Improvement of Dissolution Rates ofPoorly Water Soluble APIs UsingNovel Spray Freezing intoLiquid Technology
Jiahui Hu,1 True L. Rogers,1 Judith Brown,1
Tim Young,2,3 Keith P. Johnston,3 andRobert O. Williams III1,4
Received May 16, 2002; accepted May 30, 2002
Purpose. To develop and demonstrate a novel particle engineeringtechnology, spray freezing into liquid (SFL), to enhance the dissolu-tion rates of poorly water-soluble active pharmaceutical ingredients(APIs).Methods. Model APIs, danazol or carbamazepine with or withoutexcipients, were dissolved in a tetrahydrofuran/water cosolvent sys-tem and atomized through a nozzle beneath the surface of liquidnitrogen to produce small frozen droplets, which were subsequentlylyophilized. The physicochemical properties of the SFL powders andcontrols were characterized by X-ray diffraction, scanning electronmicroscopy (SEM), particle size distribution, surface area analysis,contact angle measurement, and dissolution.Results. The X-ray diffraction pattern indicated that SFL powderscontaining either danazol or carbamazepine were amorphous. SEMmicrographs indicated that SFL particles were highly porous. Themean particle diameter of SFL carbamazepine/SLS powder wasabout 7 �m. The surface area of SFL danazol/poloxamer 407 powderwas 11.04 m2/g. The dissolution of SFL danazol/poloxamer 407 pow-der at 10 min was about 99%. The SFL powders were free flowingand had good physical and chemical stability after being stored at25°C/60%RH for 2 months.Conclusions. The novel SFL technology was demonstrated to pro-duce nanostructured amorphous highly porous particles of poorlywater soluble APIs with significantly enhanced wetting and dissolu-tion rates.
KEY WORDS: danazol; carbamazepine; spray freezing into liquid;dissolution; stability.
INTRODUCTION
Many potentially bioactive molecules have been rejectedduring the early stages of development because they arepoorly water soluble and difficult to wet. Active pharmaceu-tical ingredients (APIs) with poor aqueous solubility oftendemonstrate low bioavailability when administrated orallydue to the dissolution rate-limiting absorption in the gastro-intestinal (GI) tract. Of particular interest is the poorly watersoluble, highly permeable APIs discussed in the Biopharma-ceutics Classification System (BCS Class II) (1), such as dana-
zol and carbamazepine. The irregular and delayed absorptionof danazol and carbamazepine has been attributed to slowdissolution rates (2,3). The physicochemical properties of theAPIs play a significant role in controlling their dissolutionfrom a dosage form. According to the Noyes-Whitney equa-tion, the aqueous solubility of an API is a factor determiningdissolution rate. Other factors include the particle size distri-bution, degree of crystallinity, as well as the presence of sur-factants and other additives. Other physical properties such aswettability, density, and viscosity contribute to particle floc-culation, flotation, and agglomeration, which further influ-ence dissolution rates.
Increasing the dissolution rate of poorly water solubleAPIs is a significant challenge to pharmaceutical scientists.Technologies that have been commonly used to improve thedissolution rate of poorly water soluble APIs include me-chanical milling, spray drying, precipitation, and freeze-drying. Spray drying is a widely used technology (4). How-ever, because the spray drying process uses elevated tempera-tures, it is not always appropriate for use with thermolabilecompounds. The process precipitation with a compressedfluid antisolvent (PCA), also referred to as SEDS or SAS, hasbeen used to produce particles containing poorly water-soluble API and water-soluble excipients (5). In some casesthis process can be limited by the lack of solvent systemscompatible with compressed carbon dioxide that dissolveboth hydrophilic and hydrophobic substances simultaneously.Although lyophilization or freeze-drying (6) is a promisingtechnique for producing pharmaceutical powders, the freez-ing rate in some cases is too slow so that the solvent crystal-lizes as it is frozen.
The objective of this study was to develop and demon-strate the use of the novel spray freezing into liquid (SFL)particle engineering technology to enhance the dissolutionrate of two poorly water soluble APIs, danazol and carba-mazepine, and to determine the physicochemical stability ofSFL powders.
SFL is a novel cryogenic atomization technology in whicheither an aqueous or an aqueous-organic cosolvent solutioncontaining an API and pharmaceutical excipient(s) is atom-ized directly into a compressed liquid, such as compressedfluid CO2, helium, propane, ethane, or the cryogenic liquidsnitrogen, argon, or hydrofluoroethers. SFL technology wascreated by the adaptation of several atomization processes.SFL is derived in part from the PCA process, which utilizesliquid-liquid impingement between an organic or organic/aqueous feed solution through a nozzle that is submerged intocompressed CO2 fluid. As the two liquids collide, the highReynolds and Webber numbers lead to intense atomizationinto micronized droplets. In PCA the solvent(s) must be mis-cible with compressed fluid CO2 to produce dry particles fromthe microdroplets. The low solubility of water in CO2 limitsmarkedly the use of solvents containing water, which are of-ten needed to dissolve hydrophilic excipients. Other limita-tions include the need for elevated pressures and the recoveryof product from a high-pressure vessel. With the novel SFLtechnology, the solvents are frozen during the spray and arenot required to be miscible with the cryogenic liquid in con-trast to PCA. Liquid nitrogen is utilized as the cryogenic fluidin this study instead of CO2 because of the ultra-rapid freez-
1 Division of Pharmaceutics, College of Pharmacy (Mail Stop A1920),The University of Texas at Austin, Austin, Texas 78712.
2 Current Address: The Dow Chemical Company, Midland, Michigan48674.
3 Department of Chemical Engineering, The University of Texas atAustin, Austin, Texas 78712.
4 To whom correspondence should be addressed. (e-mail: [emailprotected])
Pharmaceutical Research, Vol. 19, No. 9, September 2002 (© 2002) Research Paper
12780724-8741/02/0900-1278/0 © 2002 Plenum Publishing Corporation
ing rates resulting from its low boiling point of –196°C. Thefeed solution is atomized below the liquid surface in a dewarflask at atmospheric pressure. The suspended frozen powdercan then be separated from the cryogenic continuous phaseby sieve filtration or other means.
In spray freezing into vapor over liquid (SFV/L) (7–11),the feed solution is atomized through a nozzle positioned at adistance above the boiling refrigerant. The droplets may be-gin to solidify while passing through the vapor gap and thenfreeze completely as contact is made with the boiling refrig-erant liquid. Spray-freezing into halocarbon vapor over liquidhas been performed by Briggs and Maxwell and Adams et al.(7,8). The resulting particles ranged in diameter from 0.84 mmto 1.68 mm (7). Halocarbon refrigerants present problems.Chlorofluorocarbons are deleterious to the ozone layer, andthe replacement hydrofluoroalkane refrigerants are expen-sive (12). In addition, both chlorofluorocarbons and hydro-fluoroalkane are suitable solvents for lipophilic compounds,so API extraction into the halocarbon cryogen can result inlow product yield (13).
Gombotz et al. and Gusman and Johnson developedspray freezing into nitrogen vapor over liquid technologies forthe purpose of capturing frozen API particles following at-omization (14,15). As the atomized droplets pass through thevapor gap above the liquid nitrogen, they may collide andcoalesce. The solutes may precipitate and grow in the unfro-zen liquid droplets as they cool and/or coalesce. The particlemorphology is not quenched until the droplets are fully so-lidified on contact with the liquid nitrogen phase below thevapor. Each of these factors may broaden the particle sizedistribution.
In the novel SFL technology, the solution is sprayed be-low the surface of the cryogenic liquid phase to avoid particlegrowth in the vapor gap described above for the conventionalSFV/L process. The liquid-liquid impingement that occurs asthe feed solution impacts the cryogenic media results in in-tense atomization into fine microdroplets that freeze instan-taneously. A schematic representation of the SFL apparatusis shown in Fig. 1. A pressurized syringe pump is used topropel the feed solution from the solution vessel through aninsulated nozzle that is submerged beneath the surface of thecryogenic liquid. Nitrogen is the cryogen of choice because itis inexpensive, environmentally friendly, inert, and may beused at atmospheric pressure, unlike CO2. Because of theultra-rapid freezing rates achieved by atomizing the feed so-
lution directly into liquid nitrogen, a cryogenic suspensioncontaining the dispersed frozen microparticles is produced.The SFL micronized powder can then be separated from theliquid nitrogen by using a fine sieve to collect the powder. Thefrozen powder is then dried, in this study by lyophilization, toproduce the dry SFL micronized powder.
The SFL process offers a variety of advantages relative totraditional technologies mentioned above. Because of intenseatomization, the formation of high-surface area droplets andthe low intrinsic temperature of liquid nitrogen, ultra-rapidfreezing rates are achieved. As a result of the ultra-rapidfreezing rates, the time for phase separation of solutes withinthe feed solution is minimized. Therefore, the API moleculesmay be dispersed hom*ogeneously throughout the solidifiedexcipient matrix of the frozen microparticle. After lyophili-zation, the dried microparticle retains the shape of the micro-droplet, but is highly porous due to the channels created asthe solvent(s) are removed. The API is molecularly dispersedwithin a hom*ogeneous excipient matrix that composes theporous microparticle.
Aqueous media should immediately wet and dissolve theSFL microparticles due to the high surface area of the porousmicroparticles. The dissolution media will fill the pore chan-nels and dissolve the API and hydrophilic excipients, whichare utilized to enhance the aqueous dissolution of a lipophilicAPI.
MATERIALS AND METHODS
Chemicals
Danazol USP and carbamazepine USP were obtained asmicronized powders. A polyoxyethylene-b-polyoxypropylene-b-polyoxyethylene triblock copolymer (Pluronic F127; Po-loxamer 407), polyvinylpyrrolidone (PVP) K-15, sodiumlauryl sulfate (SLS) NF, sodium taurocholate, lecithin NF,potassium chloride NF, and sodium hydroxide NF were pur-chased from Spectrum Quality Products Inc. (Gardena, CA,USA). Tetrahydrofuran (THF), methanol, acetonitrile, andacetic acid were obtained from EM Industries Inc. (Gibbs-town, NJ, USA). Purified water was obtained from an ultra-pure water system (Milli-QUV plus, Millipore S. A., Mols-heim Cedex, France).
SFL Processing
SFL feed solutions were prepared according to the fol-lowing procedure. Danazol or carbamazepine was dissolvedin THF. Poloxamer 407, PVP K-15, and/or SLS were dis-solved in water. The aqueous and organic solutions were thenmixed to obtain danazol/excipient blend and carbamazepine/excipient blend SFL feed solutions.
A schematic diagram of the SFL apparatus is shown inFig. 1. The SFL feed solution was placed into the solution cell(A). A constant pressure of 4000 PSI from the ISCO syringepump (B) (Model 100 DX, ISCO Inc., Lincoln, NE, USA)provided a flow rate of 11 mL/min for the SFL feed solution.The solution cell was connected to an insulated nozzle (C),which was positioned to atomize the SFL feed solution be-neath the surface of the cryogenic liquid (D). In these experi-ments liquid N2 was used as cryogenic liquid. The atomizingnozzle used was composed of polyetheretherketone tubing
Fig. 1. Schematic diagram of the SFL apparatus illustrating the so-lution cell (A), high pressure pump (B), atomizing nozzle (C), and thecryogenic liquid cell (D).
Improvement of Dissolution Using Novel Spray Freezing into Liquid Technology 1279
with an inner diameter of 63.5 �m. The SFL feed solutionswere then sprayed through the nozzle and atomized directlyinto the liquid N2 phase. Frozen particles formed instanta-neously. The frozen particles were collected and lyophilizedin a Labconco freeze dryer (Freeze dryer 5, Labconco Cor-poration, Kansas City, MO, USA) for 48 h.
To investigate the influence of the ultra rapid freezingrate in conjunction with atomization in the SFL technology, alyophilized mixture formed directly by conventional lyophili-zation process was used as a control (LM control). The ly-ophilized mixture was made of an API and excipients frozenat −78°C and lyophilized for 48 h. The compositions of thelyophilized mixtures were identical to the SFL powders. Theinfluence of the excipients on the physicochemical propertiesof the SFL powders was further studied by comparison with aphysical mixture of API and excipients (PM control). Thephysical mixtures were of the same composition as the SFLAPI powders. The micronized APIs were also used as a con-trol. The SFL powders and control samples were stored inglass vials over desiccant in a vacuum desiccator at roomtemperature before the characterization measurement.
Powder X-Ray Diffraction
Powder X-ray diffraction (XRD) was conducted usingCuK�1 radiation with a wavelength of 1.54054 Å at 40 kV and20 mA from a Philips 1720 X-ray diffractometer (Philips Ana-lytical, Natick, MA, USA). The sample powders were placedin a glass sample holder. Samples were scanned from 5 to 45°(2�) at a rate 0.05°/s. For comparative purposes the threehighest values for relative line intensity and their correspond-ing line position 2� were compared for the API samples (16).
Scanning Electron Microscopy (SEM)
A HITACHI S-4500 field emission SEM (Hitachi Instru-ments Inc., Irvine, CA, USA) was used to examine the surfacemorphology of each sample powder. The sample was fixed toa SEM stage with double-sided adhesive tape and gold sputtercoated.
Particle Size Distribution
The particle size distribution of the sample powders wasdetermined using the method based on a time-of-flight mea-surement (Aero-Disperser®, Amherst Process Instrument,Amherst, MA, USA). The mean particle diameter and spanindex were determined for the SFL powers and controls.
Surface Area Measurement
A Nova 3000 surface area analyzer (Quantachrome Cor-poration, Boynton Beach, FL, USA) was used to determineN2 sorption at 77.40° K. The surface area per unit powdermass was calculated from the fit of adsorption data to theBrunauer, Emmett, and Teller (BET) equation (17).
Contact Angle Measurement
Compacts of sample powders were prepared at a 500 kgcompression force using a Carver Laboratory Press (ModelM, ISI Inc., Round Rock, TX, USA) with flat-faced 6 mmdiameter punches. A droplet of purified water or fed statesimulated intestinal fluid (FeSSIF) media (3 �l) was placed
onto the surface of compact and observed using a low powermicroscope. The contact angle was determined by measuringthe tangent to the curve of the droplet on the surface of thecompact using a goniometer (Model No.100-00-115, Ramè-Hart Inc., Mountain Lakes, NJ, USA).
Dissolution Studies
The amount of API dissolved, as a function of time, wasdetermined using USP Apparatus 2 method (Vankel 7000,Vankel Technology Group, Cary, NC, USA). All dissolutionstudies were conducted at sink conditions. The FeSSIF mediawas prepared according to Dressman et al. (18). SFL danazol/poloxamer 407 powders or control samples containing ap-proximately 3 mg danazol were added to 900 mL of FeSSIFmedia (37°C). The paddle speed was 50 rpm. A 5-mL aliquotwas taken at each time point and filtered through a 0.45-�mfilter then diluted with acetonitrile, filtered through a 0.45-�mfilter again, and analyzed by HPLC (19,20). Purified water(900 mL) was used for the carbamazepine as the dissolutionmedia and the same operating parameters discussed for dana-zol. Dissolution profile for SFL carbamazepine/excipientspowders were studied as well as their controls.
Stability Study
Stability studies were conducted at 25°C/60%RH.Sample powders were stored in the capped glass vials (20 mL)and characterized as a function of exposed time.
Statistical Analysis
The data were compared using a Student’s t test of thetwo samples assuming equal variances to evaluate the differ-ences. The significance level (� � 0.05) was based on the95%probability value (p < 0.05).
RESULTS AND DISCUSSION
Crystallinity
The crystallinity greatly impacts the solubility and disso-lution rate of poorly water-soluble APIs (21). The chemicalpotential of a metastable high-energy amorphous state can bemarkedly larger than that of an equilibrium crystal. Thishigher chemical potential can lead to a substantially greaterlocal solubility of the API near the interface and thus a fastermass transfer rate into the dissolution media. Therefore, thecrystallinity of an API may be reduced to enhance the disso-lution rate. Powder XRD patterns of SFL powders and con-trol samples are presented in Figs. 2 and 3. Micronized bulkdanazol had a similar X-ray diffraction pattern to that re-ported by Liversidge and Cundy (22). The high peak intensi-ties of bulk danazol indicated a high degree of crystallinity.The physical mixture and lyophilized mixture showed thecharacteristic diffraction peaks (16) of danazol at 15.8, 17.2,and 19.0 (2�). The X-ray pattern of the SFL danazol/poloxamer 407 powder exhibited a significant reduction inpeak intensity for danazol. The diffraction peaks of danazolwere not evident whereas those of poloxamer 407 at 19.3 and23.4 (2�) were present. This absence of peaks indicated thatdanazol contained in SFL danazol/poloxamer 407 powder wasin an amorphous state. Similarly, lack of crystallinity was also
Hu et al.1280
found for the SFL carbamazepine/SLS powder (Fig. 3). Theabsence of characteristic peaks of carbamazepine (23) at 15.3,25.0, and 27.6 (2�) indicated an amorphous state. The controls(either physical mixture or lyophilized mixture) exhibitedcrystalline diffraction peaks of carbamazepine. For bothdanazol and carbamazepine the powders produced by SFLtechnique exhibited very little crystallinity, in contrast withthe significant crystallinity for powders made by the conven-tional lyophilization process. Apparently, the freezing rates inthe SFL technique were fast enough to trap the API in anamorphous state without allowing time for crystallization. Ul-tra-rapid freezing may be expected from the high surface areaof the atomized droplets and the rapid heat transfer from thedroplets to the liquid nitrogen.
Particle Size Distribution and Morphology
The particle size of an API is another important param-eter controlling the dissolution rate. Decreasing the particle
size increases the surface area, which enhances the dissolu-tion rate (24). The particle size distribution for the SFLpowders and controls (Table I) were determined in an Aero-Disperser® based on time-of-flight measurements. The meanparticle diameter of the SFL danazol/poloxamer 407 powderwas 7.10 �m. In contrast, the mean particle diameters of thedanazol and lyophilized mixture were 18.27 and 18.86 �m,respectively. Similar results were found for the SFL carba-mazepine/SLS powder. The mean particle diameter of theSFL carbamazepine/SLS powder was 7.11 �m, which wassmaller than that of the controls. The span index is used todescribe the polydispersity in a given particle size distributionand is defined as (D90-D10)/D50, where D10, D50, and D90are the respective particle sizes at 10, 50, and 90%cumulativepercentage undersize (25). The span index of carbamazepinewas 2.94, indicating high polydispersity. For the SFL powderthe polydispersity was decreased markedly down to 1.31. Theconventional lyophilization process reduced the mean particlediameter only slightly compared to the micronized APIs.However, for both danazol and carbamazepine, the SFL tech-nique produced nanostructured particles with narrow particlesize distributions. Various factors contributed to the limitedparticle growth during SFL processing. The atomized dropletswere immediately frozen upon the rapid freezing rate, and therapid freezing rates limited the time for particle growth. Also,the surfactant in the unfrozen domains in the sprayed dropletsinhibits particle coalescence and crystal growth.
To further investigate the effect of the SFL technique onthe morphology of APIs, sample powders were examined bySEM. The SEM micrograph of the micronized bulk danazol(Fig. 4a) illustrates large crystalline particles. The SEM mi-crograph of the lyophilized control particles (Fig. 4b) indi-cates a smooth surface when compared with the bulk danazol.In contrast, the SFL danazol/poloxamer 407 particles (Fig. 4c)had a highly porous morphology. The surface area of the SFLdanazol/poloxamer 407 powder was 11.04 m2/g, which was asignificant increase over the lyophilized mixture (1.27 m2/g).The high surface area confirmed the highly porous structureof SFL powders observed by SEM. Such high surface areaindicated that the SFL powder particles were highly porous,similar to the SFV/L produced protein particles reported byMaa et al. (4) and Costantino et al. (26). Similar results forSFL carbamazepine/excipients powders were observed in theSEM analysis and surface area measurement. For example,the SEM micrograph of the SFL carbamazepine/SLS sampledemonstrates a fine powder consisting of porous micropar-ticles. Powder made by the conventional lyophilization pro-cess was larger and less porous relative to the SFL powder.The surface area of SFL carbamazepine/SLS powder was12.81 m2/g, which was greater than that of the lyophilizedmixture (2.33 m2/g).
Fig. 3. Powder X-ray diffraction patterns of micronized carbamaze-pine, PM carbamazepine/SLS, LM carbamazepine/SLS, and SFL car-bamazepine/SLS.
Fig. 2. Powder X-ray diffraction patterns of micronized danazol, PMdanazol/poloxamer 407, LM danazol/poloxamer 407, and SFL dana-zol/poloxamer 407.
Table I. Effect of Composition and Process on the Particle SizeDistribution
Sample D10 D50 D90 Span index
Bulk danazol 7.14 18.27 29.56 1.23Slowly frozen control 7.27 18.86 31.13 1.27SFL danazol/poloxamer 407 1.54 7.10 10.20 1.22Bulk carbamazepine 15.56 39.70 132.33 2.94Slowly frozen control 4.58 26.71 46.74 1.58SFL carbamazepine/SLS 1.33 7.11 10.61 1.31
Improvement of Dissolution Using Novel Spray Freezing into Liquid Technology 1281
Wettability
The importance of wettability on dissolution rate—thatis, the area of contact between a powder and dissolution me-dia—has been well studied (27). To characterize wettability,the contact angle was used to determine the interfacial ten-sion present between the compact of API powders and liquidintercontact angle values for the SFL powders, and controlsare reported in Table II. The limits in the contact angle are 0°for complete wetting and 180° for no wetting (28). The meanvalue of contact angle for the SFL danazol/poloxamer 407powder was only 25° against the FeSSIF media, which wassignificantly lower than that of the SFL danazol (55o), lyoph-ilized mixture (34°), and physical mixture (58o) (p < 0.05).The mean value for SFL carbamazepine/SLS powder was 24°against purified water, which was significantly lower than forthe physical mixture (52°) and SFL carbamazepine (p < 0.05).The significant reduction of contact angle � for the SFL pow-ders compared with the controls indicated the presence of ahydrophilic solid surface. Specifically, the contact angle is de-scribed by the equation: cos � � f1 cos �1 + f2 cos �2 (29),where f1 and f2 are the bulk volume fractions of the API andexcipient. The contact angle of SFL powders was well belowthat of the physical mixture, indicating that the solid surfaceof the SFL powder was enriched in the hydrophilic excipient.
This enrichment may be formed during the SFL process. Asthe API and excipient precipitated from the concentrated un-frozen aqueous phase, and the precipitate was surrounded byhydrophilic solvent. The hydrophilic solvent attracts the hy-drophilic excipient molecules preferentially to the surface ofthe particles. The lower contact angle of the SFL powderscompared with the lyophilized mixture may be due to thesurface roughness. The rougher surface for the SFL powdersrelative to the lyophilized mixture as depicted in the SEMsand the higher surface areas may be expected to decrease thecontact angle as observed. So, both preferential enrichment ofthe powder surface with the hydrophilic excipient and thesurface roughness lower the contact angle and favor wettingof the SFL micronized powder relative to the controls.
Dissolution
The profiles presented in Figs. 5 and 6 illustrate the dis-solution of the SFL powders and controls for danazol andcarbamazepine, respectively. The FeSSIF media used for thedanazol dissolution studies reportedly simulates in vivo gas-
Fig. 4. SEM micrographs of micronized danazol control (Fig. 4a), LM danazol/poloxamer 407 (Fig. 4b) and SFLdanazol/poloxamer 407 (Fig. 4c).
Table II. Effect of Composition and Process on the Contact Angle ofDanazol and Carbamazepine Powders
SamplesContact angle degrees
(FeSSIF media)
PM danazol/poloxamer 407 58LM danazol/poloxamer 407 34SFL danazol 55SFL danazol/poloxamer 407 25
Contact angle degrees(purified water)
PM carbamazepine/SLS 52LM carbamazepine/SLS 29SFL carbamazepine 45SFL carbamazepine/SLS 24
Fig. 5. Dissolution profiles of micronized danazol, PM danazol/poloxamer 407, LM danazol/poloxamer 407, and SFL danazol/poloxamer 407.
Hu et al.1282
trointestinal fluid; low levels of surfactants are recommendedto be included in the dissolution media to give a better cor-relation between in vitro and in vivo data (18). The rate ofdissolution of micronized bulk danazol (Fig. 5) was slow; only21% of the danazol dissolved in 60 min. The dissolution of thephysical mixture was also quite slow, only about 48% in 60min. The conventional lyophilization process increased thedissolution rate considerably, with 80% danazol dissolved in60 min. However, the amount dissolved reached 99% in only10 min for the SFL danazol/poloxamer 407 formulation. Ma-jor improvements in the dissolution rates were also found forSFL carbamazepine/excipient powders relative to the controls(Fig. 6). For example, the dissolved carbamazepine from SFLcarbamazepine/SLS powder in 10 min was 98%, a profoundenhancement when compared with 4% for the micronizedbulk carbamazepine. The difference in the dissolution rate ofSFL carbamazepine powders prepared by various excipientswas negligible in 60 min and only slightly different in 20 min.Each of the excipients was successful in producing extremelyrapid dissolution.
The increased dissolution rate of the SFL powders is duein part to the amorphous nature of the API, the reduction inparticle size, the increase in porosity and resulting enhancedsurface area, and intimate dispersion of the drug and hydro-philic excipient. The excipient may increase the local equilib-rium concentration of the API in the boundary layer aboutthe particles. The rapid freezing rate resulting from the in-tense atomization led to porous nanostructured particles withhigh amorphous API fraction and large surface area. Thismorphology is quite different from the low porosity, low sur-
face area semicrystalline cakes produced by the conventionallyophilization process. At the same time, the enhanced wet-tability or better contact between the surface of the solid andthe liquid increased the dissolution of the SFL powder. Also,the porous channels of the API/excipient matrix created inthe SFL process allowed the dissolution media to easily pen-etrate into the particles and facilitate dissolution.
Stability
A stability study was conducted for the SFL danazol/poloxamer 407 at 25°C/60%RH for 2 months to examine anychanges in crystallinity or other properties of the SFL pow-ders. The X-ray diffraction pattern (Fig. 2) of SFL danazol/poloxamer 407 powder exhibited no change in peak intensityof danazol, but a slightly increased peak intensity of po-loxamer 407. Despite the high humidity, the danazol crystal-linity was unchanged over 2 months. The dissolution results(Fig. 5) demonstrated that there was only a 5% difference inthe dissolution profiles between the initial and 2-monthsamples, which was calculated by the similarity factor f2 forthe dissolution profile comparison using the method reportedby Shah (30). The slightly decreased initial release of danazolmay be due to the change in crystallinity of poloxamer 407.
CONCLUSIONS
The novel SFL technology was demonstrated to producefree-flowing powder of nanostructured particles containingdanazol or carbamazepine. The SFL powders exhibited sig-
Fig. 6. Dissolution profiles of SFL carbamazepine/SLS, SFL carbamazepine/poloxamer 407, SFL carbamaz-epine/PVP K15, SFL carbamazepine/poloxamer 407/PVP K15, and their respective controls.
Improvement of Dissolution Using Novel Spray Freezing into Liquid Technology 1283
nificantly enhanced dissolution rates compared to the pow-ders formed by the conventional lyophilization process. Inaddition, an amorphous structure, high surface area and in-creased wettability of the flowable SFL powders are predomi-nant characteristics, so the novel SFL technology is an effec-tive particle engineering process for pharmaceutical develop-ment and manufacturing to improve dissolution rates ofpoorly water soluble APIs for oral delivery systems.
ACKNOWLEDGMENTS
Financial support for this study from The Dow ChemicalCompany is gratefully acknowledged.
This material is based on work supported by the STCProgram of the National Science Foundation under Agree-ment No. CHE-9876674.
REFERENCES
1. G. L. Amidon, H. Lnnernas, V. P. Shah, and J. R. Crison. Atheoretical basis for a biopharmaceutic drug classification: thecorrelation of in vitro drug product dissolution and in vivo bio-availability. Pharm. Res. 12:413–420 (1995).
2. L. Bertisson. Clinical pharmaco*kinetics of carbamazepine. Clin.Pharmaco*kinet. 3:128–143 (1978).
3. T. Loftsson and H. Frioriksdottir. The effect of water-solublepolymers on the aqueous solubility and complexing abilities of�-cyclodextrin. Int. J. Pharm. 163:115–121 (1998).
4. Y. F. Maa, P. A. Nguyen, T. Sweeney, S. J. Shire, and C. C. Hsu.Protein inhalation powders: spray drying vs. spray freeze drying.Pharm. Res. 16:249–254 (1999).
5. T. L. Rogers, K. P. Johnston, and R. O. Williams, III. A compre-hensive review: solution-based particle formation of pharmaceu-tical powders by supercritical or compressed fluid CO2 and cryo-genic spray-freezing technologies. Drug Dev. Ind. Pharm. 27:1003–1015 (2001).
6. W. Wang. Lyophilization and development of solid protein phar-maceuticals. Int. J. Pharm. 203:1–60 (2000).
7. T. H. Adams, J. P. Beck, and R. C. Menson. Novel particulatecompositions. US4323478 (1982).
8. A. R. Briggs and T. J. Maxwell. Method of preparation of lyoph-ilized biologic products. US3932943 (1976).
9. D. B. Dunn, G. J. Masavage, and H. A. Sauer. Method of freezingsolution droplets and the like using immiscible refrigerants ofdiffering densities. US3653222 (1972).
10. I. R. Buxton and J. M. Peach. Process and apparatus for freezinga liquid medium. US4470202 (1984).
11. H. A. Sauer. Method and apparatus for freeze-freeze drying.US3484946 (1969).
12. The Montreal Protocol on Substances That Deplete the OzoneLayer. 1987.
13. R. O. Williams, T. L. Rogers, and J. Liu. Study of solubility of
steroids in hydrofluoroalkane propellants. Drug Dev. Ind.Pharm. 25:1227–1234 (1999).
14. W. R. Gombotz, M. S. Healy, L. R. Brown, and H. E. Auer. Pro-cess for producing small particles of biologically active molecules.WO90/13285 (1990).
15. M. I. Gusman and S. M. Johnson. Cryochemical method of pre-paring ultrafine particles of high-purity superconducting oxides.US4975415 (1990).
16. B. D. Cullity. Chemical analysis by X-ray diffraction. In B. D.Cullity (eds.), Elements of X-ray Diffraction, Addison-Wesley,Massachusetts, 2001 pp. 397—420.
17. S. Brunauer, P. H. Emmett, and E. Teller. Adsorption of gases inmultimolecular layer. J. Am. Chem. Soc. 60:309–319 (1938).
18. E. Galia, E. Nicolaides, D. Horter, R. Lobenberg, C. Reppas, andJ. B. Dressman. Evaluation of various dissolution media for pre-dicting in vivo performance of class I and II drugs. Pharm. Res.15:698–705 (1998).
19. S. Brown, G. Rowley, and J. T. Pearson. Surface treatment of thehydrophobic drug danazol to improve drug dissolution. Int.J. Pharm. 165:227–237 (1998).
20. B. L. Pedersen, A. Mullertz, H. Brondsted, and H. C. Kristensen.A comparison of the solubility of danazol in human and simu-lated gastrointestinal fluids. Pharm. Res. 17:891–894 (2000).
21. M. Shibata, H. Kokubo, K. Morimoto, T. Ishida, and M. Inoue.X-ray structural studies and physicochemical properties of cimet-idine polymorphism. J. Pharm. Sci. 72:1436 (1983).
22. G. G. Liversidge and K. C. Cundy. Particle size reduction forimprovement of oral bioavailability of hydrophobic drugs: I. Ab-solute oral bioavailability of nanocrystalline danazol in beagledogs. Int. J. Pharm. 125:91–97 (1995).
23. M. J. Lowes, M. R. Caira, A. P. Lotter, and J. G. VanderWatt.Physicochemical properties and X-ray structure studies of thetrigonal polymorph of carbamazepine. J. Pharm. Sci. 76:744–752(1987).
24. N. Kondo, T. Iwao, H. Masuda, K. Yamanouchi, Y. Ishihara, N.Yamada, T. Haga, Y. Ogawa, and K. Yokoyama. Improved oralabsorption of a poorly water soluble drug, HO-221, by wet-beadmilling producing particles in submicron region. Chem. Pharm.Bull. 41:737–740 (1993).
25. A. H. Lefebvre. Atomization and Sprays, Hemisphere PublishingCorp., New York, 1989.
26. H. R. Costantino, L. Firouzabadian, K. Hogeland, C. Wu, C.Beganski, K. G. Carrasquillo, M. Córdova, K. Griebenow, S. E.Zale, and M. A. Tracy. Protein spray-freeze drying. Effect ofatomization condition on particle size and stability. Pharm. Res.17:1374–1383 (2000).
27. V. Bakatselou, R. C. Oppenheim, and J. B. Dressman. Solubili-zation and wetting effects of bile salts on the dissolution of ste-roids. Pharm. Res. 8:1461–1469 (1991).
28. A. P. Martin. Interfacial phenomena. In A. P. Martin (ed.), Physi-cal Pharmacy, Lea & Febiger, Philadelphia, 1993 pp. 384–386.
29. A. W. Adamson and A. P. Gast. Physical Chemistry of Surfaces,Wiley, New York, 1997.
30. V. P. Shah, Y. Tsong, P. Sathe, and J.-P. Liu. In vitro dissolutionprofile comparison-statistics and analysis of the similarity factor,f2. Pharm. Res. 15:889–896 (1998).
Hu et al.1284
