Article pubs.acs.org/molecularpharmaceutics
Influence of Compression Forces on the Structural Stability of Naproxen/PVP-VA 64 Solid Dispersions Zelalem Ayenew Worku, Jolie Aarts, and Guy Van den Mooter* Drug Delivery and Disposition, KU Leuven, Herestraat 49, 3000 Leuven, Belgium S Supporting Information *
ABSTRACT: Solid dispersions are preferentially formulated as solid dosage forms such as tablets and capsules. The structural stability of the solid dispersions has not been adequately explored during post spray drying manufacturing processes. In this paper, we describe the influence of compression forces on solid dispersions made up of naproxen and PVP-VA 64 prepared by spray drying. Compression of the solid dispersion containing 30% (w/w) of naproxen led to low intensity of the powder X-ray diffraction (PXRD) halo pattern maxima at 2θ = 16.11°, and the uncompressed samples also exhibit higher glass transition broadening than the compressed samples after 21 days storage at 75% RH at ambient temperature which indicates structural changes in the solid dispersion. The intensity of the vibration band at 1654 cm−1 originating from the interaction between the hydrogen of the carboxylic acid moiety of NAP and the amide carbonyl moiety of PVP-VA 64 was increased for the compressed samples. The consequence of compression was further amplified after a long-term stability study (5 months) where the compressed 40 and 50% (w/w) NAP/PVP-VA 64 solid dispersions showed less crystallinity than the uncompressed samples. This suggests that compression improved the physical stability of the solid dispersions as a result of enhanced drug−polymer interactions. KEYWORDS: solid dispersions, spray drying, compression, crystallization, naproxen, PVP-VA, phase separation, specific interactions, amorphous, FTIR, PXRD, hydrogen bonding
■
INTRODUCTION The majority of new chemical entities1,2 and a number of marketed drugs display poor aqueous solubility.3 This challenge often decelerates the introduction of drugs into the market.4 Amorphous solid dispersions are one of the formulation strategies to improve the solubility of drugs.5 However, amorphous systems are less stable than their crystalline counterparts; hence the long-term physical stability of the solid dispersions is crucial for their success during product development.6 Processes and storage conditions accompanied with humidity and temperature often induce and also facilitate amorphous−amorphous phase separation and/or crystallization of drugs from the solid dispersions, hence leading to loss of their solubility advantage.7−9 Processes and unit operations can impart mechanical stresses which may result in physical instability of amorphous systems. Compression can lead to nonisothermal and isothermal crystallization of amorphous indomethacin prepared by the melt-cooling method.10 Shear rate and shear strain can also induce the nucleation and also facilitate the crystallization kinetics of amorphous systems where glassy polymers and natural rubbers undergo strain induced crystallization. Li et al. also showed that shear induces crystallization of poly(butylene terephthalate) (PBT) using in situ small-angle X-ray scattering.11,12 The effect of deformation can also be extended to the solution state of the mixtures where mild solution deformation may also induce phase separation in polymer blends.13 © 2014 American Chemical Society
The success of solid dispersions is hindered due to their physical stability problems; hence most studies have been focused on investigating the effect of humidity and temperature on solid dispersions prepared by spray drying and hot melt extrusion.7,8 The major scope of this study was to investigate the effect of compression on the structural and the physical stability of the spray dried solid dispersions such as amorphous−amorphous phase separation, crystallization, and drug−polymer specific interactions. The amorphous−amorphous phase behavior was assessed based on the glass transition width from modulated differential scanning calorimetry (MDSC) measurements, and the structural changes and the phase behavior were further investigated using attenuated total reflectance Fourier transform infrared spectroscopy (ATRFTIR) and powder X-ray diffraction (PXRD).
■
MATERIALS AND METHODS Materials. Naproxen (NAP) and poly(1-vinylpyrrolidoneco-vinyl acetate) (PVP-VA 64) were purchased from CERTA Ltd. (Brainl’Allend, Belgium) and BASF (Ludwisshafen, Germany), respectively. Dichloromethane (DCM) and methanol (MeOH) (HperSolv CHROMANORM) were purchased Received: February 12, 2014 Accepted: February 17, 2014 Published: February 17, 2014 1102
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
intensity of vibration bands and shoulder for the normalized spectra. Compression. The solid dispersions were compressed using a manual tablet press RQPBA15 (Rodac International, Sittard, The Netherlands) with a die cavity of 13 mm diameter. About 100 mg of solid dispersions were placed in the die cavity and compressed using pressures of 1130, 753, and 188 MPa for 10 s dwell time. The yield pressures of the spray dried solid dispersions were determined using an 11 mm die automated Korsch single-punch tablet press (Berlin, Germany) equipped with a W+W electronic signal memory recorder (Basel, Switzerland) and a charge amplifier (Kistler Instrument Corp., Winterthur, Switzerland). The true density of dried solid dispersions was determined using a helium pycnometer (model 930, Beckman Industries Inc., Fullerton (CA), USA). The tablet thickness and hardness were measured using a Lorentzen & Wettre instrument 141 type 1-1 (AB Lorentzen & Wettre, Stockholm, Sweden) and a Schleuniger 6D tablet tester (Dr. K. Schleuniger, Zürich, Switzerland), respectively. The data were analyzed using in-house software to compute the yield pressure and the tablet porosity from the Heckel plot.14 Physical Stability. The compressed and uncompressed solid dispersions were stored in a desiccator saturated with sodium chloride solution at ambient temperature to install a 75% RH condition. Samples stored for 10 d, 21 d, and 5 months were then analyzed after drying in P2O5 for one week at room temperature. Moisture Adsorption. The moisture gain of the compressed and the uncompressed solid dispersions was determined by weighing the samples regularly after 5 and 16 days storage at 75% RH and ambient temperature during the physical stability study. Moreover, the moisture adsorption kinetics were determined using an in-house developed vapor sorption instrument. A sample on a weighing balance (Mettler Toledo Group, China) was supplied with water vapor to maintain a relative humidity of 75 ± 0.8% RH at 22.4 ± 0.1 °C. The moisture content of the samples was regularly acquired (every second) using in-house developed software. The weight gain was normalized to the weight of the dry sample and plotted against the time of exposure. Drug Content. The drug content in the solid dispersions was determined using high-performance liquid chromatography (HPLC) equipped with a Merck Hitachi pump (L7100), an ultraviolet (UV) detector (L7400), an autosampler (L7200), an interface (D7000), and a LiChrospher 60 RP Select-B C-18 (5 μm, 12.5 × 4) column (all Merck, Darmstadt, Germany). An isocratic elution was used to quantify naproxen at a detection wavelength of 270 nm using a mobile phase made up of methanol (HiPerSolv Chromanorm, Belgium) and sodium acetate buffer (pH = 3.5 and 25 mM) in a 7:3 ratio (v/v). The injection volume and the flow rate were 10 μL and 1 mL/min, respectively. Statistical Analysis. A Student’s t-test was used to compare the significance of the differences between two set of data. The two set of data were considered as independent (unpaired) groups with different variance and without prior knowledge of the direction of the prediction. Microsoft Excel 2007 (Microsoft Corp., USA) was used to perform the Student’s ttest.
from VWR International (Leuven, Belgium). All chemicals were used without further purification. Spray Drying. Solid dispersions of 10%, 20%, 25%, 30%, 40%, and 50% (w/w) NAP in PVP-VA 64 were prepared from 10% (w/v) feed solution concentration in dichloromethane using the Buchi mini spray-dryer B191(B-SDR) (Buchi, Flawil, Switzerland) using 50 °C inlet temperature, 0.56 m3/min drying air flow rate, 17 L/min nozzle air flow rate, 6 mL/min feed rate, and 0.5 mm diameter nozzle tip. The solid dispersion of 30% NAP in PVP-VA was also prepared using the Pro-CepT Micro spray dryer (P-SDR) (Zelzate, Belgium) with the same formulation and process parameters as the B-SDR except for a 0.3 m3/min drying air flow rate and 10 L/min nozzle air flow rate. Thermal Analysis. Both the compressed and the uncompressed samples were analyzed using a Q2000 MDSC (TA Instruments, Leatherhead, UK) purged with inert dry nitrogen gas at a flow rate of 50 mL/min and connected to a refrigerated cooling system (RCS90). The temperature scale was calibrated and validated using indium and octadecane standards. The enthalpy response was also calibrated and validated using indium. The heat capacity was calibrated using a sapphire disk with modulating amplitude of 0.636 °C every 40 s and 2 °C/ min underlying heating rate. The heat capacity was further validated by comparing the measured with the theoretical values at 106.85 °C. The samples were crimped in a TA Instruments aluminum pan and heated from 0 to 160 °C. All samples were analyzed in triplicate. The data were acquired using Thermal Advantage software and analyzed using Universal Analysis software (version 4.4, TA Instruments, Leatherhead, UK). The onset and the offset of the glass transition region were determined from the derivative (with respect to temperature) of the reversing heat flow signals which correspond to the glass transition step change. The melting enthalpy was determined from the endotherm corresponding to the melting of NAP in the total heat flow signal. Powder X-ray Diffraction (PXRD). An automated X’pert PRO diffractometer (PANalytical, Almelo, The Netherlands) with a Cu tube and the generator set at 45 KV and 40 mA was used to analyze the solid dispersions, the physical mixtures, and the pure drug. The samples on a zero background plates were analyzed in reflection mode for a range of 4 ≤ 2θ ≤ 40° using a 0.0165° step size and 200 s counting time. The X’pert Data Collector (PANalytical, Almelo, The Netherlands) was used for continuous data acquisition, and the X’Pert Data Viewer and X’Pert HighScore Plus (PANalytical, Almelo, The Netherlands) were used for data analysis. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR). Spectra were collected with a Nicolet Avatar 370 FTIR spectrometer with a DTGS detector and equipped with Smart Orbit single bounce ZnSe as the internal reflectance element at room temperature. The spectra were recorded (32 scans) in the range between 400 and 4000 cm−1 with 4 cm−1 spectral resolution. Origin 8.6 (OriginLab Corp., Northampton, U.S.A.) was used for further analysis of the raw data. Savitzky-Golay mathematical algorithm was used to smooth the spectra, and the subtraction spectra were generated from the differences of the normalized spectra of the compressed and the uncompressed samples. The subtraction spectra were normalized by picking a point at specific wavenumber ≈1751 cm−1 on the unsubtracted spectrum. The intensity ratios were also computed from
■
RESULTS AND DISCUSSION I. Instant Impact of Compression on Physical Structure of Solid Dispersions. The compression pressures 1103
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
FTIR was therefore used to provide deeper insight into the possible structural changes following compression. The compression of 30% (w/w) NAP in PVP-VA solid dispersion led to a specific structural change based on FTIR (ATR) spectra (Figure 3). The compression of the solid dispersions
(except for 188 MPa) and the dwell time used in this study were higher than the values used during typical pharmaceutical tablet production. The higher pressures and the prolonged dwell time were used to forecast the long-term effect of compression on the structural stability of the solid dispersions. The glass transition width of the compressed and the uncompressed 30% (w/w) NAP in PVP-VA solid dispersion were similar (p > 0.05) for samples prepared with the P-SDR. On the contrary compression slightly increased the glass transition width, but it was not directionally coherent where a decrease in glass transition width was also observed for the lowest compression pressure, 188 MPa, for solid dispersions spray-dried using the B-SDR (Figure 1). The glass transition
Figure 3. Partial FTIR spectra of the 30% (w/w) NAP in PVP/VA 64 SD (B-SDR) before and after compression (n = 3) (*similar observation for higher compression pressure and solid dispersions prepared by P-SDR). Figure 1. Glass transition width of compressed and uncompressed 30% (w/w) NAP in PVP-VA 64 SD prepared using B-SDR and PSDR.
enhanced the specific interaction between amide carbonyl of PVP-VA 64 and hydrogen of carboxylic acid functional group of naproxen. The vibration band of the amide carbonyl becomes broader after compression due to an increase in the intensity of the shoulder peak for the weak drug−polymer interaction at 1654 cm−1. Similarly, the intensity of the vibration bands at 1634 cm−1 and 1603 cm−1 were also increased after compression (Figure 3). The vibration band at 1634 cm−1 can be the cumulative effect of both the in-plane naphthalene ring C−H bending of naproxen and the strongly H-bonded drug−polymer interaction, whereas 1603 cm−1 was due to the in-plane naphthalene ring C−C stretching of naproxen. Hence the increase in the intensity at 1634 cm−1 can be either from the ring deformation of naproxen or the strong drug−polymer interaction. The intensity ratio of the vibrational bands at 1634 and 1603 cm−1 for the compressed solid dispersions is shown on Figure 4B. The intensity ratio of these two vibrational bands for the uncompressed solid dispersion was similar to the compressed one. The intensity change of these vibrational bands was affected equally and may arise from similar structural changes as depicted on the subtraction spectra (Figure 3, green line). The intensity ratio evidently showed that compression may affect the naphthalene deformation instead of the strong drug−
width relationship between the compressed and the uncompressed samples was inconclusive. Solid dispersions prepared by the B-SDR were affected to a higher extent than the samples from the P-SDR which could be resulted from differences in process parameters and also the hardware design. The halo pattern of the uncompressed samples, both in the case of the B-SDR and the P-SDR, showed higher intensity at 2θ = 16.11 halo maximum than the compressed amorphous solid dispersions. The differences in the halo pattern apparently indicate the structural dissimilarity and also the differences in the short-range orders in the solid dispersion (Figure 2A and B). Even though glass transition width was inconclusive, the halo pattern showed structural differences between compressed and uncompressed samples. It indicated that the compression of the solid dispersion may improve the drug−polymer mixing. The halo pattern may seem more conclusive and sensitive to predict the physical structure changes after the compression than the calorimetric glass transition width. However, it still provides nondirectional information about the structural changes in the solid dispersions.
Figure 2. PXRD pattern of compressed and uncompressed 30% (w/w) NAP in PVP-VA 64 SD prepared using (A) B-SDR and (B) P-SDR. 1104
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
Figure 4. Intensity ratio of weak and strong drug−polymer interactions vibrational bands to reference peaks for both the compressed and the uncompressed 30% NAP in PVP-VA 64 SD.
drug−polymer interaction. The PXRD halo pattern also showed apparent differences at 2θ = 16.11 between uncompressed and compressed solid dispersions after 15 days of storage (Supporting Information B and C). However, the solid dispersions with 30% drug loading still showed a halo pattern even after five months storage at 75% RH at an ambient temperature for both the compressed and the uncompressed samples (Supporting Information D). On the other hand, solid dispersion made up of 40% and 50% of NAP showed crystallinity. The spray-dried 40% (w/w) NAP in PVP-VA 64 solid dispersion was amorphous after preparation, but this composition is apparently less stable than 30% drug loaded dispersion due to the lower polymer composition. After five months of storage at high humidity (75% RH) and ambient temperature, naproxen underwent partial crystallization as characterized by the emergence of its characteristic Bragg peaks (Figure 6B). Interestingly, the Bragg peak intensity was higher for the uncompressed solid dispersions than for the compressed at 188 MPa. The lower crystallinity of the compressed samples can be related to the improved drug− polymer interactions upon compression. The solid dispersions with 50% drug loading showed a melting endotherm at ca. 111.23 °C which is much lower than the melting point of the pure drug (Tm = 155 °C). Melting point depression of a pure component is a common phenomenon in the presence of a second component, impurities, and for very small particle sizes.16,17 The thermal behavior clearly showed that the solid dispersion was partially crystalline (Figure 6A). Similarly, after 10 days of storage the melting enthalpy and the crystallinity of the uncompressed sample were much higher (about 2-fold) than the compressed sample due to slow crystallization of naproxen from the amorphous domain of the compressed samples (Figure 6A). The PXRD pattern also showed that the intensity of the Bragg peaks was higher for the uncompressed sample than for the compressed samples after 5 months of storage (Figure 6B). This indicates that the compression of solid dispersions can suppress the crystallization of naproxen. The strong drug−polymer interaction and the naphthalene deformation peak at 1634 cm−1 showed a distinct difference in its intensity for 50% (w/w) NAP in PVP-VA 64 solid dispersions after 5 months of storage (Figure 7). The other naphthalene ring in plane C−C stretching peak at 1603 cm−1 remained unaffected among the compressed and the uncompressed solid dispersions. This difference showed that the amorphous naproxen crystallized faster from the uncompressed solid dispersions and further led to the distortion of the strong drug−polymer interaction vibration band. The distortion of the
polymer interaction. On the other hand the intensity ratio of the free amide carbonyl of the pyrrolidone ring of the polymer at 1672 cm−1 to the weak drug−polymer interaction shoulder at 1654 cm−1 was higher for the uncompressed solid dispersions than for the compressed (Figure 4A). The intensity of the drug−polymer weak interaction shoulder peak increases with the expense of the free amide carbonyl of the pyrrolidone ring after compression. The spectroscopic study evidently showed that the compression of 30% (w/w) NAP in PVP-VA solid dispersion enhanced the weak drug−polymer interaction. Effect of Compression on the Physical Stability of the Solid Dispersions. Different compositions of spray-dried solid dispersions were stored at 75% RH at ambient temperature to assess the role of improved drug−polymer interaction of compressed solid dispersions and also to evaluate the physical stability predicting power of the halo pattern differences. Compression could theoretically improve the physical stability of solid dispersions due to the enhanced drug−polymer weak interaction. The drug−polymer intermolecular interaction is one of the stabilization mechanisms of the solid dispersions.15 Compressed and uncompressed spray-dried solid dispersions with 10, 20, 25, 30, 40, and 50% (w/w) drug loading were further stored at 75% RH at ambient temperature for the longterm stability study. Both the compressed and the uncompressed one containing 10, 20, 25, and 30% NAP remain amorphous after five months of storage (Supporting Information A). After 21 days storage the glass transition width became broader for the uncompressed solid dispersions with 30% (w/w) NAP in PVP-VA 64 compared to the compressed samples (p < 0.05) (Figure 5). The glass transition width of the solid dispersions compressed at 1130 MPa was narrower than that from the lower compression pressure (188 MPa) for samples spray-dried using the B-SDR. The homogeneity of the compressed samples thus seems better which can be likely attributed to the enhancement of the weak
Figure 5. Glass transition width of compressed and uncompressed 30% (w/w) NAP in PVP-VA SD after 21 days storage at 75% RH and ambient temperature (a* and b** = p < 0.05, a″ and b″ = p < 0.05). 1105
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
Figure 6. (A) Melting enthalpy with physical stability for 50% and (B) PXRD pattern of 40% and 50% (w/w) NAP in PVP-VA 64 SD after a 5 month stability study.
The void volume of the spray dried powders was diminished, and the solid dispersions also exhibited particle−particle bonding for all of the compression pressures. The two stages of compactions apparently decreased the effective surface area for moisture sorption (Table 1). The solid dispersions of drugs such as felodipine, quinidine, and pimozide with hygroscopic polymers like PVPVA and PVP were prone to moisture induced amorphous−amorphous demixing.7 Hence the moisture content of both the compressed and the uncompressed solid dispersions was monitored during storage at 75% RH. The solid dispersions with higher PVP-VA compositions (10% drug loading) showed higher moisture content after 5 and 16 days of storage and even higher for the uncompressed samples compared to the other compositions (Table 2).
Figure 7. Partial FTIR spectra of 50% Nap/PVP-VA 64 after a 5 month stability study.
Table 2. Moisture Gain of Compressed and Uncompressed Solid Dispersions during the Physical Stability Study
interaction can lead to apparent decrease in the intensity of the vibration band at 1634 cm−1 but not at 1605 cm−1. Löbmann et al. reported the vibration band at 1725 cm−1 as the carbonyl band of the crystalline naproxen which was in a good agreement with our observation (1724 cm−1), and the authors also described that this vibration band shifted to 1728 cm−1 for amorphous forms.18 Accordingly the uncompressed 50% naproxen solid dispersion also showed a slight shift of the carbonyl vibration band from 1728 cm−1 to 1726 cm−1 due to the formation of a higher crystalline portion of naproxen as clearly indicated on the subtraction spectra (Figure 7, green line). The compaction of powders is a two-step process with compression followed by consolidation of solid particles. Compression is the first stage of compaction which is characterized by the decrease in the void volume. Solid particles form interparticle bonding during consolidation which governs the mechanical strength of a tablet.19 The smallest compression pressure, 188 MPa, used in this study was beyond the yield pressure of the solid dispersions for all of the compositions (Table 1).
% weight gain 5d
yield pressure (MPa)
tablet porositya (%)
30% 40% 50%
73.98 68.14 46.92
12.51 10.13 6.99
uncompressed
10% 20% 30% 40% 50%
10.50 8.46 5.77 4.44 2.62
16 d 188 MPa uncompressed 8.76 8.02 5.70 4.32 3.08
11.34 9.41 6.83 5.24 3.95
188 MPa 8.84 8.22 6.15 4.67 3.65
The weight gain of the solid dispersions with lesser polymer content was expectedly lower than the moisture gain of spraydried dispersions with the higher polymer compositions. The moisture gain after 5 and 16 days was similar for both the compressed and the uncompressed solid dispersions with 30, 40, and 50% (w/w) drug loading. The moisture sorption isotherm was also determined using a custom-made gravimetric technique to determine the kinetics of the moisture sorption on the solid dispersion powders. The 40% drug loading uncompressed solid dispersion established the equilibrium moisture content within 6 h and the compressed samples absorbed 0.5% less moisture after a similar exposure time (Figure 8). Similarly, both the compressed and the uncompressed 50% (w/w) NAP/ PVP-VA 64 solid dispersions reached equilibrium moisture content within 9 h (Figure 8). However, the uncompressed samples apparently absorb moisture faster than the compressed samples, and their moisture content differs by 0.3% after 3 h. The moisture gains after 16 days were similar for the compressed and the uncompressed 30, 40 and 50% (w/w)
Table 1. Yield Pressure and the Tablet Porosity of the Binary Solid Dispersions NAP/PVP-VA 64 (% drug (w/w))
NAP/PVP-VA 64 (% drug (w/w))
a
Determined from the true density and the tablet thickness, which takes into account the elastic recovery. 1106
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
substrates often decreases the molecular relaxation on the surface of the polymer which leads to slow molecular dynamics.20 The compression pressures were high enough to trigger particle−particle consolidation in the solid dispersions (Table 1). Part of the solid particles in a solid dispersion may act as an interacting substrate as evidently shown with an increase in drug−polymer interactions that retards molecular relaxation leading to slower molecular dynamics on the interfacial region. The slower molecular dynamics of compressed samples may be the stabilization mechanism of the spray-dried solid dispersions.
■
Figure 8. Moisture sorption kinetics of 40% and 50% drug loading compressed and uncompressed solid dispersions.
CONCLUSION The impact of compression was investigated for solid dispersions of naproxen and PVP-VA 64. The glass transition width was not sensitive enough to predict the structural changes induced by compression. On the other hand PXRD data showed a halo pattern differences due to compression induced structural changes, but it lacks directional information. FTIR clearly revealed improved drug−polymer interaction as a result of compression; hence it provided real molecular level information. Therefore, the glass transition width and the PXRD halo patterns can be used in combination with vibrational spectroscopic techniques (FTIR) to predict the drug−polymer mixing and the physical stability of the solid dispersions. Compression clearly improved drug−polymer interactions, and this was further manifested as lesser crystallinity compared to the uncompressed solid dispersions during physical stability study. Even though compression and mechanical stresses often facilitate crystallization of glassy systems, spray-dried NAP/PVP-VA 64 solid dispersions was stabilized by compression due to enhanced drug−polymer interactions. This input information can be very crucial in dosage form selection for solid dispersions between tablet and capsules.
NAP containing solid dispersions since the equilibrium moisture content was established rapidly (Table 2). This indicated that the equilibrium moisture content was rapidly reached with a slightly higher rate for the uncompressed samples, and afterward the moisture contents of the compressed and the uncompressed solid dispersions were similar during the physical stability study period. Drug−polymer interactions often improve the physical stability of amorphous solid dispersions. Felodipine forms amorphous solid dispersions with PVP due to the strong drug− polymer intermolecular interactions, whereas it forms partially crystalline solid dispersions with poly(ethylene glycol) (PEG) due to the weak interactions and the low capacity of PEG to incorporate drugs.15 Moisture can lead to an irreversible decrease/disappearance in the drug−polymer interactions which further lead to amorphous−amorphous phase separations.8 The physical stability of compressed NAP/PVP-VA 64 solid dispersions can be similarly originated from the improved weak drug−polymer interaction. The higher degree of interactions restricts the molecular mobility and the molecular diffusion which will hinder nucleation and retard crystallization. A higher intermolecular interaction may not be the entire source of this slow crystallization in the compressed solid dispersions. The bulk molecules are usually confined by the surface molecules and have low degree of freedom. But molecules on the surface have low degree of confinement and a higher molecular mobility which leads to faster surface crystallization than the bulk.20 Surface crystallization is a common phenomenon for amorphous drugs. Amorphous griseofulvin undergoes 10- to 100-fold faster crystallization from its surface than the bulk.21 Yoshioka et al. also showed that amorphous indomethacin and its solid dispersions with PVP exhibited particle size dependent crystallization.22 Indomethacin crystallized faster from the smaller particle size amorphous solids than from the larger particle size materials. Indomethacin probably undergoes high surface crystallization, and the smaller particle size provides a larger surface with a higher molecular mobility.22 The hindrance toward crystallization decreases due to lower activation energy to crystallization for higher surface area particles. Similarly, the uncompressed naproxen/PVP-VA 64 solid dispersions has a much higher surface area than the compressed ones which were compressed using a pressure beyond their yield pressure (Table 1). The decrease in effective surface area due to compression probably diminished the crystallization of naproxen in the solid dispersions. Naproxen is a very poor glass former, and investigating its surface crystallization behavior was practically difficult. On the other hand interaction of glassy polymers and
■
ASSOCIATED CONTENT
S Supporting Information *
PXRD pattern of compressed and uncompressed 10, 20, and 25% NAP in PVP-VA SD after 5 months of storage at 75% RH/ambient temperature (SI-A), PXRD pattern of 30% NAP in PVP-VA 64 SD prepared by B-SDR (SI-B) and P-SDR (SIC) after 3 weeks of storage at 75% RH and ambient temperature, and PXRD pattern of 30% NAP in PVP-VA 64 SD prepared by B-SDR after 5 months of storage at 75% RH and ambient temperature (SI-D). This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: +32-16-330300; fax: +32-16-330305; e-mail: Guy.
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Z.A. acknowledges the financial support of IRO scholarship, KU Leuven. FWO-Vlaanderen (G.0764.13) and KU Leuven (OT/12/077) are acknowledged for financial support. The authors also acknowledge Department of Metallurgy and Materials Engineering (MTM), KU Leuven for providing ATR-FTIR and TGA facility. Patrick Rombaut and Danny 1107
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
Molecular Pharmaceutics
Article
experimentally determined interaction parameters. Pharm. Res. 2009, 26, 139−151. (17) Paudel, A.; Van Humbeeck, J.; Van den Mooter, G. Theoretical and experimental investigation on the solid solubility and miscibility of naproxen in poly (vinylpyrrolidone). Mol. Pharmaceutics 2010, 7, 1133−1148. (18) Löbmann, K.; Laitinen, R.; Grohganz, H.; Strachan, C.; Rades, T.; Gordon, K. C. A theoretical and spectroscopic study of coamorphous naproxen and indomethacin. Int. J. Pharm. 2012, 453, 80− 87. (19) Pore, M. Pharmaceutical tablet compaction: product and process design. Thesis (Ph.D.), Massachusetts Institute of Technology, Cambridge, MA, 2007; http://dspace.mit.edu/handle/1721.1/ 51623#files-area. (20) Yang, C.; Takahashi, I. Broadening, no broadening and narrowing of glass transition of supported polystyrene ultrathin films emerging under ultraslow temperature variations. Polym. J. 2011, 43, 390−397. (21) Zhu, L.; Jona, J.; Nagapudi, K.; Wu, T. Fast surface crystallization of amorphous griseofulvin below Tg. Pharm. Res. 2010, 27, 1558−1567. (22) Crowley, K. J.; Zografi, G. The effect of low concentrations of molecularly dispersed poly (vinylpyrrolidone) on indomethacin crystallization from the amorphous state. Pharm. Res. 2003, 20, 1417−1422.
Winant are acknowledged for PXRD and TGA sample analysis, respectively.
■
ABBREVIATIONS USED ATR-FTIR, attenuated total reflectance Fourier transform infrared spectroscopy; B-SDR, Buchi mini spray dryer; ΔHf, heat of fusion; MDSC, modulated differential scanning calorimetry; NAP, naproxen; P-SDR, Pro-C-epT Micro spray dryer; RCS 90, refrigerated cooling system; SD, solid dispersions; SI, Supporting Information; TGA, thermogravimetric analysis; Tm, melting point; PBT, poly(butylene terephthalate); PEG, poly(ethylene glycol); P2O5, phosphorus pentoxide; PVP, polyvinylpyrrolidone; PVP-VA 64, poly(1vinylpyrrolidone-co-vinyl acetate); PXRD, powder X-ray diffraction
■
REFERENCES
(1) Lipinski, C. A. Drug-like properties and the causes of poor solubility and poor permeability. J. Pharmacol. Toxicol. Methods 2000, 44, 235−249. (2) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Delivery Rev. 2001, 46, 3−26. (3) Leuner, C.; Dressman, J. Improving drug solubility for oral delivery using solid dispersions. Eur. J. Pharm. Biopharm. 2000, 50, 47−60. (4) Singh, A.; Worku, Z. A.; Van den Mooter, G. Oral formulation strategies to improve solubility of poorly water-soluble drugs. Expert Opin. Drug Delivery 2011, 8, 1361−78. (5) Chiou, W. L.; Riegelman, S. Pharmaceutical applications of solid dispersion systems. J. Pharm. Sci. 1971, 60, 1281−1302. (6) Van den Mooter, G. The use of amorphous solid dispersions: A formulation strategy to overcome poor solubility and dissolution rate. Drug Discovery Today: Technol. 2012, 9, e79−e85. (7) Rumondor, A. C.; Taylor, L. S. Effect of polymer hygroscopicity on the phase behavior of amorphous solid dispersions in the presence of moisture. Mol. Pharmaceutics 2010, 7, 477−490. (8) Marsac, P. J.; Rumondor, A. C.; Nivens, D. E.; Kestur, U. S.; Stanciu, L.; Taylor, L. S. Effect of temperature and moisture on the miscibility of amorphous dispersions of felodipine and poly(vinyl pyrrolidone). J. Pharm. Sci. 2010, 99, 169−185. (9) Damian, F.; Blaton, N.; Kinget, R.; Van den Mooter, G. Physical stability of solid dispersions of the antiviral agent UC-781 with PEG 6000, Gelucire® 44/14 and PVP K30. Int. J. Pharm. 2002, 244, 87−98. (10) Ayenew, Z.; Paudel, A.; Rombaut, P.; Van den Mooter, G. Effect of compression on non-isothermal crystallization behaviour of amorphous indomethacin. Pharm. Res. 2012, 29, 2489−2498. (11) Li, L.; de Jeu, W. H. Shear-induced crystallization of poly (butylene terephthalate): A real-time small-angle X-ray scattering study. Macromolecules 2004, 37, 5646−5652. (12) Gray, D. “Transcrystallization” induced by mechanical stress on a polypropylene melt. J. Polym. Sci., Polym. Lett. Ed. 1974, 12, 645− 650. (13) Rangel-Nafaile, C.; Metzner, A. B.; Wissbrun, K. F. Analysis of stress-induced phase separations in polymer solutions. Macromolecules 1984, 17, 1187−1195. (14) Heckel, R. Density-pressure relationships in powder compaction. Trans. Metall. Soc. AIME 1961, 221, 671−675. (15) Karavas, E.; Georgarakis, E.; Sigalas, M. P.; Avgoustakis, K.; Bikiaris, D. Investigation of the release mechanism of a sparingly water-soluble drug from solid dispersions in hydrophilic carriers based on physical state of drug, particle size distribution and drug−polymer interactions. Eur. J. Pharm. Biopharm. 2007, 66, 334−347. (16) Marsac, P. J.; Li, T.; Taylor, L. S. Estimation of drug−polymer miscibility and solubility in amorphous solid dispersions using 1108
dx.doi.org/10.1021/mp5001313 | Mol. Pharmaceutics 2014, 11, 1102−1108
