Solvent-Mediated Solid Phase Transformations of cArbamazepine: Effects of Simulated Intestinal Fluid and Fasted State Simulated Intestinal Fluid

Solvent-Mediated Solid Phase Transformations of Carbamazepine: Effects of Simulated Intestinal Fluid and Fasted State Simulated Intestinal Fluid PAULA LEHTO,1,2 JAAKKO AALTONEN,1,3 MIKKO TENHO,4 JUKKA RANTANEN,3,5 JOUNI HIRVONEN,1 VELI PEKKA TANNINEN,2 LEENA PELTONEN1 1

Division of Pharmaceutical Technology, P.O. Box 56, FI-00014, University of Helsinki, Finland

2

Orion Corporation, Orion Pharma, P.O. Box 65, FI-02101 Espoo, Finland

3 Drug Discovery and Development Technology Center, P.O. Box 56, FI-00014, University of Helsinki, Helsinki, Finland 4

Department of Physics, University of Turku, FI-20014 Turku, Finland

5

Department of Pharmaceutics and Analytical Chemistry, Faculty of Pharmaceutical Sciences, University of Copenhagen, Universitetsparken, Denmark

Received 31 December 2007; revised 7 April 2008; accepted 30 May 2008 Published online 25 July 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21490

ABSTRACT: Solvent-mediated transformations of carbamazepine (CBZ) anhydrate form III were investigated in Simulated Intestinal Fluid, a simple USP buffer medium, and in FaSSIF, which contains sodium taurocholate (STC) and lecithin, important surfactants that solubilize lipophilic drugs and lipids in the gastrointestinal tract. Raman spectroscopy (in situ) was utilized to reveal the connection between the changes in solid phase composition and dissolution rate while simultaneously detecting the solid state and the dissolved amount of CBZ. Initial dissolution rate was clearly higher in FaSSIF, while the solid phase data revealed that the crystallization of CBZ dihydrate was inhibited in both the dissolution media, albeit by different mechanisms. In SIF this inhibition was related to extensive needle growth, which impeded medium contact with the solid surface by forming a sterical barrier leading to retarded crystallization rates. Morphological changes from the needle-like dihydrate crystals to plate-like counterparts in FaSSIF, combined with the information that the transformation process was leveled off, evidenced strong hydrogen bonding behavior between the CBZ and STC molecules. These results underline the importance of biologically representative dissolution media in linking the in vitro dissolution results of solids that are capable of hydrate formation to their in vivo dissolution behavior. ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 98:985–996, 2009

Keywords: carbamazepine; polymorphs; solvates/hydrates; Raman spectroscopy; dissolution rate; SIF; FaSSIF

INTRODUCTION Correspondence to: P. Lehto (Telephone: 358-10-426-3203; Fax: 358-10-426-2024; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 98, 985–996 (2009) ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association

Drug substances that exist in various solid state forms offer remarkable challenges in pharmaceutical drug product development and manufacturing.1,2 All the more often, new chemical entities

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

985

986

LEHTO ET AL.

have very low aqueous solubility characteristics, which makes the situation even more complicated. In those cases the in vitro dissolution studies provide important information since the dissolution rate of poorly soluble drug substances is usually the rate limiting factor for drug absorption.3 Special challenge for the dissolution estimation is that solvent-mediated phase transformation may occur during the dissolution testing and also in body liquids. The main concern in these transformations is that the crystallization of a stable solid phase can interfere with the dissolution kinetics of the metastable solid phase and deplete the concentration of drug in solution that is available for absorption.4,5 Carbamazepine (CBZ), a poorly soluble compound with a narrow range of therapeutic efficacy, is one example for which the hydrate formation has been reported to be the main cause for the changes in dissolution characteristics and clinical failure.6 Although there have been investigations dealing with the dissolution and bioavailability of CBZ over 20 years, the efforts have failed to find a clear relationship between the in vitro dissolution and in vivo absorption, and further studies are still needed.7,8 The main reason for the poor predictability has been the lack of data that could describe the relationship between the solid state changes and the dissolution process of CBZ. Previous studies have combined off-line solid phase analytics with in vitro dissolution of CBZ, which, however, are incapable of accurately quantifying the conversion kinetics.7–10 Advances in Raman spectroscopy (e.g., fiber optic probes) have made it a superior method for real-time measurements of solid state changes in aqueous surroundings.11,12 However, Raman spectroscopy has not been widely applied in dissolution studies; the first in situ applications of Raman in the area of dissolution were introduced just recently.13 Another reason for incomplete understanding of CBZ dissolution may lie in the fact that instead of using physiologically based dissolution media, the dissolution is often tested in simple buffer solutions. In the cases of poorly soluble drugs, synthetic surfactant solutions, for example 1% sodium lauryl sulphate (SLS) or polysorbate 80 (Tween 80), are advised by the USP.14 While the solubilizing ability of surfactants has been well documented,15,16 little attention has been given to the fact that the synthetic surfactants, as well as the biological surfactants (such as salts of bile), can promote or inhibit the crystal growth processes depending on their solubilizing capacity JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

and interaction potential.17–19 Thus, in order to link the in vitro transformation behavior to in vivo results, the surfactants used in dissolution media should be as representative of in vivo conditions as possible. In this study we compared the dissolutionmediated transformation kinetics of CBZ anhydrate form III in a physiologically relevant dissolution medium [fasted state simulated intestinal fluid (FaSSIF)]20 that uses physiological surfactants, namely sodium taurocholate (STC; salt of conjugated bile acid) and lecithin (phospholipid), and in a simple buffer medium [simulated intestinal fluid (SIF)].14 Simultaneous measurement of drug concentrations in the dissolution medium (with UV–Vis spectrophotometry) and solid phase composition (in situ with Raman spectroscopy) was used to achieve deeper understanding and mechanistic information about the connection between the hydrate formation and intrinsic dissolution rate (IDR).

MATERIALS AND METHODS Materials Carbamazepine anhydrate form III (CBZ form III), from Fabrika Italiana Sintetici (Italy) was used as received. CBZ form III is the stable form below 708C, while in aqueous media or at high relative humidity (greater than 70%) the dihydrate form of CBZ (CBZ dihydrate) is the stable form.7,8 CBZ dihydrate was prepared by recrystallization of the CBZ from ethanol–water mixture. Both the solid forms were confirmed by X-ray powder diffraction. Samples of CBZ dihydrate and CBZ form III forms were stored at room temperature and at 75% and 0% relative humidities, respectively. The dissolution media used were SIF pH 6.8 (USP XXX, 2007, without enzymes) supplied by Reagena (Oy Reagena Ltd, Toivala, Finland) and FaSSIF pH 6.8 prepared according to Galia et al.20 The FaSSIF medium contained 3 and 0.75 mM of STC and egg phosphatidylcholine (lecithin), respectively. STC (extracted from ox, S9875 grade, crude) was purchased from Sigma–Aldrich Chemie GmbH (Munich, Germany) and Lecithin (Lipoid E PCS, n.l.t. 96% pure) was supplied by Lipoid GmbH (Ludwigshafen, Germany).

Dissolution Testing IDR is the dissolution rate for pure compounds with a constant surface area under sink condiDOI 10.1002/jps

SOLVENT-MEDIATED SOLID PHASE TRANSFORMATIONS OF CBZ

tions.14 The apparatus used was the channel flow cell (CFC),21 in which only one surface of the compact is in contact with the dissolution medium. The CFC system was modified with a quartz sight window for the Raman probe.13 Only the data obtained until
5% of CBZ dissolution was used to generate IDR results. When 5% was dissolved, the concentration was
2% of the solubility of CBZ. Compacts were formed by compressing 150 mg of drug powder using a single punch tablet machine (Korsch EK0, Erweka Apparatebau, Germany) with flat-faced punches. The surface area of the compacts was 0.64 cm2. The crystallinity and crystal form of CBZ was checked after compression by XRPD. The volume of the dissolution medium was 500 mL. The medium was continuously circulated (flow rate 9.5 mL/min) through a flow-through cuvette (closed system, 37  0.58C). Quantification was conducted with a UV–Vis spectrophotometer (Ultrospec III, Pharmacia LKB Biotechnology, Uppsala, Sweden) at the wavelength of maximum absorption, lmax, which was determined to be 285 nm. Molar absorptivities were linear over the range of concentrations obtained, indicating compliance with Beer–Lambert law.

987

to dihydrate transformation) and reference peak areas was calculated from Raman shifts 1023– 1044 cm1 and 1000–1044 cm1, respectively, and the results were correlated with the hydrate content (w/w%) in the samples. The spectra were preprocessed by standard normal variate (SNV) transformation.22 The calibration model prepared for quantification purposes produced an acceptable linear function (correlation coefficient (R2) and root mean standard error of prediction (RMSEP) values 0.989 and 4.0%, respectively). X-Ray Powder Diffraction (XRPD) The prepared anhydrate and hydrate forms of the CBZ crystals and compacts (before and immediately after the dissolution tests) were verified by X-ray powder diffraction (Philips X’Pert PRO, PANalytical, Eindhoven, Netherlands). The measurements were conducted using a RTMS X’Celerator detector and Theta–Theta goniometer. The X-ray generator was set to an acceleration voltage of 45 kV and a filament emission of 40 mA. The diffraction patterns were collected over the range of 3–308 (2u) using an aluminium sample holder. Scanning Electron Microscopy (SEM)

Solid State Analysis Raman Spectroscopy Raman spectra were collected using a Raman spectrometer (5 cm1 resolution; Control Development Inc., South Bend, IN) equipped with a fiber optic probe (laser spot size 200 mm, focal length 10 mm; InPhotonics, Norwood, MA). The laser source was an enhanced diode laser system (power 500 mW; Starbright 785 S, Torsana Laser Technologies, DK-Nivaa, Denmark) and the operational wavelength was 785 nm. The Raman signal was measured through a quartz sight window. Spectra were recorded from 200 to 2200 cm1 and each spectrum was composed of three scans with an integration time of 1 s per scan. Spectra were obtained every 300 s over the duration of the experiments. All the results were verified with XRPD. Solid-state quantification samples were prepared in triplicate by geometrically mixing CBZ form III with CBZ dihydrate in 10 or 20% (w/w) increments from 0 to 100%, with a total of 1 g per sample. The sample vial was rotated (50 rpm) during the measurement to increase the sampling volume. A ratio of test peak (distinct anhydrate DOI 10.1002/jps

The morphological changes of CBZ during the transformation were visualized with SEM. SEM micrographs were taken off-line of the compact surfaces before and after 4 and 14 h of CBZ dissolution in SIF, and after 2 and 6 h of CBZ dissolution in FaSSIF (Scanning microscope JSM840A, Jeol, Tokyo, Japan, 2.5 kV beam acceleration voltage). The excess water on the sample surfaces was gently adsorbed using a tissue. The samples were sputter coated with a layer of gold under vacuum. Micrographs were recorded using a digital imaging system (ADDA and Image AnalySIS pro software, Olympus Soft Imaging Solutions GmbH, Mu¨ nster, Germany). Molecular Modeling Interactions between the CBZ and STC/lecithin molecules were studied by docking simulations. All the calculations were performed within MOE modeling environment (Chemical Computing Group Inc., Montreal, Canada, 2007). The heavy atoms of the STC molecule, except the tail of the chain (sulphonyl-group), were fixed after the generation of molecular conformations. The CBZ molecule was moved close to the STC molecule and the geometry was optimized by searching the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

988

LEHTO ET AL.

minimum energy structures (MMFF94x force field). Hydrogen bonding capability between STC and CBZ molecules was inspected from these low energy conformations.

Statistical Analysis Transformation kinetics were evaluated using the Weibull equation and by following the nonlinear mixed effects model23 approach described in Lehto et al.24 (SAS software, version 9.1.3, SAS Institute Inc., Cary, NC). The rate of conversion functions were derived from these models and the descriptive solid conversion parameters were calculated. Since very small changes in the rate parameter could be covered by variation, the rate of 2%/h was chosen as a limit for the progression of conversion. The concentration data were analyzed by moving the fit of a simple linear regression (SAS software, version 9.1.3, SAS Institute Inc.). The series of regression coefficients were considered estimates of the rate of dissolution. Statistical significance was tested using student’s t-test.

RESULTS AND DISCUSSION Analysis of Solid Phase Changes Raman Spectroscopy The solid-state transformation was clearly visible in the region of 1000–1044 cm1 in both the dissolution media (Fig. 1). The change in the background intensity caused by the FaSSIF was insignificant in the chosen spectral region. Due to the poor water solubility of CBZ (0.4 mg mL1 at 258C for CBZ dihydrate),19 the solvent-mediated conversion was very slow (
7 h (SIF), Tab. 1). It was 20-fold slower as reported previously in the case of nitrofurantoin and 200-fold slower than the hydrate formation of highly soluble theophylline [6–8 mg mL1 (258C)25]13. X-Ray Diffraction (XRPD) Phase changes were verified with XRPD. X-ray diffractograms of CBZ form III and dihydrate form compacts before and CBZ form III compacts recovered after 19 h into the dissolution experiments are presented in Figure 2. Characteristic CBZ form III peaks at 13.1, 15.4, 15.9 and 17.28 (2u) are exemplified in the diffractograms (indiJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

Figure 1. Raman spectra of (a) calibration samples and measured during the dissolution in (b) SIF and in (c) FaSSIF. The test peak (1023–1044 cm1) and the reference peak (1000–1044 cm1) used in the quantification are illustrated.

cated by asterisks in Fig. 2). Based on the analysis performed using known crystal structures of CBZ form III and CBZ dihydrate, and utilizing Rietweld refinement, approximately one-half anhydrate residues existed in both the media. The accuracy of these measurements was, it should be noted, limited by the off-line nature of the measurements. However, since the conversion of CBZ was very slow, no remarkable phase changes were probably induced between the sampling and the X-ray measurement. The peak intensity comparison was complicated by the significant preferred orientation effect,26 which indicated that specific planes were more preferred DOI 10.1002/jps

SOLVENT-MEDIATED SOLID PHASE TRANSFORMATIONS OF CBZ

989

Table 1. Descriptive Solid Transformation (From CBZ III to CBZ Dihydrate and Liquid State Parameters for SIF and FaSSIF) Solid State Parameters CBZ form III Tinductiona (h) Tsteady stateb (h) Tslc (h) Unconverted CBZ(A) (%) Max. rate (%/h)

SIF [Mean, (95% CI)]

FaSSIF [Mean, (95% CI)]

2.8 14.8 8.5 (7.8, 9.3) 26.3 (25.4, 27.2)

0.8 7.4 3.5 (2.5, 4.5) 63.4 (62.6, 64.1)

8.1 (5.2, 11.0)

7.0 (3.5, 10.4)

Liquid State Parameters

SIF [Mean, (SD)]

FaSSIF [Mean, (SD)]

CBZ form III IDRindd (mg cm2 h1) IDRsteadye (mg cm2 h1) Tlinearizedf (h)

0.29 (0.04) 0.19 (0.02) 13.0

0.79 (0.14) 0.26 (0.04) 7.4

CBZ dihydrate IDR (mg cm2 h1)

0.22 (0.01)

0.27 (0.00)

a

Tinduction, duration of induction time. Tsteady state, start of steady state stage of conversion. c Tsl, start of conversion rate slowdown. d IDRind, intrinsic dissolution during induction period. e IDRsteady, intrinsic dissolution rate during steady state stage. f Tlinearized, start of linearized phase of intrinsic dissolution rate. b

in the studied crystals than the planes in typical randomly orientated CBZ dihydrate crystals. Based on the solved crystal structure of CBZ dihydrate, the most texturized plane in SIF [corresponding to the intense peak at 12.38 (2u)] was the plane (0 2 0). In FaSSIF, however, the most intense peak was discovered at 19.58 (2u), corresponding to the crystal plane (1 1 1). These results indicate that the habit of CBZ dihydrate crystals was different among the dissolution media.

Effects of Solid Phase Changes on the Intrinsic Dissolution Rate Simple Buffer Medium: SIF During the transformation stage the dissolution rate of initially anhydrous compact decreased (from 0.29 to 0.19 mg cm2 h1) due to the increasing relative amount of the dihydrate form (Fig. 3 and Tab. 1). The decrease in rate was, however, smaller than could be presumed according to the previously reported threefold higher

Figure 2. X-ray diffractograms of (a) dry CBZ form III compact, (b) dry CBZ dihydrate compact and CBZ form III compacts after 19 h of dissolution (c) in SIF and (d) in FaSSIF.  Characteristic CBZ form III peaks at 13.1, 15.4, 15.9, and 17.28 (2u) are indicated by in the diffractograms c and d. DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

990

LEHTO ET AL.

solubility of CBZ form III.19 During the first few hours, the dissolution rate of dihydrate compact was almost equal to the dissolution rate of initially anhydrous compact. This is a somewhat exceptional result since, thermodynamically, the anhydrous forms are more active below the transformation temperature, and they should be dissolving more rapidly than the hydrated forms.4 Morphological examination showed several layers of needle-like CBZ dihydrate growth on the surface of the dissolving CBZ form III compacts (Fig. 4a–c). It is highly probable that the changing surface area may have more than fully compensated for the greater thermodynamic activity of CBZ form III. The visible increase in particle size due to almost immediately started needle-like growth, for example, obviously decreased the effective surface area. In addition, Raman measurements indicated a 25% unconverted CBZ form III portion at the steady state stage of conversion (Fig. 5). Based on these observations, it is suggested that the extensive needle-like crystal

growth on the surface of the compacts may have impeded the SIF contact with the solid CBZ form III and resulted in a decreased dissolution rate and, correspondingly, retarded the growth of further needles. These hypotheses agrees well with the results of other recent studies.7,27,19,10 The effect of wettability related issues and kinetic factor(s), such as growth in the crystal size has, for example, been presented as most important parameter(s) controlling the dissolution rate. Furthermore, Tian et al.10 reported that an extensive needle growth may retard the transformation; a plateau of
50% conversion to CBZ dihydrate form after 210 min in aqueous dispersion was observed. Although the spectroscopic results indicate an incomplete conversion, the surface of the compacts might have been more converted. The accessible depth of Raman measurement has been reported to vary from few hundred micrometers (semitransparent materials like tissue) to one millimeter (pharmaceutical tablets),28 which

Figure 3. Intrinsic dissolution results of CBZ form III compacts in SIF. (a) amount of dissolved material (mg cm2) of CBZ form III and (b) dissolution rate (mg cm2 h1). The dissolved material of CBZ dihydrate compacts is presented as a reference. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

DOI 10.1002/jps

SOLVENT-MEDIATED SOLID PHASE TRANSFORMATIONS OF CBZ

991

Figure 4. Scanning electron micrographs taken off-line of the CBZ form III compact’s surfaces (a–c) before and after 4 and 14 h into the dissolution testing in SIF, and (d–f) before and after 2 and 6 h dissolution in FaSSIF.

implies that the Raman was not able to measure only the outermost layers where the dissolution, nucleation and growth occurred. Furthermore, the focus of the Raman measurement was undoubtedly influenced to some extent by the thick CBZ dihydrate growth on the surfaces of the compacts. The discrepancies prevailing between the results of Raman and XRPD may have originated from the fact that the estimated maximum diffraction intensity thickness for X-rays was roughly 1.5 mm while the penetration depth of conventional Raman was around 1 mm. Medium Containing Bile Salts and Lecithin: FaSSIF The rate of dissolution in FaSSIF was at its maximum (0.79 mg cm2 h1 at pretransformation stage) almost threefold higher compared to the corresponding value in SIF (0.29 mg cm2 h1; Fig. 6). Correspondingly, the pretransformation stage which consists of the time (1) to reach the supersaturated stage, (2) to form a stable nucleus, and (3) to grow to a detectable size, was shortened to one-third of the time needed for induction in SIF (Tinduction in Tab. 1 and Figs. 5 and 7). This finding is logical since the higher interfacial concentrations of CBZ, achieved on the dissolving CBZ form III surface due to the faster dissolution DOI 10.1002/jps

rate, accelerated the nucleation and growth of CBZ dihydrate by increasing the probability of intermolecular collisions. The higher dissolution rate in FaSSIF was a result of the surfactantmediated lowering in surface tension (SIF 65.35  2.33 mN/m vs. FaSSIF 39.53  0.57 mN/m), indicating that the dissolution was enhanced by increased wetting.29 Since lecithin caused a strong decrease in the critical micellar concentration (CMC) of STC (from 4 to 0.4 mM), the dissolution enhancement was also attributed to roughly planar STC-lecithin micellar aggregates capable of solubilization of aromatic ring systems such as CBZ.15,30 Unlike in SIF, in FaSSIF the dissolution rate of CBZ form III compacts was more rapid than the dissolution rate of CBZ dihydrate (Fig. 6). This diverse dissolution behavior of CBZ form III in presence and absence of STC and lecithin can be understood from the morphological studies, which showed that no needle-like CBZ dihydrate crystals were formed in FaSSIF and, principally, only plate-like shaped clusters of CBZ dihydrate were found (Fig. 4d–f). This changed morphology of CBZ dihydrate crystals supports the findings with XRPD, which showed that different crystal planes were preferred in SIF and FaSSIF. Furthermore, XRPD measurements revealed characteristic peaks for CBZ dihydrate and CBZ III forms and JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

992

LEHTO ET AL.

Figure 5. Solid phase conversion results obtained during the intrinsic dissolution tests of CBZ form III compacts in SIF. (a) amount of dihydrate (%) and (b) conversion rate (%/h).

no extra peaks were to be found (Fig. 2d). Thus, with the sensitivity of XRD method no evidence was found that the STC and/or lecithin molecules were incorporated within the CBZ dihydrate crystal lattice. Although the initial dissolution rate was greater and the onset of crystallization was shorter in FaSSIF, the maximum rate of transformation was almost equal in both the media. This, in addition to the changed morphology of CBZ dihydrate crystals, indicates that the surfactant-mediated increase in solubility was compensated by the intermolecular interactions (e.g., hydrogen bonds, electrostatic attractions) between for example STC and the specific crystal faces of CBZ dihydrate, which effectively inhibited the growth of CBZ dihydrate crystals.31,19 It can also be partly attributed to the fact that the growth material was JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

running out due to the inclusion of CBZ molecules in surfactant micelles.17 The end of conversion was reached when less than one half of the CBZ form III had been converted to the CBZ dihydrate form (Fig. 7). This is in line with the XRPD results (Fig. 2d). At this time point (t ¼ 7.4 h) the dissolution rate of CBZ form III compact was linear once again (Fig. 6). This dissolution rate was shown to be equal with that of CBZ dihydrate (Tab. 1), albeit the sample was not fully converted. Interactions between STC and CBZ Figure 8a shows the structure of a needle-shaped crystal of CBZ dihydrate with the unit cell. As discussed earlier, SEM micrographs revealed that the habit of CBZ dihydrate crystals was not DOI 10.1002/jps

SOLVENT-MEDIATED SOLID PHASE TRANSFORMATIONS OF CBZ

993

Figure 6. Intrinsic dissolution results of CBZ form III compacts in FaSSIF. (a) amount of dissolved material (mg cm2) of CBZ form III and (b) dissolution rate (mg cm2 h1). The dissolved material of CBZ dihydrate compacts is presented as a reference.

needle-like but plate-like in FaSSIF. XRPD results supported this finding by indicating that the preferred orientation plane (0 2 0) in SIF was changed to plane (1 1 1) in FaSSIF. The orientation of the molecules towards the crystal plane (1 1 1) is illustrated in Figure 8b. Docking simulations revealed that the STC molecule has four hydrogen bond donor groups and three hydrogen bond acceptor groups. A great number of low-energy conformations were observed, in which the distance between the hydroxyl oxygen atoms and sulphur atoms of the sulphonate group ˚ . This enabled the generation of is less than 7 A structures of CBZ and STC via several hydrogen bonds (distance between hydrogen and oxygen ˚) atoms that participate in hydrogen bonding
2 A of which one example is presented in Figure 8c. It is possible this kind of bonding may have occurred between STC and CBZ molecules oriented towards the crystal plane (1 1 1) and these interactions may have (1) impeded the medium contact with the CBZ and slowed down the dissolution, (2) changed the growing habit of DOI 10.1002/jps

CBZ dihydrate from needle-like to plate-like, and (3) finally limited the conversion rate to a constant level. It should, however, be noted that these simulations were conducted using isolated molecules and, thus, only a general idea about the possible interactions between the STC and CBZ is presented. In future, more detailed studies using simulated surfaces should be carried out to fulfill the understanding of the surfactant-mediated inhibition of crystal growth on certain crystal planes. The findings of this study have shown that the complex dissolution behavior of solids that are capable of hydrate formation, such as CBZ, can further be complicated by interactions between drug molecules and physiologically prevailing surfactants. In these situations Raman spectroscopic measurement in situ during intrinsic dissolution testing offers valuable information that can be of use in drug candidate selection as well as in explaining and controlling the behavior of drug substances in the final drug products. However, a few challenges related to the configuration of the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

994

LEHTO ET AL.

Figure 7. Solid phase conversion results obtained during the intrinsic dissolution tests of CBZ form III compacts in FaSSIF. (a) amount of dihydrate (%) and (b) conversion rate (%/h).

measurements need to be resolved to improve the accuracy of the results. For example, although the SEM images indicated that the whole surface of the compact was covered by dihydrate growth,

small unchanged areas may have existed among the growth centers. If the Raman laser spot (diameter 200 mm) was focused partly on the unchanged regions, an unconverted portion was

Figure 8. (a) Structure of a needle-shaped crystal of CBZ dihydrate (¼D) with the unit cell. (b) Orientation of CBZ molecules towards the crystal plane (1 1 1). (c) Possible hydrogen bonds will form between (1) (left) oxygen of the sulphonate group (STC) and hydrogen of the amino group (CBZ) and (2) (right) hydrogen of the hydroxyl group (STC) and oxygen of the carbonyl group (CBZ; colors: red ¼ oxygen, light gray/white ¼ hydrogen). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

DOI 10.1002/jps

SOLVENT-MEDIATED SOLID PHASE TRANSFORMATIONS OF CBZ

detected. Raman microscopy could help to visualize the situation in situ, whether or not the spot is focused on the site where the growth is occurring. The use of Raman probes that can be focused to a much larger area and surface mapping could also enlighten the situation. However, at the moment these tools are not directly usable with dissolution testing indicating the fact that future investigations need to be directed to further development of the measuring configuration.

CONCLUSIONS The present study showed that in situ solid-state characterization during intrinsic dissolution testing contributed to a molecular level understanding of the complex dissolution behavior of CBZ. The dissolution and anhydrate to dihydrate conversion kinetics differed distinctly between the dissolution medium without surfactants (SIF) and the medium containing STC and lecithin (FaSSIF). Regardless of the better wetting and solubilization ability of FaSSIF, the habit of the CBZ dihydrate crystals was changed from needlelike to plate-like and the transformation continued in a constant level. These effects were shown to be related to hydrogen bonding between the STC and CBZ molecules, which can selectively inhibit the incorporation of growth units into the dihydrate crystal lattice. This study has shown that the dissolution evaluation of solids that are capable of hydrate formation is far from simple, and cannot be completely understood by merely measuring the dissolved concentrations. It can also be concluded that in order to be able to relate the in vitro dissolution results of solids that convert to hydrate form in aqueous surroundings to their in vivo behavior, it is essential that the dissolution studies are initiated in a physiologically representative medium.

ACKNOWLEDGMENTS Lars Pietila¨ (Orion Pharma) and Anni Liimatainen are acknowledged for their help with docking simulations and statistical issues, respectively. Dr. Osmo Antikainen (University of Helsinki) is acknowledged for data acquisition.

REFERENCES 1. Bernstein J. 2002. Polymorphism in molecular crystals. Oxford: Oxford Science Publications. DOI 10.1002/jps

995

2. Huang L-F, Tong W-Q. 2004. Impact of solid state properties on developability assessment of drug candidates. Adv Drug Del Rev 56:321–334. 3. Amidon GL, Lennerna¨ s H, Shah VP, Crison JR. 1995. A theoretical basis for a biopharmaceutic drug classification: The correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm Res 12:413–420. 4. Shefter E, Higuchi T. 1963. Dissolution behaviour of crystalline solvated and nonsolvated forms of some pharmaceuticals. J Pharm Sci 52:781– 791. 5. Cardew PT, Davey RJ. 1985. The kinetics of solventmediated phase transformations. Proc R Soc Lond A 398:415–428. 6. Meyer MC, Straughn AB, Jarvi EJ, Wood GC, Pelsor FR, Shah VP. 1992. The bioinequivalence of carbamazepine tablets with a history of clinical failures. Pharm Res 9:1612–1616. 7. Kahela P, Aaltonen R, Lewing E, Anttila M, Kristoffersson E. 1983. Pharmacokinetics and dissolution of two crystalline forms of carbamazepine. Int J Pharm 14:103–112. 8. Kobayashi Y, Shusei I, Shigeru I, Yamamoto K. 2000. Physicochemical properties and bioavailability of carbamazepine polymorphs and dihydrate. Int J Pharm 193:137–146. 9. Kaneniwa N, Ichikawa JI, Yamaguchi T, Hayashi K, Watari N, Sumi M. 1987. Dissolution behaviour of carbamazepine polymorphs. Yakugaku Zasshi (J Pharm Soc Japan) 107:808–813. 10. Tian F, Sandler N, Aaltonen J, Lang C, Saville DJ, Cordon KC, Strachan CJ, Rantanen J, Rades T. 2007. Influence of polymorphic form, morphology and excipient interactions on the dissolution of carbamazepine compacts. J Pharm Sci 96:584– 594. 11. Jørgensen A, Rantanen J, Karjalainen M, Khriachtchev L, Ra¨ sa¨ nen E, Yliruusi J. 2002. Hydrate formation during wet granulation studied by spectroscopic methods and multivariate analysis. Pharm Res 19:1285–1291. 12. Wikstro¨ m H, Marsac PJ, Taylor LS. 2005. In-line monitoring of hydrate formation during wet granulation using Raman spectroscopy. J Pharm Sci 94: 209–219. 13. Aaltonen J, Heina¨ nen P, Peltonen L, Korteja¨ rvi H, Tanninen VP, Christiansen L, Hirvonen J, Yliruusi J, Rantanen J. 2006. In-situ measurement of solvent-mediated phase transformations during dissolution testing. J Pharm Sci 95:2730–2736. 14. USP 30. 2007. The United States Pharmacopoeial Convention, Rockville. 15. Yalkowsky SH. 1999. Solubility and solubilization in aqueous media. New York: Oxford University Press. pp. 262–266. 16. Wiedmann TC, Kamel L. 2002. Examination of the solubilization of drugs by bile salt micelles. J Pharm Sci 8:1743–1764. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

996

LEHTO ET AL.

17. Luhtala S. 1992. Effect of sodium lauryl sulphate and polysorbate 80 on crystal growth of carbamazepine. Acta Pharm Nord 4:85–90. 18. Bujan M, Sikiric M, Filipovic-Vincekovic N, Vdovic N, Garti N, Fu¨ redi-Milhofer H. 2001. Effect of anionic surfactants on crystal growth of calcium hydrogen phosphate dihydrate. Langmuir 17: 6461–6470. 19. Rodriquez-Hornedo N, Murphy D. 2004. Surfactantfacilitated crystallization of dihydrate carbamazepine during dissolution of anhydrous polymorph. J Pharm Sci 93:449–460. 20. Galia E, Nicolaides E, Ho¨ rter D, Lo¨ benberg R, Reppas C, Dressman J. 1998. Evaluation of various dissolution media for predicting in vivo performance of class I and II drugs. Pharm Res 15:698–705. 21. Peltonen L, Liljeroth P, Heikkila¨ T, Kontturi K, Hirvonen J. 2003. Dissolution testing of acetylsalicylic acid by a channel flow method-correlation methods. Eur J Pharm Sci 19:395–401. 22. Barnes RJ, Dhanoa MS, Lister SJ. 1989. Standard Normal Variate Transformation and de-trending of near-infrared diffuse reflectance spectra. Appl Spectrosc 43:772–777. 23. Davidian M, Giltinan DM. 1995. Nonlinear models for repeated measurement data. New York: Chapman & Hall. 24. Lehto P, Aaltonen J, Niemela¨ P, Rantanen J, Hirvonen J, Tanninen VP, Peltonen L. 2008. Simultaneous measurement of liquid-phase and solidphase transformation kinetics in rotating disc

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 3, MARCH 2009

25.

26.

27.

28.

29.

30.

31.

and channel flow cell dissolution devices. Int J Pharm [submitted]. Yalkowsky SH, He Y. 2003. Handbook of aqueous solubility data. Boca Raton, Florida: CRC Press. pp. 394–395. Tenho M, Heina¨ nen P, Tanninen VP, Lehto V-P. 2007. Does the preferred orientation affect the intrinsic dissolution? J Pharm Biomed Anal 12: 1315–1323. Tian F, Zeitler JA, Strachan CJ, Saville DJ, Gordon KC, Rades T. 2006. Characterizing the conversion kinetics of carbamazepine polymorphs to the dihydrate in aqueous suspension using Raman spectroscopy. J Pharm Biomed Anal 13; 40:271–280. Matousek P. 2007. Deep non-invasive Raman spectroscopy of living tissue and powders. Chem Soc Rev 36:1292–1304. Luner PE, Vander Kamp D. 2001. Wetting of bilesalt-lipid dispersions and dissolution media patterned after intestinal fluids. J Pharm Sci 96: 348–359. Naylor LJ, Bakatselou V, Rodriquez-Hornedo N, Weiner ND, Dressman JB. 1995. Dissolution behaviour of steroids in bile salt solutions is modified by the presence of lecithin. Eur J Pharm and Biopharm 41:346–353. Sinclair B. 2000. Interpreeting changes in nucleation and crystal morphology of carbamazepine dihydrate by probing intermolecular interactions between additives and crystal surfaces (Dissertation). University of Michigan, Ann Arbor, MI.

DOI 10.1002/jps