ELSEVIER
`
`International Journal of Pharmaceutics 221 (2001) 165- 174
`
`international
`journal of
`pharmaceutics
`
`www.elsevier.com/locatejijpharm
`
`Characterisation of the aggregation behaviour in a
`salmeterol and fluticasone propionate inhalation aerosol
`system
`Yonatan Michael a, Martin J. Snowden a,*, Babur Z. Chowdhry a,
`Ian C. Ashurst b, Craig J. Davies-Cutting h, Trevor Riley b
`a School of Chemical and Life Sciences, University of Greenwich, Wellington Street, Woolwich, London SE18 6PF, UK
`b Inhalation Product Development, GlaxoSmithKline R&D, Park Road, Ware, Hertfordshire SG12 ODP, UK
`
`Received 8 December 2000; received in revised form 23 March 2001; accepted 30 March 2001
`
`Abstract
`
`The nature of the drug-drug aggregation phenomena between salmeterol xinafoate and fluticasone propionate used
`in a metered-dose inhaler system has been examined. Interactions between the drugs in the solvents 1,1 ,2-trichlorotrifl(cid:173)
`oroethane (CFC-113) and 1,1,1,2-tetrafluoroethane (HFA-134a) have been characterised using a focused beam
`reflectance measurement probe by measuring the average floc size of the drug particles individually and in
`combination as a function of stirrer rate. The floc composition in the CFC-113 system, where the drug particles
`cream, was determined by high-performance liquid chromatography analysis. The aggregation behaviour of the
`individual drugs was shown to depend on the physical and chemical properties of both the drug substance and the
`media. Larger floes were observed for salmeterol xinafoate compared with fluticasone propionate, while both drugs
`formed larger aggregates in HF A-134a compared with in CFC-113. The floc composition studies demonstrated that,
`in the combined formulation in CFC-113, sahneterol xinafoate and fluticasone propionate aggregate together to form
`hetero-flocs. The interaction between the two drugs was such that they did not separate on creaming, despite having
`different densities. The average floc size of the combined drug suspension was also found to depend on the dispersion
`medium. © 2001 Elsevier Science B.V. All rights reserved.
`
`Keywords: Salmeterol xinafoate; Fluticasone propionate; HFA-134a; CFC-113; Aggregation; Focused beam reflectance
`
`*Corresponding author. Tel.: + 44-208-3319981; Fax: +
`44-208-3318305.
`E-mail address: m.j.snowden@greenwich.ac.uk (M.J. Snow(cid:173)
`den).
`
`1. Introduction
`
`Metered-dose inhalers (MDis) are the most fre(cid:173)
`quently employed dosage forms for delivering ac(cid:173)
`tive drug substances to the respiratory tract via
`the inhaled route. MDis contain fine micronised
`drugs
`in a suspension of chlorofluorocarbon
`
`0378-5173/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
`PII: 80378-5173(0 1)00678-0
`
`PLAINTIFFS'
`TRIAL EXHIBIT
`PTX0179
`
`MEDA_APTX03505071
`
`PTX0179-00001
`
`1
`
`CIP2044
`Argentum Pharmaceuticals LLC v. Cipla Ltd.
`IPR2017-00807
`
`

`

`166
`
`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`(CFC) or hydrofluoroalkane (HFA) propellants,
`which act as both a suspending medium and
`propellant. Current formulations are predomi(cid:173)
`nantly suspensions of single micronised active
`compound in the range 2-5 ~-tm in liquefied
`CFCs or HF As. The physical and chemical
`properties of the single-drug formulations are, in
`principle, easier to study. Combined drug for(cid:173)
`mulations are less well understood and, as a re(cid:173)
`significant amount of fundamental
`sult, a
`scientific research work on the subject is being
`undertaken. Previous work (Michael et al., 2000)
`on the combined inhaler formulations of both
`salmeterol xinafoate and fluticasone propionate
`in propellant CFC-113 indicates that hetero-ag(cid:173)
`gregation of the two drugs would appear to be
`taking place. The hetero-aggregation results in
`reduced drug deposition on to the MDI surfaces
`as compared with that observed with the indi(cid:173)
`vidual drug formulations. This phenomenon
`could be an advantage from a pharmaceutical
`point of view, i.e. the combination formulation
`shows a decrease in the total loss of the drugs
`due to deposition on to the internal surfaces of
`the MDI. Clinical studies have indicated that
`drug formulation of salmeterol xinafoate and
`fluticasone propionate in a single inhaler pro(cid:173)
`vides a treatment as efficacious in achieving
`asthma control and as well tolerated over a 28-
`week period as the two drugs administered indi(cid:173)
`vidually (Chapman et al., 1999). However, a
`clearer understanding of the aggregation be(cid:173)
`haviour of salmeterol xinafoate and fluticasone
`propionate will facilitate the design of combina(cid:173)
`tion aerosol systems.
`A number of methods were considered in an
`attempt
`to characterise
`the aggregation be(cid:173)
`haviour of suspensions in MDis. Microscopy is
`a commonly used technique that allows an opti(cid:173)
`cal inspection of the particles and can be used
`to judge whether a good dispersion has been
`achieved or if any aggregation is present in the
`system (Rawle, 1999). Light-diffraction
`tech(cid:173)
`niques (Sidhu et al., 1993) are also used for the
`study of floes in liquid suspensions, and the par(cid:173)
`ticles in suspension are measured by re-circulat(cid:173)
`ing the sample in front of the laser beam. This
`technique was, however, found to be unsuitable
`
`due to the high optical absorbance of the con(cid:173)
`centrated suspensions, which made scattering
`measurements difficult due to low signals and
`multiple scattering. Rheological methods (Sidhu
`et al., 1993) have also been used to study the
`in
`inhalation systems.
`extent of flocculation
`However, these methods normally make use of
`high solid-phase concentrations of the dispersed
`materials as compared with those normally en(cid:173)
`countered in commercial metered-dose inhalers.
`In addition, small-angle
`light-scattering
`tech(cid:173)
`niques have also been used to study the size and
`structural dynamics of floes (Spicer et al., 1998);
`however, floes in MDis would be too large for
`this techniques.
`The objective of the work reported herein was
`to gain a better understanding of the physical
`and chemical factors involved in the combined
`formulation of salmeterol and fluticasone propi(cid:173)
`onate in a single MDI aerosol system. The ag(cid:173)
`gregation of salmeterol xinafoate and fluticasone
`propionate
`in HFA-134a and CFC-113 was
`studied using a focused beam reflectance mea(cid:173)
`surement (FBRM), which enables floc size to be
`determined as a function of stirrer rate. In addi(cid:173)
`tion, a floc composition study of the CFC-based
`combined drug formulation was performed to
`monitor the migration behaviour of the individ(cid:173)
`ual drugs within the suspension.
`
`2. Materials and methods
`
`2.1. Reagents
`
`The test samples of salmeterol xinafoate and
`fluticasone propionate were donated by GlaxoS(cid:173)
`mithKline Research and Development (Ware,
`UK), and were used as received. The propellants
`1,1,2-trichlorotrifloroethane
`(CFC-113)
`and
`1,1,1,2-tetrafluoroethane (HFA-134a), with a wa(cid:173)
`ter content of less than 10 p.p.m., were obtained
`from ICI Chemicals and Polymers Ltd (Run(cid:173)
`corn, Cheshire, UK). Poly(vinyl chloride) (PVC)(cid:173)
`coated glass aerosol bottles were supplied by
`~-tl metering
`Wheaton Ltd (NJ, USA) and 63
`valves were purchased from Valois S.A. (Le
`Neubourg, France).
`
`MEDA_APTX03505072
`
`PTX0179-00002
`
`2
`
`

`

`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`167
`
`2.2. Focused beam reflectance measurement
`
`The aggregation behaviour of the suspensions
`was studied by FBRM, using a Lasentec Labtec
`1000 instrument (Lasentec Inc, Redmond, USA)
`(Allen and Davies, 1998). The apparatus uses a
`focused laser diode beam source that is projected
`on to the sample at a fixed velocity and rapidly
`scans across the particle structures. As the focused
`beam transcribes a particle or floc structure, light
`from the incident beam is reflected back until the
`beam reaches the opposite edge of the particle.
`The back-scattered light is collected by a pair of
`stereoscopic photodiode detectors and converted
`into an electronic signal. The chord length of the
`particle or aggregate can then be determined from
`the product of the pulse duration and the velocity
`of the scanning laser. The chord lengths measured
`over a specific time period are sorted into a
`38-chord distribution that covers the size range
`0.4-250
`J.lm. Because
`the beam does not
`intersect the particle in the same way every time,
`the so-called 'random chord length distribution'
`of the particles is measured (Fig. 1). During one
`measurement cycle, adjusted to 8 s, thousands of
`particles are identified. To produce robust statis(cid:173)
`tics, 2000 or more particles need to be sampled
`per second. An advantage of FBRM over forward
`scattering or transmission sizing techniques is that
`opaque dispersions can be analysed. Hence, ex(cid:173)
`periments can be performed at realistic concentra(cid:173)
`tions, without the need for sample dilution.
`Unlike many other particle sizing techniques,
`pressurised systems can be readily studied using
`the Lasentec Labtec 100 without the need for
`building a complex pressure cell. Instead, samples
`may be presented in cylindrical glass-coated aero(cid:173)
`sol bottles. Nevertheless, an important aspect of
`
`the Lasentec system is that the hydrodynamic
`conditions must be such that a representative
`sample flows past the probe window. This cannot
`always be guaranteed in practice, causing system(cid:173)
`atic errors or noisy signals (Dijkstra et al., 1996).
`Measurement of the floc size of the suspensions
`can be hindered by either gravitational sedimenta(cid:173)
`tion or creaming, if the densities of the solid and
`liquid phase are unequal. To avoid this problem,
`the suspensions were agitated by means of a mag(cid:173)
`netic stirrer. By varying the rotational velocity of
`the magnetic stirrer, it is possible to control the
`shear conditions within the sample. The shear
`stress in the aerosol bottle filled with HFA-134a
`at a stirrer speed of 1000 rev min- 1, was esti(cid:173)
`mated by assuming laminar flow and was found
`to be of the order of 0.025 N m - 2 .
`Data from the experiments were collected by
`analysing the samples at three equidistant posi(cid:173)
`tions on the aerosol bottle. At each position, two
`particle size determinations were performed, with
`each determination being the mean of five mea(cid:173)
`surements, averaged by the instrument software.
`The floc size measurement of one sample was,
`therefore, the average of salmeterol xinafoate and
`fluticasone propionate six particle size determina(cid:173)
`tions from
`three different positions on
`the
`bottle. In this way, errors due to imperfections in
`the plastic-coated aerosol bottles could be min(cid:173)
`imised.
`
`2.3. Floc size analysis
`
`Model suspensions in HFA-134a were prepared
`by pressure filling propellant through a valve at(cid:173)
`tached to a pressure-resistant 15 ml round clear
`PVC-coated glass aerosol bottle, to which known
`
`Scanning Laser Beam
`
`Duration of Reflection
`Measured - Chord
`
`Fig. I. An illustration of the normalised chord length distribution of a particulate species.
`
`MEDA_APTX03505073
`
`PTX0179-00003
`
`3
`
`

`

`168
`
`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`quantity of the drugs had been added. Individual
`drug suspensions over the concentration range
`0.02-0.50% w/w salmeterol xinafoate and flutica(cid:173)
`sone propionate were prepared. Suspensions con(cid:173)
`taining 0.05% w jw salmeterol xinafoate together
`with 0.07% wjw fluticasone propionate were used
`to study aggregation behaviour in the combina(cid:173)
`tion
`product.
`These
`concentrations
`are
`equivalent to the strengths of salmeterol and fluti(cid:173)
`casone propionate combination products that
`were marketed
`initially
`in
`the United
`Kingdom from July 2000. A glass-coated mag(cid:173)
`netic stirrer bar (13 x 5 mm2) was added to each
`aerosol bottle before a metering valve was
`crimped into place (Pamasol model type 2002).
`The bottles were filled with approximately 18 g
`HFA-134a via a pressure burette, which was pres(cid:173)
`surised at about 5 bar with HFA-134a at room
`temperature. Each of the formulations was pre(cid:173)
`pared in triplicate. The aerosol bottles were soni(cid:173)
`cated (Decon FS300 ultra-sonic bath) for 10 min
`to remove any entrapped air from the suspension
`and to promote dispersion. After equilibration,
`the suspensions were re-weighed to ensure that
`significant amounts of propellant had not been
`lost due to evaporation. Drug suspensions in
`CFC-113 were prepared in the same way, except
`that the CFC was filled directly into the aerosol
`bottles prior to being sealed with the metering
`valve.
`The median chord length of the individual drug
`suspensions in the two propellants was measured
`as a function of drug concentration using the
`FBRM system at a constant stirrer rate of 300 rev
`min- 1. The robustness of the floes was investi(cid:173)
`gated by measuring the average floc size of the
`drug suspension as a function of increasing stirrer
`speed up to a maximum of 1500 rev min- 1. The
`suspension was allowed to equilibrate for 5 min at
`each speed setting before measurements were
`made.
`The reproducibility of the Lasentec FBRM re(cid:173)
`sponse was checked by performing repeat mea(cid:173)
`surements of an external standard of aqueous
`graphite dispersion at a constant stirrer rate of
`250 rev min - 1 and interspersed with the samples
`throughout the measurements. The median chord
`length distribution of the graphite standard was
`
`repeatedly found to be 6.00 ± 0.33 11m. Further(cid:173)
`more, the particle size of the graphite standard
`was found to be independent of stirrer speed.
`
`2.4. Electrophoretic mobility studies
`
`The electrophoretic mobility of the suspended
`drug particles was determined using a Malvern
`Zetasizer 3000 (Malvern Instruments, UK) in con(cid:173)
`junction with a non-aqueous dip cell that has a
`narrow electrode gap of 2 mm. Individual disper(cid:173)
`sions of 0.0025% wjw salmeterol xinafoate and
`0.0035% wjw fluticasone propionate were pre(cid:173)
`pared in CFC-113 to give a suspension of suitable
`conductance. The dispersions were ultrasonicated
`for 10 min prior to analysis to ensure that the
`drug was fully dispersed.
`
`2.5. Floc composition studies
`
`Floc composition studies were performed on
`suspensions of the two drugs to examine the
`settling behaviour of the drug particles in the
`absence of shear. Individual and combined disper(cid:173)
`sion concentrations of 0.05% wjw salmeterol xi(cid:173)
`nafoate and 0.07% wjw fluticasone propionate
`were studied in CFC-113 solvent only, as the high
`vapour pressure of HFA-134a propellant makes
`sampling difficult. Suspensions were prepared as
`described in Section 2.3 except that the aerosol
`bottles were sealed with a rubber septum. Using a
`glass syringe, 1 ml suspension was slowly removed
`from a position approximately one-quarter of the
`way from the bottom of the aerosol bottle with(cid:173)
`out re-dispersing the sample. The dispersions were
`sampled at time intervals of 0, 1.5, 5 and 10 min
`and the 1 ml aliquots transferred to 50 ml volu(cid:173)
`metric flasks. Fig. 2 shows photographs taken to
`illustrate the separation behaviour of concen(cid:173)
`trated drug dispersion after a time interval of 10
`min. The collected samples were left open to the
`atmosphere to allow evaporation of the CFC
`solvent, and the drug then dissolved in 70/30 (v /v)
`methanol/water and assayed by high-performance
`liquid chromatography (HPLC) with UV detec(cid:173)
`tion at 228 nm. The details of the HPLC analyti(cid:173)
`cal method have previously been described by
`Michael et al. (2000).
`
`MEDA_APTX0350507 4
`
`PTX0179-00004
`
`4
`
`

`

`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`169
`
`Fig. 3 illustrates that the average floc size of the
`salmeterol xinafoate aggregates is larger than that
`of the fluticasone propionate aggregates, in both
`CFC and HF A media. This may be partially due
`to salmeterol xinafoate having a lower density
`(1.2 g em- 1 as determined by helium pycnometry)
`compared with that of fluticasone propionate (1.3
`g em- 1), and therefore occupying a larger phase
`volume at the same w/w concentration. However,
`the nature of the drug may also influence the
`aggregation behaviour in the suspensions. The
`electrophoretic mobilities of salmeterol xinafoate
`and fluticasone propionate in CFC-113 were
`found to be 3 X 10- 9 and 1 X 1Q- 9 m2 s- 1 y- 1 ,
`respectively, at l5°C. This reflects the fact that a
`higher charge is present on the surface of salme(cid:173)
`terol xinafoate compared with that of fluticasone
`propionate, which is believed to increase the po(cid:173)
`larity of salmeterol xinafoate relative to flutica(cid:173)
`sone propionate. It would be expected that the
`hydrophobic
`fluticasone propionate particles
`would be wetted more readily in the non-polar
`CFC-113 environment than the salmeterol xi(cid:173)
`nafoate particles. This could explain why flutica(cid:173)
`sone propionate forms more compact aggregates
`than salmeterol xinafoate.
`The results presented in Fig. 3 also illustrate
`that the suspending medium has a considerable
`effect on the floc size of the drug suspension. The
`average floc size of both salmeterol xinafoate and
`fluticasone propionate was larger in HFA-134a as
`compared with in CFC-113. For diffusion-limited
`aggregation, the aggregation kinetics of the sus(cid:173)
`pension will be dependent on the viscosity of the
`medium. Einstein's law of diffusion, states that
`the diffusion coefficient of spherical particles is
`given by (Shaw, 1992): D = (kT)/(6n1Ja), where k
`is Boltzmann's constant, Tis the absolute temper(cid:173)
`ature, 1J
`is the viscosity and a is the particle
`radius.
`Comparing the liquid viscosity of HFA-134a
`(0.20 mPa s) with that ofCFC-113 (0.66 mPa s) at
`25°C (Solvay, 1992), it can be seen that the diffu(cid:173)
`sion coefficient of drug particles in HFA-134a is
`approximately three times faster than in CFC-
`113. Hence, under equivalent shear conditions,
`the faster rate of coagulation would be expected
`to result in the observed floc size increase.
`
`X
`=~===
`
`(a)
`
`(b)
`
`Fig. 2. Photographs illustrating the settling behaviour of con(cid:173)
`centrated drug dispersion at (a) 0 min and (b) 10 min intervals.
`Samples for floc composition studies were taken from position
`x using a syringe.
`
`3. Results and discussions
`
`Characterisation of the MDI formulations ini(cid:173)
`tially concentrated on understanding the dis(cid:173)
`persibility of the individual drug suspensions in
`CFC-113 and HFA-134a aerosol propellants, and
`was then extended to examine the combined drug
`system at different drug concentrations and molar
`ratios of drugs.
`Suspension concentration is an important fac(cid:173)
`tor in determining the behaviour of suspensions of
`drugs in liquid suspensions. In the absence of
`shear, the drug particles tend to flocculate, which
`leads to a separation of the solid phase due to the
`relative density difference between the drug and
`the liquid continuous phase. When the suspension
`is subjected to stirring, random collisions of drug
`particles with the walls of the container and with
`each other can result in breakdown of the floc
`structure. Plots of the average aggregate size as a
`function of the suspension concentration of sal(cid:173)
`meterol xinafoate and fluticasone propionate in
`HFA-134a and CFC-113, at a constant stirrer
`speed of 300 rev min- 1, are shown in Fig. 3. In
`all cases, the floc size of the suspensions increased
`with increasing drug concentration. Bower et al.
`(1996) have reported that in a system of constant
`shear stress the disruptive force is unchanged and
`therefore increasing the drug concentration results
`in the equilibrium being shifted towards aggregate
`formation.
`
`MEDA_APTX03505075
`
`PTX0179-00005
`
`5
`
`

`

`170
`
`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`In addition, the dipole moment of HF A is
`known to be greater than that of CFC (Blondino
`and Byron 1998). The greater polarity of HFA
`combined with the presence of hydrogen sub(cid:173)
`stituents allows for considerable charge separation
`in this propellant molecule, owing to the elec(cid:173)
`tronegativity of the fluorine atoms (Blondino and
`Byron 1998). The presence of a layer of oriented
`dipolar molecules at the surface may make a
`significant contribution to the nature of the elec(cid:173)
`trical double layer (Shaw, 1992). This polarisation
`effect induces a temporary dipole on the drug
`molecule without changing the net surface charge
`of the drug. This may lead to an increase in the
`van der Waals force of attraction between drug
`particles and therefore an increased tendency to
`form larger floes in the HF A medium compared
`with in CFC (Fig. 3).
`Fig. 4 shows how the average floc size of the
`drugs varies with increasing stirrer speed in HFA-
`134a and CFC-113. In all cases, the average floc
`
`size of the suspensions initially decreased sharply
`as the stirrer rate was increased, before becoming
`less shear dependent at higher stirrer speeds. For
`example, the floc size of the combined drug dis(cid:173)
`persion of 0.05% wfw salmeterol xinafoate and
`0.07% w/w fluticasone propionate decreased from
`101.2 to 39.5 J.lm in HFA-134a and from 19.8 to
`8.2 J.lm in CFC-113, i.e. an approximately 60%
`reduction of the original size of the floc over the
`range 200-1000 rev min- 1• The shape of these
`profiles is attributed to the structural breakdown
`of floc structure as they are subjected to higher
`shear forces. However, it is noticeable that the
`floes persisted even at the highest shear rate stud(cid:173)
`ied, suggesting that the aggregates are not broken
`down to the size of the input drugs (2-5 J.lm). The
`average floc size of the combined formulation of
`0.05% wfw salmeterol xinafoate and 0.07% w/w
`fluticasone propionate in CFC-113 was greater
`than that of the individual drugs. This was ex(cid:173)
`pected on the basis of the higher drug concentra-
`
`i
`-.5
`1.»
`.~
`(I)
`to>
`
`..s
`
`~
`
`140
`
`120
`
`100
`
`80
`
`(j()
`
`40
`
`2() ~ ~
`
`8
`
`)<
`
`0+-----~----r----,----~-----r----.-----~----.-----r---~
`0.5
`
`()
`
`0.1
`
`0.2
`
`0.3
`
`0.4
`
`Suspension concentration (% w /w)
`
`Fig. 3. Effect of increasing suspension concentration on average floc size of salmeterol xinafoate aggregates in ( ()) HF A-134a and
`(0) CFC-113, and fluticasone propionate aggregates in {D) HFA-134a and ( x) CFC-113. All measurements were made at a stirrer
`rate of 300 rev min- 1•
`
`MEDA_APTX03505076
`
`PTX0179-00006
`
`6
`
`

`

`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`171
`
`140
`
`a
`
`120
`
`100
`
`80
`
`i
`¢
`u
`.~
`fl)
`~ 60
`.s
`~
`
`40
`
`20
`
`0
`
`0
`
`25
`
`b
`
`20
`
`5
`
`200
`
`400
`
`600
`
`800
`
`1000
`
`1200
`
`1400
`
`1600
`
`Stirrer rate (rev min"1)
`
`0
`
`200
`
`400
`
`600
`1000
`800
`Stirrer rate (rev min"1)
`
`1200
`
`1400
`
`1600
`
`Fig. 4. Effect of increasing stirrer rate on the average floc size of ( ()) 0.05% wjw salmeterol xinafoate, (D) 0.07% wjw fiuticasone
`propionate, and (0) 0.05% wjw salmeterol xinafoate/0.07% wjw fiuticasone propionate aggregates in (a) HFA-134a and (b)
`CFC-113.
`
`MEDA_APTX03505077
`
`PTX0179-00007
`
`7
`
`

`

`172
`
`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`(a)
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20 (b)~
`..........
`--~ •..... .• i!J ..........
`
`......
`!
`
`u
`.!:I
`Cll
`u .s
`
`~
`
`•
`
`•
`
`0
`
`0
`
`200
`
`400
`
`800
`1000
`600
`Stirrer rate (rev min"1)
`
`1200
`
`1400
`
`1600
`
`Fig. 5. Average floc size measurement of the combined drug formulation of 0.05% w jw salmeterol xinafoate and 0.07% w/w
`fluticasone propionate as a function of increasing and decreasing stirrer rate in (a) HFA-134a and (b) CFC-113: <>, increasing rate;
`D, decreasing rate.
`
`tion of the combined formulation. In contrast, the
`floc size of the combined formulation in HFA-
`134a appeared to be an average of that of the
`individual drugs, over the entire stirrer speed
`range. Hence, it would appear that the interaction
`between salmeterol xinafoate and fluticasone pro(cid:173)
`pionate in the combined formulation is dependent
`on the physical properties of the solvent. This in
`turn leads to differences in the aggregation be(cid:173)
`haviour of the two drugs. Wang (1992) studied
`the hetero-flocculation of particles of different
`sizes in an apatite-hematite-phlogopite disper(cid:173)
`sion system. The studies showed that particles in
`mixed suspensions were affected by each other's
`stability in a way that the size growth of the less
`stable particles increased the hetero-flocculation
`rate constant and caused the loss of flocculation
`selectivity, especially between the middle and large
`size particles. This indicates that a difference in
`floc size has an effect on the mutual interaction
`between the particles in the dispersion, which can
`
`to
`
`in preference
`
`promote hetero-flocculation
`homo-flocculation.
`In order to determine whether the breakdown
`and re-formulation of the hetero-floc structures is
`a reversible process, the effect of increasing and
`decreasing the stirrer speed on the average floc
`size was investigated. Fig. 5 shows that the aver(cid:173)
`age floc size of the combined drug formulation of
`0.05% wjw salmeterol xinafoate and 0.07% wjw
`fluticasone propionate in CFC-113 was the same
`in both increasing and decreasing shear experi(cid:173)
`ments. In the case of the combined formulation in
`HFA-134a, there was a slight hysteresis effect, but
`most of the floc structure recovered. The interpar(cid:173)
`ticulate binding between salmeterol xinafoate and
`fluticasone propionate is a reversible process.
`However, it is apparent that the drug particles
`had not completely returned to their original state
`after an equilibration period of 5 min. Further
`work performing additional measurements after
`different equilibration
`times could
`therefore
`
`MEDA_APTX03505078
`
`PTX0179-00008
`
`8
`
`

`

`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`173
`
`provide an insight into the flocculation kinetics of
`these systems.
`The composition of the floc structures of single
`and combined formulations of 0.05% wfw salme(cid:173)
`terol xinafoate and 0.07% wfw fluticasone propi(cid:173)
`onate in CFC-113 was determined as a function
`of time (see Fig. 6). Initially, post agitation, the
`suspensions are well dispersed. This is followed by
`macroscopic flocculation, which in turn leads to
`phase separation between the drug particles and
`the dispersion medium. The drug particles migrate
`to the air/liquid surface of the sample (creaming
`phenomena) and, as a result, the drug concentra(cid:173)
`tion at the bottom of the sample decreases over a
`period of time. Although all the plots in Fig. 6
`show a similar characteristic decrease in drug
`concentration with time, the data demonstrate
`that there are significant differences between the
`individual and combined suspension formula(cid:173)
`tions. The densities of salmeterol xinafoate and
`
`fluticasone propionate are approximately 1.2 and
`1.3 g em- 3, respectively. Hence, both the drugs
`cream in CFC-113, which has a density of 1.57 g
`em- 3 at 25°C. Since fluticasone propionate is
`denser than salmeterol xinafoate, it creams less
`rapidly as shown in Fig. 6. It is also interesting to
`note that, in the combined formulation, salme(cid:173)
`terol xinafoate
`and
`fluticasone propionate
`creamed at approximately the same rate. If there
`were no interparticulate interactions between the
`two drugs, then they would be expected to sepa(cid:173)
`rate on creaming as a consequence of their differ(cid:173)
`ent densities. Hence,
`this data suggests
`that
`salmeterol xinafoate and fluticasone propionate
`must be inextricably bound in a hetero-flocculated
`state. The hetero-flocs cream faster than either of
`the individual drugs. This may be attributed to
`the higher solids concentration of the combined
`formulation, giving rise to a larger floc s1ze, as
`shown in Fig. 4.
`
`12 -~ 10
`
`........
`Ill
`~
`"::!?. e.... 8
`li
`-~
`~
`
`l 4.1 l 4.1
`
`~
`4.1
`~
`
`6
`
`4
`
`2
`
`...
`. . ·· .... (b)
`... ···![ .• ...
`
`..... ...
`
`. ..
`
`...... ... . ..
`... ... .. .. .._ .. ... ... ... .. , ...
`
`...
`.• "3:
`
`. ..
`
`. •.• ·:I:
`
`0+-----~--~----~----~--~~--~----,-----~--~~--~--
`4
`6
`8
`10
`0
`2
`Time (min)
`
`Fig. 6. Percentage drug weight measurement from a flocculated suspension at different time intervals for single and combined drug
`formulations in CFC-113: (a) 0.07% w/w fluticasone propionate, (b) 0.05% w/w salmeterol xinafoate, (c) fluticasone propionate from
`a combined formulation of 0.05% w/w salmeterol xinafoate and 0.07% w/w fluticasone propionate, and (d) salmeterol xinafoate from
`a combined formulation of 0.05% w/w salmeterol xinafoate and 0.07% w/w fluticasone propionate. wE/w0 , Percentage ratio of
`experimental weight relative to the original weight of the drug.
`
`MEDA_APTX03505079
`
`PTX0179-00009
`
`9
`
`

`

`174
`
`Y. Michael et al. /International Journal of Pharmaceutics 221 (2001) 165-174
`
`4. Conclusions
`
`The focus of this paper has been to gain an
`understanding of the interaction between salme(cid:173)
`terol xinafoate and fluticasone propionate in sin(cid:173)
`gle and combined formulations. The study
`demonstrates that the formulation behaviour of
`the individual drugs is dependent on the physical
`and chemical properties of both the media and the
`drug substance itself. In CFC-113, the two drugs
`hetero-flocculate in the combination product and
`consequently do not separate under the influence
`of gravity. Furthermore, the interaction between
`the two drugs leads flocculation behaviour that
`differs from that of the individual drugs. This
`interaction again appears to depend on the physi(cid:173)
`cal properties of the medium.
`
`Acknowledgements
`
`The authors would like to thank Dr Mike
`Kaszuba, Malvern Instrumentals Ltd., for his as(cid:173)
`sistance with the experimental determinations of
`the electrophoretic mobilities of the drugs and Dr
`John F. Miller, Inhalation Product Development,
`GlaxoSmithKline, USA, for helpful discussions.
`
`References
`
`Allen, T., Davies, R., 1998. An evaluation of the Lab-Tee 100
`particle size analyser. Proceedings of the Particle Size
`
`Analysis Conference, London, P.J. Lloyd (Ltd), Wiley, pp.
`33-44.
`Blondino, F.E., Byron, P.R., 1998. Surfactant dissolution and
`water solubilization in chlorine-free liquefied gas propel(cid:173)
`lants. Drug Dev. Ind. Pharm. 24, 935 - 945.
`Bower, C., Washington, C., Purewal, T.S., 1996. The effect of
`surfactant and solid phase concentration on drug aggre(cid:173)
`gates in model aerosol propellant suspensions. J. Pharm.
`Pharmacol. 48, 342- 346.
`Chapman, K .R., Ringdal, N ., Backer, V., Palmqvist, M.,
`Saarelainen, S., Briggs, M., 1999. Salmeterol and flutica(cid:173)
`sone propionate (50/250 llg) administered via combination
`Diskus inhaler: as effective as when given via separate
`Diskus inhalers. Can. Respir. J. 6, 45- 51.
`Dijkstra, J.J.F .M ., deNie, L.H., Pot, A., 1996. In-line mea(cid:173)
`surement of crystal size -
`campaign experience with the
`Lasentec M200 system. Int. Sugar J. 98 (1174), 521 - 523 .
`Michael, Y., Chowdhry, B.Z., Ashurst, I.C., Snowden, M.J.,
`Davies-Cutting, C., Gray, S., 2000. The physico-chemical
`properties of salmeterol and fluticasone propionate in dif(cid:173)
`ferent solvent environments. Int. J. Pharm. 200, 279-288.
`Rawle, A., 1999. The Importance of Particle Size Analysis in
`the Pharmaceutical Industry, Company Report. Malvern
`Instruments, Malvern, Worchestershire, UK, pp. 1-44.
`Shaw, D.J., 1992. Introduction to Colloid and Surface Chem(cid:173)
`istry, 4th ed. Butterworths, London.
`Sidhu, B.K., Washington, C., Davis, S.S., Purewal, T.S., 1993.
`Rheology of model aerosol suspensions. J. Pharm. Phar(cid:173)
`macol. 45, 597-600.
`Solvay, 1992. Technical Service Thermodynamic properties of
`Kaltron 113 and Solkane 134a. Solvay Fluor Und Derivate
`GmbH, Hannover.
`Spicer, P.T., Pratsinis, S.E., Raper, J., Amal, R., Bushell, G.,
`Meesters, G., 1998. Effect of shear schedule on particle
`size, density and structure during flocculation in stirring
`tanks. Powder Techno!. 97, 26-34.
`Wang, Q., 1992. Studies on aggregation and heteroaggregation
`in particle separation systems. Acta Polytech. Scand.(cid:173)
`Chem. Techno!. Ser. 203, 1-48.
`
`MEDA_APTX03505080
`
`PTX01