(103a) Spray-Dried Lipid-Microparticles for Inhalation: Correlation between Solid State and Processability

Spray-drying is a versatile process in which a solution containing solid components is atomized into a hot gas promoting rapid evaporative mass transfer. Targeted particle attributes are provided by the interplay of material science and process parameters. Inhalable particles, composed of active pharmaceutical ingredients and specialized excipients, can be readily yielded. However, technical hurdles are commonly experienced when lipid-based excipients (LBE) are spray-dried. Sticky products, extensive deposition on the spray-dryer walls, or simply inability to yield a solid product have been associated to the low melting point of LBE. Comprehensive knowledge on the correlation between lipid solid state and process boundaries can lead to enhanced processability. In this work, several LBE possessing high melting points were screened in combination with ibuprofen (IBU) as a model drug. Particle attributes were envisaged to provide systemic delivery of IBU after pulmonary administration. The following LBE were employed: behenoyl polyoxyl-8 glycerides (BPG) available as Compritol® HD5 ATO, tripalmitin (PPP) available as Dynasan® 116, and triacylglycerol ester of behenic acid (PG3C22p) available as Witepsol® PMF 123.

Methods

Prior to spray-drying the solid state interactions of LBE and IBU were investigated via binary phase diagrams. Alteration of the melting points allowed the identification of temperature boundaries of spray-drying and selection of API-loading. Subsequently, solutions of LBE:IBU at 1.5%w/w solid content in tetrahydrofuran (THF) were prepared. Viscosity, surface tension and density were determined. The solutions were spray-dried. Mass and heat transfer rates and droplet drying kinetics were set constant to the most extent for all LBE:IBU by controlling the outlet temperature (T_out) and initial droplet size. Spray-drying was carried out at inlet temperature (T_in) of 71°C, using nitrogen as stagnant drying gas at 0.3m³/h. Atomization was carried out at feed rate of 3.2 g/min through a 0.2mm bi-fluid nozzle and gas to liquid mass flow ratio of 1.45. The process yield was obtained and resulting particles were characterized. Lastly, the thermal behavior and crystallization of LBE was assessed via differential scanning calorimetry (DSC) and small and wide x-ray scattering (SWAXS) and correlated to the processability.

Results

The phase diagrams LBE:IBU evidenced depression of melting points due to eutectic interaction. Liquid and solid regions were identified. The eutectic composition, having the least solid-liquid transitions at the eutectic temperature (T_eu), was selected for spray-drying. Since it is generally accepted that during spray-drying, particles experience most of their residence time at T_out, a rational criterion to guarantee processability was to achieve T_out below T_eu. This would assure the existence of a unique solid region decreasing the risk of softening and melting at the outlet. T_eu was found to be T_BPG=45.7, T_PG3C22p=60.5 and T_PPP=60.1°C at LBE:IBU ratio of 70:30.

Viscosity and surface tension of all LBE:IBU solutions were not significantly different, therefore atomization led to nearly equal initial droplet size. Semi-empirical estimation of the droplet Sauter mean diameter (DSMD) based on Weber and Onhesorge numbers yielded values between 8.931 and 9.012 µm. The combination of process parameters led to T_out between 32-35°C, which guaranteed the existence of solid region for all LBE:IBU.

Despite having equal initial droplet size, analogous mass-heat transfer and T_out below T_eu, the resulting outcome from each LBE:IBU was markedly different. Agglomeration and low yield were depicted by BPG (y=28.6%) with sauter mean diameter (SMD=107.7µm) notably larger than DSMD. PG3C22p and PPP, although having the same T_eu showed yields of 76.2 and 10.6%, respectively. PPP displayed agglomeration (SMD=19.19µm), bimodal particle size distribution and general impairments, whereas PG3C22p yielded uniform non-agglomerated particles (SMD=1.973µm).

It is reasonable to deduce that a minimum difference between T_eu and T_out should be provided. Produced particles stayed in the collection vessel at T_out until the feedstock was completely atomized. Meanwhile they followed the characteristic rotational movement of the gas in the vessel (caused by the cyclone) which potentially created friction among particles and sections of higher temperature. The closeness of the T_out to T_eu increases the risk of softening; as seen for BPG.

However, this should not be the only consideration for lipid-based spray-drying. During spray-drying, the droplet temperature reaches the saturation temperature (T_s), that is the wet bulb temperature of a nitrogen-solvent vapor (THF in this study) mixture. T_s is usually around 10-15°C below the T_out. When the particle is crystallized, its temperature is promptly increased to the dry bulb temperature of the surrounding gas. A particle recently crystallized near the atomization centerline will directly exit the drying chamber, experiencing the shortest residence time and a highest temperature equal to T_out. Whereas particles that get entrained within a recirculating eddy will experience longer residence time and exposition to the high T_in without the protective cooling effect of THF-vapor. Reaching T_in=71°C leads to instant melting for all LBE:IBU. As the time-scale of the process is very short, the temperature of the recently melted particle promptly decreases as it moves towards the outlet. The melted particle experiences a temperature profile from T_in to T_out leading to recrystallization from the melt.

PG3C22p melted particles recrystallized almost immediately, since its crystallization temperature (T_c=57.4°C) is very close to the T_eu. This is due to the characteristic absence of polymorphism and phase separation of polyglycerol esters of fatty acids (PGFAs). Whereas PPP recrystallizes from the melt at T_c=31.4°C into a different polymorph (α-form). Besides the simply relation to T_out, presented above, BPG melted in two phases and recrystallized into five, with the lowest T_c at 22.8°C. The impairments of spray-drying of BPG and PPP can be attributed to the existence of molten fractions, not being able to recrystallize at the surrounding drying gas temperature due to polymorphism or multi-phases.

PG3C22p, in combination with IBU, was effectively spray-dried. Lipid micro-particles of volume mean diameter of 6.586±0.543µm, density of 0.389±0.007 g/cm³ and corrugated surface were produced. The application of the lipid-microparticles as carrier-free dry powder for inhalation was tested using Aerolizer coupled with the next generation impactor. High emitted fraction (>90%), mass median aerodynamic diameter of 3.568±0.113µm and fine particle fraction of 45.6±1.6% were achieved. Although a LBE was employed, high yield of microparticles was obtained by a simple spray-drying process. No device modifications or material-aids were required.

Conclusion

The absence of processability of LBE via spray-drying has been associated to their low melting points. Sufficiently low T_out are typically recommended to avoid product loss. However, it was demonstrated that, LBE having the same melting temperature and spray-dried at low T_out, result in entirely different yields and particle attributes. The complex crystallization pathways of LBE strongly impact the process performance even at sufficiently low T_out. Crystallization into polymorphic forms or multi-phases negatively impact the process yield associated to the formation of liquid fractions of long residence time in the spray-drying chamber. The application of high melting LBE undergoing minimum solid state transitions, such as PG3C22p, can broaden the operational window of spray-drying leading to further possibilities of particle engineering.

Acknowledgment: IOI Oleo GmbH, the Austrian Funding Agency (FFG) and the Doctoral Academy NanoGraz, University of Graz.