0 1991 J. Pharm. Pharmacol.
`J. Pharm. Pharrnacol. 1991,43: 382-387
`Received September 27, 1990
`Evaluation of Release from Selected Thermosoftening Vehicles
`
`MICHAEL KOPCHA, *NICHOLAS G . LORDI A N D ~ K A K U J I I . TOJO
`
`Schering-Plough Research, Kenilworth. NJ 07033. USA, *Rutgers. The State University, College of Pharmacy, Piscataway,
`NJ 08854. USA and, tKyushu Institute of Technology. College of Computer Science and Systems Engineering, Lizuka, Fukuoka 820, Japan
`
`of D&C Yellow No. 10 and anhydrous theophylline have been determined from a
`Abstract-Release
`thermosoftening, hydrophilic matrix, Gelucire 50/ 13, incorporating a water-soluble additive, polyethylene
`glycol 4000. As additive level increased, release also increased. The effect of mixtures of Gelucire 50/13
`(G50/13) and Gelucire 50/02 (G50/02) on release was also investigated as a function of temperature and pH.
`As the level of G50/02 increased, release decreased and became predominantly diffusional. As temperature
`was increased, release changed from diffusion to a mixed model of both diffusion and erosion. At basic pH,
`release from these composite systems became more erosional in character, possibly reflecting partial
`hydrolysis of the ester-linked matrices. Diffusion coefficients and apparent diffusion coefficients were
`calculated in G50/02 and G50/13 matrices, respectively, and were in agreement with published data.
`
`Gelucires are inert materials derived from hydrogenated
`food-grade oils and fats, which have been developed to melt
`within specified ranges, as represented by the first number in
`their designation, and to have predetermined fat soluble or
`water dispersible characteristics, designated by their hydro-
`philic-lipophilic balance (HLB) as represented by the second
`number in their designation. Bowtle (1986) studied the
`release of four deliquescent drugs from the Gelucire class of
`excipients. Gelucires with high HLB values gave the most
`favourable release irrespective of their associated melting
`points. Thakkar et al (1987) studied the release of an
`antibiotic from G50/13, G48/09 and G46/07. Release from
`G50/13 and G48/09 was similar to that seen with a conven-
`tional powder-filled capsule. Release of active drug from
`G46/07 was not complete but the addition of hydrophilic
`additives enhanced overall release.
`Dennis & Kellaway (1987) studied the release of ketopro-
`fen from a slowly hydrating Gelucire matrix which had an
`HLB of 4.8 and a melting point of 50°C. Results suggested
`that even though release could be approximated by a square-
`root of time relationship, dissolved drug led to altered
`physical states of the matrix which showed deviations from
`the underlying assumptions surrounding strict matrix diffu-
`sion. Hence, a more appropriate model needed to be
`proposed to account for the observed behaviour.
`To address this issue, Kopcha et a1 (1990) devised a series
`of schemes to explain drug/marker release from thermosof-
`tening materials which included matrix hydration and
`erosion phenomena in addition to diffusional release. The
`schemes developed were used in this study t o evaluate the
`effect of incorporating a water-soluble excipient on release of
`D&C Yellow No. 10 (the marker dye) and anhydrous
`theophylline from G50/13. They were also used to investigate
`the effect of mixing two thermosoftening materials, G50/02
`and G50/13, which have similar melting ranges but divergent
`hydrophilic-lipophilic balances, on overall drug/marker dye
`release.
`
`Correspondence: M. Kopcha, Schering-Plough Research, Kenil-
`worth, NJ 07033, USA.
`
`Materials and Methods
`
`Materials
`All chemicals were stored over a desiccant of silica gel at a
`temperature of 22"C, and were purchased as follows:
`Gelucire 50/02 and 5Ojl3 (Gattefosse' Corporation, NY),
`D&C Yellow No. 10 (Warner Jenkinson, MO), anhydrous
`theophylline USP (Amend Drug and Chemical Company,
`NJ), and polyethylene glycol 4000 (Amend Drug and
`Chemical Company, NJ). The dissolution medium was either
`a simulated gastric fluid (pH 1.2) consisting of 2 g NaCl and 7
`mL conc HCI made up to 1 L with distilled water, or a
`simulated intestinal fluid (pH 8.1 1) consisting of 0.34 g of
`KH2P04 and 9.12 g of Na2HP04 made up to I L with distilled
`water.
`
`Release methodology
`The method and operating conditions used to monitor
`release from a stationary disc/rotating fluid system were as
`described previously (Kopcha et al 1990) for both conti-
`nuous and discrete sampling.
`Experiments were performed at a paddle speed of 50 rev
`min-' with a paddle height of 1 cm and temperature
`maintained at 30, 37 or 42°C. Results are reported as the
`average of six replicates f standard error (k s.e.).
`
`Preparation of stationary discs
`Ten g of the appropriate Gelucire was heated to 10°C above
`its melting point and the additional Gelucire (27, 50,73% w/
`w) or polyethylene glycol 4000 (10,30,50,75,100% w/w) was
`incorporated into the molten mass using a high speed, desk-
`top homogenizer. The drug/marker dye was added lastly
`under continuous high shear. The molten material was
`poured into the female half of a 6.98 cm2 or, for the
`polyethylene glycol 4000-containing system, 1.54 cm2 Milli-
`pore filter holder. The her-lock tip was sealed with a
`threaded screw to prevent leakage of the molten mass. The
`molten material was poured in three stages into the holder to
`prevent cracking on cooling. When the mass had completely
`congealed, the surface was levelled with a hot spatula.
`Prepared discs were stored in a desiccator at ambient
`temperature for up to 24 h.
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`

`

`EVALUATION OF RELEASE FROM THERMOSOFTENING VEHICLES
`
`383
`
`Differential scanning thermoanalysis
`Differential scanning calorimetry (DSC) (Model DSC-7,
`Perkin-Elmer Corporation, CT) was carried out on 1-10 mg
`of Gelucire preparations. Each sample was weighed into a
`DSC pan to the nearest 0.1 mg and the cover crimped into
`place. An empty covered sample pan was used as the
`reference. Each sample was heated to 100°C at a rate of 20°C
`min-I. The sample was then rapidly cooled to a final
`temperature of - 10°C and maintained there for 5 min. Each
`sample was reheated to 100°C to provide a second thermo-
`gram.
`
`Results and Discussion
`
`Excipient effects
`During early experiments it was noted that polyethylene
`glycol 4000 (PEG 4000) would form emulsions with molten
`Gelucires of low HLB, therefore, it was decided to mix this
`excipient with a high HLB material, G50/13.
`A representative DSC thermogram for the composite
`system of G50/13, 30% w/w PEG 4000 and 2.5% w/w D&C
`Yellow No. 10 is shown in Fig. 1. Similar thermograms were
`noted for all comparable systems irrespective of the drug/
`marker dye employed. To allow for easy visualization of the
`endothermic peaks, the curves were not normalized to
`weight.
`The thermograms show two distinct peaks, the lower
`representing the melting of G50/ 13, and the higher represent-
`ing the melting of PEG 4000. These scans indicate that G50/
`13 and PEG 4000 do not form a single, homogenous phase;
`PEG 4000 remains as a discrete molecular entity. Therefore,
`the mixture can be viewed as a dispersion of one substance in
`the other. For PEG 4000 there was a shift in the peak
`maximum from 66 to 54°C. However, as the percentage of
`PEG 4000 increased, the melting point re-established itself
`closer to that of pure PEG 4000.
`Figs 2 and 3 show the effect of increasing levels of PEG
`4000 on the release of the marker dye and anhydrous
`theophylline from G50/13. A general overview shows that as
`the level of PEG 4000 increased, release from the matrix also
`increased, regardless of drug/marker dye used.
`Table 1 shows the model coefficients (Kopcha et al 1990)
`for the release of the marker dye and anhydrous theophyl-
`line, from a stationary disc/rotating fluid system, as a
`
`30
`
`22.5
`s'
`E
`3 15
`F
`c
`m
`I
`7.5
`
`0
`
`20
`
`40
`60
`Temperature (O C)
`FIG. 1. Thermogram of G50jI 3 with a mixture of 2.5%) D&C Yellow
`No. 10 and 30% PEG 4000.
`
`80
`
`100
`
`l o (
`
`'+
`
`.;
`
`5
`
`0
`
`1
`
`3
`
`4
`
`2
`Time (h)
`FIG. 2. Effect ofincreasing levels ofPEG 4000 (O%,
`lo%, 0; 30%,
`0; 50%. 0 ; 750/, x ; loo%, A) on release of D&C Yellow No. 10
`(2.5%) from G50/13.
`
`7 I
`
`0
`
`1
`
`2
`
`3
`
`4
`
`.;
`
`5
`
`Time (h)
`FIG. 3. Effect ofincreasing levels of PEG 4000 (O%,
`lo%, 0; 30%,
`0 ; 50%, 0 ; 75%, x ; loo%, A) on release of theophylline (2.5%)
`from G50/13.
`
`function of PEG 4000 concentration. The release rate of
`marker dye was found to be less than that of anhydrous
`theophylline at any given level of PEG 4000. This is partly
`due to it having a molecular weight about twice that of
`anhydrous theophylline. Therefore, the marker dye would be
`expected to diffuse from the matrix more slowly than
`anhydrous theophylline. Also, as theophylline is partially
`soluble in the matrix, dissolution of the drug would not have
`to occur before diffusion from the matrix.
`The trends are similar for anhydrous theophylline and the
`marker dye exhibiting a dual model of release at low levels of
`PEG 4000, and a diffusional process at higher levels. As the
`level of PEG 4000 increases within the swollen matrix, from
`which release occurs, the layer becomes less dense, thereby
`allowing drug to diffuse out easily. Thus, diffusion is
`enhanced and the model shifts to a diffusion-controlled
`process, with erosion making no significant contribution.
`Only when 100% PEG 4000 was used as the matrix did a
`complete switch to an erosional process occur.
`The 100% PEG 4000 system was allowed to run for only 30
`min as the material began to erode below the disc holder
`surface after that time thereby invalidating results.
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`

`

`384
`
`Table 1. Model coefficients&s.e. for Gelucire 50/13 with PEG 4000,
`D&C Yellow No. 10 (2.5%) or theophylline (2.5%) in simulated
`gastric fluid (pH 1-2) at 37°C. Paddle speed, 50 rev min- I; paddle ht,
`1 cm.
`
`MICHAEL KOPCHA ET AL
`theophylline from mixtures of Gelucires 5Ojl3 and 50/02 are
`summarized in Table 2. It is apparent that a correlation exists
`between release and the level of G50/02 employed. As the
`level increased, the overall rate of release decreased and the
`mechanism became more pronounced as a diffusional pro-
`cess. For D&C Yellow No. 10, the profiles were all linear
`with the square-root of time indicating matrix diffusion.
`Also, as the percentage of G50/02 increased, the A term
`decreased which further confirmed this behaviour.
`A corresponding trend was also noted for theophylline
`mixtures. As the level of G50/02 increased, release decreased
`and shifted to a diffusional process. Up to the 50% mixture,
`both erosion and diffusion occurred but the A/B ratio
`indicated that diffusion was the predominant process,
`increasing from 7 to 20 to 100% diffusion. Release was found
`to be comparable between the 27% G50/02 co-mixture and
`100% G50/13 (P>O.O5).
`The decrease in release as the level of G50/02 was increased
`may be because the larger amount of lipophilic drug/marker
`dye made it harder for drug to be released. At the same time,
`the mechanism of release was changed from a dual one, if
`such was the case, to a predominately diffusional process.
`Increased levels of G50/02 may have also increased the
`internal structuring of the matrix thereby decreasing its
`extent of swelling. This, in turn, would decrease the ease by
`which drug/marker dye could elute from the matrix.
`
`A/?
`ratio
`(hiI2)
`
`-
`
`1.84
`
`9.35
`
`-
`
`-
`
`0
`
`3.89
`
`2.52
`
`9.82
`
`-
`
`-
`
`0
`
`Temperature effects. Temperature effects were explored by
`varying the temperature of the receptor fluid on co-mixtures
`of G50/02 and G50/13. The general trends were the same; as
`the percentage of G50/ 13 increased, release increased regard-
`less of the temperature studied. This was expected, since
`
`A
`(mg h-"*)
`PEG 4000
`D&C Yellow No. 10
`0.940
`0 %
`k0.017
`0.330
`& 0.096
`1.047
`- +0.104
`2.379
`f0.016
`2.826
`& 0.020
`-
`
`10%
`
`30%
`
`50%
`
`75%
`
`100%
`
`Theophylline
`0 %
`
`10%
`
`30%
`
`50 Yo
`
`75%
`
`100%
`
`1.616
`- +0.135
`1.108
`- +0.108
`+ 0.065
`1.974
`2,787
`&0.019
`- + 0.022
`3.853
`-
`
`B
`(mg h-l)
`
`-
`
`0.179
`4 0.038
`0.1 12
`& 0.041
`-
`
`-
`
`18.19
`- +0.369
`
`0.4 16
`- +0.053
`0.439
`& 0,043
`0,201
`& 0.026
`-
`
`-
`
`19.18
`f0.347
`
`C
`(mg)
`
`- 0.034
`L 0.022
`& 0.073
`- + 0.046
`0.29 1
`f 0.050
`-0.157
`- + 0.020
`-0.383
`f0.025
`0.143
`- f0.I 12
`
`- + 0.065
`0,019
`0.246
`k0.052
`- + 0.03 1
`0.1 13
`- 0,022
`- + 0.024
`-0.283
`- +0.028
`0.496
`+_0,105
`
`Release from the 10% PEG 4000 system, of the marker dye
`and anhydrous theophylline was less than release from 100%
`G50/13, although
`the difference was not significant
`( P > 0-05). Release of theophylline from 30%" PEG 4000 was
`also similar (P> 0.05). This may be due to lack of significant
`enhancement of the diffusional or erosional characteristics to
`hasten the intrinsically high diffusional release of theophyl-
`line from the matrix.
`
`Evaluation of mixtures of Gelucire excipients
`Mixture effects. The DSC thermogram for a mixture of G50/
`02 and G50/13 showed a minor endotherm at about 30°C
`with a major one at approximately 49°C. The thermogram
`did not show any major shifts in the endotherms to indicate
`an interaction between the marker dye and the mixtures of
`C50/02 and G50/13. Theophylline, on the other hand,
`showed minor shifts in the endothermic peaks which re-
`flected a minor solubility of anhydrous theophylline in G50/
`13.
`Since the DSC scans for both G50/02 and G50/ 13 overlap,
`it was difficult to determine whether the two materials
`molecularly interacted or merely formed a dispersion. Upon
`microscopic examination, and after a 5 h experimental test
`with these mixtures, discrete spots were noted within the
`matrix. These spots became more prevalent as the material
`swelled; discrete spots appeared in the swollen layer as non-
`hydrated regions. This indicated that G50/02 was dispersed,
`as discrete packets, throughout the entire matrix.
`Model coefficients for release of D&C Yellow No. 10 and
`
`Table 2. Model coefficients&s.e. for Gelucires 50/13 and SOj02 with
`theophylline (2.5%) or D&C Yellow No. 10 (23%) in simulated
`gastric fluid (pH 1.2) at 37°C. Paddle speed, 50 rev min - I ; paddle ht,
`I cm.
`
`Percent
`G50/02:
`G50/13
`Theophylline
`Ojloo
`
`A
`(mg hWij2)
`
`5,847
`& 0.439
`7.076
`- + 0.494
`6.51 1
`- +0,31 I
`6.276
`f 0.068
`2.674
`k 0.095
`D&C Yellow No. 10
`Ojl00
`4.183
`*0.253
`2.728
`k 0,047
`2.412
`+0.112
`2.208
`f 0.032
`0,668
`& 0.039
`
`27/73
`
`50/50
`
`73/27
`
`lOOj0
`
`27/73
`
`50j50
`
`73/27
`
`loO/O
`
`'
`
`C
`(mg)
`
`0.57 I
`- +0,211
`0. I65
`L0.238
`0.092
`f0.150
`0.047
`k0.087
`0.08 1
`- +0.121
`
`-0.892
`k0.323
`0,445
`- + 0.060
`- 0.029
`- +0.143
`0~010
`- + 0.04 1
`0.026
`- + 0.050
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`

`

`EVALUATION OF RELEASE FROM THERMOSOFTENING VEHICLES
`Table 3. Model coefficients ks.e. for anhydrous theophylline and D&C Yellow No. 10 (2.5%) in simulated gastric fluid (pH 1.2) as a function of
`temperature. Paddle speed, 50 rev min-I; paddle ht, 1 cm.
`
`385
`
`30°C
`
`Percent
`G50/02:
`G50/13
`Theophylline
`O/lOo
`
`27/73
`
`50/50
`
`73/27
`
`lOoj0
`
`D&C Yellow No. 10
`O j l o o
`
`27173
`
`50150
`
`73/27
`
`1 Ooio
`
`A
`(mg h-'")
`
`8,352
`k 0.044
`- + 0,407
`4.242
`5.236
`k 0.224
`4.2 I5
`- f0.213
`0.529
`- +0.014
`
`4.415
`k0.145
`1.705
`k 0.023
`1.487
`k0.017
`- + 0.085
`1.073
`0.314
`- +0.015
`
`C
`(mg)
`
`-0.643
`k 0,056
`0.670
`- +0.196
`0.262
`- +0.108
`0.210
`- +0.102
`0.039
`- f0.018
`
`- 0.565
`k0.185
`0.327
`& 0.030
`0.278
`
`* 0.022
`
`0.335
`- f0.041
`0.306
`k0.019
`
`A
`(mg h-Ii2)
`
`11.51
`kO.088
`7.003
`- f0.954
`5.1 12
`k0.583
`5.124
`k0.493
`1.562
`k 0.208
`
`2.268
`- + 0.348
`- + 0.074
`3.199*
`3.104*
`k0.272
`1.640
`k0.196
`1.362
`- + 0.345
`
`42°C
`B
`(mg h-'1
`-
`
`1.850
`k0.379
`1.669
`k 0.23 1
`1.452
`k0.196
`0.705
`- f0.083
`
`0.900
`k0.138
`-
`
`-
`
`0,786
`- f0.078
`0.645
`k0.137
`
`C
`(mg)
`
`- 1.228
`- + O . l l
`0.054
`k0.45
`1.082
`k0.28
`0.742
`k0.23
`-0.249
`*0.10
`
`0.1 15
`- +0.170
`- + 0.060
`0.21 I
`-0.350
`k0.210
`0.335
`*0.010
`-0.223
`k0.170
`
`*Calculated for the first 2 h. See text for explanation.
`
`temperature was approaching the softening range of the
`mixtures. At about 30"C, the first softening peak of G50/02
`and G50/13 is reached. At 37"C, the peak is overcome and
`the major melting endotherm is approached. Thus, the bases
`continued to soften and became less rigid, allowing water to
`penetrate and drug/marker dye to diffuse outward. At 4 2 T ,
`the ascending part of the major melting endotherm is
`approached; further softening occurred and the rigidity of
`the matrix continued to decrease.
`From Tables 2 and 3, it is apparent that the mechanism of
`release changes as temperature increases. At 30°C the release
`of anhydrous theophylline from G50/02 is diffusion con-
`trolled. At 37"C, the mechanism is still diffusional but the A
`coefficient has increased five-fold. At 42"C, which is close to
`the onset of melting of the matrix, the release mechanism has
`converted to a mixed one of both diffusion and erosion. This
`was expected for several reasons: (1) as temperature in-
`creases, the diffusion coefficient of theophylline would also
`increase, reflecting the dependency of diffusion o n tempera-
`ture; (2) as temperature increases, the matrix softens and
`poses a weaker resistance to drug diffusion; and (3) as
`temperature approaches the onset of melting, the matrix
`becomes more susceptible to erosion as an additive process
`for release.
`For the 27% G50/02 co-mixture, with theophylline, as
`temperature was increased from 30 to 37"C, release became
`more diffusional; the A/B ratio increased. As temperature
`was further increased to 42"C, release became more depen-
`dant on erosion; the A/B ratio decreased. Similar trends were
`noted for the 50 and 73% G50/02 co-mixtures supporting the
`hypothesis that from 30 to 37°C it is the diffusional release
`is enhanced. When
`temperature
`is
`mechanism which
`
`increased to 4 2 T , the shift is to a combined model with the
`emphasis on erosion.
`At 30"C, the release of theophylline from G50/13 was
`diffusional. Going from 30 to 37"C, the matrix softened and
`erosion became a minor component of release; the A/B ratio
`decreased. However, as temperature increased to 42"C, only
`diffusion was seen. This may be due to the absence of G50/02
`from the matrix, which would have contributed to erosion of
`the matrix at this temperature. G50/02 does not swell but, as
`temperature increases, it does soften. Also, since it lacks the
`gelling ability of G50/13, it cannot maintain the integrity of
`the disc at elevated temperatures. As temperature increased,
`G50/13 remained pliable and retained disc integrity afford-
`ing drug release by a diffusional process without a significant
`contribution from erosion.
`The release of the marker dye as a function of temperature
`(Tables 2, 3), followed in an expected fashion. As the level of
`G50/13 increased, the amount of marker dye released also
`increased. Release from the 27,50 and 73% C50/13 matrices
`was similar demonstrating that these levels of G50/13 are not
`sufficient to enhance overall release. Since D&C Yellow No.
`10 is a charged molecule, it may have difficulty diffusing
`through a matrix which has become more lipophilic by
`addition of G50/02. It should be kept in mind, however, that
`as the level of G50/02 increased, release did decrease but not
`as dramatically as expected. Thus, release was predominately
`diffusional at both 30 and 37°C with an increase in rate as
`temperature was increased. Erosion only became significant
`when temperature was increased to 42°C. The effect of
`temperature was seen noticeably for the 27 and 50% G50/02-
`containing systems. Their profiles were best represented by
`an initial square-root of time relationship which became
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`

`

`386
`Table4. Model coefficients + s.e. for theophylline (2.5%) and D&C Yellow No. 10 (2.5%) in simulated intestinal fluid (pH 8.1 1) as a function of
`G50/02 and G50/13 composition at 37°C. Paddle speed, 50 rev min-'; paddle ht, 1 cm.
`
`MICHAEL KOPCHA ET AL
`
`Percent
`G50102:
`G50/13
`o/ 1 00
`27/73
`
`50/50
`
`73/27
`1 oojo
`
`Theophylline
`B
`(mg h - 9
`-
`
`0.705
`2 0.09 1
`1.040
`f0.113
`0.393
`- + 0.047
`-
`
`A
`(mg h-'")
`- + 0.042
`7.964
`5.568
`- +0.230
`4.846
`f0.284
`4.794
`- f 0 . 1 I9
`2.966
`- + 0.061
`
`C
`(mg)
`- + 0.054
`-0.167
`0,361
`- +0.111
`0.653
`- +0.137
`-0.076
`k0.057
`0.552
`0.078
`
`D&C Yellow No. 10
`
`A
`(mg h-'j2)
`- + 0.356
`1,605
`2.115
`- +0.280
`1.922
`- +0.150
`1.535
`k0.050
`- + 0.074
`0.145
`
`B
`(mg h-')
`0.99 1
`k0.141
`0.045
`- f0.111
`0.231
`- + 0.060
`0.437
`- + 0.020
`0.244
`- + 0.029
`
`C
`(mg)
`0.498
`f0.17
`0.386
`- +0.14
`0.301
`f 0.07
`0.199
`f 0.02
`0.288
`- + 0.04
`
`linear with time. This reflected a predominately erosional
`mechanism after 2 h. These observations support the
`hypothesis that G50/02 is the excipient which contributes
`mostly to matrix erosion.
`
`p H efsects. In simulated intestinal fluid (pH 8.1 1) release of
`theophylline and the marker dye correlates with increasing
`levels of G50/13 (Table 4).
`For the theophylline-containing system, release from the
`50 and 27% G50/02 preparations was similar. Similar
`behaviour was seen for the marker dye and theophylline-
`containing systems in simulated gastric fluid (pH 1.2). We
`suggest that two opposing factors prevent any noticeable
`change in release at these levels of G50/02. As the level of
`G50/02 increased, the extent of swelling of the hydrated layer
`decreased. This would be expected to enhance diffusion of
`the drug marker dye from the system because the diffusional
`pathlength decreased. However, as the level of G50/02
`increased, the system became more lipophilic which thwarted
`drug/marker dye release by preventing water from easily
`permeating the matrix.
`The following discussion will make comparisons between
`the various mixtures of G50/02 and G50/13 in both simu-
`lated gastric (Table 2) and intestinal fluid (Table 4) for the
`theophylline-containing systems. For the G50/02 system, the
`two curves were essentially the same (P>O.O5); release into
`either medium was diffusion-controlled.
`For G50/13, although the two profiles were found to be
`coincidental over the test period, the mechanisms of release
`were not identical. In gastric fluid, a dual mechanism of
`erosion and diffusion was noted, whereas in the basic
`intestinal fluid only diffusion was seen. This may be due to a
`pH partitioning effect of theophylline into the basic medium;
`pH was close to the pK,.
`For the 73% G50/02 system, the profiles were found to be
`the same (P> 0.05). In the acidic media, a diffusional model
`was noted while in the basic solution, a mixed model was
`fitted. Erosion became more noticeable in the basic medium.
`For 27% G50/02, A/B ratios of 7.90 and 7.18 were noted
`for the basic and acidic mediums, respectively, which are not
`statistically different ( P > 0.05). Hence, erosion and diffusion
`are seen to effect release from this system.
`For the 50% G50/02 mixture, even though overall release
`from this system was similar at both pH values, the
`
`mechanisms were different. The A/B ratio for the basic
`system was 4.66 while that for simulated gastric fluid was
`20.41. Hence, the basic environment allowed for a more
`erosional process to occur while the acidic solution allowed
`for a more diffusional process to occur.
`The release of the marker dye in simulated gastric fluid, is
`by diffusion, whereas in simulated intestinal fluid, release is
`by a dual mechanism; erosion appears to occur more
`noticeably in a basic environment, which may be attributed
`to the partial hydrolysis of the ester-linked matrices.
`
`Predictive ability
`To evaluate the predictive nature of the models described
`previously (Kopcha et a1 1990), the diffusion coefficient for
`the drug/marker dye in G50/02 and apparent diffusion
`coefficient in G50/13 were determined as follows: for D&C
`Yellow No. 10 in G50/02, 2 . 6 6 ~ lo-' and in G50/13,
`and
`and for theophylline in G50/02, 1.98 x
`8.82 x
`in G50/13, 1.72 x
`For GS0/02, the A term was evaluated for a rigid matrix as
`described by Paul & McSpadden (1976).
`
`1+H
`M =--[C,(Dt)]''*
`(3H)'I2
`
`where:
`
`M=amount of solute released per unit area, W,=initial
`drug loading per unit volume, C,=equilibrium drug solu-
`bility, D=drug diffusion coefficient, and t = time. This
`approximate analytical solution is valid for all W& values.
`Note that strict matrix diffusion was assumed. No hydro-
`dynamic diffusion boundary layer was considered in this
`calculation, as a first approximation. Ignoring this effect, the
`results were still in agreement with published data. This
`reflects the hypothesis that the matrix, and not drug/marker
`dye dissolution, was the predominate resistance to drug
`release. Thus, the effect of a boundary layer can be consi-
`dered insignificant (ix.
`cm).
`The apparent diffusion coefficient for drug/marker dis-
`persed in G50/1 3 was calculated from:
`
`(3)
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`

`

`387
`
`EVALUATION OF RELEASE FROM THERMOSOFTENING VEHICLES
`References
`where S is the effective diffusional area, V is the effective
`volume of the hydrated matrix, D’ is the apparent diffusion
`Bowtle, W. (1986) The application of semi-solid capsule technology
`to antibiotic formulation. Boll. Chim. Farm. 125: 72-74
`coefficient of drug in the hydrated matrix which takes into
`Dennis, A. B., Kellaway, I. W. (1987) Drug release from a slowly
`account both tortuosity and porosity of the hydrated matrix,
`hydrating semi-solid matrix. J. Pharm. Pharmacol. 39 (Suppl.):
`and the other terms are as defined previously.
`40P
`Results were compared with those calculated by Harland
`Harland, R. S., Gazzaniga, A. M., Sangalli, M. E., Colombo, P.,
`Peppas, H. A. (1988) Drug/polymer matrix swelling and dissolu-
`et a1 (1988). The order of magnitude for these coefficients is in
`tion. Pharm. Research 5 488494
`agreement with the published data. The calculated coeffi-
`Kopcha, M., Tojo, K. J., Lordi, N. G. (1990) Evaluation of
`cients for G50/13 are somewhat larger than those seen with
`methodology for assessing release characteristics of thermosoften-
`G50/02; the system is less lipophilic which allows drug to
`ing vehicles. J. Pharm. Pharmacol. 42: 745-751
`Paul, D. R., McSpadden, S. K. (1976) Diffusional release of a solute
`diffuse more easily through the matrix. Thus, it appears that
`from a polymer matrix. J. Membr. Sci. 1: 33-48
`the models used to quantitate release from these excipients
`Thakkar, A. L., Gibson, L. L., Quay, J. F. (1987) Semi-solid matrix
`are adequate and can realistically predict the processes
`capsule in formulations of cephalexin: comparative bioavailabi-
`associated with release.
`lity in the dog. J. Pharm. Sci. 76: S301
`
`Purdue 2036
`Collegium v. Purdue, PGR2018-00048
`
`