Article pubs.acs.org/molecularpharmaceutics
Accelerated Physical Stability Testing of Amorphous Dispersions Mehak Mehta and Raj Suryanarayanan* Department of Pharmaceutics, University of Minnesota, Minneapolis, Minnesota 55455, United States S Supporting Information *
ABSTRACT: The goal was to develop an accelerated physical stability testing method of amorphous dispersions. Water sorption is known to cause plasticization and may accelerate drug crystallization. In an earlier investigation, it was observed that both the increase in mobility and decrease in stability in amorphous dispersions was explained by the “plasticization” effect of water (Mehta et al. Mol. Pharmaceutics 2016, 13 (4), 1339−1346). In this work, the influence of water concentration (up to 1.8% w/w) on the correlation between mobility and crystallization in felodipine dispersions was investigated. With an increase in water content, the α-relaxation time as well as the time for 1% w/w felodipine crystallization decreased. The relaxation times of the systems, obtained with different water concentration, overlapped when the temperature was scaled (Tg/T). The temperature dependencies of the α-relaxation time as well as the crystallization time were unaffected by the water concentration. Thus, the value of the coupling coefficient, up to a water concentration of 1.8% w/w, was approximately constant. Based on these findings, the use of “water sorption” is proposed to build predictive models for crystallization in slow crystallizing dispersions. KEYWORDS: felodipine, water, molecular mobility, crystallization, relative humidity, coupling model, prediction, dielectric spectroscopy, PVP
■
the dispersion.11 This plasticization effect of water allows the accelerated stability studies to be conducted at or close to room temperature. In several model systems, a correlation between mobility and crystallization (both onset and kinetics)3,12 has been observed. In itraconazole, there was an excellent coupling between the αrelaxation and crystallization onset times.13 In amorphous dispersions, hydroxypropyl methylcellulose acetate succinate (HPMCAS) substantially increased the α-relaxation time, with an attendant increase in crystallization onset time demonstrating the correlation between α-relaxation and physical stability.5 However, these studies were restricted to systems with negligible water content. Therefore, a logical extension is to determine the effect of water content on the correlation between mobility and crystallization. The combined influence of temperature and sorbed water content on the crystallization behavior of amorphous pharmaceuticals has been investigated. The crystallization behavior of supercooled nilvadipine was investigated with different water contents.14 The plots of crystallization time vs Tg/T overlapped, and this observation was attributable to the plasticizing effect of water. Similar observations were reported in lamotrigine mesylate15 and in an amorphous dispersion.16 In our earlier investigation, the effect
INTRODUCTION A major challenge in the development of amorphous dispersions is the reliable prediction of long-term physical stability. Studying degradation kinetics at elevated temperatures and extrapolating to room temperature, an approach classically used for predicting chemical stability, cannot be used for physical stability (crystallization) prediction. Molecular mobility has been used to build predictive models for crystallization in both dry amorphous drugs1−4 and dispersions.5−7 Such an approach may not be practically feasible in systems with high polymer loading, since crystallization is extremely slow. It will therefore be useful to develop an accelerated stability (physical) testing method enabling crystallization prediction under practical storage conditions. The goal, much like in chemical degradation studies, is to accelerate drug crystallization from amorphous dispersions. This can be accomplished by monitoring crystallization under elevated temperature or by increasing the water content in dispersions. Exposing to elevated temperatures is not a desirable approach for at least two reasons. (i) There is potential for chemical degradation, particularly if the temperature is substantially higher than room temperature. (ii) If the studies are conducted in the supercooled state (as is likely to be the case), extrapolation of the results to the glassy state may lead to erroneous conclusion. This is because of the discontinuity in the temperature dependence of the properties of interest.8,9 Whereas water sorption can facilitate crystallization by decreasing the glass transition temperature (Tg)10 of © 2016 American Chemical Society
Received: Revised: Accepted: Published: 2661
March 11, 2016 May 16, 2016 June 14, 2016 June 14, 2016 DOI: 10.1021/acs.molpharmaceut.6b00218 Mol. Pharmaceutics 2016, 13, 2661−2666
Article
Molecular Pharmaceutics
Figure 1. Dielectric loss spectra of FEL−PVP dispersion, (a) in the dry state and containing (b) 1.5% and (c) 1.8% w/w water, respectively. The loss curves have been normalized to the maximum loss value.
of water concentration on the glass transition temperature and the molecular mobility in NIF dispersions was investigated.4 The increase in molecular mobility was attributed to the plasticization effect of water. Interestingly the temperature dependence of the α-relaxation time was independent of water content. A single linear profile described the relationship between crystallization time and Tg/T in both dry systems and systems containing water. Thus, the temperature dependencies of the α-relaxation time as well as the crystallization time were unaffected by the water concentration. We therefore hypothesize that the extent of coupling between molecular mobility and crystallization time would remain unaf fected in dispersions with low water contents. If the hypothesis is valid, then the coupling coefficient obtained from studying the stability of dispersions with different water contents can be used to predict physical stability (drug crystallization) in dry systems. In other words, it could serve as an accelerated testing method to potentially predict physical stability under storage conditions of pharmaceutical interest. This could be particularly useful in dispersions with high polymer loading which may be stable over short time periods (time scale of months) but may not have the desired shelf life (typically 24 months).
Dielectric Spectroscopy (DES). A broadband dielectric spectrometer (Novocontrol Alpha-AK high performance frequency analyzer, Novocontrol Technologies, Germany) was used to measure molecular motions over a frequency range of 10−2 to 107 Hz. The details of the sample preparation and data analysis are provided elsewhere.4 X-ray Diffractometry (XRD). The powder samples were exposed to Cu Kα radiation (40 kV and 40 mA) in a powder Xray diffractometer (D8 ADVANCE; Bruker AXS, Madison, WI, USA) equipped with a variable temperature stage (TTK 450; Anton Paar, Graz-Straßgang, Austria) and Si strip onedimensional detector (LynxEye; Bruker AXS). The diffraction patterns were obtained over an angular range of 5 to 40° 2θ with a step size of 0.05° and a dwell time of 1 s. Synchrotron XRD (SXRD; Transmission Mode). Isothermal crystallization studies were carried out, in situ, in the beamline (17-BM-B) at Argonne National Laboratory (Argonne, IL, USA). A monochromatic X-ray beam [wavelength 0.73040 Å; beam size 250 μm (horizontal) × 160 μm (vertical)] was used. The details of the sample preparation, experimental method, and quantification of XRD data have been described earlier.3
EXPERIMENTAL SECTION Preparation of Amorphous Dispersion. FEL (C18H19Cl2NO4) and PVP (K17 grade) were used as received. The amorphous dispersions were prepared by solvent evaporation followed by melt-quenching. Details of the experimental procedure are provided in an earlier report.4 The water content of the samples was consistently
