Freeze-drying above room temperature

Freeze-Drying above Room Temperature
MARC S. TESCONI, KIARESH SEPASSI, AND SAMUEL H. YALKOWSKY* Contribution from Department of Pharmaceutical Science, College of Pharmacy, University of Arizona, Tucson, Arizona 85721. Received December 10, 1998. Accepted for publication February 22, 1999.
environment for water-sensitive compounds, (3) be rapidly Abstract 0 This study investigates the use of solid, organic
and completely removed from the product under vacuum, compounds to lyophilize drugs without conventional freeze-drying and (4) produce cakes that are easily reconstituted using equipment. The aim of the investigation is to find a pharmaceutically pharmaceutically acceptable cosolvents.
acceptable solvent or solvent combination that is appropriate for freeze-drying on the basis of its ability to (1) solubilize hydrophobic drugs,(2) provide a stable environment for water-sensitive compounds, (3) Materials
be rapidly and completely removed from the product under vacuum,and (4) produce cakes that are readily reconstituted. A eutectic formed Solvents were selected from the Handbook of Pharmaceutical from 1,1,1-trichloro-2-methyl-2-propanol (chlorobutanol) hemihydrate Excipients2 and The Merck Index3 on the basis of their having lowtoxicity, melting points between 35 and 120 °C, and relatively low and dimethyl sulfone (DMSO2) is determined to be a suitable medium.
molecular weights. The compounds chosen for the study are listedin Table 1.
All compounds were >98% pure and were used as received from their respective suppliers except for anhydrous chlorobutanol,which was prepared by dehydration of a chlorobutanol hemihy- Introduction
The formulation of certain parenteral drugs as recon- stitutable freeze-dried products offers the advantages of Approach
improved stability, dissolution rate, dosing accuracy, andsterility. However, freeze-drying is generally restricted to The individual solvents in Table 1 were experimentally drugs that are water soluble and/or stable to hydrolysis determined to be either too high melting to ensure solute for at least the time required for freezing. Compounds that stability or too nonpolar to solubilize semipolar compounds.
hydrolyze rapidly are poor candidates because the amount Solvent combinations that formed eutectics were sought of drug that remains in the dried product is likely to depend for their intermediate polarities and lower melting tem- on the time required to freeze the sample. Hydrophobic peratures. The selection of a eutectic medium for freeze- nonelectrolytes are poor candidates because they require drying involved seven steps: (1) screening solvent combi- either large volumes of water or significant amounts of nations for eutectic formation; (2) determination of the cosolvent to yield the desired quantity of drug in a eutectic compositions; (3) measurement of sublimation reasonably sized container. Although cosolvents usually rates; (4) assessment of solubilizing ability; (5) selection of improve both stability and solubility, they can cause an optimal medium; (6) freeze-drying sample compounds; incomplete freezing and a drying stage that involves the and (7) analysis of freeze-dried products. Each of these simultaneous sublimation of ice and evaporation of liquid residue.1 The resultant dried product is often very denseand difficult to reconstitute. Furthermore, since a surfaceskin is often produced when an incompletely frozen me- Experimental and Results
dium evaporates, drying times can be extended beyond thepoint of practicality.
1. Eutectic FormationsThe solvents in Table 1 were
It is also important to consider that if sublimation combined in a 1:1 mole ratio in quantities sufficient to yield temperatures are significantly lower than normal to keep approximately 2 g of material. The mixtures were melted, the medium frozen, then the condensing temperature and those systems that exhibited solvent miscibility and required to recover the solvents may be lower than the -55 stability were allowed to cool at room temperature for 2 h.
These were refrigerated for 24 h at 4 °C to ensure complete The current study investigates the feasibility of using solidification and then brought to room temperature.
organic solvents that are solids at room temperature to Samples that appeared “wet” were deleted from the study.
lyophilize hydrophobic and water-sensitive compounds Core samples of approximately 5 mg were taken from the without conventional freeze-drying equipment. The aim of remaining solids and analyzed with a DuPont Instruments this investigation is to find a solvent or solvent combination model 910 differential scanning calorimeter using closed that is suitable for freeze-drying on the basis of its ability sample pans and a heating rate of 10 °C/min. Solvent to (1) solubilize hydrophobic drugs, (2) provide a stable combinations that exhibited a single, well-defined eutecticmelt between 35 and 75 °C were considered further.
2. Eutectic CompositionsThe eutectic compositions
* To whom correspondence should be addressed. Tel: (520) 626- 1289. Fax: (520) 626-4063. E-mail: [email protected].
were estimated using the van’t Hoff equation 1999, American Chemical Society and Journal of Pharmaceutical Sciences / 501
American Pharmaceutical Association Table 1sCompounds Used in This Study
a Prepared by dehydration of chlorobutanol hemihydrate.
6. Freeze-DryingsThe five compounds shown in Figure
2, progesterone, coronene, Fluasterone, phenytoin, and 1,1- dicyano-3-nitrobenzyl-1-propene, were lyophilized from thechlorobutanol hemihydrate-DMSO2 eutectic. Samples were with the molar heat of fusion, ∆fusHA, determined by prepared in 10 mL freeze-drying vials with orifices of ∼2 differential scanning calorimetry (DSC). In eq 1, XA rep- cm2 by dissolving 30 mg of solute in 5 g of the eutectic melt resents the mole fraction of component A, Tm is the melting at 60 °C. These were allowed to solidify for 2 h at 25 °C.
point of pure component A, and Teut is the eutectic melting The samples were then placed in a large Erlenmeyer flask temperature. Once the calculated mole fractions were that was connected to a vacuum pump. The base of the 1 - XA), solvent ratios slightly above flask was submerged in a water bath that was heated by and below the estimated value were prepared and a closer a hot plate. The solvents were sublimed at approximately approximation of the eutectic composition was made ex- 0.3 mmHg pressure using two drying stages: perimentally using DSC. The eutectic composition wasobtained as the mole ratio that produced a single melting endotherm. Hot stage microscopy was used to confirm the 3. Sublimation RatessSamples of the eutectics weigh-
stage 2: 60 °C, 6.5 h (-150 mg/cm2-h: DMSO ) ing ∼50 mg were sublimed at 10 °C below their eutecticmelting points at ∼0.3 mmHg pressure. Evaporation rates The stage times were roughly equal in this study because were measured using a DuPont Instruments model 951 both the mass and evaporation rate ratios are 3:1. However, thermogravimetric analyzer. Two evaporation rates were the evaporation rate ratio is likely to vary with the observed in most cases. The greater first rate is due to the sublimation of both solvents. The second rate is due to the 7. Freeze-Dried Product AnalysissResidual Solvents
sublimation of the less volatile component after the more Thermogravimetric (TG) analysis was used to measure the volatile solvent has been removed. The results for the most amount of residual solvent remaining in the cakes. Samples readily sublimed solvent combinations are given in Table of ∼10 mg were placed under vacuum and heated at a rate of 10 °C/min. The TG scans of the cakes and the solvents 4. Solubilizing AbilitysTo test the ability of the binary
are shown in Figure 3. The analysis indicates that the systems to solubilize both semipolar and nonpolar com- cakes contain less than 1% residual solvent. The profiles pounds, 20 mg of anthracene, caffeine, coronene, phenytoin, for pure chlorobutanol-hemihydrate (-×-) and pure DMSO2 progesterone, and urea were added separately to 1 mL (-+-) in Figure 3 indicate the temperatures at which weight volumes of the eutectic melts at temperatures within 5 °C loss due to residual solvent would be expected.
above their respective melting temperatures. All of the Solute StabilitysThe investigational compound 1,1- solvent systems were able to solubilize this quantity of the dicyano-3-nitrobenzyl-1-propene has a half-life in water of solutes. Solubility parameters were also calculated for the approximately 20 min, which precludes its freeze-drying eutectic melts using the method of Adjel et al.4 for combined by conventional methods. High-performance liquid chro- solvent systems with group contribution values of cohesive matography was used to analyze the product immediately energy density and molar volume from Fedors.5 The following lyophilization from the eutectic and after 15months of storage at room temperature. Assays were solubility parameters shown in Table 3 are in the range of performed using a Beckman System Gold HPLC with a most drugs, which suggests that the melts are good Pinnacle ODS amine column (5 µm, 250 mm × 4.6 mm i.d., Restek, Bellefonte, PA) and a mobile phase consisting of 5. Premier Eutectic for Freeze-DryingsThe chlo-
500 parts water, 500 parts acetonitrile, and 3 parts acetic robutanol hemihydrate-dimethyl sulfone (DMSO2) eutectic acid at a pH of 3.45. A Beckman model 110A pump was was found to be the most suitable media for freeze-drying used to maintain a flow rate of 1.0 mL/min. Ultraviolet on the basis of nontoxicity, solubilizing ability, and rate of detection was made using a Kratos Analytical Spectroflow solvent removal. As shown in Table 2, the eutectic is formed model 757 detector at a wavelength of 300 nm. Quantifica- at a mole ratio of 60% chlorobutanol hemihydrate and 40% tion of drug in the effluent was made using a Hewlett- dimethyl sulfone, and it has a melting temperature of 50 Packard model 3394 integrator. The relative retention °C. Figure 1 shows a phase diagram constructed from DSC times of 1,1-dicyano-3-nitrobenzyl-1-propene and its major degradation product are 3.4 and 2.8 min, respectively. No 502 / Journal of Pharmaceutical Sciences
Table 2sCompositions, Melting Temperatures, and Relative Evaporation Rates (at 10 °C below the Melting Temperatures) of the Eutectics
a Expressed as a fraction of the estimated sublimation rate of ice at −10 °C and 0.3 mmHg (∼10 mg/mm2 min).
Table 3sSolubility Parameters for the Most Promising Solvent
dicyano-3-nitrobenzyl-1-propene and Fluasterone cakes Combinations
were reconstituted using 8 mL of a nonaqueous 50%ethanol and 50% poly(ethylene glycol) 400 solution due to their low water solubilities, which are on the order of 1 µg/mL. For hydrophobic drugs such as these, the composi-tion of the vehicles that are used for reconstitution, whether aqueous solutions or cosolvent concentrates, will depend In all of the above cases, dissolution of the cakes was complete with 1-2 s of shaking. Note that coronene, which is not a therapeutic agent, was not soluble in pharmaceuti-cal cosolvents.
degradation was detected in the freeze-dried cake im-mediately after lyophilization or after 15 months of storage Discussion
Cake Structure and Particle SizesFigure 4a,b contains Solid organic solvents were successfully used to lyo- photomicrographs of the freeze-dried cakes of phenytoin philize hydrophobic and water-sensitive compounds. Be- and coronene, respectively, that were taken using a SPOT cause the solvents used for the freeze-drying medium were camera and a Leica DMLP polarizing microscope. The solids at room temperature, the process could be conducted fiberlike structures in the phenytoin cake are loose ag- without refrigeration and without conventional freeze- gregates of crystals that are approximately 1-2 µm long.
drying equipment. This also facilitated solvent collection, The coronene cake is composed of bundles of roughly 5-10 which was accomplished without using a cooled condenser.
crystals. The coronene crystals are the largest in the test A eutectic comprised of two solvents was used in order set, having an average length of approximately 50 µm. The to achieve good solubilization of drugs in a medium with a cakes of the other three compounds are similar in overall moderate melting temperature. The latter property is appearance and contain particles that range from 5 to 15 required because drugs must be dissolved in a medium above its melting temperature and drug stability is com- ReconstitutionsThe phenytoin cake was reconstituted promised if that temperature is high. Since a eutectic is using Millipore water adjusted to pH 12. Dissolution was composed of two pure solid phases, these solvents are complete with a single shake of the vial. With the exception removed relatively independently of one another according of coronene, the other compounds were reconstituted using to their vapor pressures. Consequently, two drying stages ethanolic solutions in which they were soluble. The proges- were used: one at 40 °C to remove (primarily) the more terone cake was reconstituted using 8 mL of a 60% ethanol, volatile chlorobutanol-hemihydrate and one at 60 °C to 10% dimethylacetamide, and 30% water solution. The 1,1- remove the DMSO2. Note that these stages do not cor- Figure 1sChlorobutanol hemihydrate−DMSO2 phase diagram.
Journal of Pharmaceutical Sciences / 503
Figure 2sCompounds that were freeze-dried from the chlorobutanol hemihydrate−DMSO2 eutectic.
Figure 3sTG analysis of the pure solvents and the freeze-dried cakes: (×)
chlorobutanol; (+) DMSO2; (b) 1,1-dicyano-3-nitrobenzyl-1-propene; (2)
Fluasterone; (9) phenytoin; (]) progesterone; (O) coronene.
respond to the drying stages used for ice in which the firstremoves frozen water and the second removes bound water.
Thermogravimetric analysis showed that the freeze-dried products did not contain significant amounts of residualsolvent. This indicates that the 60 °C used to sublime thebulk DMSO2 was sufficient to remove bound solvent.
However, 60 °C is a relatively high temperature comparedto what is normally used in freeze-drying, and care mustbe taken to avoid drug loss. It may be necessary to uselower temperatures and longer drying times when freeze-drying volatile compounds.
One of the major concerns in this study was that the structure of the eutectic should not be significantly altered Figure 4s(a) Phenytoin cake. (b) Coronene cake.
by the presence of solute. The eutectic contains ∼3.6%water that is believed to exist primarily in the form of 50 °C, small, needlelike phenytoin crystals remain in the chlorobutanol-hemihydrate crystals. If a dissolved drug melt. These dissolved by 70 °C on heating the melt at a caused the anhydrous form of chlorobutanol to crystallize on cooling, it could lead to incomplete freezing or variable Evidence that a hydrated form of chlorobutanol exists solvent sublimation rates. Analysis of a eutectic containing in the eutectic is provided by the different thermal proper- 5% phenytoin indicated that the eutectic structure is ties of the solid formed from the anhydrous chlorobutanol- largely unaffected by the presence of this solute. Figure DMSO2 mixture at a 60:40 mole ratio. DSC and hot-stage 5a,b shows the DSC traces of the eutectic and the eutectic microscopic analysis of the anhydrous chlorobutanol- with 5% phenytoin, respectively. The similarity of the DSC DMSO2 solid indicate a fusion event at 35 °C that is not traces suggests that the eutectic microstructures are the exhibited by the solid formed from the hemihydrate.
However, when 3.6% water is added to the anhydrous melt, The likeness of the photomicrographs of the eutectic and the solid that is produced on cooling exhibits the same the eutectic with 5% phenytoin in Figure 6a,b, respectively, thermal properties as the solid prepared from the hemi- provides additional evidence for a stable eutectic structure.
hydrate, i.e., a melting point of 50 °C without a 35 °C Hot-stage microscopy indicated that the apparent coarse- transition. Physically, the anhydrous solid is much softer ness in the eutectic with phenytoin is due to presence of than the hydrous form. The former also appears translu- drug crystals in the solid. After fusion of the solvents at cent, whereas the hydrous form (the eutectic) is a white 504 / Journal of Pharmaceutical Sciences
Figure 5s(a) DSC scan of the eutectic. (b) DSC scan of the eutectic + 5%
phenytoin.
Figure 6s(a) Eutectic viewed with polarized light after crystallization between
a glass slide and coverslip. (b) Eutectic with 5% phenytoin.
solid. Both properties may be due to the anhydrous formbeing incompletely frozen at room temperature. The physi- cal properties of the hydrous eutectic also return with the 2 was first identified as a natural constitu- ent of biological systems in 1934 when Pfiffner and Vars addition of water to the anhydrous melt.
isolated it from beef adrenals.9 It was later isolated from Thermogravimetric analysis indicated that the cakes beef blood and milk, and it has been indicated that human produced in this study contained less than 1% residual intake from these sources may be as high as a few solvent. Although human toxicity data are scarce, the milligrams daily.8 The latter figure is consistent with available data suggest that trace amounts of the solvents Williams et al.’s finding of 4-11 mg of DMSO which may remain in the freeze-dried products are well samples of human urine10 and a reported circulating blood below harmful levels. Borody6 found that plasma concen- concentration of 0.2-0.5 ppm in the adult male.11 trations of chlorobutanol as high as 100 µg/mL produced sluggishness and slowed speech, but did not produce liver 2 has recently been tested for efficacy in the treatment of interstitial cystitis11 without findings of any disease or abnormal electrolyte levels. Others have reported adverse effects. Bertken8 also reported that a man who adverse reactions to high doses such as cardiovascular effects following the intravenous administration of heparin suffered no ill effects. The only (related) report of toxicity preserved with chlorobutanol, neurological effects following that was found in the literature, other than the detrimental the administration of large doses of morphine with the preservative, and hypersensitivity reactions.2 However, chlorobutanol is currently used in the U.S. as a parenteral However, as indicated by the authors of that paper, there preservative in concentrations of up to 0.5% in injectable is no established precedent for the direct conversion of (e.g., methadone, epinephrine, oxytocin, thiamine), opthalm- ic (e.g., pilocarpine, epinephrine, phospholine iodide), otic (e.g., Cresylate, Cerumenex), and cosmetic products. It isalso the active ingredient in a nonprescription sleep aid Conclusion
(Seducaps: chlorobutanol 150 mg, salicylamide 300 mg)that is available in several countries outside of the United The chlorobutanol hemihydrate-dimethyl sulfone (DM- States.2,7 The lethal dose of chlorobutanol is estimated to SO2) eutectic was determined to be a suitable medium for lyophilizing drugs that are poorly water soluble and/or Dimethyl sulfone is an oxidation product of dimethyl highly water sensitive. The eutectic is formed at a 60:40 sulfoxide (DMSO) and is its major metabolite in all species mole ratio (75:25 weight ratio) and has a melting point of Journal of Pharmaceutical Sciences / 505
50 °C. Lyophilization can be accomplished without refrig- Approach: Solubility of Caffeine in Dioxane-Water Mix- eration, and thus, without conventional freeze-drying tures. J. Pharm. Sci. 1980, 69, 659-661.
5. Fedors, R. F. A Method for Estimating Both the Solubility equipment, because the melting point is above room Parameters and Molar Volumes of Liquids. Polym. Eng. Sci. temperature. The solvents are rapidly and thoroughly 1974, 14, 147-154.
removed under vacuum with moderate heating. Two drying 6. Borody, T.; Chinwah, P. M.; Graham, G. G.; Wade, D. N.; stages were used in the current study, both at a pressure Williams, K. N. Chlorobutanol Toxicity and Dependence.
Med. J. Australia. 1979, 7, 288.
of 0.3 mmHg. The first stage was at 40 °C to primarily 7. Nordt, S. P. Chlorobutanol Toxicity. Ann. Pharmacother.
remove the chlorobutanol hemihydrate, and the second was 1996, 30, 1179-1180.
at 60 °C to remove the bulk of the DMSO2. Each drying 8. Bertken, R. “Crystalline DMSO”: DMSO2. Arthritis Rheum.
stage required approximately 6.5 h using standard 10 mL 1983, 26, 693-4.
freeze-drying vials and solvent volumes of ∼5 mL. The 9. Pfiffner, J. J.; Vars, H. M. Extraction Studies on Adrenal Cortical Hormone; Disrtibution Studies. J. Biol. Chem. 1934,
cakes produced from the eutectic contain <1% residual solvent and are rapidly reconstituted. Trace amounts of 10. Williams, K. I. H.; Burstein, S. H.; Layne, D. S. Dimethyl residual solvent are below harmful levels.
Sulfone: Isolation forom Human Urine. Arch. Biochem.
Biophys. 1966, 113, 251-252.
11. Childs, S. J. Dimethyl Sulfone (DMSO2) in the Treatment of Interstitial Cystitis. Urol. Clin. North Am. 1994, 21, 85-88.
References and Notes
12. Goldstein, P.; Magnano, l.; Rojo, J. Effects of Dimethyl Sulfone (DMSO2) on Early Gametogenesis in Caenorhabditis 1. Seager, H.; Taskis, C. B.; Syrop, M.; Lee, T. J. Structure of Elegans: Ultrastructural Aberrations and Loss of Synap- Products Prepared by Freeze-Drying Solutions Containing tonemal Complexes from Pachytene Nuclei. Reprod. Toxicol.
Organic Solvent. J. Parenter. Sci. Technol. 1985, 39, 161-
1992, 6, 149-159.
13. Grant, P. M.; Haas, J. S.; Whipple, R. E.; Andrensen, B. D.
2. The Handbook of Pharmaceutical Excipients; Wade, A., A Possible Chemical Explanation for the Events Associated Weller, P. J., Eds.; The Pharmaceutical Press: London, 1994.
with the Death of Gloria Ramirez at Riverside General 3. The Merck Index, 12th ed.; Budavari, S., Ed.; Merck Research Hospital. Forensic Sci. Int. 1997, 87, 219-237.
Laboratories: Whitehouse Station, NJ, 1996.
4. Adjei, A.; Newburger, J.; Martin, A. Extended Hildebrand 506 / Journal of Pharmaceutical Sciences

Source: http://perso.centrale-marseille.fr/~tyang/visible/pt/poivre/CAMPHENE%20SOLVANT.pdf