Available online at www.sciencedirect.com
`
`SCIENCE@DIRECT•
`
`ELSEVIER
`
`International Journal of Pharmaceutics 299 (2005) 55-64
`
`international
`journal of
`pharmaceutics
`
`www.elsevier.com/locate/ijpharm
`
`Significance of lipid matrix aging on in vitro
`release and in vivo bioavailability
`
`Y.W. Choy , Nurzalina Khan *, K.H. Yuen
`
`School of Pharmaceutical Sciences, University of Science Malaysia , 11800 Penang, Malaysia
`
`Received 11 November 2004; received in revised form 7 April 2005; accepted 23 April 2005
`Available online 13 June 2005
`
`Abstract
`
`A polyglycolised glyceride ca1Tier, Gelucire 50/13 , was incorporated with paracetamol as a model drug, filled into hard
`gelatin capsules and stored at three different temperatures for various lengths of time. The resultant solidified matrix within the
`capsule was subjected to thermal analysis using differential scanning calorimetry (DSC) to ascertain its supramolecular structure.
`Polymorphic transformations towards more stable gelucire forms were observed upon aging the matrices , with samples stored at
`a temperature near the melting range of the lower temperature gelucire melting fraction showing the most profound changes. The
`increase in the rate of drug release from aged samples could be correlated to the alterations to the supramolecular structure of the
`gelucire. Accelerated drug release from aged samples could also be seen from in vivo studies using healthy human volunteers,
`although the extent of absorption was not affected . Therefore, even though the sustainability of release may be compromised by
`aging the gelucire matrices , the bioavailability of the incorporated drug is unlikely to be affected .
`© 2005 Elsevier B.V. All rights reserved.
`
`Keywords: Lipid matrices; Gelucire®; Aging; In vivo; Sustained release; DSC
`
`1. Introduction
`
`Lipid based materials are fast becoming the carri(cid:173)
`ers of choice for oral delivery of new chemical entities
`(NCE) and well-established active agents (Mueller et
`al., 1994; Constantinides , 1995; Chambin et al. , 2004).
`NCEs that had been developed are mostly hydropho(cid:173)
`bic , thus presenting associated problems such as low
`and erratic bioavailability. Solubilisation into lipid car-
`
`* Corresponding author. Tel.: +604 6569040; fax: +604 6570017 .
`E-mail address: nurza@ usm.my (N. Khan).
`
`0378-5173/$ - see front matter © 2005 Elsevier B .V. All rights reserved.
`doi: JO .1016/j .ijpharrn.2005 .04.030
`
`riers followed by intraluminal processing circumvent
`these inherent limitations (Humberstone and Charman ,
`1997). However, more hydrophilic drugs which are
`already in existence can also be reformulated into cer(cid:173)
`tain lipid based carriers for modification of their release
`profiles (Saraiya and Bolton , 1990; Esquisabel et al. ,
`1996; Hamdani et al. , 2002). These carriers include the
`gelucires, which are a family of polyglycolised glyc(cid:173)
`eride bases , consisting of polyethylene glycol (PEG)
`esters of various fatty acids, tri-, di-, and monoglyc(cid:173)
`erides of the fatty acids, with some of the corresponding
`free fatty acids and PEGs present in small quantities.
`
`COLL 1094084
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`56
`
`Y.W. Choy et al. / International Journal of Pharmaceutics 299 (2005) 55-64
`
`Each type of gelucire is characterized by two numbers ,
`the first being the nominal melting point and the second
`being the hydrophile-lipophile balance (HLB) number.
`In general, the higher melting bases with a bigger pro(cid:173)
`portion of lipophilic components in them are used as
`coating and matrix agents for sustained release formu(cid:173)
`lations , whilst those with more hydrophilic components
`within are suitable as bioavailability enhancers. The
`amphiphilicity of the base conferred by the long hydro(cid:173)
`carbon chain and the alcohol moieties means that both
`hydrophilic and hydrophobic drugs can potentially be
`incorporated into these carriers.
`However, lipid based carriers are known to be
`affected by stability issues , particularly those associ(cid:173)
`ated with the changes to the matrix structure upon
`aging. These changes have been attributed to polymor(cid:173)
`phic transformations within the lipid component and
`also to the progressive increase in the crystallinity of the
`matrix components (Sutananta et al., 1994; Hamdani et
`al., 2002) , which in turn could result in a modification
`of the in vitro and in vivo releases of the incorporated
`drugs. In a system such as gelucire whereby the charac(cid:173)
`terization of the glycerides is made even more complex
`by esterification with PEG, the poor understanding of
`the phase changes has made it very difficult to predict
`its release behaviour. Variations in the results of drug
`release studies upon aging is a function of the com(cid:173)
`ponents within the particular gelucire, conditions of
`storage and incorporated drugs (Remunan et al., 1992;
`Sutananta et al. , 1995 ; San Vicente et al. , 2000) . Even
`though these investigations were vital in establishing
`the role of aging in altering drug release in vitro , stud(cid:173)
`ies that examine the relevance of these findings in vivo
`are scarce. Dennis et al. (1990) found that aged formu(cid:173)
`lations ofketoprofen dispersed in G50/13:G50/02 base
`mixture gave an increased rate of drug release in vitro
`but this observation was discovered to be insignificant
`in vivo .
`In our current study, Gelucire 50/13 (G50/13) ,
`which has a sufficiently high melting point to form
`sustained release matrices and a balance of compo(cid:173)
`nents that allows both hydrophobic and hydrophilic
`drugs to be incorporated within was chosen as the
`model gelucire. G50/l 3 has the capacity to alter its
`dimensions through its gelling and swelling abilities
`up to the extent necessary for sustaining its controlled(cid:173)
`release properties (Kopcha et al ., 1991; Prapaitakul et
`al ., 1991 ; Esquisabel et al., 1996). Most dosage forms ,
`
`including lipid containing ones , are usually stored for
`a certain length of time before being used . The stor(cid:173)
`age conditions that these forms are subjected to on
`leaving the manufacturing plant are not always easily
`controlled and it would be unduly optimistic to pre(cid:173)
`dict an absence of temperature fluctuations within their
`environs . As these factors have the potential of altering
`the performance of the dosage form, it is essential to
`attempt to elucidate any mechanism of change within
`the solid-state structure , especially in relation to its
`consequent behaviour in vivo . Therefore, our current
`study aimed to address the above issue by investigat(cid:173)
`ing the release behaviour of a relatively hydrophilic
`drug from a lipid matrix before and after storage at
`different temperatures, which represent the aging pro(cid:173)
`cess . The association of structural modifications to the
`matrix after aging, as indicated by thermal analysis, to
`changes in drug release was investigated. More impor(cid:173)
`tantly, the significance of such modifications as shown
`in in vitro systems , to drug release in healthy human
`volunteers was determined.
`
`2. Materials and methods
`
`2 .1. Materials
`
`Gelucire 50/13 was purchased from Gattefosse
`(Saint-Priest, France) . The model drug used was parac(cid:173)
`etamol (Wenzhou Pharmaceutical, China) and the hard
`gelatin capsules used were of size 0 . The paraceta(cid:173)
`mol was sieved to give particles of :S300 µm in size .
`All other chemicals and reagents used were either of
`analytical reagent (AR) grade or of high performance
`liquid chromatography (HPLC) grade.
`
`2 .2 . Preparation of liquid.filled hard gelatin
`capsules
`
`10% w/w of the drug was weighed accurately, added
`to the molten gelucire at 75 °C and mixed using a mag(cid:173)
`netic hot plate stirrer for 10 min to ensure homogeneity
`and removal of thermal history. The drug did not com(cid:173)
`pletely dissolve in the molten carrier and remained
`as a dispersion . This was also observed by Khan and
`Craig (2003) at the same drug loading . The mixture
`was then filled into hard gelatin capsules size O using
`preheated pasteur pipettes and left to set under ambient
`
`COLL 1094085
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Y. W. Choy et al. I International Journal of Pharmaceutics 299 (2005) 55-64
`
`57
`
`conditions. The capsules were stored in amber bottles
`containing silica gel for 24 h at room temperature for
`samples designated as freshly prepared , and in a refrig(cid:173)
`erator at 4 °C, at room temperature (24.0 ± 0.5 °C), or
`in a temperature-controlled oven at 3 7 °C until required
`for samples designated as aged. Before any analysis
`was perlormed , the aged capsules were equilibrated at
`room temperature for 24 h. The mean fill weight of the
`capsules was 682.6 ± 19.1 mg with the coefficient of
`variation being 2.8% . The prepared capsules fulfilled
`the criteria for the uniformity of dosage units as speci(cid:173)
`fied in US Pharmacopeia 24 (2000) .
`
`2 .3. Differential scanning calorimetry
`
`Differential scanning calorimetry (DSC) scans were
`performed on Perkin-Elmer Pyris 6 DSC (Beacons(cid:173)
`field, UK) equipped with an intracooler, at the rate of
`2 °C/min. Purge gas used was helium flowing at the rate
`of 20 ml/min and the instrument was calibrated using
`indium and zinc standards. The 9.0-10.5 mg of sam(cid:173)
`ples obtained from the content of the capsules were
`put into non-hermetic aluminium pans (Perkin-Elmer,
`UK) and sealed. The melting point was taken to be
`the temperature where the rate of melting was the
`highest, that is at the tip of the endothermic curve .
`The heat or enthalpy of fusion was calculated from
`the area under the curve from the first inflection of
`the baseline to the temperature where the endotherm
`returned to the baseline (Khan and Craig, 2003).
`Replicates of four scans were performed for each
`sample.
`
`2 .4. In vitro dissolution studies
`
`The dissolution of the liquid filled hard gelatin cap(cid:173)
`sules was determined following the basket method of
`the USP 24 (2000) in a Type PTW S III apparatus
`(PharmaTest, Germany). The test was run under sink
`conditions with 900 ml distilled water maintained at
`37 °C being used as the dissolution medium and the
`basket rotation speed set at 100 rpm. Samples were
`collected from the dissolution of six capsules using
`SDX automated fraction collector (Sadex, Malaysia)
`up to 14h. Concentration of the dissolved drug was
`measured spectrophotometrically at 243 nm (Hitachi,
`Japan). Mean dissolution time (MDT) is the mean ratio
`of the first to zero moments of the dissolution rate-time
`
`curve and was calculated using the following equation
`(Brockmeier and Von Hattinberg , 1982):
`
`ACC
`MDT=(cid:173)
`Moo
`
`(1)
`
`where ACC is the area complimentary to the area under
`the accumulated dissolution curve and M 00 is the accu(cid:173)
`mulated amount dissolved at maximum time. MDT has
`also been described as the sum of different periods of
`time during which fractions of the dose stay in the poly(cid:173)
`mer matrix before release, divided by the total dose
`(Lindner and Lippold , 1995).
`
`2 .5. In vivo study protocol
`
`Twelve healthy adult male volunteers not receiv(cid:173)
`ing any medication prior to and during the sampling
`period participated in the study. The study followed a
`three treatment, three periods , three sequence crossover
`design with a washout period of 1 week. The sam(cid:173)
`ples for this study were the liquid filled hard gelatin
`capsules prepared according to Section 2 .2 , which had
`been freshly prepared or subjected to the different stor(cid:173)
`age temperatures of 4 and 37 °C. The volunteers were
`randomly divided into three groups of four and were
`administered the samples, according to the schedule
`shown in Table 1 .
`Five capsules containing a total amount of 300 mg
`of drug were administered after an overnight fast with
`240 ml of water. Drinks were available ad libitum
`2 h after dosing whilst standardised meals were
`given at 4 and 10 h after dosing. Blood samples of
`5 ml in volume were collected at 0 (predose), 0 .5,
`1, 2, 3, 4, 6, 8, 10, 14, 18 and 24h after dosing
`via an in-dwelling cannula placed in the forearm,
`centrifuged for 15 min at 3500 rpm, and the plasma
`transferred into glass containers to be kept frozen until
`analysis.
`
`Table I
`The schedule for the administration of samples to the volunteers at
`the three different study periods
`
`Group
`
`Period I
`
`Petiod II
`
`Period III
`
`2
`3
`
`Freshly prepared
`Storage at 37 °C
`Storage at 4 °C
`
`Storage at 4 °C
`Freshly prepared
`Storage at 37 °C
`
`Storage at 37 °C
`Storage at 4 °C
`Freshly prepared
`
`COLL 1094086
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`58
`
`Y.W. Choy et al. / International Journal of Pharmaceutics 299 (2005) 55-64
`
`2 .6. Analysis of the drug concentration in plasma
`
`The plasma samples were analysed using areversed(cid:173)
`phase HPLC method described by Yuen et al. (1997) .
`
`2.7. Data analysis
`
`The pharmacokinetic parameters, namely the extent
`of absorption (AUCo-oo), maximum plasma concen(cid:173)
`tration (Cmax), time to reach maximum concentra(cid:173)
`tion (T max) and mean residence time (MRT) were
`calculated from the individual plasma drug concen(cid:173)
`tration data. Cmax and Tmax were obtained directly
`from the plasma values (Weiner, 1981). AUC0_00
`was obtained by adding the area under the plasma
`drug concentration- time curve from time O to the
`last measurable concentration (AUCo- r) and the area
`from the last measurable concentration to infinite
`(AUC1_ 00 ) . The former was calculated using the trape(cid:173)
`zoidal formula whilst the latter was calculated by
`dividing the last measurable concentration with the
`apparent elimination rate constant (ke)- ke was esti-
`
`mated from the terminal slope of the individual plasma
`drug concentration-time curves after logarithmic trans(cid:173)
`formation of the plasma drug concentration values
`and the application of linear regression (Gibaldi and
`Perrier, 1982) .
`The Wagner-Nelson equation, which can be used
`for one as well as two compartments open model was
`employed to calculate in vivo absorption from plasma
`drug concentration data (Wagner and Nelson , 1963 ;
`Wagner, 1975):
`
`C1 + keAUCo-t
`% absorbed at time t = - - - - - -
`keAUCo-oo
`
`(2)
`
`where C1 is the plasma concentration at time t.
`MRT was calculated from the individual plasma
`drug concentration-time curves using the Statistical
`Moment Theory, according to the equation:
`
`AUMC
`MRT = -
`AUC
`
`(3)
`
`where AUMC is the first moments curve and AUC is the
`area under the plasma drug concentration-time curve.
`
`Temperature (0 C)
`50.0
`45.0
`25.0
`40,0
`65.0
`60.0
`55.0
`30.0
`35.0
`2.0 +---~~--~ - - -~ - - -~ - -~ - - -~ - - -~ - ---,
`
`Fresh
`
`4°c
`
`RT
`
`" 10c
`
`endo down •
`
`6.0 i---- ---
`
`10.0
`
`~ 14.0
`~
`3:
`0 ,::
`.;
`"'
`::c
`
`18.0
`
`22 .0
`
`26.0
`
`30.0
`
`Fig . I . DSC thermal profiles of Gelucire 50/ 13 samples which had been fresh ly prepared , or stored at 4 °C , room temperature (RT) or 37 °C for
`22 weeks.
`
`COLL 1094087
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`(cid:173)
`

`

`Y. W. Choy et al. I International Journal of Pharmaceutics 299 (2005) 55-64
`
`59
`
`These parameters were calculated using the trapezoidal
`rule.
`
`3. Results
`
`3 .1 . Physical characterization of the matrices
`using DSC
`
`Fig. 1 shows the thermal profiles obtained from the
`DSC scans of the samples containing G50113 which
`had been freshly prepared, or stored at 4 °C, room tem(cid:173)
`perature or 37 °C. For clarity, only the thermal profiles
`for the aged samples as stored for 22 weeks are shown.
`Several peaks could be seen within the broad melting
`endotherm of the gelucire and this is due to the het(cid:173)
`erogeneity of components within the carrier. The first
`prominent peak is designated the lower melting gelu(cid:173)
`cire fraction from hereon, and the second prominent
`peak which commenced from about 40 °C together
`with the succeeding shoulder are designated the higher
`melting gelucire fraction. Out of the three aged sam-
`
`ples , the ones stored at 4 °C showed the least variation
`to the thermal profiles of the freshly prepared samples.
`However, a new shoulder could be detected at the
`lower temperature end of the higher melting gelucire
`fraction, preceeding the second prominent peak, which
`was apparent for all the aged samples. An increase in
`the melting temperature of the higher melting gelucire
`fraction was observed for all the aged samples, with
`the samples stored at 37 °C showing the biggest
`change.
`Elevations in the heat of fusion values were also
`observed upon aging, with 160.7 ± 6.3 Jig for the sam(cid:173)
`ples that were freshly prepared to 176.0±8.SJlg and
`172.5 ± 6.8 Jig for the samples stored for 22 weeks
`at 4 °C and room temperature, respectively. A more
`drastic change was observed for the samples stored
`at 37 °C with the heat of fusion value after 22 weeks
`being 187.4 ± 8.2 Jig . This is a 16.6% increase from the
`value of freshly prepared samples, compared to 9.5%
`and 7 .3% increases for the samples stored at 4 °C and
`room temperature, respectively. These differences were
`found to be statistically significant (p<0.01).
`
`Temperature (0 C)
`50.0
`40.0
`45.0
`35.0
`25 .0
`60.0
`30.0
`65.0
`55 .0
`2.0 -1 - - - -~ - - - - ' - - - - - ' - - - - - - - ' - - - - - - ' - - - - - - '~ - - - - ' - - - - - - '
`
`6.0 r----
`
`10.0
`
`l 14.0
`:I,'.
`0
`..
`i::
`~ 18.0
`==
`
`22.0
`
`26.0
`
`30.0
`
`Fresh
`
`4°c
`
`RT
`
`31°c
`
`Fig. 2. DSC thermal profiles of 10% paracetamol in gelucire 50/13 dispersion samples which had been freshly prepared, or stored at 4 °C, room
`temperature (RT) or 37 °C for 22 weeks.
`
`COLL 1094088
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`60
`
`Y.W. Choy et al. / International Journal of Pharmaceutics 299 (2005) 55-64
`
`100.0
`
`80.0
`
`60.0
`
`40.0
`
`20.0
`
`"O
`
`"' "' ..
`..
`"'
`-.;
`OD
`?
`"O
`l;;'<
`
`0.0
`0.0
`
`...
`
`.. • ,.-·
`
`.. t ··
`
`wcck:22
`~
`-+- week 14
`---Ir- week LO
`
`-& - wcck 8
`
`- - wcck3
`-+- weck 2
`
`____,.__ week I
`
`--•-· Fresh
`
`2.0
`
`4.0
`
`6.0
`
`8.0
`Time (hr)
`
`10.0
`
`12.0
`
`14.0
`
`Fig . 3. In vitro dissolution profiles of paracetamol in GS0/13 dispersion samples which had been freshly prepared or stored at 37 °C.
`
`Upon the incorporation of 10% wlw paraceta(cid:173)
`mol into GS0/13 , the lower melting gelucire fraction
`became more prominent . The stabilisation of this frac(cid:173)
`tion by paracetamol was previously described by Khan
`and Craig (2003) . Similar to the samples of G50113
`on its own, the total heat of fusion values for the drug
`
`dispersion samples were also elevated when stored at
`4 °C,room temperature and 37 °C,from 126.0 ± 9.0 Jig
`when freshly prepared to the after storage values of
`173.8 ± 7 .7, 178.5 ± 10.5 and 196 .7 ± 8.3 Jig signify(cid:173)
`ing increases of37 .9%,41.7% and 56 .0%,respectively.
`It can be seen that there were little changes to the ther-
`
`100.0
`
`80.0
`
`"' .. 60.0
`"O .,
`.,
`-.; .. OD
`.. "O
`
`;:s
`
`~ 40.0
`
`20.0
`
`0.0
`0.0
`
`..1··
`
`.. . ,-··
`
`.-1· ·
`
`- - 37°C
`
`---- RT
`
`-+- 4°C
`
`•• • · · Fresh
`
`2.0
`
`4.0
`
`6.0
`
`8.0
`Time (hr)
`
`10.0
`
`12.0
`
`14.0
`
`Fig . 4. In vitro dissolution profiles of paracetamol in GS0/13 dispersion samples which had been freshly prepared, or stored at 4 °C, room
`temperature (RT) or 37 °C for 22 weeks .
`
`COLL 1094089
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Y. W. Choy et al . I International Journal of Pharmaceutics 299 (2005) 55-64
`
`61
`
`(J
`
`i 4.0
`ell
`5
`.§ 3.0
`~ C
`~ C 2.0
`8 .. s
`"' £ 1.0
`
`o.o ~ - - - - - - - - - - - ---=::::::::::::;;~;;;;;;;;;;===s,a;;J
`
`12,0
`Time (hr)
`
`14.0
`
`16.0
`
`18.0
`
`20.0
`
`22.0
`
`24,0
`
`0.0
`
`2.0
`
`4.0
`
`6.0
`
`8.0
`
`10.0
`
`Fig. 5. Mean plasma drug concentration (µ g/ml) against time (h) profiles after the oral administration of paracetamol in G50/13 dispersion
`samples which had been freshly prepared , or stored at 4 or 37 °C for 22 weeks.
`
`mal profiles of the samples stored at 4 °C and room
`temperature (Fig. 2), but the thermal profile of the
`sample stored at 37 °C was vastly different from the
`freshly prepared sample. This modification was appar(cid:173)
`ent even after only 1 week of storage at the higher
`temperature.
`
`3 .2. In vitro dissolution studies
`
`In general, increasing the storage period under the
`three test conditions accelerated the release of parac(cid:173)
`etamol from GS0/13 matrices, although this trend was
`less clearly observed in the samples which had not been
`subjected to prolonged storage, as exemplified in Fig . 3
`by the dissolution profiles of samples stored at 37 °C.
`Storage at 37 °C resulted in the highest elevation in the
`release rates compared to storage at 4 °C and room tem(cid:173)
`perature. Fig. 4 shows the comparison between the stor(cid:173)
`age conditions , using the dissolution profiles of freshly
`prepared samples and samples stored for 22 weeks
`only for clarity. The mean dissolution time (MDT)
`values calculated reflected the changes stated above,
`with the MDT decreasing from 8 .92 ± 0 .11 h for the
`freshly prepared samples to 8.78±0.12, 8.73±0.14
`and 8 .4 7 ± 0 .14 h for the samples stored for 22 weeks at
`4 °C, room temperature and 37 °C, respectively. How(cid:173)
`ever, only the value for the sample stored at 37 °C was
`significantly different statistically (p < 0.01) from the
`value of the freshly prepared sample.
`
`3 .3. In vivo study
`
`Fig . 5 shows the mean plasma concentration-time
`profiles after the oral administration of 300 mg parac(cid:173)
`etamol as the drug dispersion samples which had
`been freshly prepared, or stored at 4 or 37 °C for 22
`weeks. The AUCo-oo values of 26 .2 ± 7 .9, 26 .2 ± 8 .3
`and 26.4 ± 8.9 µg/ml for the freshly prepared, stored at
`4 and 37 °C samples respectively were found to be not
`significantly different from each other (p>0 .9) indi(cid:173)
`cating that the extent of bioavailability of the drug
`was not affected by the storage conditions imposed
`on the matrices. Similarly, no statistically significant
`difference (p>0.05) was observed among the Tmax
`values of the samples , that were 3 .2 ± 0 .9 , 3 .0 ± 1.2
`and 2.8 ± 1.1 h for the samples in the same order as
`above. Cmax values for the samples were 3 .3 ± 1.6,
`3.1 ± 0.9 and 4.4 ± 1.5 µg/ml respectively and a statis(cid:173)
`tically significant difference was observed among the
`logarithmically transformed Cmax values of the three
`samples (p<0.01) .
`The mean in vivo paracetamol absorption-time pro(cid:173)
`files (Fig . 6) showed that the absorption of the drug ,
`as calculated from Eq . (2) , was the fastest from the
`matrices stored at 37 °C. This is in accord with the
`in vitro dissolution profile of the same samples which
`showed the fastest release compared to release from
`freshly prepared samples or samples stored at 4 °C.
`However, the in vivo absorption profiles of the lat-
`
`COLL 1094090
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`62
`
`Y.W. Choy et al. / International Journal of Pharmaceutics 299 (2005) 55-64
`
`100.0
`
`90.0
`
`80.0
`
`..
`,-...
`-;;!if_
`.._,
`-,:,
`,.Q ... Q
`~ 70.0
`"' 60.0
`,.Q
`C'I
`01) 50.0
`::, ...
`-,:,
`~ 40.0
`01)
`.... = 30.0
`C'I
`"' ...
`
`~
`
`~ 20.0
`Q.;
`
`....... 31°c
`
`--- 4°c
`
`..._ Fresh
`
`2.0
`
`4.0
`
`6.0
`
`8.0
`Time (hr)
`
`10.0
`
`12.0
`
`14.0
`
`10.0
`
`0.0
`0.0
`
`Fig. 6. Mean percentage drug absorbed(%) against time (h) profiles after the oral administration of paracetamol in GS0/13 dispersion samples
`which bad been freshly prepared, or stored at 4 or 37 °C for 22 weeks.
`
`ter two samples were similar even though the in vitro
`release indicated a notable difference with the release
`from the samples stored at 4 °C being much faster than
`release from freshly prepared samples. When the mean
`residence time (MRT) values were calculated, it was
`revealed that the samples stored at 37 °Chad the lowest
`value at 5 .50 ± 0 .66 h, followed by the freshly prepared
`samples at 6 .69 ± 0 .94 h and samples stored at 4 °C at
`7.18±0.86h.
`
`4. Discussion
`
`Gelucire 50/13 is made up of many lipid compo(cid:173)
`nents such as glycerides and free fatty acids, as well
`as more hydrosoluble components such as PEG esters
`and residual PEG 1500. These numerous components
`generate the endotherms observed on the thermal pro(cid:173)
`files of the samples, which in turn could be the result
`of several overlapping transitions. However, it may
`be too simplistic to suggest that the thermal profile is
`the simple sum of the transitions caused by individual
`components. The composition of the gelucire is made
`even more complex by the probable presence of the
`different polymorphic forms of the lipid components
`and the PEG esters of fatty acids. The different
`constituents may form mixed crystals with each other,
`leading to what had been described previously as
`
`segregations of components which melting gave rise
`to the endothermic peaks (Sutananta et al., 1994).
`Upon ageing under all three conditions, the gelucire
`components transformed into their more stable poly(cid:173)
`morphic forms , as evidenced by the elevated heats of
`fusion and the progression towards the formation of
`the higher melting fraction. This transformation was
`more pronounced in the samples stored at the highest
`temperature, 37 °C, which is also in the vicinity of
`the endotherm of the lower melting fraction. At this
`temperature, the solidity of the matrix is lowered due
`to a portion of the gelucire existing as a molten mass ,
`which results in the more stable polymorphic form
`being preferred during lipid recrystallisation. This
`was also noted for other lipid mixtures, when the low
`solid content of the freshly prepared sample resulted
`in the crystals becoming more prone to polymorphic
`changes (deMan and deMan, 1994).
`As gleaned from the in-vitro dissolution studies, the
`pattern of change in the release rates was not consistent
`in the earlier stages of storage (less than or at 14 weeks) ,
`which could be attributed to the rearrangement of the
`gelucire molecules prior to achieving a stable structure .
`Further storage after 14 weeks did not bring about sig(cid:173)
`nificant alterations to the release rates, suggesting that
`transformations into the most stable gelucire structure
`possible under each condition were completed . When
`the MDTs were calculated from the samples that had
`
`COLL 1094091
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`Y. W. Choy et al . I International Journal of Pharmaceutics 299 (2005) 55-64
`
`63
`
`achieved this equilibrium state, it was found that the
`drug release from the samples stored at 37 °C was sig(cid:173)
`nificantly faster than from freshly prepared samples.
`The in vitro dissolution profiles together with the
`thermal profiles obtained from DSC studies indicate
`an association between the supramolecular structure of
`the matrices and drug release . Storage at 37 °C resulted
`in the highest degree of transformation to a more sta(cid:173)
`ble form of G50/13 and also the fastest drug release ,
`with both changes found to be significantly different
`statistically from the freshly prepared samples. Such
`an association may be explained by the formation of
`larger crystals in the stored samples causing a disrup(cid:173)
`tion to the previously compact matrix . The change in
`the lipid crystal structure, generally to larger sizes, is
`known to occur during the polymorphic transforma(cid:173)
`tion of a metastable to a more stable form (Hachiya et
`al. , 1990; Heathcock, 1993). The structural alteration
`throughout the matrix may have increased its porosity,
`allowing faster and easier dissolution medium penetra(cid:173)
`tion into the bulk of the samples. Even though some
`paracetamol may have been solubilised in the molten
`gelucire during manufacture , the drug did not recrys(cid:173)
`tallise out in its pure form during solidification, but
`instead induced a structural modification within the
`carrier matrix which was further altered during aging
`(Khan and Craig , 2003 , 2004) .
`The in vivo studies demonstrated that even though
`the bioavailability of the drug was not altered by the
`aging process, Cmax was higher for samples stored
`at 37 °C. Bioavailability was not affected because
`the chemical composition remained the same for all
`samples and in addition , the model drug used is not
`known to have a poor absorption profile. The structural
`changes caused by the aging process led to more drug
`being released in the first few hours as demonstrated by
`the in vitro dissolution and the higher maximum con(cid:173)
`centration achieved in vivo. These results suggest that
`although the extent of absorption will not be adversely
`affected by storage conditions , the sustainability of the
`release over time may be compromised especial] y when
`the gelucires are kept at temperatures near its lower
`melting fraction. This change in release profile may
`be predicted by studying the matrix structure as deter(cid:173)
`mined by thermal analyses.
`There is no rank order correlation between MDT
`values obtained from the in vitro dissolution tests and
`MRT values obtained from the in vivo study. On closer
`
`inspection of the results , it could be seen that the sam(cid:173)
`ples stored at 37 °C had both the lowest MDT and
`MRT values , suggesting that correlation could only be
`seen for matrices that underwent substantial changes.
`As stated before , there may be disruption to the bulk
`of the matrix due to the formation of larger lipid
`crystals brought about by polymorphic transformation .
`The high release rates observed in vitro and in vivo
`for these samples may then be due to the increased
`preponderance of erosion as the mechanism of drug
`release. Shameem et al. (1995) demonstrated that the
`release of paracetamol was faster from one controlled
`release dosage form than from another, both in vitro
`(50-150 rpm rotating basket methods) and in vivo , due
`to enhanced erosion . A higher Cmax was attributed to a
`"burst release effect" caused by the erosion of HPMC(cid:173)
`Gelucire core, which was also reflected in in vitro
`paddle-beads dissolution tests (Mehuys et al. , 2004). It
`has been shown previously in an in vitro study (Khan
`and Craig , 2003) that even though paracetamol was
`released from G50/13 through both erosional and dif(cid:173)
`fusional mechanisms , the latter was more dominant for
`freshly prepared matrices. The current study suggests
`that even though the in vitro release for the freshly
`prepared samples was the slowest due to the more diffu(cid:173)
`sional release , the hydrodynamic conditions in the gas(cid:173)
`trointestinal tract is sufficiently destructive to promote
`the faster erosional process, leading to the obscuration
`of any minor differences such as those seen between
`the freshly prepared samples and those stored at 4 °C.
`Hence , this led to the poor rank correlation between the
`MDT and MRT values of these two samples. However,
`this finding suggests that keeping gelucire containing
`preparations in a low temperature environment such
`as the refrigerator could ensure drug releases that are
`consistent with those prior to storage.
`
`5. Conclusion
`
`The extent of the structural supramolecular changes
`seen upon aging was greatly affected by the storage
`conditions, with storage temperature , which was close
`to the melting range of the unstable gelucire form caus(cid:173)
`ing the biggest modifications . The structure with the
`greatest polymorphic stability and crystallinity gave
`the highest rate of release. However, in practical terms ,
`such changes as detected by physical characterization
`
`COLL 1094092
`
`Purdue 2034
`Collegium v. Purdue, PGR2018-00048
`
`

`

`64
`
`
`
`
`
`Y.W. Choy et al. I International Journal of Pharmaceutics 299 (2005) 55-64
`
`
`
`References
`
`Acknowledgement
`
`and in vitro techniques did not cause major alterations
`
`Khan, N., Craig, D.Q.M., 2003. The influence of drug incorporation
`
`
`
`
`
`
`on the structure and release properties of solid dispersions in lipid
`
`
`to in vivo performance. This study indicates that the
`
`
`matrices. J. Contr. Rel. 93, 355-368.
`
`
`
`extent of bioavailability can be maintained at least for
`Khan, N., Craig, D.Q.M.,2004. Role of blooming in determining the
`
`
`
`
`
`
`hydrophilic drugs that are incorporated into polygly­
`of lipid based dosage forms. J. Phann. Sci. 93,
`storage stability
`
`
`
`colised glycerides, even when the lipid carriers have
`2962-2971.
`
`
`
`
`been subjected to drastic environmental changes. How­
`Kopcha, M., Lordi, N.G., Tojo, K.J., 1991. Evaluation of release
`
`
`
`
`
`
`from selected thermosoft