Microcalorimetric Measurement of the Interactions between Water Vapor and Amorphous Pharmaceutical Solids David Lechuga-Ballesteros, 1,3 Aziz Bakri, 2 and Danforth P. Miller 1 Received July 26, 2002; accepted October 22, 2002 Purpose. Use a microcalorimetric technique to measure the interac- tions between water vapor and amorphous pharmaceutical solids and describe the relationship between long-term physical stability and the storage relative humidity (RH) at constant temperature. Methods. A thermal activity monitor was used to characterize inter- actions of water vapor with spray-dried amorphous sucrose, lactose, raffinose, and sodium indomethacin. Differential scanning calorim- etry was used to measure glass transition temperature, T g . X-ray powder diffraction was used to confirm that the spray-dried samples were amorphous. Scanning electron microscopy was used to examine particle morphology. Specific surface area was determined by BET analysis of nitrogen and krypton adsorption isotherms. Results. The moisture-induced thermal activity traces (MITATs) of the materials in this study exhibit general behavior that helps explain the effect of moisture content on the physical stability of the glassy phase at a given storage temperature. At some RH threshold, RH m , the MITAT exhibits a dramatic increase in the energy of interaction between water vapor and the glass that cannot be explained by a phase or morphology change. Calorimetric data indicate that water vapor-solid interactions are reversible below RH m ; above RH m , en- ergetic hysteresis is observed and water-water interactions predomi- nate. In addition, the MITAT was deconvoluted into sorptive and nonsorptive components, making it possible to assign the observed heat flow to unique thermal events. Samples stored at a RH just below RH m for more than 2 months show no evidence of morphology or phase change. In addition, the MITAT can be deconvoluted into sorptive and nonsorptive components by using a twin-calorimeter arrangement. This analysis provides specificity to the microcalorimet- ric analysis and helps explain the nature of the physical changes that occur during the hydration glassy phase. Conclusions. The MITAT is a useful tool to determine the onset of moisture-induced physical instability of glassy pharmaceuticals and may find a broad application to determine appropriate storage con- ditions to ensure long-term physical stability. KEY WORDS: microcalorimetry; MITAT; glass transition tempera- ture; BET monolayer; hydration limit; stability. INTRODUCTION Chemical reactivity or physical changes of glassy phar- maceuticals (i.e., structural collapse, crystallization) occur be- cause of disorder and increased molecular mobility. Methods are available to estimate average relaxation times, a measure of molecular mobility, and meaningful correlations to stability have been proposed (1,2). Most techniques use the dry state (i.e., zero water content) as the reference state and therefore do not provide information on the effect of water content on average relaxation times. To our knowledge, there are only two cases in which the effect of absorbed water on the glass transition temperature, T g , and average relaxation times has been studied (3,4). Andronis and Zografi (4) report that mo- lecular mobility, determined from either the heating rate de- pendence of T g or dielectric measurements of the average relaxation time, is significant below T g . This helps explain why amorphous indomethacin crystallizes at temperatures well below its T g . Kajiwara et al. (3) report the decrease in the “zero mobility temperature” T o , of raffinose, measured by the heating rate dependence of T g . However, neither of these studies addresses the relationship between T o and a critical water content. Water molecules are always mobile within a hydrated glassy solid. In fact, it has been shown that the diffusivity of water molecules within a glassy matrix is finite and substantial below T g (5). However, for small molecule glasses, amor- phous polymers and proteins alike, water mobility decreases at water contents below the hydration limit, W m (6–9). It has also been shown that the degrees of freedom of water mobil- ity within a glass are restricted below W m . For example, di- electric resonance spectroscopy experiments on lysozyme show that water molecules are irrotationally bound to a pro- tein surface at such hydration levels (10). The use of W m as a moisture threshold for stability is similar to treatments based on W g (11), the amount of water required to depress T g to the storage temperature of the surroundings. However, W g is a less-rigorous criterion than W m since instability can occur at temperatures far below T g . Furthermore, in many instances, the T g of formulations cannot be determined because of in- herent limitations of traditional experimental methods. How- ever, it is safe to assume that the T g of a protein-rich formu- lation, even at its hydration limit, is much higher than typical storage temperatures (i.e., T g >>40°C). This is an important consideration since it demonstrates that molecular motions implicated in the decomposition processes of protein-rich for- mulations at low water contents may be independent of T g , possibly because of decoupling of water mobility and protein mobility. The hydration limit has also been linked to the long-term chemical stability of glassy formulations. Below W m , im- proved chemical stability of foodstuffs (12) and therapeutic proteins has been extensively reported (13–15). In general, freeze-dried protein formulations are more stable at lower water contents. It is generally acknowledged that the molecu- lar mobility of water and the glassy matrix at hydration levels above W m are among the main factors that affect overall stability. Generally, the minimum in the rate of bimolecular reac- tions is found near the hydration limit (12,13,16). It has been proposed that above W m , water-water interactions increase, thereby favoring the formation of microscopic regions of con- densed water. These regions can promote chemical instability since chemical species can readily dissolve, diffuse, and react (12). In addition to the many references to the relationship between chemical stability and the hydration limit, W m has 1 Nektar Therapeutics (formerly Inhale Therapeutics Systems, Inc.), 150 Industrial Road, San Carlos, California 94070. 2 Université Joseph Fourier at Grenoble, Formulation and Process Engineering, Faculty of Pharmacy, ISERM 008. 3 To whom correspondence should be addressed. (e-mail: dlechuga@ ca.nektar.com) Pharmaceutical Research, Vol. 20, No. 2, February 2003 (© 2003) Research Paper 308 0724-8741/03/0200-0308/0 © 2003 Plenum Publishing Corporation