Michael J. Pikal

Design of freeze-drying processes for pharmaceuticals: practical advice.

Design of freeze-drying processes is often approached with a “trial and error” experimental plan or, worse yet, the protocol used in the first laboratory run is adopted without further attempts at optimization. Consequently, commercial freeze-drying processes are often neither robust nor efficient. It is our thesis that design of an “optimized” freeze-drying process is not particularly difficult for most products, as long as some simple rules based on well-accepted scientific principles are followed. It is the purpose of this review to discuss the scientific foundations of the freeze-drying process design and then to consolidate these principles into a set of guidelines for rational process design and optimization. General advice is given concerning common stability issues with proteins, but unusual and difficult stability issues are beyond the scope of this review. Control of ice nucleation and crystallization during the freezing step is discussed, and the impact of freezing on the rest of the process and final product quality is reviewed. Representative freezing protocols are presented. The significance of the collapse temperature and the thermal transition, denoted Tg`, are discussed, and procedures for the selection of the “target product temperature” for primary drying are presented. Furthermore, guidelines are given for selection of the optimal shelf temperature and chamber pressure settings required to achieve the target product temperature without thermal and/or mass transfer overload of the freeze dryer. Finally, guidelines and “rules” for optimization of secondary drying and representative secondary drying protocols are presented.

The Effects of Formulation Variables on the Stability of Freeze-Dried Human Growth Hormone

Formulation often has a dramatic effect on degradation of proteins during the freeze-drying process as well as impacting on the “shelf-life” stability of the freeze-dried product. This research presents the results of a formulation optimization study of the “in-process” and shelf-life stability of freeze-dried human growth hormone (hGH). Chemical decomposition via methionine oxidation and deamidation of asparagine residues as well as irreversible aggregation were characterized by HPLC assay methodology. In-process degradation and stability of low moisture freeze-dried solids were studied at 25 and 40°C in a nominal nitrogen headspace (≈0.5% O2). Formulation variables included pH, level of salts, and the nature of the lyoprotectant. Studies of the effect of shear on aggregation in solutions indicated that shear comparable to that experienced during filtration does not induce aggregation. Irreversible changes in hGH during the freeze-drying process were minimal, but chemical decomposition via methionine oxidation and asparagine deamidation and aggregation did occur on storage of the freeze-dried solid. Decomposition via methionine oxidation was significant. A combination of mannitol and glycine, where the glycine remains amorphous, provided the greatest protection against decomposition and aggregation. It is postulated that an excipient system that remains at least partially amorphous is necessary for stabilization. However, the observation that dextran 40 formulations showed poor stability toward aggregation demonstrates that an amorphous excipient system is not a sufficient condition for stability. Stability of the solid was optimal when produced from solutions in the pH range, 7–7.5, with severe aggregation being observed at high pH. The level of sodium phosphate buffer affected stability of the solid, although this relationship was complex. Freeze-drying in the presence of NaCl produced severe aggregation and precipitation during the freeze-drying process as well as acceleration of oxidation and/or deamidation.

The collapse temperature in freeze drying: Dependence on measurement methodology and rate of water removal from the glassy phase

Accurate determination of the maximum allowable product temperature during primary drying is critical to optimization of the freeze drying process. For an amorphous solute system, this maximum temperature is normally the collapse temperature. Methodologies for determining the collapse temperature involve direct microscopic observation of collapse during freeze drying and methods which, in effect, determine the glass transition temperature of the amorphous phase. While one might be tempted to assume that the collapse temperature is a property only of the material, and independent of details of the measurement method, both theoretical concepts and limited experimental observations suggest that this assumption may not be wholly correct. The main objective of this research is to determine the magnitude of variations in measured collapse temperature caused by variations in experimental methodology. The approach taken is both experimental, using moxalactam di-sodium formulated with 12% mannitol as a model, and theoretical. The theoretical analysis is based on two fundamental concepts. Firstly, for collapse to be observed, viscous flow of the amorphous phase must occur over a finite distance during the measurement time. Secondly, during a freeze drying process, water is removed from the amorphous phase once the ice-vapor boundary recedes past the region of interest. Since water removal increases viscosity, viscous flow, and therefore, collapse is partially arrested, and the effective collapse temperature will be increased, the effect being greater the faster the sublimation rate. A quantitative model based on these concepts is developed with key parameters being evaluated by experimental studies. The observed variation in collapse temperature of moxalactam di-sodium with sublimation rate is quantitatively predicted by the theory. The theory is used to investigate differences between collapse temperatures determined by laboratory procedures and the observation of collapse in production processes. The collapse temperature will increase as the sublimation rate increases (i.e., as the solute concentration decreases), and at constant sublimation rate, the collapse temperature may increase as the surface area of the solid increases. In general, product freeze drying in a vial will collapse at a slightly higher temperature than collapse measured by the microscopic method. However, the calculated variations in collapse temperatures are modest (1–3°C). Collapse temperature and glass transition temperature, T′g, are not identical, the latter being slightly lower when measured at low rates of temperature increase. A secondary but important experimental result is that, contrary to some opinion in the literature, water in a glassy system has sufficient mobility to be in approximate ‘equilibrium’ with the ice phase during the relatively slow temperature changes relevant to freeze drying operations.

Solubility advantage of amorphous pharmaceuticals: I. A thermodynamic analysis

In recent years there has been growing interest in advancing amorphous pharmaceuticals as an approach for achieving adequate solubility. Due to difficulties in the experimental measurement of solubility, a reliable estimate of the solubility enhancement ratio of an amorphous form of a drug relative to its crystalline counterpart would be highly useful. We have developed a rigorous thermodynamic approach to estimate enhancement in solubility that can be achieved by conversion of a crystalline form to the amorphous form. We rigorously treat the three factors that contribute to differences in solubility between amorphous and crystalline forms. First, we calculate the free energy difference between amorphous and crystalline forms from thermal properties measured by modulated differential scanning calorimetry (MDSC). Secondly, since an amorphous solute can absorb significant amounts of water, which reduces its activity and solubility, a correction is made using water sorption isotherm data and the Gibbs-Duhem equation. Next, a correction is made for differences in the degree of ionization due to differences in solubilities of the two forms. Utilizing this approach the theoretically estimated solubility enhancement ratio of 7.0 for indomethacin (amorphous/gamma-crystal) was found to be in close agreement with the experimentally determined ratio of 4.9.

Protein stability during freezing : Separation of stresses and mechanisms of protein stabilization

Although proteins are often frozen during processing or freeze-dried after formulation to improve their stability, they can undergo degradation leading to losses in biological activity during the process. During freezing, the physical environment of a protein changes dramatically leading to the development of stresses that impact protein stability. Low temperature, freeze-concentration, and ice formation are the three chief stresses resulting during cooling and freezing. Because of the increase in solute concentrations, freeze-concentration could also facilitate second order reactions, crystallization of buffer or non-buffer components, phase separation, and redistribution of solutes. An understanding of these stresses is critical to the determination of when during freezing a protein suffers degradation and therefore important in the design of stabilizer systems. With the exception of a few studies, the relative contribution of various stresses to the instability of frozen proteins has not been addressed in the freeze-drying literature. The purpose of this review is to describe the various stages of freezing and examine the consequences of the various stresses developing during freezing on protein stability and to assess their relative contribution to the destabilization process. The ongoing debate on thermodynamic versus kinetic mechanisms of stabilization in frozen environments and the current state of knowledge concerning those mechanisms are also reviewed in this publication. An understanding of the relative contributions of freezing stresses coupled with the knowledge of cryoprotection mechanisms is central to the development of more rational formulation and process design of stable lyophilized proteins.

The Stability of Insulin in Crystalline and Amorphous Solids: Observation of Greater Stability for the Amorphous Form

AbstractPurpose. Generalizations based upon behavior of small molecules have established that a crystalline solid is generally much more stable toward chemical degradation than is the amorphous solid. This study examines the validity of this generalization for proteins using biosynthetic human insulin as the model protein. Methods. Amorphous insulin was prepared by freeze drying the supernate from a suspension of zinc insulin crystals adjusted to pH 7.1. Storage stability at 25°C and 40°C were compared for the freeze dried material, the dried suspended crystals, and the starting batch of crystals. Samples were equilibrated at selected relative humidities between zero and 75% to obtain samples at various water contents. Assays for dimer formation were performed by size exclusion HPLC and assays for deamidated product were carried out by reverse phase HPLC. Degradation was found to be linear in square root of time, and the slopes from % degradation vs. square root of time were used to define the rate constants for degradation. Differential scanning calorimetry (DSC) and Fourier-transform infrared spectroscopy (FTIR) were used to characterize the state of the protein in the solids. Results. As expected based upon previous results, the primary degradation pathways involve deamidation at the AsnA21 site and co-valent dimer formation, presumably involving the A-21 site. Contrary to expectations, amorphous insulin is far more stable than crystalline insulin under all conditions investigated. While increasing water content increases the rate of degradation of crystalline insulin, rate constants for degradation in the amorphous solid are essentially independent of water content up to the maximum water content studied (≈15%). Conclusions. Based upon the FTIR and DSC data, both crystalline and amorphous insulin retain some higher order structure when dried, but the secondary structure is significantly perturbed from that characteristic of the native solution state. However, neither DSC nor FTIR data provide a clear interpretation of the difference in stability between the amorphous and crystalline solids. The mechanism responsible for the superior stability of amorphous insulin remains obscure.

Mechanisms of protein stabilization in the solid state

The solid state is preferred for many proteins due to their marginal stability in solution. However, even in the solid state, chemical and physical degradation can occur on the time scale of the drying process, distribution and use. This review summarizes the major degradation pathways of proteins in the solid state and the corresponding stabilization strategies. Specially, this review discusses the mechanisms of protein stabilization in the solid state. Understanding the mechanisms of protein stabilization is critical to efficient protein formulation development in the pharmaceutical industry.

Heat and mass transfer scale-up issues during freeze drying: II. Control and characterization of the degree of supercooling

This study aims to investigate the effect of the ice nucleation temperature on the primary drying process using an ice fog technique for temperature-controlled nucleation. In order to facilitate scale up of the freeze-drying process, this research seeks to find a correlation of the product resistance and the degree of supercooling with the specific surface area of the product. Freeze-drying experiments were performed using 5% wt/vol solutions of sucrose, dextran, hydroxyethyl starch (HES), and mannitol. Temperature-controlled nucleation was achieved using the ice fog technique where cold nitrogen gas was introduced into the chamber to form an “ice fog”, there-by facilitating nucleation of samples at the temperature of interest. Manometric temperature measurement (MTM) was used during primary drying to evaluate the product resistance as a function of cake thickness. Specific surface areas (SSA) of the freeze-dried cakes were determined. The ice fog technique was refined to successfully control the ice nucleation temperature of solutions within 1°C. A significant increase in product resistance was produced by a decrease in nucleation temperature. The SSA was found to increase with decreasing nucleation temperature, and the product resistance increased with increasing SSA. The ice fog technique can be refined into a viable method for nucleation temperature control. The SSA of the product correlates well with the degree of supercooling and with the resistance of the product to mass transfer (ie, flow of water vapor through the dry layer). Using this correlation and SSA measurements, one could predict scaleup drying differences and accordingly alter the freeze-drying process so as to bring about equivalence of product temperature history during lyophilization.

The secondary drying stage of freeze drying : drying kinetics as a function of temperature and chamber pressure

Abstract Secondary drying involves removal of water which did not freeze. This report emphasizes the phenomenological description of the effects of temperature and chamber pressure on the kinetics of drying. A crystalline solute (mannitol) and two amorphous solutes (moxalactam di-sodium and povidone) were selected for study. Drying kinetics were determined gravimetrically using a vacuum microbalance and by Karl Fischer assay of vials sealed at selected times during secondary drying experiments conducted in a laboratory scale freeze dryer. The main observations may be summarized as follows: (1) the water content decreases rapidly during the first few hours of drying and then appears to approach a plateau level of residual water which far exceeds the equilibrium water content calculated from the desorption isotherm data and the measured partial pressure of water in the drying chamber; (2) this plateau level of water sharply decreases as the drying temperature is increased; (3) the drying rate increases as the product specific surface area increases; and (4) variations in chamber pressure (0–0.2 mmHg) and dried product thickness have little or no effect on drying rate. We conclude that the rate-limiting mass transfer process for drying an amorphous solid is either evaporation at the solid/vapor interface or diffusion in the solid, probably the former. The ‘plateau level kinetics’ appears to be consistent with amorphous particle size heterogeneity superimposed on a simple model based on Fickian diffusion with rate controlling surface evaporation.

Transport Mechanisms in Iontophoresis. III. An Experimental Study of the Contributions of Electroosmotic Flow and Permeability Change in Transport of Low and High Molecular Weight Solutes

The objective of this research was to provide in vitro transport data designed to clarify the relative importance of permeability increase and electroosmotic flow in flux enhancement via iontophoresis, Iontophoretic fluxes were measured with both anode and cathode donor cells, and passive fluxes were measured both before iontophoresis (Passive 1) and after iontophoresis (Passive 2). Data were generated for three uncharged low molecular weight solutes (glycine, glucose, and tyrosine) and two high molecular weight anionic species (carboxy inulin and bovine serum albumin). Flux enhancement is greater for anodic delivery than for cathodic delivery, even for the negatively charged molecules, and anodic flux of glucose decreases as the concentration of NaCl increases. Both observations are consistent with a mass transfer mechanism strongly dependent on electroosmotic flow. Steady-state anodic flux at 0.32 mA/cm2, expressed as equivalent donor solution flux (in µl/hr cm2), ranged from 6.1 for glycine to about 2 for the large anions. As expected, iontophoretic flux is higher at 3.2 mA/cm2 than at 0.32 mA/cm2, and passive flux measured after iontophoresis is about a factor of 10 greater than the corresponding flux measured before the skin was exposed to electric current. There are two mechanisms for flux enhancement relative to passive flux on “fresh” hairless mouse skin: (1) the effect of the voltage in increasing mass transfer over the passive diffusion level, the effect of electroosmotic flow dominating this contribution in the systems studied in this report; and (2) the effect of prior current flow in increasing the “intrinsic permeability” of the skin. Both effects are significant. Based on theoretical results given elsewhere, theoretical values for flux were calculated and compared with the experimental data. While agreement between theory and experiment was only qualitative in several cases, most of the data are predicted quantitatively by the theory.