 to 1700 cm(cid:150)1), as well as in the fingerprint region.
`FT Raman spectra of AnH and MH were recorded at ambient conditions (Fig. 6).
`Reproducible differences can be observed in the aromatic and aliphatic C(cid:150)H stretch-
`ing bands between 3250 and 2750 cm(cid:150)1 (AnH: 3067, 2948 cm(cid:150)1; MH: 3065,
`2943 cm(cid:150)1), and in the range of C=O and C=C stretching vibrations (AnH: 1651,
`1617 cm(cid:150)1; MH: 1647, 1610 cm(cid:150)1). Distinct shifts and intensity differences are found
`in the region of aliphatic and N(cid:150)H bending vibrations (AnH: 1495, 1449 cm(cid:150)1; MH:
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`369
`
`1481, 1451weak cm(cid:150)1) indicating an involvement of the dihydropyridine side chains in
`position 2 and 3 in the different molecular arrangements of AnH and MH. Although
`the differences of the Raman spectra are less striking than FTIR results, highly repro-
`ducible spectra are obtained for these two crystal forms.
`
`X-ray powder diffractometry
`
`X-ray powder diffractograms of AnH and MH (Fig. 7) show distinct differences in the
`positions and relative intensities of the reflections (Table 2), clearly indicating different
`crystal lattices. Unfortunately, no crystal structures of amlodipine besylate have been
`solved up to now because of their small size. In order to record any changes in the crystal
`
`Fig. 7 X-ray powder patterns of amlodipine besylate AnH and MH
`
`Fig. 8 Left side: variable temperature X-ray powder diffraction at 0% RH, starting with
`the MH (bottom); Right side: TG- and DSC-curves of corresponding crystal
`forms
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`370
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`lattice of the MH upon heating, its XRPD pattern between 4 and 15(cid:176) 2q was recorded at
`different temperatures and 0% RH (Fig. 8, left side from bottom up). For correlation, the
`different patterns are shown together with the corresponding TG- and DSC-curves
`(Fig. 8, right side). On heating up MH (70(cid:176)C, 2 h), the crystals release their water essen-
`tially retaining the three-dimensional order of the parent hydrate, which is apparent
`through similarities in XRPD patterns. However, there are some considerable shifts of
`the peak positions of MH toward higher 2q
`(labeled with asterisks in Fig. 8, +0.26(cid:176) and
`+0.48(cid:176) 2q differences, respectively) in the diffraction pattern of DeH, caused by a distinct
`contraction of the corresponding unit cell dimensions. On the other hand, the 2q positions
`between 10 and 15(cid:176) 2q are more or less unaffected by the dehydration process and vary
`only between (cid:150)0.1(cid:176) and +0.1(cid:176) 2q
`. This clearly indicates an anisotropic shrinkage of the
`lattice upon dehydration. Further heating up to 100(cid:176)C leads to the melting of the isomor-
`phic DeH followed by a crystallization process of the melt at about 140(cid:176)C. The resulting
`pattern is characterized by definite peak positions of AnH. For comparison, the powder
`pattern of AnH is depicted on the top of Fig. 8. The obtained results show that the trans-
`formation from DeH to AnH does not occur in the solid state, but via the melt.
`
`Table 2 Two theta positions (2q ), d-spacings (d) and relative intensities (I) of X-ray powder dif-
`fraction patterns of AnH and MH
`
`2q/(cid:176)
`5.907
`
`10.583
`
`11.358
`
`11.705
`
`13.105
`
`13.369
`
`14.368
`
`15.223
`
`AnH
`
`d/¯
`
`14.9499
`
`8.3525
`
`7.7838
`
`7.5542
`
`6.7503
`
`6.6174
`
`6.1593
`
`5.8156
`
`4.5438
`
`I/%
`
`100.0
`
`9.08
`
`9.31
`
`33.45
`
`14.52
`
`4.87
`
`7.67
`
`3.75
`
`4.64
`
`2q/(cid:176)
`4.889
`
`9.691
`
`11.209
`
`13.985
`
`14.283
`
`17.262
`
`18.712
`
`19.346
`
`20.789
`
`MH
`
`d/¯
`
`18.0616
`
`9.1195
`
`7.8871
`
`6.3274
`
`6.1958
`
`5.1328
`
`4.7382
`
`4.5843
`
`4.2693
`
`I/%
`
`100.00
`
`69.83
`
`13.22
`
`76.12
`
`32.36
`
`13.67
`
`22.70
`
`28.08
`
`13.76
`
`19.520
`
`19.654
`
`20.145
`
`21.378
`
`21.967
`
`22.753
`
`23.128
`
`23.445
`
`24.399
`
`25.414
`
`31.006
`
`4.5131
`
`4.4042
`
`4.1529
`
`4.0429
`
`3.9049
`
`3.7912
`
`3.7912
`
`3.6452
`
`3.5018
`
`2.8819
`
`5.79
`
`7.79
`
`5.83
`
`4.29
`
`6.41
`
`11.43
`
`11.43
`
`16.70
`
`9.52
`
`6.79
`
`21.570
`
`21.848
`
`22.371
`
`23.146
`
`23.786
`
`24.099
`
`26.271
`
`26.944
`
`28.193
`
`34.133
`
`4.1164
`
`4.0647
`
`3.9709
`
`3.8396
`
`3.7376
`
`3.6899
`
`3.3895
`
`3.3064
`
`3.1627
`
`2.6246
`
`25.07
`
`68.64
`
`21.60
`
`24.43
`
`14.04
`
`16.50
`
`31.27
`
`19.42
`
`23.61
`
`14.49
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`371
`
`Discussion
`
`XRPD experiments indicate that there is hardly any reorganization of the crystal lat-
`tice during dehydration of MH apart from an anisotropic crystal shrinkage. This sug-
`gests (a) a not very rigid crystal lattice, and (b) the existence of voids in the form of
`channels or planes in MH, which are exclusively occupied by water molecules. Such
`phenomena are often observed in hydrates and solvates, subsumed under the term
`isomorphic desolvate or dehydrate [10, 12]. However, a destructive process upon wa-
`ter loss of amlodipine besylate takes place while heating up MH at ambient condi-
`tions. The dehydration and the collapse of the three dimensional order occur simulta-
`neously. On the other hand, the exposition of MH over phosphorus pentoxide at 25(cid:176)C
`or to elevated temperatures (65 to 70(cid:176)C) at 0% RH (nitrogen 5.0 purge or dried silica
`gel) results in a cooperative departure of water [13, 14] and formation of an isomor-
`phic dehydrate (DeH). The loss of water and the creation of void spaces in crystal lat-
`tice finally lead to a reduced packing efficiency and a less stable crystal lattice [15],
`as demonstrated by the low enthalpy of fusion of DeH (9.7 kJ mol-1). In order to in-
`crease the packing density and to gain a higher stability, two possible lattice compen-
`sations are commonly observed [12]: the relaxation in the sense of a crystal shrinkage
`and/or the incorporation of solvent, which mainly corresponds to an extensively high
`hygroscopicity of the desolvated material. In the case of amlodipine besylate, both
`stabilization processes are clearly visible by variable temperature XRPD- and water
`sorption-desorption investigations. The results indicate that the water molecules are
`tightly held in MH over the whole range of RH at 25(cid:176)C, not showing any tendency for
`an additional uptake at higher RH values or loss of 3% water at 0% RH. Therefore,
`MH is characterized by a stoichiometric relationship of 1:1, amlodipine besylate to
`water. At room temperature extremely dry conditions are required for the dehydration
`of MH, which can only be achieved with phosphorus pentoxide, but not with silica
`gel. At ambient conditions DeH stabilizes its crystal structure within a few minutes
`by reabsorbtion of water from the atmosphere (Fig. 4c).
`
`Conclusions
`
`Amlodipine besylate crystallizes from organic solvents as AnH, which represents the
`thermodynamically stable crystal form of this compound. However, from aqueous
`solutions or suspensions a stable monohydrate (MH) is obtained. Under extreme con-
`ditions (0% RH, 50 to 70(cid:176)C), this pseudopolymorphous crystal form is enabled to
`completely release its water in retaining its crystal lattice and forming an iso-
`morphous dehydrate (DeH). It has been shown that thermal analysis in combination
`with variable temperature XRPD and moisture sorption-desorption measurements are
`valuable tools for the investigation of the dehydration behavior.
`Both, AnH and MH, show a high stability at ambient conditions and are easily
`identified by vibrational spectroscopy and XRPD. These features are requirements
`for a well-directed application of a respective crystal form in solid dosage forms. Be-
`cause of the stability at high moisture conditions and a thermal stability up to 190(cid:176)C,
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`372
`
`ROLLINGER, BURGER: AMLODIPINE BESYLATE
`
`AnH should be the crystal form of choice for manufacturing. In addition, anhydrates
`generally show a higher aqueous solubility and dissolution rate than the respective
`hydrates. Since amlodipine besylate is a slightly soluble drug substance, the usage of
`AnH is also advisable from biopharmaceutical point of view.
`
`* * *
`
`The authors thank Solvias AG, Basel, Switzerland, for supplying (–)-amlodipine besylate. Helpful
`discussions with Dr. Ulrich Griesser are gratefully acknowledged.
`
`References
`
`1 Martindale, The Extra Pharmacopoeia, Royal Pharmaceutical Society, 1999.
`2 RxList: The Top 200 Prescriptions for 2000 by Number of US Prescriptions Dispensed,
`http://www.rxlist.com (2001).
`3 A. Grunenberg, B. Keil and J.-O. Henck, Int. J. Pharm., 118 (1995) 11.
`4 A. Burger and K. T. Koller, Sci. Pharm., 64 (1996) 293.
`5 A. Burger, J. M. Rollinger and P. Br(cid:252)ggeller, J. Pharm. Sci., 86 (1997) 674.
`6 F. Hirayama, M. Honjo, H. Honjo, H. Arima, K. Okimoto and K. Uekama, Eur. J. Pharm. Sci.,
`11 (2000) 81.
`7 J. M. Rollinger and A. Burger, J. Pharm. Sci., 90 (2001) 949.
`8 K. R. Morris, S. L. Nail, G. E. Peck, S. R. Byrn, U. J. Griesser, J. G. Stowell, S.-J. Hwang and
`K. Park, PSTT, 1 (1998) 235.
`9 A. Burger and R. Ramberger, Mikrochim. Acta, 1979 II (1979) 259.
`10 S. Byrn, R. Pfeiffer, M. Ganey, C. Hoiberg and G. Poochikian, Pharm. Res., 12 (1995) 945.
`11 U. J. Griesser and A. Burger, Int. J. Pharm., 120 (1995) 83.
`12 G. A. Stephenson, E. D. Groleau, R. L. Kleemann, W. Xu and D. R. Rigsbee, J. Pharm. Sci.,
`87 (1998) 536.
`13 S. Petit and G. Coquerel, Chem. Mat., 8 (1996) 2247.
`14 C. Habare, S. Petit and G. Coquerel, PhandTA 4, Karlsruhe Workbook KC 5 (1999).
`15 A. I. Kitaigorodsky, Molecular Crystals and Molecules, Academic Press, New York, London
`1973, p. 184.
`16 E. Shefter and T. Higuchi, J. Pharm. Sci., 52 (1963) 781.
`17 A. Burger and U. J. Grie(cid:223)er, Eur. J. Pharm. Biopharm., 37 (1991) 118.
`
`J. Therm. Anal. Cal., 68, 2002
`
`Merck Exhibit 2239, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`