- Email: [email protected]
Analysis of matrix dosage forms during dissolution testing using raman microscopy
Analysis of Matrix Dosage Forms During Dissolution Testing Using Raman Microscopy MIRIAM HAASER,1 MAIKE WINDBERGS,2 CUSHLA M. MCGOVERIN,3 PETER KLEINEBUDDE,4 THOMAS RADES,1 KEITH C. GORDON,3 CLARE J. STRACHAN1 1
School of Pharmacy, University of Otago, Dunedin 9054, New Zealand
2
School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts 02138
3
Department of Chemistry and MacDiarmid Institute for Advanced Materials and Nanotechnology, University of Otago, Dunedin 9054, New Zealand 4
¨ Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University, 40225 Dusseldorf, Germany
Received 25 January 2011; revised 6 April 2011; accepted 19 April 2011 Published online 10 May 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22609 ABSTRACT: Matrix dosage forms are widely used for sustained drug release. As both the distribution of the matrix components and physical changes during dissolution can impact drug release behavior, a comprehensive investigation of these phenomena is required during matrix development. In this study, Raman microscopy was used to investigate different extrudate formulations in terms of component distribution and structural changes during dissolution testing. Two systems containing the model drug theophylline anhydrate were investigated: a binary system, based on a tripalmitin matrix, and a ternary system, containing tripalmitin and polyethylene glycol. The distribution of the drug and the soluble and insoluble matrix components were mapped during dissolution testing. Although a receding drug boundary was observed, it was not uniformly distant from the matrix edge. The lipid structure remained intact, whereas the water-soluble polymer rapidly dissolved and diffused from the matrix leaving a more extensive network of channels through which the dissolution medium could penetrate and the drug could diffuse. Raman mapping can be considered a useful aid in the direct analysis of multiple matrix components during drug release, and therefore a deeper understanding of factors affecting drug release can be obtained during the development of sustained-release matrices. © 2011 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 100:4452–4459, 2011 Keywords: Raman spectroscopy; Raman mapping; lipids; controlled release; solid state; spatial resolution; solid lipid extrusion; dissolution; theophylline; polyethylene glycol
INTRODUCTION Matrix dosage forms, in which drugs are dispersed in lipidic or polymeric materials, are commonly used for sustained-release drug delivery systems. To obtain the desired drug release profiles from such dosage forms, there has been much research into understanding the mechanisms of drug release, as well as how the release is affected by dosage form changes.1,2 Drug release mechanisms are usually studied by fitting modCorrespondence to: Clare J. Strachan (Telephone: +64-3-4797324; Fax: +64-3-479-7034; E-mail: [email protected]) Miriam Haaser and Maike Windbergs have contributed equally to the manuscript. Journal of Pharmaceutical Sciences, Vol. 100, 4452–4459 (2011) © 2011 Wiley-Liss, Inc. and the American Pharmacists Association
4452
els to data obtained by measuring the drug concentration in solution over time.2 In the case of noneroding matrices, in which the dissolution medium penetrates the matrix and the drug dissolves and diffuses through the matrix, Higuchi kinetics is commonly applied. This approach assumes a linear concentration gradient in the solution between the receding drug boundary in the matrix (the interface between the undissolved drug and dissolution medium) and the outside of the matrix.3,4 In eroding systems, drug release is controlled by erosion of the matrix, which makes the drug surface accessible to the dissolution medium, and drug release is commonly described mathematically with empirical models assuming a single zero-order release profile.5 These and most other modeling approaches assume steady-state
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
diffusion/erosion profiles, and do not consider the lag time and/or burst effect that may be observed at the beginning of drug release. More recently, Laaksonen et al.6 have simulated these three phases of drug release from matrices using a cellular automata approach, in which a two-dimensional grid space represents the release device. Each unit cell of this grid represents water, drug, or polymer/lipid and is connected to four adjacent cells. Different states with different physicochemical parameters can be modeled. This modeling approach is related to percolation theory, in which a percolation threshold is defined as the volume-to-volume ratio at which an interpenetrating network of a component is formed.7,8 Laaksonen et al.6 have shown their modeling to be applicable to both diffusion- and erosion-controlled release mechanisms as well as a combination of the two. Nevertheless, to fully understand drug release from sustained-release matrices, the dosage forms themselves should be directly analyzed. With the recent advent of chemically selective analysis approaches with appropriate spatial resolution, analysis of individual components within matrices has become possible.9,10 Raman microscopy has several features that make it well suited to analyzing matrix dosage forms during dissolution testing.11 First, it is highly chemically selective, and therefore multiple components in the matrix can be resolved.12 In addition, different solid-state forms of the same chemical species can be resolved.13 Second, it is possible to obtain a high spatial resolution (in the order of about 1–2 :m). Third, the technique is relatively insensitive to water, and therefore the matrix components can be examined in the presence of aqueous dissolution medium.10,14,15 Although Raman microscopy has been used to image component distribution in various matrix systems,9,11 to the best of our knowledge, apart from some preliminary work by ourselves,16 the technique has not been used to image how drug and matrix component distributions in oral dosage forms change during the drug release process. In this study, the potential of Raman microscopy for the analysis of matrix dosage forms during drug release was investigated. The specific objectives were to optimize the microscope measurement parameters, and then analyze drug and excipient distribution changes during dissolution testing and relate these changes to the drug release behavior. The matrices analyzed were solid lipid extrudates. The production of these extrudates involves extruding lipid below its melting temperature with solid drug.17,18 In these systems, the drug is dispersed in particulate form.16 As long as the lipid concentration is above its percolation threshold,7 the lipid remains intact during drug release in water or buffered dissolution media, and drug release has been found to follow Higuchi kinetics following an initial burst DOI 10.1002/jps
4453
release.19 Two formulations have been imaged, both containing theophylline as the drug embedded in a matrix. For the first system, the lipid tripalmitin was used alone to form the matrix, whereas the matrix of the second system contained equal parts of tripalmitin and the water-soluble polymer, polyethylene glycol.
MATERIALS AND METHODS Materials Theophylline anhydrate (BASF, Ludwigshafen, Germany) in powdered form was used as the model drug. The powdered lipid was the pure monoacid R ) provided by triglyceride tripalmitin (Dynasan 116 Sasol GmbH (Witten, Germany). The polymer was polyethylene glycol in powdered form with a mean R ) molecular weight of 10,000 (Polyglykol 10000 P and was provided by Clariant (Waalwijk, the Netherlands). Extrusion Extrusion was performed with physical powder mixtures of either tripalmitin or tripalmitin and the hydrophilic polymer polyethylene glycol in equal parts as matrix formers and the model drug theophylline anhydrate. Each mixture was weighed in a ratio of 50:50 (w/w %) matrix former(s) to drug and was blended in a laboratory mixer (LM20; Bohle, Ennigerloh, Germany) at 25 rpm for 15 min. A gravimetric dosing device (KT20; K-Tron, Niederlenz, Switzerland) fed the powder mixtures into the barrel of a corotating twin-screw extruder (Mikro 27GL-28D; Leistritz, Nuremberg, Germany) with a feeding rate of 40 g min−1 . The screw speed was kept constant at 30 rpm at a processing temperature of 55◦ C. In previous studies, these processing conditions were shown to avoid solid-state structure changes in the resulting extrudates.18,19 The cylindrical extrudates had a diameter of 1 mm. Raman Mapping Transverse sections of extrudates were prepared with a razor blade and mounted on to microscope slides in an upright position. A dispersive Raman microscope (Senterra, Bruker Optics, Ettlingen, Germany) was used to collect the mapping data on the cross-sections of the extrudates. To control the microscope and collect the spectra, OPUS software version 6.5 (Bruker Optics, Ettlingen, Germany) was used. Wavenumber stability and transferability were ensured using the R (Bruker Optics, Ettlingen, Germany) and Sure Cal spectral shape correction options. The Raman signal was generated with an excitation wavelength of 785 nm and a 100 mW power before the objective. The spectrograph was operated with a 1200 groove per millimeter grating providing a resolution of 3–5 cm−1 . JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4454
HAASER ET AL.
Each Raman dataset comprised about 4700 spectra, wherein each spectrum was the coaddition of two 5 s exposures collected from 65 to 1520 cm−1 . Both 20× (numerical aperture (N.A.) 0.4) and 50× (N.A. 0.75) objectives (Olympus Corporation, Shinjuku-ku, Tokyo, Japan), with a 50 :m confocal pinhole, were used to collect the Raman spectra. A serial set of spectra from a silicon wafer–glass transect was used to measure sample area diameter. The sample diameter was calculated as the distance along the wafer–glass transect corresponding with a 50% decrease in the maximum intensity of the 520 cm−1 band. Sample diameters were 4 and 14 :m under 50× and 20× objectives, respectively. A second serial set of spectra was produced by moving the silicon wafer into and out of the laser focal plane. The 520 cm−1 band intensities from each spectrum were collated into an intensity profile (520 cm−1 intensity vs. sampling height), with sampled depth calculated as the full width at half maximum height of the intensity profile. The sampled depth based on this calculation was 8 :m (50× objective) and 45 :m (20× objective) through the 50 :m confocal pinhole. Rectangles of approximately 450 × 950 :m2 , with a step size of 10 :m in both the x and y directions, were collected using the 20× objective, which constituted 50% of the cross-section of the extrudates. Two adjacent rectangles of approximately 100 × 420 :m2 , with a step size of 3 :m in both the x and y directions, were collected using the 50× objective, which constituted 12.5% of the extrudate crosssections. Chemical maps were constructed from peak areas characteristic to each component (Fig. 1). For theophylline anhydrate, the 554 cm−1 band was used20 ;
for tripalmitin, the 1100 cm−1 band21 was used; and for polyethylene glycol, two close peaks at 844 and 860 cm−1 were used.22 In the absence of polyethylene glycol in the system being investigated, the peak for tripalmitin at 1130 cm−1 was used because its signal is much stronger. The peak areas were normalized as a percentage of the average maximum signal from the 10 largest peak areas observed for each component in the analyzed systems before dissolution. An 8% threshold was used to create the Raman maps, as established in a previous study.16 The resulting data were imported into Origin 7.5 (Origin Lab Corporation, Northampton, Massachusetts) to construct maps. These single component maps were combined using Adobe Photoshop 7.0 (Adobe Systems Incorporated, San Jose, California). Mapping of Samples After Dissolution Testing The United States Pharmacopeia 29 method 1 was used to perform the dissolution of the samples used for mapping (Erweka DT 600; Erweka GmbH, Heusenstamm, Germany). In each basket, a total of 140 mg was weighed after cutting the extrudates into approximately 1 cm long cylinders. Dissolution testing was performed at 37 ± 0.5◦ C with a stirring speed of 50 rpm in 900 mL of purified water containing 0.001% (w/v) polysorbate 20. The extrudate samples without polyethylene glycol were removed from the dissolution media after 120 min, whereas the extrudate samples containing the polyethylene glycol were removed after 10 and 30 min of dissolution. Excess moisture was removed from the samples by patting with a nonlinting tissue, and cross-sections were cut perpendicular to their surface with a razor blade. These extrudates were then mapped as described above. The drug concentration in solution during dissolution was determined using an ultraviolet–visible (UV–Vis) spectrometer with a detection wavelength of 244 nm (Ultrospec 2000; Pharmacia Biotech, Cambridge, UK).
RESULTS AND DISCUSSION Optimization of Measurement Conditions for the Analysis of Spatial Component Distribution in Matrix Dosage Forms
Figure 1. Raman spectra of theophylline anhydrate (a), tripalmitin (b), and polyethylene glycol (c). The band at 554 cm−1 was used for theophylline anhydrate; for tripalmitin, the 1100 cm−1 band or the 1130 cm−1 band (in absence of polyethylene glycol) was used; and the two overlapping bands at 844 and 860 cm−1 were used for polyethylene glycol. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
Irrespective of the system under investigation, Raman mapping is always affected by the inherent compromise between spatial resolution and acquisition time.16 In this study, two different objective sizes were probed to determine appropriate measurement conditions for pharmaceutical solid dosage forms. The aim was to increase the sampling area of each measurement and hence total map area as much as possible while retaining sufficient spatial resolution to analyze the processes of interest. DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
4455
Figure 2. Cross-section of an extrudate containing tripalmitin (blue, 1130 cm−1 ) and theophylline anhydrate (red, 554 cm−1 ). Raman maps of both components recorded with a 20× objective (a), and recorded with a 50× objective (b) and optical microscopy image (c), rectangles depict area mapped.
The extrudate cross-sections analyzed had a diameter of 1 mm. Recording the Raman signals with a 50× objective and a step size of 3 :m between pointto-point measurements in both the x and y directions enabled the analysis of 12.5% of the cross-sectional area, whereas with a 20× objective and a step size of 10 :m, half of the cross-sectional area could be investigated in half of the time needed for the area analyzed with the 50× objective. First, the binary component matrix system containing tripalmitin and theophylline anhydrate was mapped before dissolution with a 20× and a 50× objective as described above. The Raman maps and an optical microscope image with two red rectangles indicating the mapped areas of the extrudate crosssection are shown in Figure 2. The chemically selective Raman maps (Figs. 2a and 2b) were obtained by overlaying two separate Raman maps of tripalmitin (blue) and theophylline anhydrate (red). From those images, it is evident that, compared with the 20× objective, the spatial resolution obtained with a 50× objective leads to more detailed information within the same area and smaller particles can be detected. However, within the area mapped using both objectives, the component distribution is similar, suggesting that the influence of the depth of focus difference on the resulting images is limited in these systems. In the ternary system, consisting of equal amounts of tripalmitin and polyethylene glycol together with theophylline anhydrate, the false colors for the Raman maps were blue for tripalmitin, green for polyethylene glycol, and red for theophylline (Figs. 3a and 3b). The images suggest a wide range of particle sizes exposed at the cross-section surface for all three components. As was the case for the binary system, DOI 10.1002/jps
Figure 3. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ). Raman map of all three components recorded with a 20× objective (a), and recorded with a 50× objective (b) and optical microscopy image (c), rectangles depict areas mapped.
the maps created from the data collected with the 50× objective revealed more detailed information regarding the distribution and particle size of the components, but the overall component distribution in the matrices was similar in the area mapped using both objectives. Figure 3c depicts an optical micrograph of the surface with the investigated area indicated by two red rectangles. Although the 50× objective gave more detailed Raman maps within the same area than the 20× objective, the spatial resolution of the images created from data collected with the lower magnification was considered sufficient to observe the component distribution and changes in distribution during dissolution. As the total area mapped is important for the analysis of component distribution and physical changes during dissolution, the 20× objective was chosen to image extrudates after dissolution testing, with supplementation of some images using the 50× objective when more detailed information was of interest. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4456
HAASER ET AL.
Investigation of the Receding Drug Boundary in a Two-component Matrix System Extrudates containing equal amounts of the matrix former tripalmitin and the drug theophylline anhydrate were subjected to dissolution testing. Previously, the release kinetics of theophylline from these matrices under the same dissolution conditions has been investigated using continuous UV analysis.18 As the lipid matrix stays intact in the dissolution medium,17,18 the release process is diffusion controlled. After 120 min, the extrudates were removed from the medium, and the cross-sections were investigated with Raman microscopy to map the remaining drug in the lipid matrix. The Raman maps of the single components and the overlaid map are shown in Figures 4a–4c, with the optical microscopy image shown in Figure 4d indicating the sampled areas. According to solution analy-
sis of the medium using UV–Vis spectroscopy, after 120 min, 17% (w/w) of the drug had been released, which is in line with previous studies on the same extrudates.16,19 Assuming a uniformly receding drug boundary based on Higuchi kinetics, this amount of drug release would lead to an absence of the drug signal 45 :m into the cross-section of the extrudates. This is shown in Figures 4a–4c by two black semicircles, representing the edge of the extrudate and the theoretical drug boundary. Although a zone of drug loss can be observed in the maps, the drug boundary is not uniformly receding. The reason for this behavior can be explained by the varying tortuosity of the pores and channels that form in the matrix during drug dissolution, through which the dissolved drug diffuses out of the matrix. The tortuosity of the channels varies due to the random distribution of the drug particles and the variability in particle size.
Figure 4. Cross-section of an extrudate containing tripalmitin (blue, 1130 cm−1 ) and theophylline anhydrate (red, 554 cm−1 ) after 120 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), and both components (c) recorded with a 20× objective, optical microscopy image (d) and small area recorded with a 50× objective (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
In addition, some of the theophylline anhydrate particles are likely to be totally encapsulated by tripalmitin and therefore would never dissolve during the dissolution testing. In a previous study, 100% drug release could not be achieved for a binary system containing equal amounts of tripalmitin and theophylline anhydrate.18 The map recorded with the 50× objective, shown in Figure 4e, reveals drug particles that are completely surrounded by lipid in two dimensions. As theophylline anhydrate has been observed to convert to the monohydrate form during intrinsic dissolution testing15 and Raman spectroscopy can be used to resolve these two solid-state forms, the Raman spectra used to construct the Raman maps were also inspected for any sign of monohydrate formation. As has previously been observed in these extrudates using macroscopic and microscopic Raman measurements during dissolution testing in water at 37◦ C,16 there was no evidence of any monohydrate formation.
4457
Investigation of Polyethylene Glycol as a Release Modifier in Solid Lipid Extrudates A ternary system containing the same ratio of drug to matrix, where the matrix contained equal amounts of tripalmitin and the highly water-soluble polymer polyethylene glycol, was investigated. In contrast to the binary system with the pure lipid matrix that stayed intact during dissolution, the matrix of the ternary system changed with the watersoluble polyethylene glycol dissolving and the lipid component remaining intact. Extrudates were removed from the dissolution medium after 10 and 30 min. According to analysis of the dissolution medium, 19% of drug had been released after 10 min dissolution time. In the Raman mapping data (Figs. 5a–5d), the drug signal can still be detected near the edge of the extrudate in the area mapped. If there were a uniformly
Figure 5. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ) after 10 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), polyethylene glycol (c), and all three components (d) recorded with a 20× objective and optical microscopy image (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4458
HAASER ET AL.
receding drug boundary, then there should be a calculated absence of signal 50 :m into the cross-sectional area. As for the binary systems, there was no evidence of any theophylline monohydrate in the spectra used to construct the Raman maps. Interestingly, the signal for polyethylene glycol is less than 30% of the defined maximum peak area of the polyethylene glycol signal recorded from the extrudate before dissolution. Thus, the polymer, with a water solubility (530 mg/mL at 20◦ C)23 44-fold higher than that of theophylline anhydrate (12 mg/mL at 25◦ C)24 seems to start dissolving and diffusing out of the matrix first. This rapid polymer release increases the exposed surface area of the drug particles to the dissolution media, resulting in faster drug release for the ternary system compared with the binary system. The analyzed area of the cross-section is depicted by a red rectangle in the optical microscopy image (Fig. 5e).
The ternary system was also investigated after 30 min of dissolution testing, after which 50% (w/w) of drug had been released. The Raman maps of the system after 30 min drug release (Figs. 6a–6d) do not show any signal for polyethylene glycol (above the 8% threshold). This indicates that most or all the polyethylene glycol in the matrix had been dissolved and diffused into the dissolution medium (Fig. 6c). The map of the extrudate shows a receding boundary for the drug, although this is not uniformly distant from the matrix edge (Fig. 6a). On the basis of the assumption of a uniformly receding drug boundary, an absence of the drug signal approximately 150 :m from the edge of the sample would be expected, given the calculated drug release of 50% (w/w) by UV spectroscopy. The area between the two black semicircles only contains tripalmitin signal (with different intensities).
Figure 6. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ) after 30 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), polyethylene glycol (c), and all three components (d) recorded with a 20× objective and optical microscopy image (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
CONCLUSION Drug distribution and drug release in matrices were successfully mapped using Raman microscopy, with appropriate measurement parameters being developed for both binary and ternary systems. Although in these systems, a receding drug boundary was observed during dissolution testing, it was not uniformly distant from the matrix edge. This study also showed that Raman microscopy can be used to examine soluble and insoluble excipient distribution during dissolution testing, which, in turn, affects drug release behavior. For the ternary system, consisting of insoluble lipid and soluble polymer matrix, the polymer polyethylene glycol dissolved and diffused out first, leaving pores that increased the exposed surface area of the drug to the dissolution medium and thus accelerated the drug release. Overall, Raman mapping can be considered a useful method to investigate the distribution of multiple components in matrices during dissolution, which is necessary to obtain a deeper understanding of factors affecting drug release behavior during the development of matrix dosage forms.
ACKNOWLEDGMENTS This work was financially supported by a University of Otago Research Grant (to C. Strachan, K. Gordon, P. Kleinebudde, and Cushla McGoverin) and a University of Otago PhD Scholarship (to M. Haaser).
REFERENCES 1. Grassi G, Lapasin R, Grassi M, Colombo I. 2006. Drug release from matrix systems. Understanding drug release and absorption mechanisms: A physical and mathematical approach. CRC Press, Boca Raton, FL. 2. Costa P, Sousa Lobo JM. 2001. Modeling and comparison of dissolution profiles. Eur J Pharm Sci 13(2):123–133. 3. Higuchi T. 1963. Mechanism of sustained-action medication. Theoretical analysis of rate of release of solid drugs dispersed in solid matrices. J Pharm Sci 52:1145–1149. 4. Higuchi T. 1961. Rate of release of medicaments from ointment bases containing drugs in suspension. J Pharm Sci 50:874–875. 5. Siepmann J, G¨opferich A. 2001. Mathematical modeling of bioerodible, polymeric drug delivery systems. Adv Drug Deliv Rev 48(2–3):229–247. ¨ L. 6. Laaksonen TJ, Laaksonen HM, Hirvonen JT, Murtomaki 2009. Cellular automata model for drug release from binary matrix and reservoir polymeric devices. Biomaterials 30(10):1978–1987. 7. Leuenberger H, Rohera BD, Haas C. 1987. Percolation theory—a novel approach to solid dosage form design. Int J Pharm 38:109–115.
DOI 10.1002/jps
4459
8. Leuenberger H, Bonny JD, Kolb M. 1995. Percolation effects in matrix-type controlled drug release systems. Int J Pharm 115:217–224. 9. Pivonka DE, Chalmers JM, Griffiths PR, Eds. 2007. Applications of vibrational spectroscopy in pharmaceutical research and development. John Wiley & Sons, Chichester, UK. ˇ si´c S, Zhang L. 2007. Phar10. Clark D, Henson M, LaPlant F, Saˇ maceutical applications of mapping and imaging. In Applications of vibrational spectroscopy in pharmaceutical research and development; Pivonka DE, Chalmers JM, Griffiths PR, Eds. John Wiley & Sons, Chichester, UK, pp 309–335. 11. Gordon KC, McGoverin CM. Raman mapping of pharmaceutics. Int J Pharm. doi:10.1016/j.ijpharm.2010.12.030. In Press. ˇ si´c S. 2007. An in-depth analysis of Raman and near12. Saˇ infrared chemical images of common pharmaceutical tablets. Appl Spectrosc 61(3):239–250. 13. Heinz A, Strachan CJ, Gordon KC, Rades T. 2009. Analysis of solid-state transformations of pharmaceutical compounds using vibrational spectroscopy. J Pharm Pharmacol 61(8):971–988. 14. McGoverin CM, Rades T, Gordon KC. 2008. Recent pharmaceutical applications of Raman and terahertz spectroscopies. J Pharm Sci 97(11):4598–4621. ¨ ¨ 15. Aaltonen J, Heinanen P, Peltonen L, Kortejarvi 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(12):2730–2737. 16. Windbergs M, Haaser M, McGoverin CM, Gordon KC, Kleinebudde P, Strachan CJ. 2010. Investigating the relationship between drug distribution in solid lipid matrices and dissolution behaviour using Raman spectroscopy and mapping. J Pharm Sci 99(3):1464–1475. 17. Reitz C, Kleinebudde P. 2007. Solid lipid extrusions of sustained release dosage forms. Eur J Pharm Biopharm 67(2):440–448. 18. Windbergs M, Strachan CJ, Kleinebudde P. 2009. Tailormade dissolution profiles by extruded matrices based on lipid polyethylene glycol mixtures. J Control Release 137(3):211– 216. 19. Windbergs M, Strachan CJ, Kleinebudde P. 2009. Understanding the solid-state behaviour of triglyceride solid lipid extrudates and its influence on dissolution. Eur J Pharm Biopharm 71(1):80–87. 20. Amado AM, Nolasco MM, Ribeiro-Claro PJA. 2007. Probing pseudopolymorphic transitions in pharmaceutical solids using Raman spectroscopy: Hydration and dehydration of theophylline. J Pharm Sci 96:1366–1379. 21. Bresson S, El Marssi M, Khelifa B. 2005. Raman spectroscopy investigation of various saturated monoacid triglycerides. Chem Phys Lipids 134:119–129. ¨ 22. Kozielski M, Muhle M, Blaszczak Z. 2004. The Raman scattering study of selected polyoxyethyleneglycols. J Mol Liq 111(1–3):1–5. 23. Clariant International Ltd. Functional Chemicals Division BL Personal Care. 2007. Your universally applicable polymer. Accessed August 12, 2010, at www.clariant.com: Please provide the Web address for Ref. 23. Done. 24. Rodr´ıguez-Hornedo N, Lechuga-Ballesteros D, Wu H. 1992. Phase transition and heterogeneous/epitaxial nucleation of hydrated and anhydrous theophylline crystals. Int J Pharm 85:149–162.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
School of Pharmacy, University of Otago, Dunedin 9054, New Zealand
2
School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts 02138
3
Department of Chemistry and MacDiarmid Institute for Advanced Materials and Nanotechnology, University of Otago, Dunedin 9054, New Zealand 4
¨ Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University, 40225 Dusseldorf, Germany
Received 25 January 2011; revised 6 April 2011; accepted 19 April 2011 Published online 10 May 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22609 ABSTRACT: Matrix dosage forms are widely used for sustained drug release. As both the distribution of the matrix components and physical changes during dissolution can impact drug release behavior, a comprehensive investigation of these phenomena is required during matrix development. In this study, Raman microscopy was used to investigate different extrudate formulations in terms of component distribution and structural changes during dissolution testing. Two systems containing the model drug theophylline anhydrate were investigated: a binary system, based on a tripalmitin matrix, and a ternary system, containing tripalmitin and polyethylene glycol. The distribution of the drug and the soluble and insoluble matrix components were mapped during dissolution testing. Although a receding drug boundary was observed, it was not uniformly distant from the matrix edge. The lipid structure remained intact, whereas the water-soluble polymer rapidly dissolved and diffused from the matrix leaving a more extensive network of channels through which the dissolution medium could penetrate and the drug could diffuse. Raman mapping can be considered a useful aid in the direct analysis of multiple matrix components during drug release, and therefore a deeper understanding of factors affecting drug release can be obtained during the development of sustained-release matrices. © 2011 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 100:4452–4459, 2011 Keywords: Raman spectroscopy; Raman mapping; lipids; controlled release; solid state; spatial resolution; solid lipid extrusion; dissolution; theophylline; polyethylene glycol
INTRODUCTION Matrix dosage forms, in which drugs are dispersed in lipidic or polymeric materials, are commonly used for sustained-release drug delivery systems. To obtain the desired drug release profiles from such dosage forms, there has been much research into understanding the mechanisms of drug release, as well as how the release is affected by dosage form changes.1,2 Drug release mechanisms are usually studied by fitting modCorrespondence to: Clare J. Strachan (Telephone: +64-3-4797324; Fax: +64-3-479-7034; E-mail: [email protected]) Miriam Haaser and Maike Windbergs have contributed equally to the manuscript. Journal of Pharmaceutical Sciences, Vol. 100, 4452–4459 (2011) © 2011 Wiley-Liss, Inc. and the American Pharmacists Association
4452
els to data obtained by measuring the drug concentration in solution over time.2 In the case of noneroding matrices, in which the dissolution medium penetrates the matrix and the drug dissolves and diffuses through the matrix, Higuchi kinetics is commonly applied. This approach assumes a linear concentration gradient in the solution between the receding drug boundary in the matrix (the interface between the undissolved drug and dissolution medium) and the outside of the matrix.3,4 In eroding systems, drug release is controlled by erosion of the matrix, which makes the drug surface accessible to the dissolution medium, and drug release is commonly described mathematically with empirical models assuming a single zero-order release profile.5 These and most other modeling approaches assume steady-state
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
diffusion/erosion profiles, and do not consider the lag time and/or burst effect that may be observed at the beginning of drug release. More recently, Laaksonen et al.6 have simulated these three phases of drug release from matrices using a cellular automata approach, in which a two-dimensional grid space represents the release device. Each unit cell of this grid represents water, drug, or polymer/lipid and is connected to four adjacent cells. Different states with different physicochemical parameters can be modeled. This modeling approach is related to percolation theory, in which a percolation threshold is defined as the volume-to-volume ratio at which an interpenetrating network of a component is formed.7,8 Laaksonen et al.6 have shown their modeling to be applicable to both diffusion- and erosion-controlled release mechanisms as well as a combination of the two. Nevertheless, to fully understand drug release from sustained-release matrices, the dosage forms themselves should be directly analyzed. With the recent advent of chemically selective analysis approaches with appropriate spatial resolution, analysis of individual components within matrices has become possible.9,10 Raman microscopy has several features that make it well suited to analyzing matrix dosage forms during dissolution testing.11 First, it is highly chemically selective, and therefore multiple components in the matrix can be resolved.12 In addition, different solid-state forms of the same chemical species can be resolved.13 Second, it is possible to obtain a high spatial resolution (in the order of about 1–2 :m). Third, the technique is relatively insensitive to water, and therefore the matrix components can be examined in the presence of aqueous dissolution medium.10,14,15 Although Raman microscopy has been used to image component distribution in various matrix systems,9,11 to the best of our knowledge, apart from some preliminary work by ourselves,16 the technique has not been used to image how drug and matrix component distributions in oral dosage forms change during the drug release process. In this study, the potential of Raman microscopy for the analysis of matrix dosage forms during drug release was investigated. The specific objectives were to optimize the microscope measurement parameters, and then analyze drug and excipient distribution changes during dissolution testing and relate these changes to the drug release behavior. The matrices analyzed were solid lipid extrudates. The production of these extrudates involves extruding lipid below its melting temperature with solid drug.17,18 In these systems, the drug is dispersed in particulate form.16 As long as the lipid concentration is above its percolation threshold,7 the lipid remains intact during drug release in water or buffered dissolution media, and drug release has been found to follow Higuchi kinetics following an initial burst DOI 10.1002/jps
4453
release.19 Two formulations have been imaged, both containing theophylline as the drug embedded in a matrix. For the first system, the lipid tripalmitin was used alone to form the matrix, whereas the matrix of the second system contained equal parts of tripalmitin and the water-soluble polymer, polyethylene glycol.
MATERIALS AND METHODS Materials Theophylline anhydrate (BASF, Ludwigshafen, Germany) in powdered form was used as the model drug. The powdered lipid was the pure monoacid R ) provided by triglyceride tripalmitin (Dynasan 116 Sasol GmbH (Witten, Germany). The polymer was polyethylene glycol in powdered form with a mean R ) molecular weight of 10,000 (Polyglykol 10000 P and was provided by Clariant (Waalwijk, the Netherlands). Extrusion Extrusion was performed with physical powder mixtures of either tripalmitin or tripalmitin and the hydrophilic polymer polyethylene glycol in equal parts as matrix formers and the model drug theophylline anhydrate. Each mixture was weighed in a ratio of 50:50 (w/w %) matrix former(s) to drug and was blended in a laboratory mixer (LM20; Bohle, Ennigerloh, Germany) at 25 rpm for 15 min. A gravimetric dosing device (KT20; K-Tron, Niederlenz, Switzerland) fed the powder mixtures into the barrel of a corotating twin-screw extruder (Mikro 27GL-28D; Leistritz, Nuremberg, Germany) with a feeding rate of 40 g min−1 . The screw speed was kept constant at 30 rpm at a processing temperature of 55◦ C. In previous studies, these processing conditions were shown to avoid solid-state structure changes in the resulting extrudates.18,19 The cylindrical extrudates had a diameter of 1 mm. Raman Mapping Transverse sections of extrudates were prepared with a razor blade and mounted on to microscope slides in an upright position. A dispersive Raman microscope (Senterra, Bruker Optics, Ettlingen, Germany) was used to collect the mapping data on the cross-sections of the extrudates. To control the microscope and collect the spectra, OPUS software version 6.5 (Bruker Optics, Ettlingen, Germany) was used. Wavenumber stability and transferability were ensured using the R (Bruker Optics, Ettlingen, Germany) and Sure Cal spectral shape correction options. The Raman signal was generated with an excitation wavelength of 785 nm and a 100 mW power before the objective. The spectrograph was operated with a 1200 groove per millimeter grating providing a resolution of 3–5 cm−1 . JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4454
HAASER ET AL.
Each Raman dataset comprised about 4700 spectra, wherein each spectrum was the coaddition of two 5 s exposures collected from 65 to 1520 cm−1 . Both 20× (numerical aperture (N.A.) 0.4) and 50× (N.A. 0.75) objectives (Olympus Corporation, Shinjuku-ku, Tokyo, Japan), with a 50 :m confocal pinhole, were used to collect the Raman spectra. A serial set of spectra from a silicon wafer–glass transect was used to measure sample area diameter. The sample diameter was calculated as the distance along the wafer–glass transect corresponding with a 50% decrease in the maximum intensity of the 520 cm−1 band. Sample diameters were 4 and 14 :m under 50× and 20× objectives, respectively. A second serial set of spectra was produced by moving the silicon wafer into and out of the laser focal plane. The 520 cm−1 band intensities from each spectrum were collated into an intensity profile (520 cm−1 intensity vs. sampling height), with sampled depth calculated as the full width at half maximum height of the intensity profile. The sampled depth based on this calculation was 8 :m (50× objective) and 45 :m (20× objective) through the 50 :m confocal pinhole. Rectangles of approximately 450 × 950 :m2 , with a step size of 10 :m in both the x and y directions, were collected using the 20× objective, which constituted 50% of the cross-section of the extrudates. Two adjacent rectangles of approximately 100 × 420 :m2 , with a step size of 3 :m in both the x and y directions, were collected using the 50× objective, which constituted 12.5% of the extrudate crosssections. Chemical maps were constructed from peak areas characteristic to each component (Fig. 1). For theophylline anhydrate, the 554 cm−1 band was used20 ;
for tripalmitin, the 1100 cm−1 band21 was used; and for polyethylene glycol, two close peaks at 844 and 860 cm−1 were used.22 In the absence of polyethylene glycol in the system being investigated, the peak for tripalmitin at 1130 cm−1 was used because its signal is much stronger. The peak areas were normalized as a percentage of the average maximum signal from the 10 largest peak areas observed for each component in the analyzed systems before dissolution. An 8% threshold was used to create the Raman maps, as established in a previous study.16 The resulting data were imported into Origin 7.5 (Origin Lab Corporation, Northampton, Massachusetts) to construct maps. These single component maps were combined using Adobe Photoshop 7.0 (Adobe Systems Incorporated, San Jose, California). Mapping of Samples After Dissolution Testing The United States Pharmacopeia 29 method 1 was used to perform the dissolution of the samples used for mapping (Erweka DT 600; Erweka GmbH, Heusenstamm, Germany). In each basket, a total of 140 mg was weighed after cutting the extrudates into approximately 1 cm long cylinders. Dissolution testing was performed at 37 ± 0.5◦ C with a stirring speed of 50 rpm in 900 mL of purified water containing 0.001% (w/v) polysorbate 20. The extrudate samples without polyethylene glycol were removed from the dissolution media after 120 min, whereas the extrudate samples containing the polyethylene glycol were removed after 10 and 30 min of dissolution. Excess moisture was removed from the samples by patting with a nonlinting tissue, and cross-sections were cut perpendicular to their surface with a razor blade. These extrudates were then mapped as described above. The drug concentration in solution during dissolution was determined using an ultraviolet–visible (UV–Vis) spectrometer with a detection wavelength of 244 nm (Ultrospec 2000; Pharmacia Biotech, Cambridge, UK).
RESULTS AND DISCUSSION Optimization of Measurement Conditions for the Analysis of Spatial Component Distribution in Matrix Dosage Forms
Figure 1. Raman spectra of theophylline anhydrate (a), tripalmitin (b), and polyethylene glycol (c). The band at 554 cm−1 was used for theophylline anhydrate; for tripalmitin, the 1100 cm−1 band or the 1130 cm−1 band (in absence of polyethylene glycol) was used; and the two overlapping bands at 844 and 860 cm−1 were used for polyethylene glycol. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
Irrespective of the system under investigation, Raman mapping is always affected by the inherent compromise between spatial resolution and acquisition time.16 In this study, two different objective sizes were probed to determine appropriate measurement conditions for pharmaceutical solid dosage forms. The aim was to increase the sampling area of each measurement and hence total map area as much as possible while retaining sufficient spatial resolution to analyze the processes of interest. DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
4455
Figure 2. Cross-section of an extrudate containing tripalmitin (blue, 1130 cm−1 ) and theophylline anhydrate (red, 554 cm−1 ). Raman maps of both components recorded with a 20× objective (a), and recorded with a 50× objective (b) and optical microscopy image (c), rectangles depict area mapped.
The extrudate cross-sections analyzed had a diameter of 1 mm. Recording the Raman signals with a 50× objective and a step size of 3 :m between pointto-point measurements in both the x and y directions enabled the analysis of 12.5% of the cross-sectional area, whereas with a 20× objective and a step size of 10 :m, half of the cross-sectional area could be investigated in half of the time needed for the area analyzed with the 50× objective. First, the binary component matrix system containing tripalmitin and theophylline anhydrate was mapped before dissolution with a 20× and a 50× objective as described above. The Raman maps and an optical microscope image with two red rectangles indicating the mapped areas of the extrudate crosssection are shown in Figure 2. The chemically selective Raman maps (Figs. 2a and 2b) were obtained by overlaying two separate Raman maps of tripalmitin (blue) and theophylline anhydrate (red). From those images, it is evident that, compared with the 20× objective, the spatial resolution obtained with a 50× objective leads to more detailed information within the same area and smaller particles can be detected. However, within the area mapped using both objectives, the component distribution is similar, suggesting that the influence of the depth of focus difference on the resulting images is limited in these systems. In the ternary system, consisting of equal amounts of tripalmitin and polyethylene glycol together with theophylline anhydrate, the false colors for the Raman maps were blue for tripalmitin, green for polyethylene glycol, and red for theophylline (Figs. 3a and 3b). The images suggest a wide range of particle sizes exposed at the cross-section surface for all three components. As was the case for the binary system, DOI 10.1002/jps
Figure 3. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ). Raman map of all three components recorded with a 20× objective (a), and recorded with a 50× objective (b) and optical microscopy image (c), rectangles depict areas mapped.
the maps created from the data collected with the 50× objective revealed more detailed information regarding the distribution and particle size of the components, but the overall component distribution in the matrices was similar in the area mapped using both objectives. Figure 3c depicts an optical micrograph of the surface with the investigated area indicated by two red rectangles. Although the 50× objective gave more detailed Raman maps within the same area than the 20× objective, the spatial resolution of the images created from data collected with the lower magnification was considered sufficient to observe the component distribution and changes in distribution during dissolution. As the total area mapped is important for the analysis of component distribution and physical changes during dissolution, the 20× objective was chosen to image extrudates after dissolution testing, with supplementation of some images using the 50× objective when more detailed information was of interest. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4456
HAASER ET AL.
Investigation of the Receding Drug Boundary in a Two-component Matrix System Extrudates containing equal amounts of the matrix former tripalmitin and the drug theophylline anhydrate were subjected to dissolution testing. Previously, the release kinetics of theophylline from these matrices under the same dissolution conditions has been investigated using continuous UV analysis.18 As the lipid matrix stays intact in the dissolution medium,17,18 the release process is diffusion controlled. After 120 min, the extrudates were removed from the medium, and the cross-sections were investigated with Raman microscopy to map the remaining drug in the lipid matrix. The Raman maps of the single components and the overlaid map are shown in Figures 4a–4c, with the optical microscopy image shown in Figure 4d indicating the sampled areas. According to solution analy-
sis of the medium using UV–Vis spectroscopy, after 120 min, 17% (w/w) of the drug had been released, which is in line with previous studies on the same extrudates.16,19 Assuming a uniformly receding drug boundary based on Higuchi kinetics, this amount of drug release would lead to an absence of the drug signal 45 :m into the cross-section of the extrudates. This is shown in Figures 4a–4c by two black semicircles, representing the edge of the extrudate and the theoretical drug boundary. Although a zone of drug loss can be observed in the maps, the drug boundary is not uniformly receding. The reason for this behavior can be explained by the varying tortuosity of the pores and channels that form in the matrix during drug dissolution, through which the dissolved drug diffuses out of the matrix. The tortuosity of the channels varies due to the random distribution of the drug particles and the variability in particle size.
Figure 4. Cross-section of an extrudate containing tripalmitin (blue, 1130 cm−1 ) and theophylline anhydrate (red, 554 cm−1 ) after 120 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), and both components (c) recorded with a 20× objective, optical microscopy image (d) and small area recorded with a 50× objective (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
In addition, some of the theophylline anhydrate particles are likely to be totally encapsulated by tripalmitin and therefore would never dissolve during the dissolution testing. In a previous study, 100% drug release could not be achieved for a binary system containing equal amounts of tripalmitin and theophylline anhydrate.18 The map recorded with the 50× objective, shown in Figure 4e, reveals drug particles that are completely surrounded by lipid in two dimensions. As theophylline anhydrate has been observed to convert to the monohydrate form during intrinsic dissolution testing15 and Raman spectroscopy can be used to resolve these two solid-state forms, the Raman spectra used to construct the Raman maps were also inspected for any sign of monohydrate formation. As has previously been observed in these extrudates using macroscopic and microscopic Raman measurements during dissolution testing in water at 37◦ C,16 there was no evidence of any monohydrate formation.
4457
Investigation of Polyethylene Glycol as a Release Modifier in Solid Lipid Extrudates A ternary system containing the same ratio of drug to matrix, where the matrix contained equal amounts of tripalmitin and the highly water-soluble polymer polyethylene glycol, was investigated. In contrast to the binary system with the pure lipid matrix that stayed intact during dissolution, the matrix of the ternary system changed with the watersoluble polyethylene glycol dissolving and the lipid component remaining intact. Extrudates were removed from the dissolution medium after 10 and 30 min. According to analysis of the dissolution medium, 19% of drug had been released after 10 min dissolution time. In the Raman mapping data (Figs. 5a–5d), the drug signal can still be detected near the edge of the extrudate in the area mapped. If there were a uniformly
Figure 5. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ) after 10 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), polyethylene glycol (c), and all three components (d) recorded with a 20× objective and optical microscopy image (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4458
HAASER ET AL.
receding drug boundary, then there should be a calculated absence of signal 50 :m into the cross-sectional area. As for the binary systems, there was no evidence of any theophylline monohydrate in the spectra used to construct the Raman maps. Interestingly, the signal for polyethylene glycol is less than 30% of the defined maximum peak area of the polyethylene glycol signal recorded from the extrudate before dissolution. Thus, the polymer, with a water solubility (530 mg/mL at 20◦ C)23 44-fold higher than that of theophylline anhydrate (12 mg/mL at 25◦ C)24 seems to start dissolving and diffusing out of the matrix first. This rapid polymer release increases the exposed surface area of the drug particles to the dissolution media, resulting in faster drug release for the ternary system compared with the binary system. The analyzed area of the cross-section is depicted by a red rectangle in the optical microscopy image (Fig. 5e).
The ternary system was also investigated after 30 min of dissolution testing, after which 50% (w/w) of drug had been released. The Raman maps of the system after 30 min drug release (Figs. 6a–6d) do not show any signal for polyethylene glycol (above the 8% threshold). This indicates that most or all the polyethylene glycol in the matrix had been dissolved and diffused into the dissolution medium (Fig. 6c). The map of the extrudate shows a receding boundary for the drug, although this is not uniformly distant from the matrix edge (Fig. 6a). On the basis of the assumption of a uniformly receding drug boundary, an absence of the drug signal approximately 150 :m from the edge of the sample would be expected, given the calculated drug release of 50% (w/w) by UV spectroscopy. The area between the two black semicircles only contains tripalmitin signal (with different intensities).
Figure 6. Cross-section of an extrudate containing tripalmitin (blue, 1100 cm−1 ), theophylline anhydrate (red, 554 cm−1 ), and polyethylene glycol (green, 844 and 860 cm−1 ) after 30 min of dissolution testing. Raman map of theophylline anhydrate (a), tripalmitin (b), polyethylene glycol (c), and all three components (d) recorded with a 20× objective and optical microscopy image (e), rectangles depict areas mapped; semicircles indicate the edge and the calculated uniformly receding drug boundary. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
DOI 10.1002/jps
ANALYSIS OF MATRIX DOSAGE FORMS DURING DISSOLUTION TESTING
CONCLUSION Drug distribution and drug release in matrices were successfully mapped using Raman microscopy, with appropriate measurement parameters being developed for both binary and ternary systems. Although in these systems, a receding drug boundary was observed during dissolution testing, it was not uniformly distant from the matrix edge. This study also showed that Raman microscopy can be used to examine soluble and insoluble excipient distribution during dissolution testing, which, in turn, affects drug release behavior. For the ternary system, consisting of insoluble lipid and soluble polymer matrix, the polymer polyethylene glycol dissolved and diffused out first, leaving pores that increased the exposed surface area of the drug to the dissolution medium and thus accelerated the drug release. Overall, Raman mapping can be considered a useful method to investigate the distribution of multiple components in matrices during dissolution, which is necessary to obtain a deeper understanding of factors affecting drug release behavior during the development of matrix dosage forms.
ACKNOWLEDGMENTS This work was financially supported by a University of Otago Research Grant (to C. Strachan, K. Gordon, P. Kleinebudde, and Cushla McGoverin) and a University of Otago PhD Scholarship (to M. Haaser).
REFERENCES 1. Grassi G, Lapasin R, Grassi M, Colombo I. 2006. Drug release from matrix systems. Understanding drug release and absorption mechanisms: A physical and mathematical approach. CRC Press, Boca Raton, FL. 2. Costa P, Sousa Lobo JM. 2001. Modeling and comparison of dissolution profiles. Eur J Pharm Sci 13(2):123–133. 3. Higuchi T. 1963. Mechanism of sustained-action medication. Theoretical analysis of rate of release of solid drugs dispersed in solid matrices. J Pharm Sci 52:1145–1149. 4. Higuchi T. 1961. Rate of release of medicaments from ointment bases containing drugs in suspension. J Pharm Sci 50:874–875. 5. Siepmann J, G¨opferich A. 2001. Mathematical modeling of bioerodible, polymeric drug delivery systems. Adv Drug Deliv Rev 48(2–3):229–247. ¨ L. 6. Laaksonen TJ, Laaksonen HM, Hirvonen JT, Murtomaki 2009. Cellular automata model for drug release from binary matrix and reservoir polymeric devices. Biomaterials 30(10):1978–1987. 7. Leuenberger H, Rohera BD, Haas C. 1987. Percolation theory—a novel approach to solid dosage form design. Int J Pharm 38:109–115.
DOI 10.1002/jps
4459
8. Leuenberger H, Bonny JD, Kolb M. 1995. Percolation effects in matrix-type controlled drug release systems. Int J Pharm 115:217–224. 9. Pivonka DE, Chalmers JM, Griffiths PR, Eds. 2007. Applications of vibrational spectroscopy in pharmaceutical research and development. John Wiley & Sons, Chichester, UK. ˇ si´c S, Zhang L. 2007. Phar10. Clark D, Henson M, LaPlant F, Saˇ maceutical applications of mapping and imaging. In Applications of vibrational spectroscopy in pharmaceutical research and development; Pivonka DE, Chalmers JM, Griffiths PR, Eds. John Wiley & Sons, Chichester, UK, pp 309–335. 11. Gordon KC, McGoverin CM. Raman mapping of pharmaceutics. Int J Pharm. doi:10.1016/j.ijpharm.2010.12.030. In Press. ˇ si´c S. 2007. An in-depth analysis of Raman and near12. Saˇ infrared chemical images of common pharmaceutical tablets. Appl Spectrosc 61(3):239–250. 13. Heinz A, Strachan CJ, Gordon KC, Rades T. 2009. Analysis of solid-state transformations of pharmaceutical compounds using vibrational spectroscopy. J Pharm Pharmacol 61(8):971–988. 14. McGoverin CM, Rades T, Gordon KC. 2008. Recent pharmaceutical applications of Raman and terahertz spectroscopies. J Pharm Sci 97(11):4598–4621. ¨ ¨ 15. Aaltonen J, Heinanen P, Peltonen L, Kortejarvi 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(12):2730–2737. 16. Windbergs M, Haaser M, McGoverin CM, Gordon KC, Kleinebudde P, Strachan CJ. 2010. Investigating the relationship between drug distribution in solid lipid matrices and dissolution behaviour using Raman spectroscopy and mapping. J Pharm Sci 99(3):1464–1475. 17. Reitz C, Kleinebudde P. 2007. Solid lipid extrusions of sustained release dosage forms. Eur J Pharm Biopharm 67(2):440–448. 18. Windbergs M, Strachan CJ, Kleinebudde P. 2009. Tailormade dissolution profiles by extruded matrices based on lipid polyethylene glycol mixtures. J Control Release 137(3):211– 216. 19. Windbergs M, Strachan CJ, Kleinebudde P. 2009. Understanding the solid-state behaviour of triglyceride solid lipid extrudates and its influence on dissolution. Eur J Pharm Biopharm 71(1):80–87. 20. Amado AM, Nolasco MM, Ribeiro-Claro PJA. 2007. Probing pseudopolymorphic transitions in pharmaceutical solids using Raman spectroscopy: Hydration and dehydration of theophylline. J Pharm Sci 96:1366–1379. 21. Bresson S, El Marssi M, Khelifa B. 2005. Raman spectroscopy investigation of various saturated monoacid triglycerides. Chem Phys Lipids 134:119–129. ¨ 22. Kozielski M, Muhle M, Blaszczak Z. 2004. The Raman scattering study of selected polyoxyethyleneglycols. J Mol Liq 111(1–3):1–5. 23. Clariant International Ltd. Functional Chemicals Division BL Personal Care. 2007. Your universally applicable polymer. Accessed August 12, 2010, at www.clariant.com: Please provide the Web address for Ref. 23. Done. 24. Rodr´ıguez-Hornedo N, Lechuga-Ballesteros D, Wu H. 1992. Phase transition and heterogeneous/epitaxial nucleation of hydrated and anhydrous theophylline crystals. Int J Pharm 85:149–162.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011