83
`
`Poly( ethylene oxide) (pEG) and
`different molecular weight PEG blends
`monolithic devices for drug release
`
`A. Apicella, B. Cappello', M.A. Del Nobile, M.l. La Rotonda-,
`G. Mensitieri and L. Nicolais
`Department of Materials and ProductIOn Engineering. ~Department of Pharmaceut;c(J{ and TOJ(/coJogicel
`Chemistry University of Naplss Federico 1/ 80131 Naples. Italy
`
`An interpretation of the drug release from monolithic water-swell able and soluble polymer
`tablets IS presented. A convenient palameter, a, which compares the drug-diffuSive conductance
`in the gel layer with the swelling and dissolving characteristics of the un penetrated polymer was
`blseg t9 describe t~e release t;)e~al'iQIH af p tlydmxyett:lyl ttl8Q~hylliR8 (etQfyIliRe) frGm
`compression-moulded tablets of hydrophilic pure semicrystalline poly(ethylene oxides) of mol wt
`600000 and 4000 000 and of two blends of the two molecular weights of poly(ethylene oxides).
`The wa.ter swelling and dissolution characteristics of two polymers and two blends were
`analysed, monitoring the thickness increase of the surface-dissolving layer a.nd the rates of
`waler swellillQ alid peilehalioll ill IIle tablets. Tile dl u9 dirfusiyilies ill lIle watel-pellell aled
`polymer gels were measured by carrying out permeation tests. finally, drug release tests were
`performed to investigate the release kinetics of the different systems in an aqueous environment
`at 37°C. The drug release from the high molecular weight poly(ethylene oxide) is prinCipally
`related to trte..materjal $.weIJLog rather than pOlymer diSSOlution, leading to a progressive
`decrease of the drug s diffUSive conductance In the growing swollen layer, and hence to a non(cid:173)
`constant release induced by the prevailing diffusive control. ConversElly, drug release from the
`low molecular weight poly{ethylene oxide) is strictly related to the polymer dissolution
`mee,laliisrn. Tne aenie~enlent of stational 9 conditions, in ."nien tne Fate of swelling equals tlge
`rate of dissolution, ensures R. constant release rate, even in the case of very low drug-diffusive
`conductance in the external gel layer. Intermediate behaviours were detected in the case of the
`two blends.
`
`Keywords' Pnly(ethy/ene mOdel controlled release dUI9 delivery diffllsion
`
`Received 12 December 1991: accepted 10 June 1992
`
`Diffusiwt-CDlltlOlied release technology, based all po19IileI
`barrier charl:lcteristics, is a gnod alternative to con(cid:173)
`ventional delivery systems. Complex reservoir systems
`and monolithic matriK systems are two important
`applications. For the former, a zero-order kinetic release
`fiay be mamtamed until the orug acuvity ill the reservoll
`can be kept constant. However, environmentally passive
`monolithLc matrix systems, containing dispersed or
`d1ssolvcd active ingredients, arc not able to give a
`constant delivery rate, at least for simple device geometries,
`Tn snr:h r:H~es, the diffusion control gf!nt~rally leads to a
`square root of time dependency of the drug delivery,
`AlLernatively, the use of environmentally interactive
`monohtluc devlces made wnll nydrophihc polymers wa.s
`first proposed by Hopfenberg et 8.1.1. 2, and extensively
`investig:atedJ-B, As the penetrant enlers the drug-entrapping
`matrix, the polymer swells and the active ingredient
`diffuses from the swollen parl. This relaxation-controlled
`
`Correspondence to Dr A. Apicella.
`
`I.E) 19~3 Buil"'Tworth·Heimjllli:lnn.ltd
`0142-9212/931020083-08
`
`SUiptiOlI is govemed by tire solvent COIIceiItlaliuII at the
`interface separating the swollen from the unpenetrated
`polymer. The polymer at the interface relaxes and swells
`at a constant rate as long as the penetrant concentration
`at the moving boundary remains constant. Zero-order
`release from this type of deVIce requires a constant
`surface area and a constant swelling rale of the polymer
`matrix (limiting Case II sorption) as well as a high
`diffusivity of the cntrappcd species. These conditions,
`generally, cannot be protracted for long release times:'!,
`sinCp. they fire strongly rlependFmt on thp. timp. evolution
`of the interactions involving polymer, penetrant and
`solute. It is only in the early stages of polymer swelling
`that, due to the small IhLCkness of the swollen layer, the
`diffusiVE! conductances of both the solvent toward the
`unpenetrated core and the drug moving outside are much
`higher than the swelling rate, and the constant delivery is
`brought about by the constant rate polymer relaxation.
`However, as the swollen thickness mcreases, lower
`values of the diffusive conductance are attained and both
`
`BiomaLerials 1993, Vol. 14 No. Z
`
`SUBJDG-0007605
`
`RBP_TEVA05017784
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`84
`
`PEO devices for drug release: A. Apicella et al.
`
`swellins and release rates BfB progressively reduced.
`Hopfanberg at a1. Z and Pep pas and Franson::! proposed
`similar dimensionless parameters to identify zero-order
`release conditions in terms of initial solvent penetration
`late, swelling thickness, and diffusivity of the active
`ingredient in the Bwollen polymer. To guarantee zero-
`order releases from swelling systems of increasing
`thicknesses, the drug should present the proper high
`diffusivity in the swollen polymer. The addition of water-
`soluble components lD the polymer fias been proposed4to
`enhance the diffusivity of the drug in the more parous
`swollen layers created.
`Poly(ethylene oxide] (PEO]-based systems have been
`proposed as drug delivery devices. In particular, Graham
`et 8.1. reported the constant release of prostaglandin Ez
`from crass-linked crystalline-rubbery hydrogel matrices
`based on PEO which were undergoing solvent sorption
`and from crystalhtesmeItmg In aqueous eDvlronments"~-ll.
`In the case of uncross-linkedPEO matrices, the solubility
`of the polymer can aiter the characteristics of the
`penetrated layer, leading to different behaviours in
`systems presenting different dissolution features. To
`control the release of the active agent, there should be a
`balance between diffusion of the active agent and
`solubilization of the polymer matrix. The diffusivity of
`the drug through the matrix. the swelling of the polymer,
`and its solubilization rate can be bjased by changing the
`molecular weight of the poLymer or blending polymer
`fractions with different molecular weights.
`Mucoadhesive capability is another important property
`of a polymer to be used as matrix for monolithic drug
`III faGt, in the dev 819pmsIlt 9f 91'al
`rieliuery G9JT.:i88S
`controlled release devices, considerable benefit may
`result from the use of bioadhesive polymers providing
`relativelJ": short-term a.dhesion between the drug deliver»:
`system and the epithelial surface of the gastrointestinal
`tract. Due to the linear flexible structure of the PEO
`macromolecule, this polymer shows a particular ability
`to form entangled physical bonds by interpenetrating
`deeply ud rapidly ipto mucous !mbs*r;ttuJP networks
`The mucoadhesive properties of PEO reported in the
`literature ere strongly dependent on the polymer molecular
`weight and are more pronounced in the case of the high
`14
`molecular weight materials12-
`• In particular PEO shows
`a behavioul 14 langiuk flom no bioad-hesion at mol wt
`20 000 to very good bioadhesion at mol wt 4 000 000.
`Consequently, both the release and the mucoadhesive
`properties of PED-based systems are expected to be
`finely tuned by blending PED fractions with different
`molecular weIghts,
`We were therefore interested in the analysis of the
`capability of PEO to be used as mucoa.dhesive control
`release devices. In Ihe present investigation, we relate
`the sorption. melting and disflolution characteristics of
`two PEds of different molecular weight to the dlfluslvltiea
`in the swollen polymer of an active ingredient (etofylline]
`and the release behaviour observed.
`
`RELEASE MECHANISMS FROM
`MONOLITHIC DEVICES
`
`The release rate of a dissolved or dispersed drug from a
`polymeric film or tablet introduced in a specific environ-
`
`Biomllleria1s 1993, Vol. 14 No.2
`
`ment, strongly depends on the nature of the diffusion and
`sorption processes involving the polymer/environment
`system and the polymer/drug system.
`
`DiHusion-control1od deVices
`The dissolved species will diffuse from a matrix which
`does not actively interact with the external environment
`according to an ordinary diffusion law [Figure 1a). [n
`such a caae, the concentratioll plofile ill tbe slab
`decreases with time leading to the progressive reduction
`of the release rate (Le. the slope of the fractional release
`versus lime curve).
`
`Swelling-controlled device.
`Completely different release behaviour is observed for
`h~,..:Il'ophilic polymers ~~[bell Inlier IiOrpt:iOD i5 £DIID~lI1€!d
`by significant polymer swelling. Limiting Case II sorption
`occurs when constant rate water absorption is associated
`with a front advancing at a constant rate into the confines
`of the glassy polymer. A sharp boundary separates the
`esseBti;::tlI~ l:I:H:fItloehattld ~tlre from 1:10 tlnHt:Jrmb 8 ~~oHen
`shell (Figure lb). The polymer relaxation and swelling Is
`driven by the osmotic stresses generated at the moving
`boundary by the presence of the penetr8nt"~-17 and
`remains constant as long as a constant local concentration
`persists. Ellag release is controlled qUantItatIvely by the
`invasion of the swelling sol vent and by the solute counter
`diffusion in the swollen polymer. Zero-order release
`kinetics may be achieved from a polymer which swells at
`a constant rate and with a constant penetration surface
`area, but only H the counter dlffuslOn of the solute
`molecules is rapid compared with the swelling rate.
`
`Intermnediate cases
`There are intermediate cases in which both the swelling
`and the diffusive control can be important during drug
`release. The swelling front advancing rate (v) and the
`dtffusl·.e eef.latletanee (the rat"it) bet", 8en the solute
`diffusivity and the shell thickness at a given time DI5(t))
`have been used.2 to define a convenient dimensionless
`parameter:
`a VlIDI') X vJ
`which accounts for the relative contribution of the solute
`counter diffusion and of the penetrant uptake rate to the
`overall rate of release.
`In the early stages of swelling. the diffusive con-
`ductance is hiah, due to the small value of the swollen
`shell thickness (,sit)). Release is controlled by the
`polymer swelling rate (Figure 1) and values of a :> 1 are
`measured. Conversely, the diffusion of the flolute
`molecules through the outer shell will increaSingly
`control the release kinetic observed as the swollen layer
`progressivp.ly thickP.ns" In this casp., values of a srnallp.r
`than unity are observed. Diffusive control is reflected by
`the fractIonal release versus time curve as a progrESSive
`reduction of the release rate (Figure 1b). For low values
`of a, the polymer rapidly swells. after which the drug is
`depleted by an exclusively diffusive mechanisrnz. As a
`conseqUEnce. a zero~order release rate is expected when
`a(t) > 1 and the penetration rate of the swelling agent is
`constant.
`
`SUBJDG-0007606
`
`
`
`RBP_TEVA05017785
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`PEO devices for drug release: A. Apicella at all.
`
`Fickian diffusion
`
`Release from monolithic devices
`
`Limiting Case 11 sorption
`Penetrating
`front
`
`Co
`
`ix = D/(6(t)-v)
`Swelling rate
`[constant)
`v =cmfs
`Diffusion conductance
`(variable)
`Difilti =cmIs
`
`Release rate
`decreases
`
`LIZ
`
`Penetrated
`""'V”‘°'
`
`Diffusion
`control
`a (I
`
`Tlme
`
`Time
`
`Figure 1 Release mechanisms from monolithic devices. a, Fickian diffusion; b, limiting Case ll sorption.
`
`A theoretical framework of penetrant uptake and
`solute release was developed by Peppas et .-113'” and by
`Davidson and Peppes” 2°
`to characterize monolithic
`swellable systems for controlled drug release. It has been
`shown that
`two dimensionless parameters should be
`used to predict the release behaviour of these systems.
`The swelling interface number. Sw (1/:1 in the present
`context), and the diffusional Deborah number, De, have
`been introduced. The former represents the ratio of the
`penetrant uptake rate to the rate of solute diffusion. The
`latter represents the ratio of the characteristic swelling
`time of polymer chains, related to the presence of the
`swelling penetrant, to the characteristic diffusion time of
`the penetrant into the polymer. Zero-order release rates
`should be expected if both the solute diffusion through
`the swollen polymer layer is rapid compared to the
`penetrant uptake rate (Sw << 1] and the pengtrant uptake
`is controlled by polymer relaxation [De = 1]. As con-
`sequence the solute release kinetics cannot be uniquely
`related to a (or 1/SW) and also the value of De has to be
`taken into account.
`
`Polymer swelling and dissolution-controlled
`devices
`
`Thermoplastic polymers which are sufficiently hydro-
`philic are also water soluble. A sharp advancing front
`divides the unpenetrated core from a swollen and
`dissolving shell. Under stationary conditions, a constant
`thickness surface layer [5] is formed by the swollen
`polymer and by a high concentration polymer solution“.
`In fact, once the hydrodynamic external conditions
`are defined, a stationary state is reached where the rate of
`penetration of the moving boundary {V} equals the rate of
`removal of the polymer at the external surface. The time
`lapse until the quasi stationary state is reached is swelling
`time“.
`Figure 2 reports the typical polymer concentration in
`
`Penetrating
`front
`
`Dlssolvingpolymer
`
`
`
`Unpenetrateclpolymer
`
`When:
`
`Swelling rate = dissolving
`rate
`
`Git) = constant
`
`Figua-e2 Polymer concentration profile for swelling and
`dissolving materials.
`
`the
`If
`the surface layer of u dissolving polymer.
`dissolution occurs normally. the steady-state surface
`layer consists of
`four different sublayersz‘:
`liquid
`sublayer[adjacent to the pure solvent). gel sublayer, solid
`swollen sublayer and infiltration sublayer (adjacent to
`the polymer base into which the solvent has not yet
`migrated). If the test temperature is higher than the glass
`transition temperature of the polymer, the surface layer
`consists of only liquid and gel sublayers.
`At steady state, the dissolution rate is constant and can
`be defined equally by either the velocity of the retracting
`front of the polymer or the velocity of the front
`separating the pure penetrant and the liquid dissolving
`sublayer. Thus both fronts are synchronized.
`The dissolution rate strongly depends on hydrodynamic
`conditions, temperature, polymer molecular weight and
`crystallinity level. Close analogies have been found
`between crystallization and dissolution behaviours for
`semicrystalline polymers“.
`In fact, dissolution and
`crystallization rates are maximal when plotted as
`
`Biomaterials 1993, Vol. 14 No. 2
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`
`
`RBP_TEVA05017786
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`86
`
`PEO devices for drug release: A. Ap{cefla et al.
`
`function of temperature and are both expected to
`decrease with increasing polymer molecular weight.
`In the case of a dissolving polymer, a dimensionless
`parameter aCt} can still be defined as the ratio of the
`diffusive conductance DJb (t) and dlSsoluhon (or penetra(cid:173)
`tion) rate v. Consistent with the previous discussion
`concerning swellable polymers, a zero-order release rale
`is to be expected only if a > 1 and the dissolution rate is
`constant with time. Two different dissolution stages can
`be identified: the initial time lapse (swelling time) and the
`steady-state conditions. During the initial transient,
`neither the thickness of the dissolving surface layer or
`the dissolution rate are yet constant. A lime-dependent
`diffusive conductance (Dlolt)) and a time dependent
`dissolution rate (v(t}) should be considered for the
`evaluation of the parameter a (t). Conversely, under
`steady-state conditionR, a comtanl diffmive cmnrluctaor:p.
`(constant gel layer thickness) and a constant dissolution
`rate are attained. As a consequence, in the first stage a
`changes during Ihe dissolution; release process and
`release rate can be a function of time. During the second
`stage however a time-independent concentration profile
`develops into the external surface layer. A constant
`release rate is obtained. determined by the penetrating
`front rate or equally by the rate of advancement of the
`liquid sublayer-pure solvent boundary, synchronized to
`the J3eIletrating Hant. 1' ... similar e:fsmple ef zere order
`release kinetics resulting from the synchronization of
`front velocities of the identical velocities of diffusing and
`eroding fronts in erodible polymer matrix delivery
`systems12
`•
`
`MATERIALS AND ME'tHODS
`
`Materia1s
`PEo of average mol wt of 600 000 (Aldrich Chimica S.r.!.,
`Catalogue No. 18,£82 8) and 4 Baa Baa ({Uorieb Ghimiea
`S.r.l., Catalogue No. 18,946-4) were used. Analytical
`grade
`(purity 99.6%) p-hydroxyethyl-theophyiline
`(etofylline) was supplied by Sigma Chemical. The
`materials were used as received.
`
`Melhod.
`
`Thblet re sration
`The polymer and the etofyllinc powders were first
`desiccated under vacuum, next mixed in the desired
`proportions, then dissolved in chloroform. The solution
`was stirred well to assure a homogeneous mixing of the
`components. Polymer films containing the drug were
`obtained by casting. After desiccation, several film
`layers were then compression moulded at 75t)C to form
`sheets from which were cut circular tablets (25 mm
`diameter) with thicknesses of 2.0 or 3.3 mm. Four
`different kinds of tablets were produced. all containing
`10% etofylline. The four different polymer matrices used
`were: pure PEO (mol wt = 600000), pure PEO (mol
`wt = 4000 DOD), 50% h.w. blend of the two PEDs and
`87% b. w. mol wt of 4 000 006 and 13% b. w. mol WI of
`600000 blend of the two PEGs (50% mol blend).
`
`Biomaterials 1993. vol. 14 NO.2
`
`Calorimetric analysis
`In order to characlerize PEDs amI PEOs/etu[ylLine
`mixtures, 8 Du Pont differential scanning calorimeter
`(DSe instrument 910J operating under nitrogen flux and
`at a beating rate of lO°C/mm was used.
`
`Gel layer thickness meEisurement
`Water swelling and penetration depths, and gel layer
`evolution were OptiCBJIy measured during the water
`conditioning at fixed times. The samples were cleaved on
`glass plates and placed in thermostated distilled water
`held at 37°C. The water was continuously stirred and
`measurements were taken using a cathetometer.
`
`Etofy/line permeBtion tests through the swollen
`polymers
`An apPIH'flhlfl F!guipJlF!rl with a cell for liquid permeation
`measurements in membranes was used for the permeability
`measurements [absorption simulator made by Sartorius
`AG). The polymers. free of drug, were confined in the Gell
`between two semipennsable cellulose acetate membranes
`and equilibrated witb distilled water at 37"C before the
`start of the permeability measurements. The equilibrium
`water swelling thicknesses of the four different polymeric
`materials confined in the cell were measured and used
`in the calculation of the permeability values. The con(cid:173)
`G811tratiGB iBGreaS8 Qf the etefyllia8 in the eis'w'nstream
`chamber of the permeability cell was monitored. The
`amount of drug passing through Ihe polymer was then
`evaluated as function of time.
`Reference permeation tests were performed to detennine
`the influence of the buppmLing cellulose acetate ltIefil-
`branes on the druB permeation kinetics. The resistance to
`the drug transport due to the supporting membranes was
`found to be negligible comp .. e<l to the "aiu., llot.,t.<I
`during Ihe permeation tests performed on swollen
`polymers. The drug concentration into the swollen
`polymer on the upstream side was evaluated from the
`permeation tests by means of the calculated water(cid:173)
`snollen pO!llmer etofJlline partition coe!fficienl. 'Fhe
`resulting concentration was approximately ten times
`lower than the tablet drug loading. Accordingly, a.
`different drug diffusion rate could occur in the tablet
`dissolVing layer. Nevertheless, a drug com:::entration
`lower than the drug loadmg IS to be expected In the
`dissolving layer of Ihe tablet, due to the swelling, For this
`reason, the diffusion constant evaluated by means of the
`permeation tests was used in the evaluation of 1:11e
`dimensionless numbers introduced in the previous
`sections. Eventua composition epen ence 0 eto y ine
`diffusivity through the surface tablet layer not measured.
`
`Drug release kinetics analysis
`A dissolution apparatus fErweka D.T.) operating at
`50 rev min 1 and at a constant temperature was used for
`the evaluation of the etofylline release kinetics from the
`polymer tablets. Each tablet was placed in a container
`filled with a known amount of distilled water and the
`increasing concentration of etofylline in the aqueous
`conditioning environment was measured as function of
`time.
`The etofyl1inp. concentrations in the aqueous solutions
`during the permeability and the release tests were
`detennined by means of UV spectroscopy using a
`
`SUBJDG-0007608
`
`
`
`RBP_TEVA05017787
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`PEO devices for drug release: A. Apicella at al.
`
`87
`
`E
`oS
`" " c
`
`~
`u
`i'
`t-
`
`2.0
`
`1.5
`
`0.5
`
`•
`
`Beckman DU-40 spectrometer operating at 262 run. All
`test.'! wel'p. carried out at 37°C.
`
`Drug release system charat::terization
`To characterize properly the drug release from the
`
`penetration, polymer swelling and dissolution behaviour,
`drug solub1l1ty in the host polymer and diffusion in the
`swollen and gelled layer were examined.
`
`Differential scanning calorimetry {DSC}
`The DSC thermograms of etofylline, PEO and of tablets
`made of the four different PEQ matrices containing 10%
`etefylline .. ElFe taken The thEll'tllQSfam relative tg the
`etofylline shows a well-defined melting peak around
`170°C while that of the semicrystalline pure PEOs and
`blends present a melting peak in the range 63-65°C, Only
`thF! mA1ting pAR.k of thf! polymer matrix is evident in the
`theIlliogldlII of tablets made of PEGs blended wUb
`etofylline. The absence of the melting peak of the
`etofylline indicates complete dissolution of the drug in
`the amorphous regions of the polymer. The crystalline
`fraction of the tablet matrices was about 0.65. This value
`a!';summg an en a pya crys a lza Ion
`WRf; eVil llR. F!
`of PEO equal to -210 Jig.
`
`Polymer swelling and dissollltion properties
`The photo reproduced in Figure 3 shows the initial
`swelling conditions observed for the two homopnlymf!r
`tablets of different molecular weights. Figures 4 and 5
`report the penetration depth evolution kinetics in the
`case 0
`average rna
`e a e s rna e 0 pure
`s WI
`wt of 600 000 and 4 000 000 containing 10% etofylline.
`The PEG of mol wt of 800 000 (Figure 4) shows an almost
`linear shape of the penetration depth curve as a function
`of time which indicates that. after a very short initial
`transient, the penetration front moves into the pulym~r al
`a constant rate. The higher molecular weight PEO tablet
`(Figure 5) is characterized by an initial rapid water
`penetration rate. A sharp front moves at a constant rate
`through the unnenetrated tablet core after a swelling time
`signiflcantly larger than the time lapse detected for lower
`
`Figure 3 Water swelling of PEO tabs of mol wt of 600 000 (left)
`and 4 000 000 (right).
`
`,
`--------- ;-----
`
`/
`
`/
`
`/
`
`199
`
`2{lQ
`
`39g
`
`Tl:ne (min)
`
`Figure 4 PenetratlGn depth kinetics of water at 37"C in mol wt
`of 600 000 PEO: the horizontal dotted line represents the time
`corresponding to the total penetration in the case of a 3.3 mm
`thick tablet. v = 1.3 X 10-4 mm/s.
`
`E
`oS
`• • 0
`~
`0
`:i'
`
`1.2
`
`1.0
`
`0.2
`
`0
`
`300
`200
`Time (min)
`
`'00
`
`soo
`
`Figure 5 Penetration depth kinetiCS of water at 37 Q C in mol wt
`of 4 000 000 PEO: the horizontal dotted line represents Ihellme
`corresponding to the total penetration In the case of a 2 mm
`thick tablet. v = 5.0 X 10-5 mm/s.
`
`molecular weight PEG. The behaviour of PED blends is
`similar to that of pure PEU with a mol wt 01 4000 000.
`The corresponding values of the steady-state penetration
`rates far the four kinds of polymer matrices, VS, are
`reported in Thble 1. As
`v decreases with
`
`rate is
`weight
`Hence,
`
`X 10'.
`was
`
`When the dissolution rate equals the penetration rate,
`a constant thickness surface layer should be observed.
`The dissolving layer evolution during water conditioning
`should reflect the different dissolution charactsristics of
`the materials. The surface layer thicknesses as a function
`of lime are compared in Figure 8. It is evident that a
`
`Diomaterials 1993, Vol. 14 No.2
`
`SUBJDG-0007609
`
`
`
`RBP_TEVA05017788
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`88
`
`PEO devices for drug release: A. Apicella et aJ.
`
`Table 1 Swelling and diffusion parameters of PEDs
`
`Polymer
`
`v(cm/s)
`
`Gel layer (em)
`
`D (cm 2Js)
`
`pea 600 000
`PEa blend (50% b.w.)
`PEa blend (87% b.w.)
`PEO 4000 000
`
`130XlO-~
`7.0 X 1(}~
`7.6 X 1O-B
`6.0 X 10-"
`
`01
`0.1-0.5
`0.1-0.5
`0.1 0.5
`
`70XlO-~
`100.0 X 10 8
`150.0 x 10-8
`140.0 X 10-8
`
`(± 20 x 10-8)
`(± 20 X 10-8)
`(± 20 x lO-a}
`
`E .s
`•
`• • c
`u :c
`
`f-
`
`"
`
`3
`
`•
`
`•
`
`0
`
`, ..
`
`a
`
`0
`
`200
`
`300
`
`Time (min)
`
`Figure 6 Gel layer development kinetic at 3rC in water_
`penetrated PEa and PEO blends tablets. 0, PEa mol wt
`4 000 000;., PEa blend: 87% b.w. 4 000 000: e, PEa blend:
`500/0 b.w. 600000; 0, PEO mol wt 600 000.
`
`stational) thickness (about t liUll) is IeadHs leached Oldy
`in the low molecular weight polymer, In the case of the
`high molecular weight PED and of the two blends, it
`al::!ruf)Uy NlIlGHS an initial "lime (1 mm} in tile Her}' €:ill'ly
`stages of water sorption, then increases continuQusly (up
`to 5 mm in our test). It can be concluded that. in the case
`of the chosen tablet thicknesses, a steady-state dissolution
`rate (constant 5) is detected only for PEO with an average
`lIlol wt of 666 600. The aliJa llil6P. IYi'M of illBli ices do
`not reach the quasi stationary dissolution stage. A zero(cid:173)
`order release rate is thus likely to be expected only in the
`case of tablets made of PEQ with average mol wt of
`600000. Conversely, the other three matrices should
`snow more complex release oenavlOur.
`
`Drug diffusivities in the polymer gel
`Permeation tests of etofylline in water-swollen PEO
`were performed to evaluate the time lag. FiSl.U'e 7 shows
`the drug amount released in the downstream chamber tor
`tests made at 37°C for the hlgh mol wt (4 000 ODD) PEO.
`The time lag is defined by the intercept on the time axis of
`the stationary part of the permeability curve. Assuming
`that the first and second Fick's law for diffusive
`transport hold true, the time lag is equal to L 2/6 D where
`L is the equilibrium swelling thickness of the polymer
`sample and D is the etofylline diffusivity, From the time
`lag D can De reaOllY evaluateu. Ine values oDtamea are
`reported in '!able 1,
`
`o
`
`0.04
`
`1),03
`
`• 0 -
`~
`'" •
`
`0.01
`
`100
`
`300
`
`"00
`
`Figure 7 Etofylline permeation test in distilled water swollen
`mol wt of4000000 PEO. D = L2/(6 x n = 1.4 X 10-6 crWls.
`t* = L 2/6 O.
`
`PEO tablets show an almost constant release rate over
`the entire period of the test [Figure 8). Under stationary
`swelling and dissolution conditions, a constant thickness
`dissolving surface layer, possessing a time-independent
`drug distribution profile, moves toward the inner part of
`the tablet. The constant delivery rate is determJncd by the
`constant penetration [or dissolution) rate.
`More complex behaviour has been observed for thp.
`
`1.0
`
`~. v
`
`0.8
`
`0.6
`
`0
`~
`
`\
`I
`
`f/
`i / 0. "
`
`". ,
`~
`
`100
`
`200
`
`'00
`
`"00
`
`Time/I {minfmml
`
`Drug release rate
`As already pointed out a zero-order drug release is to be
`expected in the case of the tablets made of PED with
`average mol wt of 600 000, The lower molecular weight
`
`Figure8 Release kinetics of etotylline at 37°C in distilled
`water from mol 'Nt of 600 000 and 4000 000 pEa tablets
`containing 10% by weight etofyiline. D. mol wt 600 000; 0, mol
`wt 4 000 000.
`
`Biomaterials 199:3. Vol. 14 No.2
`
`SUBJDG-0007610
`
`
`
`RBP_TEVA05017789
`
`TEVA EXHIBIT 1028
`TEVA PHARMACEUTICALS USA, INC. V. MONOSOL RX, LLC
`
`

`
`PEO devices for drug release: A. Apicella et al.
`
`89
`
`other three matrices adopted in the present investigation.
`In the case of tablr:ts prcparr:d with thr: higher molecular
`weight polymer (4000000), the value of a is likely to
`change with time due to both the continuous increase of
`the thIckness (0) of the dIssolVIng surface layer andto the
`decrease of the dissolving rate in the initial transient. In
`fact, a(t) evaluated from experimental da.ta ranges
`initially from values> 1 to values < 1 in the final stages.
`The development of drug diffusive resistance in the
`dissolving swollen layer becomes signihcant only when
`a thickness> 2.5-3.0 mm is formed. As a consequence,
`the drug release is governed by the dissolution rate during
`the initial stages of release, then controlled by the solute
`diffusion through the swollen dissolvinslayer in the final
`stage. The release kinetics from a tablet made of PEO
`with an average mol wt of 4000000 is also reported in
`Figure 8. A roughly constant release rate stage is
`followed by a decreasing rate delivery.
`Release tests performed on tablets made of blends of
`the two PEOs are reported in Figure g. In both cases, the
`values o( a ranges during the test from values> 1 to
`value); < 1. A hf!havlour similar to the case of pure PED
`with average mol wt of 4000000 should hence be
`expected. However, an almost constant release, very
`dose to the beha.viour of tablets made of pure PED with
`average mol wl of 600000 was detected in the case of the
`blend containing the lower amollnt of high molecular
`weight PEQ (5U% b.w.). A release behaviour close to that
`of the tablets made nf PEa with an average mol wt of
`4000000 was observed only in the case of the blend
`containing 87% by weight of the high molecular weight
`fraetiem. For tablets made o£ PEa blends, the pHameter fl
`does not provide an interpretation of the data which
`cannot be uniquely correlated with the parameter
`valueS
`
`CONCLUSIONS
`
`Our results show that it is possible to obtain different
`release aeaaviBuFs b~j l;;Rauging lhe mgLeswar weight 9f
`
`1.0
`
`I ~
`
`0.8 1/
`0.' j
`
`0.'
`
`0
`0
`
`;0
`
`p
`
`0.2 '{
`
`o
`
`100
`
`ZOO
`TimeJI (min/mm)
`
`300
`
`figure 9 Release kinetics of etofylline at 37"C in distilled
`water from 50% b.w .• mol wt of 4000 000 and 87% b.w. mol wi
`sf 4009 ggg P~Q slaRes tablets sSFItaiAlng 19% by '""slOAt
`etofylllne.:1, mol w14 000 000 (87% b.w.); 0, mol wt 4 000 000
`(50% b.w.).
`
`the soluble polymeric matrix-forming monolithic: drug
`delivery devices. The low mol wt (600 000) PED matrix
`showed a nearly constant drug delivery rate due to the
`synchronization of penetrating and dissolving fronts.
`N on~constant release behavIOur has been detected m the
`case of the high mol wt (4000 000] PEQ matrix. The
`tablets made of a PEO blend containing 87% by weight of
`the high molecular weight fraction showed behaviour
`similar to the tablets made of pure PED with mol wt of
`4000000. The best compromise between release and
`potential mucoadhesive properties was obtained from
`tablets made of 50% by weight blends of the two PEO
`fractions. The nearly constant release rate is coupled
`wiLh good polential bioadhesive properties imparted by
`the high molecular weight fraction.
`
`REFERENCES
`
`1
`
`2
`
`3
`
`4
`
`fj
`
`6
`
`7
`
`8
`
`9
`
`11
`
`13
`
`Hopfl3obsrg, H.R. ~ml HlllI. K.r. .. RwelLins-controlled,
`constant rate deli.ery s.ystems, PolJm. Eng. Sci. 1978, 18,
`1186
`Hopfenberg, H.B., Apicella. A. and Saleeby. D.E.,
`Sorption affecting wattll" sorptiuJl in auu solute release
`from glassy
`ethylene· vinyl
`alcohol
`copolymers,
`r Membrane Sci 1981 8 273
`Peppas, N.A. and Franson, M.N., The swelling interfHGt:
`number as a criterion for prediction of diffusional solute
`relea.se mechaniBms in