poorly water-soluble mefenamic acid with polyvinylpyrrolidone K12.

http://informahealthcare.com/ddi ISSN: 0363-9045 (print), 1520-5762 (electronic) Drug Dev Ind Pharm, Early Online: 1–11 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/03639045.2014.920024

RESEARCH ARTICLE

Cryomilling-induced solid dispersion of poor glass forming/poorly water-soluble mefenamic acid with polyvinylpyrrolidone K12 Drug Dev Ind Pharm Downloaded from informahealthcare.com by UMEA University Library on 04/06/15 For personal use only.

Naewon Kang1, Jangmi Lee1, Ji Na Choi1, Chen Mao2, and Eun Hee Lee1 1

College of Pharmacy, Korea University, Sejong, Korea and 2Genentech Inc., South San Francisco, CA, USA

Abstract

Keywords

The effect of mechanical impact on the polymorphic transformation of mefenamic acid (MFA) and the formation of a solid dispersion of mefenamic acid, a poor glass forming/poorly-water soluble compound, with polyvinylpyrrolidone (PVP) K12 was investigated. The implication of solid dispersion formation on solubility enhancement of MFA, prepared by cryomilling, was investigated. Solid state characterization was conducted using powder X-ray diffraction (PXRD) and Fourier-transform infrared (FTIR) spectroscopy combined with crystal structure analysis. Apparent solubility of the mixtures in pH 7.4 buffer was measured. A calculation to compare the powder patterns and FTIR spectra of solid dispersions with the corresponding physical mixtures was conducted. Solid state characterization showed that (1) MFA I transformed to MFA II when pure MFA I was cryogenically milled (CM); and (2) MFA forms a solid dispersion when MFA was cryogenically milled with PVP K12. FTIR spectral analysis showed that hydrogen bonding facilitated by mechanical impact played a major role in forming solid dispersions. The apparent solubility of MFA was significantly improved by making a solid dispersion with PVP K12 via cryomilling. This study highlights the importance of cryomilling with a good hydrogen bond forming excipient as a technique to prepare solid dispersion, especially when a compound shows a poor glass forming ability and therefore, is not easy to form amorphous forms by conventional method.

Ball-milling, cryomilling, mefenamic acid, PVP K12, solid dispersion

Introduction One of the most well-known non-steroidal anti-inflammatory drugs is mefenamic acid (MFA) (Figure 1a). Three polymorphic forms of MFA are known: the first crystal structure of MFA I was reported in 1976 by McConnell and Company1. After three decades, the crystal structure of MFA II was reported by Lee et al.2. Recently, the crystal structure of MFA III was reported by SeethaLekshmi and Row3. MFA is a typical example of a BCS class II drug that shows low solubility and high permeability. Techniques such as making solid dispersions with modified starch4, a complex with cyclodextrin5, or a co-solute system using a structural analogue have been attempted to improve the solubility/dissolution rate of MFA6. In addition, the solubility behavior of MFA I and II are of interest due to the improved solubility/dissolution rate by using MFA II, the meta-stable form7,8. Therefore, a method for preparing MFA II and the physical stability/polymorphic transformation from MFA II to MFA I has also been investigated widely to take advantage of the desired properties of the metastable form9,10. Amorphous form can increase solubility of a given compound by as much as 1000 times11. An amorphous form can be prepared by processes such as supercooling the melt, solvent evaporation, Address for correspondence: Eun Hee Lee, PhD, College of Pharmacy, Korea University, 2511 Sejong-ro, Sejong 339-700, Korea. Tel: (+82) 044-860-1620. Fax: (+82) 044-860-1606. E-mail: [email protected]

History Received 9 December 2013 Revised 24 April 2014 Accepted 26 April 2014 Published online 22 May 2014

precipitation from solution or intense milling12. Among them, milling is also a common technique used in the pharmaceutical industry to reduce particle size and thus, increase dissolution rate of poorly soluble compounds13. The milling process can also induce polymorphic transformation, disorders in the crystalline lattice, or amorphization14–16. Transformation from the stable form to the unstable form such as the metastable polymorph or amorphous form by mechanical milling would occur due to the introduction of energy into the system. For example, mechanical stress can increase mechanical energy that can cause local heating; the mechanical impact can cause lattice instability, softening of the lattice vibrations or lattice defects that lead to lattice collapse and, thus, amorphization17–19. Cryomilling, which is also known as cryogenic grinding, has been used to generate amorphous form of pharmaceutical compounds. Normally, cryomilling works better than milling at ambient temperature in terms of amorphization20. Milling, in general, induces disordering process of crystal lattice. However, milling at ambient temperature can also introduce thermally activated restoration process by the heat generated during the milling process. Cryomilling that is conducted in liquid nitrogen can inhibit the restoration process. Descamps et al.21 suggested that milling conducted below Tg can result in amorphization, while milling above Tg can result in polymorphic transformation. Often, amorphous forms prepared by cryomilling exhibit reduced physical stability, as well as distinct polymorph selection behavior as compared to amorphous forms prepared by

2

N. Kang et al.

Drug Dev Ind Pharm, Early Online: 1–11

Drug Dev Ind Pharm Downloaded from informahealthcare.com by UMEA University Library on 04/06/15 For personal use only.

conventional methods, such as undercooling of the melt12,14,22–25. There are several explanation for this phenomena: (1) milling process destroys long-range order but retain short-range order, while quenching process makes molecules more randomly distributed as compared to milling26, or (2) amorphous solids with large surface area induced by the milling process can provide high energy sites for nucleation as well as enhance the rate of crystal growth and thus, show the reduced physical stability27.

Physical instability hinders the use of the amorphous form. In general, solid dispersion is a technique that takes advantage of high solubility while avoiding the physical instability of amorphous form28. Traditionally, solid dispersions are prepared by cooling the melted drug and matrix with various cooling rates depending on the properties of a given compound or evaporating a solution containing both drug and matrix29,30. Recently, alternative ways of stabilizing amorphous materials have been introduced, such as making co-amorphous mixtures containing small molecules or using mesoporous silicon or silica-based carriers31. It is challenging to produce amorphous solid or solid dispersion of MFA by undercooling the melt because MFA melts, sublimes and degrades at a similar temperature; it is equally challenging to produce amorphous MFA by solvent crystallization because MFA has a high melting point indicating high crystallinity. Generally, the solvent evaporation method does not produce a complete X-ray amorphous solid dispersion of MFA32. In light of these challenges, we investigated the formation of a solid dispersion of MFA using polyvinylpyrrolidone (PVP) K12 (Figure 1b)33. We cryogenically milled or ball-milled (BM) MFA I for various lengths of time to investigate the effect of cryomilling or ball-milling on the phase transformation of MFA. Additionally, we prepared solid dispersions by cryogenically milling or ball-milling of MFA I with PVP K12 in different ratios and compared them with their respective physical mixtures. Solid state characterization of these samples was conducted using powder X-ray diffraction (PXRD) and Fourier-transform infrared (FTIR) spectroscopy, and a calculation was conducted to distinguish solid dispersions from physical mixtures from X-ray powder patterns. The structural aspects of solid dispersions were also discussed. The advantage of preparing MFA in an amorphous phase was evaluated through apparent solubility measurements of the systems produced using the cryomilling processes.

Materials and methods Materials

Figure 1. Molecular structures of (a) MFA and (b) a repeating unit of polyvinylpyrrolidone (PVP) K12.

MFA was purchased from Sigma-Aldrich (St. Louis, MO). PVP K12 (mw 2000–3000) was obtained from BASF (BASF Corporation, Seoul, South Korea). Ethanol and N,N-dimethylformamide (DMF) were purchased from Tokyo Chemical Industry Co., Ltd. (Chuo-ku, Tokyo, Japan). Sodium hydroxide and monobasic potassium phosphate were purchased from Sigma-Aldrich (St. Louis, MO). Water was double-distilled and filtered with a Milli-QÕ ultrapure water purification system (Billerica, MA).

Figure 2. Resulting polymorphic forms of MFA prepared by different methods.

DOI: 10.3109/03639045.2014.920024

Preparation of MFA I and MFA II The resulting polymorphic forms of MFA prepared by different methods are illustrated in Figure 2. We started with MFA I as received. Recrystallization from ethanol resulted in MFA I. Recrystallization from dimethylformamide (DMF) resulted in a DMF solvate of MFA; desolvation of the DMF solvate of MFA resulted in MFA I when the solvate was stored at room temperature; however, desolvation resulted in MFA II when the solvate was stored in a 65 C oven for 1 h. The desolvated DMF solvate of MFA was used as the MFA II reference.


Preparation of the cryogenically milled (CM MFA) and ball-milled MFA I (BM MFA) The milling effect was further investigated by extending the milling time of MFA I powder. Samples were grounded for 120 min. At 30, 60 and 120 min, solid samples were taken and analyzed by using PXRD and FTIR spectroscopy. Preparation of the mixtures of MFA and PVP K12 Preparation of cryogenically milled samples A total of 450 mg of solids including MFA and PVP K12 at different ratios was placed in a polycarbonate center cylinder. Figure 3. Powder X-ray diffraction patterns of (a) MFA cryogenically milled for 30, 60 and 120 min, respectively; and (b) MFA ball-milled for 30, 60 and 120 min.

Cryomilling-induced solid dispersion of mefenamic acid

3

Ratios of MFA to PVP K12 included 1/9, 3/7, 5/5, 7/3 and 9/1 (w/w). The solids were cryogenically milled for 20 min using a Spex SamplePrep Freezer Mill model 6770 (Spex CertiPrep Group L.L.C., Metuchen, NJ) with a stainless steel impactor. Milling time was set to 10 min intervals, and total time was set to 20 min. After cryomilling, a cylinder containing the sample was placed in desiccators over P2O5 for 10 min, and the mixtures were then removed for further analysis. Preparation of physical mixtures of cryogenically milled samples Each of the 450 mg MFA or PVP K12 samples was cryogenically milled following the same procedures used for the cryomilled samples. After cryomilling, each powder at different ratios was placed in a scintillation vial and mixed with a spatula for 5 min. The ratios of MFA to PVP K12 included 1/9, 3/7, 5/5, 7/3 and 9/1 (w/w). Preparation of ball-milled samples For each experiment, 300 mg of total solids including MFA and PVP at different ratios was placed in a grinding jar. The ratios of MFA to PVP K12 included 1/9, 3/7, 5/5, 7/3 and 9/1 (w/w). A polystyrene ball was added to the grinding jar. Samples were grounded for 20 min using a

4

N. Kang et al.

heavy-duty Wig-L-Bug grinding mill (New Era Enterprise, Vineland, NJ).


Table 1. Infrared spectra peak assignments for ball-milled MFA, cryogenically milled MFA, and cryomilled MFA: PVP (7:3) (7/3 CM).

Preparation of physical mixtures of ball-milled samples Each of the 300 mg MFA or PVP samples was ball-milled following the same procedure used for ball-milled samples. Then, each powder at different ratios was placed in a scintillation vial and mixed with a spatula for 5 min. The ratios of MFA to PVP K12 included 1/9, 3/7, 5/5, 7/3 and 9/1 (w/w).


Powder X-ray diffraction Powder X-ray diffraction analysis was performed on samples using a D8 ADVANCE with Davinci (Bruker AXS Inc., GmbH, Karlsruhe, Germany). The samples were illuminated with Cu-Ka radiation ( ¼ 1.5418) at a tube voltage of 40 kV and a tube current of 40 mA. The samples were analyzed over a 2y range of 4–40 with an increment of 0.02 at a rate of 6 /min.

Assignments

Cryogenically 7/3 CM (Solid milled MFA Ball-milled Dispersion) (MFA II) MFA (MFA I)

nNH nC¼O

3311 1650

3307 1648

dNH Benzene ring stretching CH deformation overlapped with CC and CO stretching, and COH deformation vibration of COOH CH3 rocking vibrations Ring deformation with CO2 wagging CH out-of-plane bending vibrations

1577 1511 1259

1567 1496 1249

3342 1652 1675 1577 1506 1220

892 777

916 775

917 769

755

744

750

Figure 4. Powder X-ray diffraction patterns of (a) the cryogenically milled mixtures of MFA and PVP and (b) calculated and experimentally obtained powder patterns of the ball-milled mixtures of MFA and PVP at ratios of 1/9, 3/7, 5/5, 7/3 and 9/1 (from top to bottom).

DOI: 10.3109/03639045.2014.920024


Calculated powder patterns

spectrum was obtained by co-adding 128 scans. A smart multibounce horizontal attenuated total reflectance (HATR) sample accessory (Thermo-Nicolet) with a ZnSe cell (International Crystal, Garfield, NJ) was used for these measurements. OMNIC32 (Thermo Fischer Scientific Inc.) was used for the data analysis.

The pure MFA or pure PVP powder pattern was obtained. The powder pattern of the mixture was calculated by obtaining the weighted average of that of the individual pure material. The resulting powder patterns were normalized to compare with the experimentally obtained powder patterns.

5

Calculated FTIR spectra Fourier transform infrared spectroscopy

Each of the pure MFA or pure PVP spectra was obtained. The spectrum of the mixture was calculated by obtaining the weighted average of that of the individual pure material. The resulting spectrum was normalized to compare with the experimentally obtained spectra.


Fourier transform infrared measurements were conducted using a Nicolet 6700 FTIR with a DTGS detector and KBr beam splitter (Thermo Fischer Scientific Inc., Madison, WI). The scan range was set from 650 to 4000 cm1 with a 4 cm1 resolution, and each

Figure 5. Comparisons of calculated and experimentally obtained attenuated total reflectance FTIR spectra of the mixtures of MFA and PVP in the region ranging from 1450 to 1800 cm1: (a) cryomilled (CM) and (b) ball-milled (BM) mixtures at ratios of 9/1, 7/3, 5/5, 3/7, 1/9 (from top to bottom). Theoretical spectra by calculation are shown as dotted lines and experimental spectra are shown as solid lines.

6

N. Kang et al.



Figure 6. Attenuated total reflectance FTIR spectra of the mixtures of MFA and PVP in the region from 3200 to 3600 cm1: (a) cryomilled (CM) and (b) ball-milled (BM) mixtures at ratios of 9/1, 7/3, 5/5, 3/7, 1/9 (from top to bottom).

Apparent solubility measurements The apparent solubility of the mixtures containing MFA/PVP (3/7) prepared by four different methods were measured. Excessive amounts of the mixtures were placed in a tube containing pH 7.4 phosphate buffer. The tube was placed in a shaking incubator (JEIO TECH Co., LTD., Seoul, Korea) at 37 C and shaken horizontally at 200 rpm. 1 mL of aliquots were withdrawn at a predetermined time interval, and filtered using a nylon filter with a pore size of 0.2 mm. The filtered solution was diluted appropriately with pH 7.4 phosphate buffer and analyzed using UV/VIS spectrophotometer (Optizen POP, Megasys Co., LTD., Daejeon, Korea).

Results and discussion Effect of cryomilling and ball-milling on phase transition of MFA I Cryogenically milled versus ball-milled The powder pattern of ball-milled MFA I was similar to that of MFA I (Figure 3). The reduction in peak intensity was observed

for both cryogenically milled and ball-milled samples indicating that some disorder/particle size reduction occurred during the milling process (Figure 3a and b). Interestingly, cryomilling induced a solid state polymorphic transformation from MFA I to MFA II (Figure 3a). The powder pattern of cryogenically milled MFA I was similar to that of MFA II. However, cryomilling for an extended period of time up to 120 min did not induce amorphization or further reduction in peak intensity. Ball-milling for an extended period of time up to 120 min induced neither a polymorphic transformation nor amorphization (Figure 3b). It is noteworthy that the powder patterns remained after 30 min to 120 min of milling, indicating that the milling for the extended time did not induce further reduction in peak intensities. Characterization of MFA with PVP Powder X-ray diffraction Cryomilled MFA with PVP K12 (CM) versus ball-milled MFA with PVP K12 (BM). The mixtures of MFA and PVP with five different compositions were prepared and cryogenically milled


7


DOI: 10.3109/03639045.2014.920024

Figure 7. Comparisons of theoretical and experimentally obtained attenuated total reflectance FTIR spectra of the mixtures of MFA and PVP in the region from 1000 to 1450 cm1: (a) cryomilled (CM) and (b) ball-milled (BM) mixtures at ratios of 9/1, 7/3, 5/5, 3/7, 1/9 (from top to bottom). Calculated spectra are shown as dotted lines and experimental spectra are shown as solid lines.

(CM) and ball-milled (BM). The complete X-ray amorphous patterns of CM samples were observed in the composition when 470% PVP K12 was added (Figure 4a). However, samples containing 50% or 30% PVP showed largely amorphous-like X-ray patterns in which only very small diffraction peaks were observed on the amorphous halo. Samples containing 90% MFA showed a powder pattern that resembled that of MFA II. However, the peak which is characteristic for MFA I, was also observed at 6.3 2 y. It was often difficult to determine whether the amorphous halo came from PVP of the physical mixture or from the solid

dispersion. We compared the calculated powder pattern with the experimentally obtained powder pattern. Figure 4(b) showed that the BM samples containing 450% PVP generated an amorphous halo with greater intensity as compared to an amorphous halo generated by pure PVP, suggesting that the ball-milled samples contain small amounts of solid dispersion. Formation of solid dispersions was confirmed by comparing the powder patterns of the cryogenically milled/ball-milled mixtures to those of the same compositions but physically mixed samples. The calculated powder patterns of PM CM/PM BM samples matched well with the experimentally obtained

N. Kang et al.



8

Figure 8. Comparisons of theoretical and experimentally obtained attenuated total reflectance FTIR spectra of the mixtures of MFA and PVP in the region from 650 to 1100 cm1: (a) cryomilled (CM) and (b) ball-milled (BM) mixtures at ratios of 9/1, 7/3, 5/5, 3/7, 1/9 (from top to bottom). Calculated spectra are shown as dotted lines and experimental spectra are shown as solid lines.

powder patterns. It implies that MFA formed solid dispersions with PVP K12 at certain ratios by cryomilling or milling but PM CM/PM BM samples did not (data not shown). The amorphous halo for cryogenically milled PVP is shown for comparison (Figure 4a). The first amorphous halo for cryogenically milled PVP appeared at lower 2y as compared to the amorphous halo for the cryomilled mixture, and the first halo moved toward the higher 2y as the concentration of MFA increased. This is a typical observation for solid dispersions in which the amorphous halo of the pure polymer looks different from that of solid dispersions containing both the drug and the polymer.

FTIR spectroscopic analysis The FTIR spectra were analyzed to investigate interactions and structural changes occurring in the MFA and PVP solid dispersions. Assignments of MFA FTIR spectra have been investigated in many studies, and we adopted part of assignments from other studies33–36. Some important absorption bands of the spectra of MFA I powder as received, ball-milled MFA and cryogenically milled MFA are summarized in Table 1. The FTIR spectra of CM/BM obtained experimentally and by calculation are shown in Figures 5–8. In the cases of PM CM and PM BM samples, the calculated spectra generally matched with the experimentally obtained spectra (data not shown). However,

DOI: 10.3109/03639045.2014.920024

discrepancies between the calculated spectra and experimentally obtained spectra were clearly seen in the CM (MFA: PVP ¼ 1/9, 3/7, 5/5 and 7/3) and BM (MFA/PVP ¼ 1/9 and 3/7) samples indicating the formation of solid dispersions.


Crystal structure analysis combined with FTIR spectroscopic analysis A crystal structure analysis combined with FTIR spectroscopic analysis was conducted to better understand the state of MFA molecules in solid dispersions. Since FTIR spectrum of MFA III was not obtainable, we limited our discussion to the crystal structure analysis as well as FTIR spectra analysis of MFA I and MFA II. In the solid state, MFA I and MFA II form intermolecular hydrogen bonds through carboxylic acid homodimers. In addition, an intramolecular hydrogen bond exists between the carboxylic acid and the imino moiety. PVP has two possible hydrogen bond acceptors, the carbonyl oxygen and the nitrogen on pyrrolidone ring. MFA has two possible sites, hydrogen bond donors, for interacting with PVP; one is the imino moiety and the other is the carboxylic acid moiety (Figure 1a). One possible interaction between MFA and PVP as the result of milling could occur through hydrogen bonding between the hydroxyl moiety of carboxylic acid of MFA and the carbonyl

Figure 9. Overlaid hydrogen bonding motifs of MFA I (in blue, Cambridge Structural Database (CSD) refcode: XYANAC) and MFA II (in element color, CSD refcode: XYANAC02) generated using Mercury 2.3.

Figure 10. Apparent solubilities of cryogenically milled (CM), ball-milled (BM), cryogenically milled physical (PM CM) mixtures and ball-milled physical (PM BM) mixtures at a ratio of 3/7 (MFA/PVP) and the cryogenically milled mefenamic acid (MFA CM) and the ball-milled MFA (MFA BM).


9

oxygen on the pyrrolidone of PVP (Figure 5). Both MFA and PVP show a strong absorption in the region around 1650 cm1 (Figure 5). Two absorption bands were clearly observed in the 7/3 (MFA/PVP) CM sample (Figure 5a). It would be possible to assign one absorption band for MFA and the other for PVP. The C¼O absorption bands appear at 1650 and 1648 cm1 for MFA I and MFA II, respectively, and 1652 cm1 for the 7/3 CM sample. The difference in wave number is negligible indicating that the carboxylic acid homodimers may be relatively well preserved in solid dispersion. However, the wave number of the C¼O stretching for PVP shifted significantly from 1666 cm1 to 1675 cm1. The significant shift of wave number strongly indicates the formation of new bond. Since the carbonyl oxygen of PVP is a hydrogen bond acceptor, it could form a hydrogen bond with one of the hydrogen bond donors of MFA. Another possible interaction would be the hydrogen bonds between the NH of imino group of MFA and the carbonyl oxygen on the pyrrolidone of PVP. In the crystalline state, the imino moiety is on the same plane with benzoic acid moiety, whereas the 2,3-dimethylphenyl moiety attached to the amino benzoic acid formed torsion angles of 71.66 (65%), 76.09 (35%) for MFA II, and 119.99 for MFA I (Figure 9), and 80.82 for MFA III1–3. Two distinct absorption bands at 1571 cm1 and 3305 cm1 can be assigned to the NH bending and stretching vibrations of MFA (Figures 5 and 6). Both NH bending and stretching vibrations of MFA I and MFA II show significantly different absorption wave numbers. Similarly, the CM samples (MFA/PVP ¼ 1/9, 3/7, 5/5 and 7/3) show new absorption bands appearing at higher wave number 1577 cm1 as compared to the wave number of the 9/1 CM sample (Figure 5). The NH stretching vibration can be clearly seen in region at 3311 and 3346 cm1 for MFA I and II, respectively. Since PVP does not show any distinct absorption band in that region, the experimentally observed spectra are only shown in Figure 6. In the case of the CM samples (MFA/PVP ¼ 1/9, 3/7, 5/5 and 7/3), the distinct trace of the NH stretching vibration from MFA was not observed. Instead, new distinct absorption bands appeared at 3305 cm1. Jabeen et al.36 reported that the absorption frequency changes between MFA polymorphs at 3311 and 3346 cm1 are related to the conformational changes associated with rotation about the N–C bond. It is plausible that the newly formed hydrogen bonding between MFA and PVP in the solid dispersion would lead to the conformational changes associated with rotation about the N–C bond. The 1/9 BM sample did not show a distinct absorption band at 3311 cm1, and the 3/7 BM sample showed a shift of the absorption band


10

N. Kang et al.

from 3311 to 3305 cm1. However, FTIR spectra of the PM CM and PM BM samples clearly showed the absorption bands at 3311 and 3346 cm1. The observation indicates the interaction between the NH of imino group of MFA and the carbonyl oxygen on the pyrrolidone of PVP. Interestingly, the 7/3 CM sample showed the strongest absorption band in that region. For a solid dispersion at a weight ratio of MFA/PVP (7/3), the number of the NH of MFA is approximately equimolar to the number of the C¼O of PVP. This results in the most efficient hydrogen bonding interaction between the NH of MFA and the C¼O of PVP among the ratios we prepared and thus, the strongest absorption band is observed in that ratio. On the basis of this observation, it can be derived that one molecule of MFA interacts with one monomeric unit of PVP. The structural changes from MFA molecules in the crystalline state to MFA molecules in solid dispersion can also be elucidated by comparing the FTIR spectra changes in other regions (Figures 7 and 8). The significant changes in FTIR spectra between MFA I and MFA II were also observed in the regions at 755, 892 and 1259 cm1 (Table 1). The absorption bands in those regions are all related to the CH bending and the CH stretching of 2,3-dimethylphenyl moiety, the benzene ring deformation and the CH3 rocking vibration. The CM samples showed significant downfield shifts in vibrations including CH deformation overlapped with CC and CO stretching, and COH deformation vibration of COOH at 1257 cm1. Comparisons of the crystal structures of MFA I and MFA II show a significant difference in the torsion angles between aminobenzoic acid and 2,3-dimethylphenyl moiety. The torsion angle difference seems to significantly affect the stretching and bending vibrations of 2,3-dimehtylphneyl moiety. This is also the case when MFA forms the solid dispersion with PVP. These spectral changes suggest that the amorphous form of MFA has structural disorder on the 2,3-dimethylphenyl ring with respect to the aminobenzoic acid moiety of MFA, as a result of the newly formed hydrogen bonding between the imino moiety of MFA and the carbonyl oxygen of PVP. Apparent solubility The apparent solubilities of the mixtures are shown in Figure 10. The apparent solubilities of cryogenically milled MFA and ball-milled MFA were used as references. As mentioned previously, cryomilling induces the polymorphic transformation from MFA I to MFA II which is the metastable form. Therefore, the apparent solubility of cryogenically milled MFA is supposed to be higher than ball-milled MFA. The rank order of apparent solubility of the mixtures was CM (high)4BM4PM CM4PM BM4cryomilled MFA4ball-milled MFA. These results agree well with the energy states of the systems we tested. Namely, the solids maintained at the glassy state exhibit higher solubility than those at the crystalline state, and that the metastable form shows a higher solubility than the stable form. BM sample shows higher solubility than PM BM samples indicating that small amounts of solid dispersion were formed. The result is consistent with PXRD and FTIR spectroscopy data (Figure 4b). The apparent solubility of MFA was improved by at least four times as compared to cryogenically milled pure MFA. The solid dispersion dissolved rapidly such that the maximum apparent solubility was reached less than 10 min, and the maximum apparent solubility last for at least 1 h.

Conclusions We showed that MFA formed solid dispersions with PVP K12 via cryomilling method. Four types of mixtures prepared by cryogenical milling of the binary mixtures, ball-milling of the binary mixtures, physical mixing of the cryomilled components,


and physical mixing of the ball-milled components were prepared, and solid state characterizations were conducted. The formation of a solid dispersion can be assessed by comparing the experimentally obtained powder pattern with the weighted average of that of individual components. FTIR spectra analysis clearly indicated the formation of hydrogen bonding between the C¼O moiety of PVP and the NH moiety of MFA. The apparent solubility of MFA was significantly improved by forming cryomilling-induced solid dispersion. In order to form hydrogen bonds between PVP and MFA, it is necessary to juxtapose the interacting components. Mechanical milling can provide a favorable situation where MFA and PVP are brought close to each other. This study highlights the importance of cryomilling with a good hydrogen bond forming excipient as a technique to prepare solid dispersion, especially when a compound shows a poor glass forming ability and therefore, is not easy to form amorphous forms by conventional method.

Declaration of interest The authors report no conflicts of interest. Financial support for this project was provided by the Korea University Research Foundation (Korea University, Republic of Korea).

References 1. McConnell JF, Company FZ. N-(2,3-xylyl) anthranilic acid, C15H15NO2 mefenamic acid. Cryst Struct Commun 1976;5:861–4. 2. Lee EH, Byrn SR, Carvajal MT. Additive-Induced metastable single crystal of mefenamic acid. Pharm Res 2006;10:2375–80. 3. SeethaLekshmi S, Row TNG. Conformational polymorphism in a non-steroidal anti-inflammatory drug, mefenamic acid. Cryst Growth Des 2013;12:4283–9. 4. Vikram A, Firoz S, Kishore D, et al. Enhancement solubility and dissolution of mefenamic acid by modified starch. Int J Pharm Sci Tech 2012;7:85–92. 5. Hladon T, Pawlaczyk J, Szafran BJ. Stability of mefenamic acid in the inclusion complex with b-cyclodextrin in the solid phase. Incl Phenom Macro 1999;35:497–506. 6. Lee EH, Byrn SR, Pinal R. The solution properties of mefenamic acid and a closely related analogue are indistinguishable in polar solvents but significantly different in nonpolar environments. J Pharm Sci 2012;101:4529–39. 7. Romero S, Escalera B, Bustamante P. Solubility behavior of polymorphs I and II of mefenamic acid in solvent mixtures. Int J Pharm 1999;178:193–202. 8. Aguiar AJ, Zelmer JE. Dissolution behavior of polymorphs of chloramphenicol palmitate and mefenamic acid. J Pharm Sci 1969; 58:983–7. 9. Kato F, Otsuka M, Matsuda Y. Kinetic study of the transformation of mefenamic acid polymorphs in various solvents and under high humidity conditions. Int J Pharm 2006;321:18–26. 10. Otsuka M, Kato F, Matsuda Y. Effect of temperature and kneading solution on polymorphic transformation of mefenamic acid during granulation. Solid State Ionics 2004;172:451–3. 11. Huang LF, Tong WQ. Impact of solid state properties on developability assessment of drug candidate. Adv Drug Deliver Rev 2004;56:321–34. 12. Surana R, Pyne A, Suryanarayanan R. Effect of preparation method on physical properties of amorphous trehalose. Pharm Res 2004;21: 1167–76. 13. Jardin L, Williams II RO. Recent advancements in mechanical reduction methods: particulate systems. Drug Dev Ind Pharm 2014;40:289–300. 14. Sheth AR, Lubach JW, Munson EJ, et al. Mechanochromism of piroxicam accompanied by intermolecular proton transfer proved by spectroscopic methods and solid-phase changes. J Am Chem Sco 2005;127:6641–51. 15. Takahashi Y, Nakashima K, Ishihara T, et al. Polymorphism of fostedil: characterization and polymorphic change by mechanical treatments. Drug Dev Ind Pharm 1985;11:1543–63. 16. Miyamae HK, Kawabata T, Yasuda T, et al. X-ray powder diffraction study on the grinding effect of the polymorphs of a

DOI: 10.3109/03639045.2014.920024

17. 18. 19. 20. 21. 22.


23. 24.

25. 26.

novel and orally effective uricosuric agent FR76505. Drug Dev Ind Pharm 1994;20:2881–97. Kubelka PJ. New contributions to the optics of intensely light scattering materials. Part I. Opt Soc Am 1948;38:448–57. Lindemann FA. Uber die berechnung molekularer eigenfrequenzen. Physik Z 1910;11:609–12. Born M, Huang K. Dynamical theory of crystal lattices. Oxford, London: Clarendon Press; 1954. Chieng N, Zujovic Z, Bowmaker G, et al. Effect of milling conditions on the solid-state conversion of ranitidine hydrochloride form 1. Int J Pharm 2006;327:36–44. Descamps M, Willart JF, Dudognon E, et al. Transformation of pharmaceutical compounds upon milling and comilling: the role of Tg. J Pharm Sci 2007;96:1398–407. Patterson JE, James MB, Foster AH, et al. The influence of thermal and mechanical preparative techniques on the amorphous state of four poorly soluble compounds. J Pharm Sci 2005;94:1998–2012. Bhugra C, Shmeis R, Pikal MJ. Role of mechanical stress in crystallization and relaxation behavior of amorphous indomethacin. J Pharm Sci 2008;97:4446–58. Crowley KJ, Zografi G. Cryogenic grinding of indomethacin polymorphs and solvates: assessment of amorphous phase formation and amorphous phase physical stability. J Pharm Sci 2002;91: 492–507. Graeser KA, Strachan CJ, Patterson JE, et al. Physicochemical properties and stability of two differently prepared amorphous forms of simvastatin. Cryst Growth Des 2008;8:128–35. Yonemochi E, Onoue Y, Buckton G, et al. Differences in crystallization behavior between quenched and ground amorphous ursodeoxycholic acid. Pharm Res 1999;16:835–40.


11

27. Zhu L, Wong L, Yu L. Surface-enhanced crystallization of amorphous nifedipine. Mol Pharm 2008;5:921–6. 28. Deng W, Majumdar S, Singh A, et al. Stabilization of fenofibrate in low molecular weight hydroxypropylcellulose matrices produced by hot-melt extrusion. Drug Dev Ind Pharm 2013;39: 290–8. 29. Serajuddin AT. Solid dispersion of poorly water-soluble drugs: early promises, subsequent problems, and recent breakthroughs. J Pharm Sci 1999;88:1058–66. 30. Maniruzzaman M, Rana MM, Boateng JS, et al. Dissolution enhancement of poorly water-soluble APIs processed by hot-melt extrusion using hydrophilic polymers. Drug Dev Ind Pharm 2013; 39:218–27. 31. Laitinen R, Lo¨bmann K, Strachan CJ, et al. Emerging trends in the stabilization of amorphous drugs. Int J Pharm 2013;453: 65–79. 32. Rao ChVP, Nagabhushanam MV. Enhancement of dissolution profile of mefenamic acid by solid dispersion technique. Int J Res Pharm Chem 2011;1:1127–34. 33. Gilpin RK, Zhou W. Infrared studies of the thermal conversion of mefenamic acid between polymorphic states. Vib Spectrosc 2005; 37:53–9. 34. Topacli A, Ide SJ. Molecular structures of metal complexes with mefenamic acid. Pharm Biomed Anal 1999;21:975–82. 35. Bellamy LJ. The infrared spectra of complex molecules. Vol. 1. London: Chapman & Hall; 1975. 36. Jabeen S, Dines TJ, Leharne SA, et al. Raman and IR spectroscopic studies of fenamates – conformational differences in polymorphs of flufenamic acid, mefenamic acid and tolfenamic acid. Spectrochim Acta A 2012;96:972–85.

Steatorrhoea complicating therapy with mefenamic acid.

Letter: Steatorrhoea complicating therapy with mefenamic acid.

Mefenamic acid poisoning and epilepsy.

Mefenamic acid poisoning and epilepsy.

Mefenamic acid: an under-rated antirheumatic?

Mefenamic acid in prevention of premature labour.

Mefenamic Acid induced nephrotoxicity: an animal model.

Nano-proniosomes enhancing the transdermal delivery of mefenamic acid.

Manganese-mefenamic acid complexes exhibit high lipoxygenase inhibitory activity.

Accessing Mefenamic Acid Form II through High-Pressure Recrystallisation.

Mefenamic acid-induced neutropenia and renal failure in elderly females with hypothyroidism.

Abducens palsy after lumbar myelography with watersoluble contrast media.

Cardiovascular reactions to myelography with watersoluble contrast media.

Molecularly imprinted polymer coated magnetite nanoparticles as an efficient mefenamic acid resonance light scattering nanosensor.

Flufenamic acid, mefenamic acid and niflumic acid inhibit single nonselective cation channels in the rat exocrine pancreas.

A comparison of tiaprofenic acid, mefenamic acid and placebo in the treatment of dysmenorrhoea in general practice.

Comparing the effect of mefenamic Acid and vitex agnus on intrauterine device induced bleeding.

In Vitro Cytotoxic Effects of Celecoxib, Mefenamic Acid, Aspirin and Indometacin on Several Cells Lines.

Caffeine potentiation of mefenamic acid-induced lesions in the rat renal medulla.

Drug targeting using non-magnetic and magnetic albumin-globulin mix microspheres of mefenamic acid.

Caffeine potentiates the nephrotoxicity of mefenamic acid on the rat renal papilla.

Whitening Effect of Watersoluble Royal Jelly from South Korea.

Mefenamic acid taste-masked oral disintegrating tablets with enhanced solubility via molecular interaction produced by hot melt extrusion technology.

Dissolution Profile of Mefenamic Acid Solid Dosage Forms in Two Compendial and Biorelevant (FaSSIF) Media.