To meet the need of the ever-growing demand of increased oral bioavailability of Biopharmaceutics Classification System (BCS) class II active pharmaceutical ingredients (APIs), many formulators are turning to amorphous spray dried dispersions, given their track record of successful bioavailability enhancement. However, some molecules exhibit a high propensity to phase separate and recrystallize from the amorphous state.

These APIs present a challenge that may be overcome by focusing on variables in the SDD manufacturing process, including condenser temperature, particle size and dryer chamber temperature, that can affect SDD stability. This approach connects the materials science with the process science to stabilize amorphous spray dried dispersion formulations and enhance the bioavailability of low-solubility APIs.

Note: Amorphous spray dried dispersion technology has a demonstrated track record of success in increasing the oral bioavailability of compounds that present absorption challenges due to low solubility. One of the principal “critical to quality attributes” of an SDD is physical stability.¹

Characterizing SDD and Crystallization Risk

A drug’s propensity to crystallize can be predicted by several physicochemical properties of the API, one of which is the ratio of the melting point (Tm) to the glass transition temperature (Tg) in degrees Kelvin. The propensity of a SDD with a high Tm/Tg to crystallize is typically mitigated by excipient or polymer selection to optimize the formulation for physical stability.

However, for drugs with a very low Tg and consequently a high Tm/Tg, the drug loading may be reduced to lower the propensity for recrystallization. The lower drug loading changes the material properties (i.e., Tg) of the SDD to be more similar to the polymer which decreases the mobility of the system when the Tg of the matrix polymer is high relative to the API.

Decreasing the drug loading also limits the diffusion of the drug within the polymer matrix, slowing the kinetics of phase separation and crystallization during processing and/or storage of the SDD. However, lowering the drug loading in the SDD decreases manufacturing throughput and increases tablet burden, potentially lowering patient compliance.

For these reasons, SDDs are usually formulated to maximize drug loading as much as possible. For formulations that are prone to crystallization, the optimization of spray drying process conditions may afford a higher drug loading, while preserving physical stability. These relationships have been extensively studied and general formulation maps developed to serve as SDD formulation guidelines.

Figure 1 is one such map depicting the relationship between lipophilicity and the ratio of the melting point to the glass transition temperature.

Characterization of the SDD and modeling of process conditions can be used to minimize or eliminate the risk of solid state recrystallization if an API demonstrates or is suspected to be at risk for crystallization. This is done by characterizing the SDD, then using that information to define process conditions with appropriate constraints to manufacture a dispersion with the desired attributes.

An important first step in this strategy is to define the Tg of the SDD as a function of residual solvent content, since excess residual solvent within the SDD acts to plasticize the matrix and depress the Tg, increasing the mobility of the API and the propensity for phase separation and crystallization.

Dynamic vapor sorption (DVS) experiments provide the equilibrium solvent content on the SDD as a function of the solvent relative saturation in the spray dryer atmosphere. Representative DVS results are shown in Figure 2.

In practice, the solvent levels in the SDD upon exiting the spray dryer are slightly higher than the equilibrium defined by the DVS due to increased resistance in the drying kinetics as a droplet transitions to a particle; so while a conservative interpretation should be made of these data, this method provides good guidance for understanding the SDD properties under the conditions realized in the spray dryer chamber. The Tg may then be measured or modeled as a function of solvent content.

There are several models for predicting the Tg as function of solvent content based on the properties of the components. One of the simplest that can be used with easily obtainable parameters is the Fox Equation, which has shown to provide adequate accuracy under the process conditions of interest for many materials.

Where x1 and x2 represent the mass fractions of the solvent and solid SDD, Tg1, and Tg2 are the corresponding glass transition temperatures of the dry SDD and pure solvent. Representative data showing the Tg for a SDD and the prediction afforded by the Fox Equation are shown in Figure 3. These data are then used to select process conditions by applying our model-based methodology for spray-drying process development.²

Process Variables That Can Optimize SDD Stability
Establishing the proper set points and allowable operating range for process parameters is a challenging exercise when using a purely empirical approach to determine spray drying conditions, and this task is only confounded when the physical stability of the product is a risk.

One of the first priorities for defining a robust manufacturing process is to reduce the solvent content within the SDD from initial particle formation through packaging and storage of the final product.

With an understanding of the SDD properties, a model based approach enables mapping of the available process space as a function of critical processing parameters to optimize physical stability, while maintaining the desired attributes for downstream processing (e.g., tableting) and efficacy. This methodology enables process optimization while minimizing the use of additional raw materials and the expense of trials at the laboratory or pilot scales.

Liquid to Gas Ratio
The solution and drying gas flow rates determine the liquid to gas ratio (L/G), which is a primary process variable that may be used to control the relative saturation of solvent and consequently, the solvent content in the SDD coming out of the spray dryer. The drying gas flow rate is typically run constant, at the maximum possible rate, but is limited by the spray dryer scale and the equipment configuration. This leaves the solution flow rate as the only degree of freedom to reduce the L/G.

However, the tradeoff of decreasing the solution flow rate is a decrease in throughput. Other process variables can be adjusted to influence SDD stability without sacrificing process capacity.

Condenser Temperature
Spray drying at a typical commercial scale requires a closed loop, in which the drying gas is recycled in process by adding an inline condenser to remove the solvent from the gas stream as it recirculates through the dryer.² The relative saturation of the solvent in the gas stream under these conditions is in part a function of condenser performance. The temperature of the condenser is limited by design and capability of the chiller used.

Figure 4 shows the difference between the predicted Tg and the temperature of the dryer chamber (T_out) vs. the condenser temperature for a typical spray drying process at the PSD-2 scale using acetone as the spray solvent with all other parameters held constant. Tg – T_out provides a simple metric that affords the relative stability as a function of process conditions; as the difference between these temperatures decreases, the process approaches conditions that raise the probability of phase separation and crystallization.

The results in Figure 4 are color coded to show the change in relative saturation over a range of condenser temperatures from -20 C to 20 C. The relative saturation of the solvent in the dryer atmosphere rises with the condenser temperature, resulting in a decrease between the difference in the glass transition and outlet temperature, increasing the risk for crystallization within the dispersion.

Figure 4 illustrates the importance of considering the condenser temperature during process design and scale up under closed loop spray drying conditions. The investment to improve condenser performance will increase process viability and final product stability when physical stability of the SDD is at risk.

Atomization Controls
SDD particle size and morphology are additional attributes that influence drying kinetics and therefore, residual solvent levels in the SDD. Particle size is controlled via selection of nozzle geometry and atomization pressure and also dependent on the spray solution composition. Larger particles typically contain higher levels of residual solvent for a given formulation; therefore, droplet and particle size should be minimized without sacrificing performance or downstream processability.³

Dryer Chamber Temperature (T_out)
The outlet temperature of the dryer is a key parameter in spray drying, as it effects the drying kinetics, particle morphology and the relative saturation.² With respect to SDD physical stability, increasing the outlet temperature decreases the relative saturation and increases the rate of drying in the chamber, and both of these effects contribute to lower solvent levels in the SDD upon exiting the spray dryer.

However, there is a risk in increasing the outlet temperature because phase separation, nucleation and crystal growth are temperature dependent events. In addition, as the outlet temperature is increased, it approaches the Tg (low Tg-T_out) and the gains in reducing solvent content from higher outlet temperature diminish, at the expense of exposing the dispersion to conditions that could lead to crystallization.

These constraints are addressed by using the predicted Tg as function of solvent content with the results from process modeling to identify an optimized outlet temperature that maximizes product stability by remaining below the in-process Tg while minimizing the residual solvent on the SDD as it exits the dryer chamber.

Combining Variables to Optimize SDD Formulation
The process map in Figure 5 shows how several of the key variables discussed come together to form an applied model for process exploration and optimization. This plot assumes a constant condenser temperature of -20 C. The L/G is plotted vs. T_out with colored contours representing the solvent relative saturation according to the legend. The solid contours are lines of constant Tg-T_out with graduations of 4 C.

Consideration of a process with a L/G = 0.1 and an outlet temperature in the range of 40 to 45 C would be a reasonable choice, as this keeps the relative saturation in a range where the Tg is predicted to be approximately 20 C higher than T_out while maintaining a reasonable process capacity.

It may be tempting to consider a lower outlet temperature to increase Tg –T_out within the process. For example, considering a lower outlet temperature and/or increasing the L/G along a constant Tg-T_out contour to maintain consistent in process stability. However, under these conditions, the residual solvent content on the SDD coming out of the dryer will go up, decreasing the Tg and the stability of the intermediate under storage conditions prior to secondary drying or downstream processing.

More detailed studies to define the crystallization kinetics with excess residual solvent are necessary to define and de-risk the acceptable hold times and conditions of the SDD under this scenario. A more prudent approach may be to tune the process conditions to manufacture a SDD with improved stability during collection and storage.

Conclusions
SDD technology is increasingly utilized for improving oral bioavailability of BCS II compounds that present absorption challenges due to low solubility. Though some compounds demonstrate a high propensity to crystallize from the amorphous state, this risk can usually be managed during SDD formulation by adjusting drug loading and excipient choice.

Characterization of the SDD and modeling of process conditions can be used to minimize or eliminate the risk of solid state recrystallization if an API demonstrates or is suspected to be at risk for crystallization. This is done by characterizing the SDD, then using that information to define process conditions with appropriate constraints to manufacture a dispersion with the desired attributes.

For some compounds, a residual risk of the physical stability may require a closer examination of the product properties and manufacturing conditions. These APIs present a challenge that may be overcome by focusing on variables in the SDD manufacturing process that can affect SDD stability.

This approach connects the materials science with the process science to stabilize amorphous spray dried dispersion formulations and enhance the bioavailability of low-solubility APIs.

Pharmaceutical companies can benefit from accessing specialist particle engineering service providers for addressing solubility issues. Extensive studies of poorly bioavailable compounds and solid dispersion processing have been used to develop models and formulation reference maps connecting the materials science with the processing science for stabilized SDD formulations.

These tools facilitate meeting target product profiles, desired downstream processing attributes and optimization, while minimizing the need for laboratory/pilot scale trials and raw material requirements.

References

1. Hydroxypropyl Methylcellulose Acetate Succinate-Based Spray-Dried Dispersions: An Overview. Friesen, D.T., et al. 6, 2008, Mol. Pharm., Vol. 5, pp. 1003-1019.
2. A Model-Based Methodology for Spray-Drying Process Development. Dobry, D.E. et al. 2009, J. Pharm Innov, Vol. 4, pp. 133-142.
3. Efficient Scale-Up Strategy for Spray-Dried Amorphous Dispersions. Dubose, D., et al. 8, October 2013, Drug Dev. and Delivery, Vol. 13.

About the Authors
Tim Elwell is Principal Engineer and Zach Martin is Principal Scientist and Manager, Late Stage Formulation Development at Lonza Pharma & Biotech. They are based in Bend, Oregon.