An understanding of modified release matrix tablets behavior duringdrug dissolution as the key for prediction of pharmaceutical productperformance – case study ofmultimodal characterization of quetiapinefumarate tablets

Piotr Kulinowski a,*, Krzysztof Woyna-Orlewicz b, Gerd-Martin Rappen c,Dorota Haznar-Garbacz d, Władysław P. Weglarz e, Przemysław P. Doro _zy�nski b

a Institute of Technology, The Pedagogical University of Cracow, ul. Podchora _zych 2, 30-084 Kraków, PolandbDepartment of Pharmaceutical Technology and Biopharmaceutics, Pharmaceutical Faculty, Jagiellonian University, ul. Medyczna 9, 30-688 Kraków, Polandc Physiolution GmbH, Walther-Rathenau-Strasse 49a, 17489 Greifswald, GermanydDepartment of Biopharmaceutics and Pharmaceutical Technology, Center of Drug Absorption and Transport (C_DAT), Felix-Hausdorff-Str. 3, 17487Greifswald, GermanyeDepartment of Magnetic Resonance Imaging, Institute of Nuclear Physics PAN, ul. Radzikowskiego 152, 31-342 Kraków, Poland

A R T I C L E I N F O

Article history:Received 6 December 2014Received in revised form 11 February 2015Accepted 16 February 2015Available online 18 February 2015

Keywords:Pharmaceutical generic productQuality by design (QbD)Magnetic resonance imaging (MRI)X-ray microtomography (mCT, Micro-CT)Texture analysisBiorelevant dissolution

A B S T R A C T

Motivation for the study was the lack of dedicated and effective research and development (R&D) in vitromethods for oral, generic, modified release formulations. The purpose of the research was to assessmultimodal in vitro methodology for further bioequivalence study risk minimization.Principal results of the study are as follows: (i) Pharmaceutically equivalent quetiapine fumarate

extended release dosage form of Seroquel XR was developed using a quality by design/design ofexperiment (QbD/DoE) paradigm. (ii) The developed formulation was then compared with originatorusing X-ray microtomography, magnetic resonance imaging and texture analysis. Despite similarity interms of compendial dissolution test, developed and original dosage forms differed in micro/mesostructure and consequently in mechanical properties. (iii) These differences were found to be the keyfactors of failure of biorelevant dissolution test using the stress dissolution apparatus.Major conclusions are as follows: (i) Imagingmethods allow to assess internal features of the hydrating

extended release matrix and together with the stress dissolution test allow to rationalize the design ofgeneric formulations at the in vitro level. (ii) Technological impact on formulation properties e.g., on poreformation in hydrating matrices cannot be overlooked when designing modified release dosage forms.

ã 2015 Elsevier B.V. All rights reserved.

1. Introduction

Medicinal products differ significantly from other ordinaryconsumer products. First of all, there are no medicines that arecompletely safe and patients cannot evaluate the risks and benefitsof their application. Therefore, the process of registration of drugs,also referred to as product authorization is a multistage, highlyregulated procedure that focuses on the quality, safety and efficacyof the drug product (Rägo, 2008).

The common technical document (CTD) that organizes themodern structure of registration dossiers is based on more than

fifty specific guidelines prepared and recommended by theInternational Conference of Harmonization (ICH), which describethe content and structure of particular sections of the drug productdossier. The pharmaceutical development – an important aspect ofdrug product quality – is characterized by the ICH Q8 (R2)guideline (ICH, 2009). The pharmaceutical development is focusedon designing “a quality product and its manufacturing process toconsistently deliver the intended performance of the product”(ICH, 2009). The guideline emphasizes the gaining of scientificknowledge necessary to ensure the highest possible quality of thedrug product and to establish the methods of its constant control.The concept of systematic approach to pharmaceutical develop-ment is called quality by design (QbD). It assumes that the qualityof drug products should not be tested post hoc, but should bedesigned and built into the product.

* Corresponding author. Tel.: +48 12 6626333; fax: +48 12 637 22 43.E-mail address: pkulino@up.krakow.pl (P. Kulinowski).

http://dx.doi.org/10.1016/j.ijpharm.2015.02.0400378-5173/ã 2015 Elsevier B.V. All rights reserved.

International Journal of Pharmaceutics 484 (2015) 235–245

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journal homepage: www.elsev ier .com/ locate / i jpharm

The most important aspects of the pharmaceutical develop-ment are the determination of the quality target product profile(QTPP) and identification of critical quality attributes (CQA) thataffect the product quality, safety and efficacy. Comprehensiveknowledge of the mechanistic relationship between processingparameters and drug CQA lead to reduction of variability andachievement of desired quality in a repeatable manner. Theoptimization process through understanding of technologicalattributes of drug product as well as structural and functionalconsequences of the composition andmanufacturing process givesobvious benefits for both, patients and industry, and may create abasis for reducing the overall risk and implementing flexibleregulatory approaches.

In 2004, the US Food and Drug Administration (US FDA)initiated the implementation of the process analytical technology(PAT) in the pharmaceutical industry through “GMPs for the 21stCentury” (Hinz, 2006; PAT, 2010). The PAT approach introduced anumber of tools for enabling scientific understanding of drugformulation. Since drug formulae are considered a complex,multifactorial systems, the use of statistical design of experiments(DoE) is recommended for studying the effects of process variableson drug product attributes.

The above-mentioned strategies of pharmaceutical develop-ment could be applied equally for innovative medicines as well astomultisource (generic) products. However, in the case of generics,QTTP is predefined by the quality of the originator.

The drug dissolution test is one of the most important toolsapplied in the pharmaceutical development procedures. In the lastfew decades, dissolution testing has become equally a routine toolfor quality control as well as a prerequisite for biopharmaceuticalcharacterization of different products (Dickinson et al., 2008). Thecompendial dissolution tests have a relatively simple constructionand provide well-definable conditions by implying continuousexposition of the dosage form to a sufficient amount of dissolutionmedium and mechanical agitation (Garbacz and Klein, 2012;Garbacz et al., 2010). The dissolution equipment represents highlystandardized tools for quality control and with appropriateexperimental settings the simulation of the physicochemicalconditions in the gastrointestinal tract (GIT) is also possible.However, these well definable and continuous conditions duringthe dissolution test do not reflect the physiological circumstancesalong the gastrointestinal (GI) tract. The design of the officialdissolution test apparatus does not provide the possibility ofsimulating GI mechanical stress conditions in a realistic way anddoes not reflect the volumes, discontinuous distribution and flowpatterns of the gastro-intestinal fluids (Garbacz and Klein, 2012;Schiller et al., 2005).

For modified release dosage forms the application of compen-dial dissolution methods that reliably allow comparing theformulations, brings additional challenges. It is known that therelease behavior of solid oral dosage forms during the GI transitmay be affected by physicochemical conditions and mechanicalstress (Garbacz and Klein, 2012). It has been recognized that the GItransit is characterized by highly variable conditions with long restphases and short but intensive events of transport (Weitschieset al., 2005). During GI transport events, dosage forms are movedwith high velocities of up to 30–50 cm/s for short periods. Suchintensivemovements aremainly triggered by gastric emptying andtransition through the ileocaecal junction as well as colonic massmovement. During GI transit, monolytic dosage forms such ascapsules or tablets are also exposed tomechanical pressure causedby GI motility events. Maximum pressures are registered in theantropyloric region of the stomach and reach up to 350mBar in thecase of monoliths like modified release tablets (Kamba et al., 2000;Kuo et al., 2008). It has already been demonstrated that the releasebehavior of modified release formulations can be affected by

mechanical stress in the GIT (Garbacz and Klein, 2012; Garbaczet al., 2010, 2008). Due to the complex physiology of the GIT,standard dissolution methods are not necessarily capable tosimulate realistically the GI transit conditions of solid oral dosageforms. By use of the bio-relevant dissolution methods, such as thestress test device, the impact of mechanical stress on drug releaseof modified release products can be investigated (Garbacz et al.,2014, 2010; Garbacz and Klein, 2012). The device provides thesimulation of essential physiological stress parameters includingdiscontinuous dosage formmovement and GI motility forces usingphysiology-based test algorithms. By this, the stress test device candemonstrate the mechanical conditions of the GI transit in arational way (Garbacz and Klein, 2012).

In physiological conditions, the structural and compositionalfactors play an important role in the drug release during thepassage throughout the whole gastrointestinal tract henceselection of appropriate methodology reflecting the physiologicalconditionsmay be problematic (Dickinson et al., 2008). The need ofscientific characterization of processes occurring during drugrelease frommodified release dosage forms induces the increasingdemand for techniques that could provide additional informationabout the mechanism of action of the dosage form and thetemporal changes of its properties during the drug delivery (Chenet al., 2010). Even biorelevant dissolution testing gives onlyindirect information concerning structural/morphological andphysicochemical properties of the modified release matrices. Forthis reason, during the last two decades, various, new analytic andimaging methods were introduced to investigate hydratedpolymeric matrices (Doro _zy�nski et al., 2012). It was shown thatthey have great potential, but they were not used as a tool for arational dosage form development so far. These methods, some ofthem destructive, were mainly used to study properties of modeldosage forms. Moreover, most of the methods have restrictionsconcerning size and shape of the matrix and their application isoften limited to the characterization of modified release mono-lithic dosage forms such as tablets and capsules. Some previouslyperformed studies on swelling dosage forms have drawn attentionto important new aspects of matrix properties and structureevolution during hydration e.g., tomatrix porosity (Karakosta et al.,2006; Laity and Cameron, 2010; Laity et al., 2010), potentialpresence of drug depletion zone (Chen et al., 2014), formulationdependent differences in physicochemical properties of matricesand different layer formation (Kulinowski et al., 2014).

Magnetic resonance imaging (MRI) and X-ray microtomogra-phy (micro-CT) can be used to study intact hydratedmatrix dosageform, virtually of any size and shape. Most of the MRI methods aresensitive towater proton properties inside the hydrated polymericmatrix – images reflect directly or indirectly molecular dynamicsand proton density (Doro _zy�nski et al., 2012; Mantle, 2011, 2013).MRI of dosage forms in flow-through cell was found to be a verypromising tool for evaluation of matrix systems (Doro _zy�nski et al.,2012). However, it inherently suffers from relatively low spatialresolution (0.2–0.5mm) (Chen et al., 2014; Kulinowski et al., 2011;Zhang et al., 2011). X-raymicrotomography offers better resolutionand mainly density based contrast, but cannot be performedduring dissolution (Laity and Cameron, 2010; Laity et al., 2010). Todate, only few studies were performed on commercial productsusing MRI and/or X-ray computed microtomography (micro-CT)(Chen et al., 2014; Doro _zy�nski et al., 2014; Kulinowski et al., 2011;Yin et al., 2013; Zhang et al., 2011).

The first and only micro-CT studies during matrix tablethydration combined with an MRI study was reported by Laityet al. (2010). In the subsequent work, the authors applied asynchrotron X-ray source to achieve shorter scan time and higherspatial resolution (Laity and Cameron, 2010). They studiedplacebo tablets composed of hydroxypropylmethylcellulose

236 P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245

(HPMC), microcrystalline cellulose (MCC) and lactose (Laity et al.,2010) or HPMC and MCC/pre-gelatinized starch (PGS) (Laity andCameron, 2010). X-raymicrotomography applied by Laity allowedthe observation of the lower density zone (Laity and Cameron,2010; Laity et al., 2010), as well as crack formation in the core ofthe matrix for some formulations (Laity and Cameron, 2010).Concerning the lower density zone, the presence of air micro-bubbles was suggested, but there was no direct evidence of theirexistence. The authors concluded that presence of air bubblesmight affect water penetration – the fact that was omittedpreviously when modeling processes occurring inside hydratedmatrices.

The understanding of the mechanisms of drug dissolution is aprerequisite for predicting in vivo performance of the dosage form.That is why, our previous works aimed at understanding andparameterization of the polymeric matrices (mainly HPMC basedmatrices) under hydration (Doro _zy�nski et al., 2014, 2011, 2010,2012; Kulinowski et al., 2008, 2011, 2012, 2014).

This paper is focused on some technological and analyticalaspects of reference and generic modified release drug productscomparison. The aim of the study was the development ofprocedures that allow comparing the reference modified releasedosage forms with the formulations intended to be used forbioequivalence studies.

The work was focused mainly on the identification ofmethodology necessary for effective development of modifiedrelease formulations. As an illustrative example, the developmentprocess of modified release formulationwith quetiapine fumarate,referenced to Seroquel XR medicinal product, was chosen.However, the development of a generic formulation was beyondthe scope of the study.

Quetiapine (QTP) is an atypical, biopharmaceutical classifica-tion system (BCS) class II, antipsychotic compound, indicated fortreatment of schizophrenia and bipolar disorders (DeVane andNemeroff, 2001). Quetiapine has an excellent risk/benefit profileand is a suitable first-line option for the treatment of schizophrenia(Cheer andWagstaff, 2004). Quetiapine is commonly available as afumarate salt and is usedmainly as amodified release formulation.The formulation is intended to release the drug in a controlled waywith the aim to increase compliance of schizophrenia patients andto reduce side effects.

The goals of the study have been completed in a multistageprocedure that covered:

� development of the formulation using design of experiment(DoE) methodology,

� characterization and comparison of reference vs. developed(generic) formulations with several methods (multimodalstudy), i.e., X-ray microtomography, MR imaging includingimaging in a USP4 apparatus and measurement of mechanicalrobustness of hydrated tablets,

� analysis of the influence of the structural characteristics on drugrelease using dissolution stress test apparatus in conditions thatmimics in vivo GIT conditions (Garbacz et al., 2008).

2. Materials and methods

2.1. Materials

In the present work the following materials were used:Seroquel XR 400mg (AstraZeneca), quetiapine fumarate (IpcaLaboratories Ltd., India), hydroxypropylmethylcellulose (HPMC)Metolose type 2910 of viscosity 50mPas and type 2809 ofviscosities 100 and 4000mPas (Shin-Etsu, Japan), Tablettose 80(Meggle, Germany), Vivapur type 102 (JRS Pharma, Germany),

Aerosil 200 Pharma (Evonik, Germany),Magnesium stearate (POChS.A., Poland), Sodium Citrate (POCh S.A., Poland). All othermaterials used in the study were of analytical grade.

2.2. Design of experiment and tablet preparation

The formulation development was carried out according to themethodology of statistical experimental design using a (3,3)simplex-lattice design for mixtures implemented in Statistica 9.0,quality control, design of experiments package (StatSoft Inc., USA).Three grades of HPMC were selected, i.e., 2910 type of viscosity50mPa s and type 2809 of viscosities 100 and 4000mPa s ascomponents of the matrix controlling drug release. The HPMCquantities were applied at four different levels. According to thestudy plan, ten laboratory batches of tablets of different HPMCgrades and quantities were prepared. The tablets were preparedafter wet granulation. Quetiapine fumarate and sodium citratewere mixed evenly (Erweka Cube Mixer KB15S, Germany) andloaded into a low-shear granulator (LK5 Erweka, Germany). Thepowders were wetted with 15% hypromelose 6 cP solution andmassed for 10min under high speed of agitation. The granulatewasdried in a tray drier (KCW 100, Premed, Poland) until 2.0–2.5% losson drying was achieved and then screened through a 1.0mmmesh(Erweka Wet Granulator FGS, Germany). At the next stage, thegranulate was blended with the extra-granular excipients in thecube mixer and compressed into oblong bi-convex tablets of20mm length (Korsch Pressen 103, Germany).

2.3. Dissolution studies and comparison of dissolution profiles

The dissolution studies were carried out using USP apparatus II(Hanson SR 8 Plus, Hanson Research Corp., USA) in 1000mL offollowing media: 0.1 N HCl (pH 1.2), USP acetate buffer (pH 4.5),USP phosphate buffer (pH 6.8) and distilled water – the stirringratewas 50 r.p.m. at a temperature of 37 �C�0.5 �C. The duration ofthe study was 12h. The samples (5mL volume) were withdrawn at0.5,1, 2, 4, 6, 8,10 and 12h. The sample volumewas replaced by theblank medium. The amount of the drug dissolved was determinedspectrophotometrically at 289nm (V-530 spectrophotometerJasco, Japan). The tests were performed for n =6 parallels in eachmedia. The dissolution profiles were compared to the referenceproduct by model independent approach using similarity (f2)factor (EMEA, 2010; Shah et al., 1998).

2.4. Time resolved MRI and dissolution study in USP apparatus 4

Time resolved MRI in a flow-through cell was performed asdescribed in Kulinowski et al. (2008, 2011),) using an MRI researchsystem consisting of a 4.7 T superconducting magnet (Bruker,Germany), actively shielded gradient coil set of 290mm ID, digitalTMX console (IBD NRC, Canada) and a non-magnetic setup fordissolution/imaging with a USP4 flow-through cell. Followingparameters of a spin echo (SE) sequence were used: field of view(FOV) =35�35mm, echo time (TE) = 19ms, repetition time (TR) =625ms, slice thickness 1mm, matrix size of 256�256. The drugrelease studies were performed in a closed loop configurationusing 1000mL of distilled water as dissolution medium. Themedium was circulated with a flow rate of 40mL/min, thetemperature in the dissolution cell wasmaintained at 37 �C�0.5 �Cusing a thermostatic water bath (LW 502M, AJL Electronic, Poland).The sampling schedule and analytical procedures were identical tothe methodology described above for the dissolution study in USPapparatus II. The images were acquired after the mediumcirculation was temporarily stopped.

P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245 237

2.5. MRI studies at the single time point

The tablets were placed in a flow-through cell in the speciallydesigned holder. The conditions of experiment in terms of flowspeed, temperature and solution type were identical to theconditions established for the dissolution study experiment withapparatus 4. After 2h, the holders were withdrawn from the flowthrough cell and placed in a 9.4 T Bruker Biospin magnet. For thestudy purpose multi-slice-multi-echo (MSME) sequence wasapplied, using the following sequence parameters: number ofechoes (NE) =64, echo time (TE) = 10ms, TR=1.5 s, FOV=30�30mm, matrix size of 256�256.

2.6. X-ray microtomography (micro-CT)

Microtomograms of dry tables and tablets at 2 h of hydrationwere obtained. The procedure of tablet hydration was identical tothe procedure described above for MRI experiments (seeSection 2.5). Hydrated samples withdrawn from the flow-throughcell were placed in the Benchtop X-Tek CT 160Xi (Nikon MetrologyInc.) X-ray microtomograph. The number of acquired projectionswas 3010, the acquisition time was 17.8min and the spatialresolution of the method was 16.9mm. The image reconstructionwas performed with Avizo 6.3 (VSG – Visualization SciencesGroup) software. The images were processed and analyzed usingpublic domain software Fiji (Schindelin et al., 2012), distribution ofImageJ v.1.44. Pore segmentation was performed using histogram-based segmentation (manual thresholding), after Gaussian blur-ring (sigma 1.7). Pore/grain/swollen polymer segmentation for T5formulation was performed using a Trainable Weka Segmentationmodule. Visualization and pore space renderingwas donewith theVolume Viewer plug-in.

2.7. Texture analysis of hydrated tablets

The samples of the tablets were hydrated in 30mL of distilledwater for 2h at 37 �C. For the study a Compact Tabletop, UniversalTester EZ-SX (Shimadzu, Japan) equipped with a 500NSX load celland a type B toothed push rod was used. The machine operatedunder Trapezium X software (Shimadzu, Japan). The hydratedtablets were fixed on the lower metal plate and the test wasinitiated at a crosshead speed of 0.5mm/min. The load wasdirected perpendicularly to the long axis of the tablet. Themeasurement duration was 150 s at sampling time of 0.01 s.

2.8. Drug release studies using dissolution stress test apparatus

The dissolution stress test was performed using a devicedescribed in details in Garbacz et al. (2008). The arrangement ofthe dissolution stress test procedure is presented in Fig. 1.

The gastric residence was imitated as an incubation phase in thesimulated gastric fluid (SGF)without pepsin (pH 1.8) for one hour. Astress phase of high intensity composed of three consecutive

inflations of the balloons with a duration of 6 s and fortitude of300mBar mimicking peristaltic pressure waves, was arranged tosimulate gastric emptying. These pressure events were followed by1min of rotation of the apparatus axle with 100 rpm mimickingaccelerated tablet transport as observed during gastric emptying(GarbaczandKlein,2012;Garbaczetal., 2010). Subsequently, 50mMphosphate buffer (pH 6.8) supplied with 0.1% Tween 80was used tobuildup the conditions in the gastric intestinal tract. Themechanicalagitationof the intestinal passagewas simulatedas three sequences,each consisting of two pressure waves of 150mBar fortitude,followed by 30s rotatorymovementwith 50 rpmat 2, 3 and 4h. Thepurpose of the stress sequenceswas the improvement of themixingconditions inside the probe chamber as well as the check of theintegrity and robustness of the swollen matrices. The imitation ofileocecalpassageof thedosage formswasdoneat5hinthesamewaylike gastric emptying (Garbacz et al., 2010, 2008). For a roughmimicryof thecolonpassagetwostressphases, identicalwithgastricemptying, were executed at 9 and 12h.

3. Results

3.1. Development and selection of formulation for further studies

In the early development stage for the study’s purpose tenformulations were prepared. The comparison of the test andreference formulations was performed in the USP apparatus II inthree standardmedia (pH 1.2, 4.5, 6.8) and inwater. The dissolutionprofiles were compared to reference formulation using a f2similarity factor (Shah et al., 1998). For further studies only theformulation T5 was chosen, which fulfilled criteria of similarity inall testing media (f2 value above 50). The qualitative andquantitative composition of T5 formulation is presented in Table 1.

[(Fig._1)TD$FIG]

Fig. 1. Overview of dissolution testing procedure in the bio-relevant dissolution stress test device.

Table 1Composition of T5 formulation.

No. Substance Amount(%)

Amount (mg/tablet)

Function

Granulate1. Quetiapine fumarate 52.2 459.4 Active2. 2910 HPMC 6cP 3.0 26.4 Binder3. Sodium citrate 13.0 114.4 pH

modifier4. Water purified q.s.a -/-

Extra-granular excipients5. 2910 HPMC 50mPa s 8.3 73.4 CR agent6. 2809 HPMC 100mPa s 0.0 0.0 CR agent7. 2809 HPMC 4000mPas 16.7 146.6 CR agent8. Lactose monohydrate 2.3 20.2 Filler9. Cellulose microcrystalline

type 1023.0 26.4 Filler

10. Colloidal silica dioxide 0.5 4.4 Glidant11. Magnesium stearate 1.0 8.8 Lubricant

In total 100.0 880.0

a Removed during drying of granules.

238 P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245

The T5 tablets contained ca. 70% of granulate. The low-shearprocess resulted in formation of large and dense granules, which isclearly visible in sieve analysis results, i.e.,ca. 50% agglomeratesretained with a 0.8mm sieve. The extra-granular fraction iscomposed mainly of water-soluble controlled release hypromel-lose (25%).

The results of dissolution studies are presented in Fig. 2. Due tothe alkaline nature of quetiapine the highest cumulative concen-trations of active substance were released in acidic media – at 12hcomplete drug dissolution from T5 formulation was observed(Fig. 2A), 93% of the drug was released from Seroquel XR(reference). In pH 4.5 and 6.8 buffers at 12h 61% and 48% of thedrug was released, respectively (Fig. 2B and C). At the same time60% and 51% of QTP was released from the reference product. Drugrelease studies in water (Fig. 2D) also showed similarity ofcumulative dissolution profiles.

3.2. Simultaneous MRI and dissolution study in USP apparatus 4

The dissolution of quetiapine fumarate was incomplete inapparatus 4 for both formulations (Fig. 3). At 12h, 47% of QTP wasreleased from T5 formulation and from Seroquel XR only 41% ofactive substance was released.

The MR imaging showed that up to 4h, the referenceformulation formed three distinct layers of different watermolecular properties (see Figs. 4A and 5A ). For details of MRIcharacteristics of the reference product see thework byKulinowskiet al. (2011). Therefore, a three-layer morphology of the hydratedmatrix was also expected for the developed T5 formulation. But it

turned out that the multilayer pattern in case of T5 formulationwas different. Apparently, up to four layers could be observed inthe MR images at the corresponding hydration times (see Figs. 4Band 5B). An additional layer of lower image intensity implies alower proton density (water concentration) in this part of thematrix, regardless to the fact that this part of thematrix is expectedto be fully hydrated.

[(Fig._2)TD$FIG]

A B

CD

Fig. 2. Comparative dissolution profiles (n =3) using an USP apparatus II: A – 0.1M HCl pH=1.2, B – acetate buffer pH=4.5, C – phosphate buffer of pH=6.8, D – water(reference – blue diamonds, T5 – orange squares). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

[(Fig._3)TD$FIG]

Fig. 3. Comparative dissolution profiles (n =3) using an USP apparatus IV in water(reference – blue diamonds, T5 – orange squares). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version ofthis article.)

P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245 239

Moreover, T5 matrix promotes water ingression – at 2h ofhydration, thewhole volume of thematrix is hydrated. At the samehydration time the reference matrix retains its core.

Obtained results, strongly imply different characteristics of thematrices under hydration in termsofmicro/meso-structure (whichreflect physicochemical properties) as well as mechanical proper-ties. To clarify the nature of observed phenomena, we appliedadditional imaging methods at 2h of hydration. At this time point,the internal structure of both formulations was most compound.

3.3. Characteristics of the structure of hydrated matrices at 2 h withhigh field MRI, micro-CT and texture analysis

The first appliedmethodwasmulti-echoMR imaging techniqueat 9.4 T.

Results obtained for samples measured directly in the flow-through cell at 4.7 T (Figs. 4 and 5) and removed from the cell asobtained at 9.4 T (Fig. 6) are consistent, but MR images obtained at9.4 T revealmore details. As for imaging inside the USP apparatus 4,multilayer structure of the hydrated reference matrix is evident.The hydrated part of the matrix consists of two layers of differentconstitutions (see Fig. 6A). Moreover, a narrow line of lower imageintensity suggests existence of pores (cracks) between layers insidethe hydrated part of the matrix (see grayscale fragment in Fig. 6A).If existing, the pores cannot be resolved unambiguously due toMRIresolution. The first echo was acquired at 10ms, and even at thefirst echo image of MSME data set there is no measurable signal inthe core of the matrix. It implies no hydration or minimalhydration of the core. When comparing qualitatively withreference one, it can be observed that the T5 matrix promotessolvent penetration – at 2h the whole volume of the matrix ishydrated (Figs. 6B and 5B). The images acquired at 4.7 T in USPapparatus 4 revealed a zone of lower image intensity inside thefully hydrated part of the matrix. Images obtained at 9.4 T withslightly higher spatial resolution (117mm vs. 137mm at 4.7 T) andshorter echo time (10ms vs.19ms at 4.7 T), suggest the presence oflarge amount of air in the form ofmicrocracks/microbubbles of sizelower or comparable with pixel size. As in case of reference dosageform, single voids cannot be resolved unambiguously.

Magnetic resonance imaging at 9.4 T gives some hints related tothe presence and the nature of air voids inside both matrices. Theonly indisputable fact is that the only origin of voids can be airincorporated within the matrix during manufacturing (compres-sion, granulation). That is why X-ray microimaging was alsoapplied. Due to the mainly density-based contrast, X-ray micro-CTis well suited to detect voids in the matrix, providing highresolution images (Karakosta et al., 2006; Laity and Cameron, 2010;Laity et al., 2010).

Therefore, a somewhat different structure aspect of thehydrated reference matrix can be observed by micro-CT

[(Fig._4)TD$FIG]

Fig. 4. Temporal changes of reference (A) and T5 (B) matrices as studied by MRI inUSP4 flow-through cell (longitudinal, coronal cross sections).

[(Fig._5)TD$FIG]

Fig. 5. MRI in USP4 flow-through cell – details of micro/mesostructure at 1 and 2h for reference (A) and T5 (B) formulations: 1 – dissolution medium, 2 – gel-like layer, 3 –

interface (swollen glassy) layer, 4 – matrix core. Question marks denote ambiguities in region identification.

240 P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245

(Figs. 7C and D, 8A and C, 9 ). The only region, which matchesmatrix morphology obtained by MRI at 2h is core (for comparisonsee Fig. 6A). The image intensity (related to the density of thematrix) of the core region is slightly lower than external, hydratedpart of the matrix. Additionally, a narrow zone of higher densityaround the core of the matrix can be observed. At this scale(resolution of 16.9mm), external, hydrated region of the matrix aswell as internal (core) part of the matrix are relatively homoge-nous, which suggests an intraregional homogenous density of thematrix. Two regions, as observed in MR images, cannot bedistinguished by micro-CT in the hydrated part of the matrix.

Micro-CT reveals that density of the hydrated part of the matrix isuniform, but water properties as reflected in MR images changespatially, resulting in layered structure of hydrated region.

One of the most interesting observations presented in thearticle, is detection of three types of pores (voids inside the

[(Fig._7)TD$FIG]

Fig. 7. Longitudinal (coronal) micro-CT cross sections of reference (A, C, E) and T5(B, D, F) tablets – the results are presented for drymatrices (A and B) andmatrices at2h of hydration (C, E, D, F): 1 – crack inside the hydrated part of the matrix, 2 – corecrack, 3 – quetiapine grain with adjacent short cracks, 4 – air bubbles.

[(Fig._6)TD$FIG]

Fig. 6. Longitudinal (coronal) MR cross section images of reference and T5 tabletsover 2h of hydration with echo time of 10ms.

[(Fig._8)TD$FIG]

Fig. 8. Micro-CT pore rendering for reference (A and C) (9.78�9.87�4.24mm) and T5 (B and D) (11.02�10.46�4.24mm) formulation in selected axial slice of 4.24mm.

P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245 241

hydrated matrix). Inside the reference matrix following types ofpores/voids can be distinguished:

� internal cracks in the core of the matrix (Figs. 8C and 9A).� air bubbles in the hydrated part of the matrix (Fig. 8A).� macro cracks in the hydrated part of the matrix (Figs. 8C and 9B).

These pores differ in size and shape, and generally cannot bedirectly observed by MRI, due to the limited resolution (16.9mmformicro-CT vs. 137mmand 117mm for 4.7 Tand 9.4 TMR imaging,respectively). The only exception was cracks inside the core of thematrix – they were visible clearly at longer hydration times (>5h)(see Fig. 3) as described by Kulinowski et al. (2011). At 2h, in MRimages, cracks in the core of the matrix were not detected – thecore remained “dry”. But core internal cracks as detected bymicro-CT suggest that the core should be minimally hydrated, which isconsistent with our results obtained for model HPMC matricesusing MR microscopy (Kulinowski et al., 2012). The planes of corecracks are roughly parallel to compression direction. Cracks insidethe hydrated part of thematrix form, sometimeswide, planesmoreor less perpendicular to the compression direction (see Figs. 8C and9B). Only sparse air bubbles can be observed in the hydrated part ofthe matrix.

On the contrary, for T5 formulation, X-ray microtomographyreveals, that the whole volume of the matrix is highly inhomoge-neous. As is indicated in Figs. 7D and F and 10B , at 2 h of hydration,distinct quetiapine grains are visible in the whole volume ofmatrix: in the central part of the matrix with some short cracksbetween grains as well as in highly hydrated, external part of thematrix. In the external part of the matrix, grains are surrounded byhydrated polymer. This region also contains a high concentration ofair voids (bubbles), which can be distinguished in Figs. 8B and Dand 10A. These bubbles are responsible for the presence of a zoneof low signal intensity observed in MR images.

Additionally to micro-CT of hydrated matrices, microtomo-grams of dry matrices are presented in Fig. 7A and B. In T5formulation quetiapine granules are clearly visible, while referenceproduct is homogenous (strips of slightly lower intensity in theimages are artifacts originating from holder).

The results of the texture analysis at 2 h of hydration are givenin Fig 11. In the case of reference formulation a progressive increaseof the force required for penetration of the hydrated tablet up to22.5Nwas observed (then the tabletswere broken). It indicated thepresence of the hard solid core within the tablet. For formulationT5, only small increase (�0.8N) of the force necessary forpenetration of the push rod was observed – in this case errorbars were lower than line width of the graph (solid orange line).

3.4. Biorelevant drug release using dissolution stress test apparatus

The results of the dissolution stress test experiments are shownin Fig. 12. The dissolution profiles of both formulations wereslightly affected by the simulated gastric emptying and the pHchange of the dissolution media. The stress events of low intensitysimulated at 2 and 3h resulted in an increase in the dissolution rateof the test formulation. The tablet’s matrices disintegrated atapproximately 5h and the dissolution was completed. In the case

[(Fig._9)TD$FIG]

Fig. 9. Micro-CT pore rendering for reference dosage form: A – core cracks (3.23�2.31�7.93mm), B – cracks inside the hydrated part of the matrix (7.27�3.06�7.93mm).

[(Fig._10)TD$FIG]

Fig. 10. Micro-CT air bubbles (A) and grains (B) rendering for T5 for a fragment ofthe matrix (0.42�3.28�4.61mm).

[(Fig._11)TD$FIG]

Fig. 11. Mechanical strengthof reference and T5 tablets (n =3) after 2 h of hydration(reference – blue line, reference error bar envelope – dashed blue line, T5 – orangeline). (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

242 P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245

of the reference product the dissolution characteristics did notchange up to 5h. The stress phase simulated at 5h provoked anaccelerated dissolution of approximately 20% of the drug andresulted in a subsequent increase in the dissolution rate. The stressphase simulated at 9h caused disintegration of the referenceproduct and an accelerated dissolution of approximately 20% of thedrug (i.e., 80mg).

4. Discussion

In a case of modified release formulations drug releasecharacteristics is a synergetic effect of number of factors such asdrug and excipients properties, solvent type, solvent hydrody-namic condition and technological factors. Modified releasematrices evolve during hydration, changing their structure, size,shape, mechanical resistance and what is most important, theirinternal physicochemical parameters. Due to thematrix hydration,discrete structures can be formed, which determine both thewateruptake and drug/polymer dissolution rates. The properties of thehydrated part of the tablet matrices are often influenced by thesolubility of the active pharmaceutical ingredient (API) and can bemodified by diverse excipients. Examples can be found inKulinowski et al. (2014), where layer formation was dependenton API solubility. The understanding of the hydrating matrixproperties and release mechanisms is essential for a successfulproduct development. For this reason, it is necessary to clearlydistinguish between: (i) biopharmaceutical features of themodified release dosage form, which could be obtained in thedissolution experiments and (ii) the intrinsic physicochemical,structural and morphological properties of the matrix itselfobtained using various imaging and analytical techniques.

In this studywe showed that both kinds ofmatrix properties areequally important and the intrinsic properties of the modifiedrelease matrices explain their biopharmaceutical properties.

4.1. Intrinsic properties of hydrated matrix

Magnetic resonance images, especially those obtained insidethe flow-through cell (USP apparatus 4) during dissolution testshould be very cautiously used for interpretation of the matrix(dosage form) behavior because of the relatively low sensitivityand low resolution (typically in-plane resolution 100–500mm at a

slice thickness of 1mm or more) (Chen et al., 2014; Kulinowskiet al., 2011; Zhang et al., 2011). On the other hand, MRI results areinvaluable. The contrast is sensitive to proton concentration andproton mobility and allows for dosage form imaging in vitro evenunder flow conditions. Therefore, other characterization techni-ques aswell as their combinations thereofmay be useful in order toprovide detailed insight intomatrix morphology and drug deliverymechanisms. In our case these are multi-echo magnetic resonanceimaging, X-ray microimaging and texture analysis. These methodscan be used complementary at representative time points of thedissolution test (in our case at 2h of hydration) to visualize thedifferences among the formulations. Moreover, some advancedMRI techniques allow quantitative measurements (relaxometry,diffusometry, magnetization transfer) as reported by Doro _zy�nskiet al. (2012).

X-ray microtomography is an imaging technique complemen-tary to MRI. It provides other contrast mechanisms and muchhigher resolution. It allowed mainly for pore space characteriza-tion. Pores inside the hydrated matrix, as detected by X-raymicrotomography, should modify both, the water penetration andthe drug diffusion. It is very likely that the wide, plane pores(cracks) between hydrated layers in reference formulation shoulddecrease water penetration kinetics. On the other hand, bubbles inT5 formulation can modify drug diffusion from the matrix due tolonger diffusion path. Our results are supported by works of Laityet al. (Laity and Cameron, 2010; Laity et al., 2010), where theauthors concluded that the presence of bubbles appears to havebeen overlooked in previous attempts tomeasure diffusion rates inthe swollenmatrix. The authors only hypothesized the presence ofbubbles inside the hydrated part of the matrix. They observed aregion of lower image intensity (density), but they did notobserved bubbles directly. Interestingly, despite the high concen-tration of air bubbles in T5 formulation it promotes solventpenetration. The high volume of gas inclusions in the externalhydrated part of thematrix reduce themechanical resistance of thegel layers and increase their susceptibility towards mechanicalstress of biorelevant intensity and consequently dose dumping.Reduction of mechanical resistance of T5 formulation wasconfirmed by texture analysis (see Fig. 11). Our results are inaccordance with the conclusions by Laity: pores and voids that arecreated in hydrated part of the matrix should have consequencesfor a priori (theoretical) modeling, which were never taken intoconsideration (Laity and Cameron, 2010).

Imaging and texture analysis results show that the biopharma-ceutical performance of modified release products is determinedby the composition but also by the manufacturing technology. Themanufacturing technology highly influences the structural, mo-lecular and mechanical matrix properties/parameters. It is evidentcomparing micro-CT images of dry matrices and after 2h ofhydration (see Fig. 7). Consequently, manufacturing technologyalso influences the in vivo drug delivery characteristics, which wasindirectly shown using in vitro dissolution stress test. Therefore,influence of technology on thematrix properties during hydration/dissolution cannot be neglected during formulation development.In our particular case, low shear wet granulation resulted indifferent internal structure and properties of the dry and hydratedT5matrix comparing to reference pharmaceutical product. Despitethese differences, it was possible to adjust the formulation tomatch the compendial in vitro dissolution characteristics of thereference product.

Influence of technology on the structure and physicochemicalproperties of hydrating modified release formulations was neverstudied before using imaging techniques. The results of our workalso show that ideas originating from Colombo and Bettini works(Bettini et al., 2001; Colombo et al., 1999), i.e., fronts and regions,

[(Fig._12)TD$FIG]

Fig. 12. Comparative dissolution profiles (n =3) using the bio-relevant dissolutionstress test apparatus with dissolution procedure presented in Fig. 1 (reference –

blue diamonds, T5 – orange squares). (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)

P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245 243

work well when dealing with model formulations, however,cannot always be easily applied to real ones.

4.2. Characterization of biopharmaceutical properties

Application of general R&D procedures for modified releasegeneric dosage form design, as presented in the first part of ourstudy, resulted in the development of a product characterized bysimilar in vitro drug release behavior/characteristics under well-defined conditions of compendial dissolution test. Despite thesimilarity of reference and developed T5 formulation in terms ofstandard criteria, striking differences were observed with imaging(MRI, micro-CT) and texture analysis based approach. Afterinspection of the developed T5 formulation using MRI, micro-CTand texture analysis, the biorelevant dissolution test using a stressdissolution apparatus allowed for a final assessment of developedproduct quality. Unlike the compendial dissolution study, results ofstress dissolution test are consistent with imaging and textureanalysis.

It leads to the conclusion that the design of genericformulations solely based on results of the compendial dissolutiontesting can lead to potential bioequivalence pitfalls. The reasonthereof is most probably related to the specifics of the standarddissolution apparatuses, which ability to simulate the GI physiol-ogy is very limited. Dissolution tests provide indirect informationconcerning intrinsic matrix properties. Using complementary(imaging/analytical) techniques it is possible to assess internalproperties of modified release matrices and provide guidelines forrational redesign or optimization of the MR formulation.

4.3. R&D process rationalization

Applying the integrated approach we can, with a highprobability, indicate a risk of bioincompatibility of the testedproduct in vivo. The study results show that there is a strong needto develop novel research and development strategies adequate formodified release dosage forms. Our paper presents a possiblesolution to this problem. According to Chen et al. (2010),understanding of the formulation gives the possibility torationalize the R&D process. A rational way of formulationcomparison should be proposed for each specific type offormulation (composition + technology). It is necessary to chooseproper imaging and analytical methods. In some cases, theevolution of the layers in the matrix, as proposed by Kulinowskiet al. (2011), can provide useful information. But in other cases, forexample when grain size and its distribution are critical factors,other ways of matrix parameterization should be established.Similar idea of rational approach to pharmaceutical productdevelopment using QbD and different analytical methods includ-ing pore assessment with scanning electronmicroscopy (SEM)waspresented very recently by Sauri et al. (2014).

Returning to our current results, we hypothesize, that whendesigning generic formulation, in vitro stress dissolution testtogether with magnetic resonance imaging, micro-CT and othertechniques can be discriminative enough to check whetherformulations are similar in terms of compositions, morphologyof hydrating matrix and their physicochemical properties. Thisinformation may effectively support the rational formulationdevelopment. Each of the introduced methods provides uniqueinsight into the structure and properties of formulation andcontributes towards better product characterization.

5. Conclusions

Advantages of in vitro pharmaceutical studies are obvious. Theygive results not disturbed by physiological factors and can be

performed at relatively low cost compared to the biological studiesi.e., bioequivalence test. They offer the possibility to test theformulation in a systematic way, free from in vivo inter-subjectvariability. The main goal of the pharmaceutical availability studyin the generic drug development process is theminimization of thebioequivalence risk. To ensure it, due to the lack of dedicated andeffective R&D in the area of in vitro methods for oral, generic,modified release formulations, we have introduced a multimodal,experiment-based approach. It assumes strong separation be-tween biopharmaceutical and intrinsic properties of the hydratingmatrix. Intrinsic matrix properties include structural and physico-chemical properties (porosity, specific layer formation, moleculardynamics) and can be studied using various imaging and analyticalmethods. Assessment of intrinsic matrix properties allowsexplaining its biopharmaceutical properties.

The proposed workflow of the R&D process consists of threesteps: (i) DoE using compendial dissolution test; followed by (ii)application of several, selected, imaging/analytical methods(magnetic resonance imaging, X-ray microtomography andtexture analysis) and finally by (iii) biorelevant stress dissolutiontest. The complete path, or their fragments, can be repeatediteratively. We are convinced, that the proposed methodologyallows for bioequivalence test risk reduction. Besides presentingnovel approach to generic modified release dosage forms R&D,several (new) important aspects concerning the study of matrixformation during hydration were touched upon, i.e., (i) criticalevaluation of application of MRI and micro-CT imaging methods,(ii) importance of technological aspect that was never addressedthis way, (iii) detection of three types of pore inside the hydratedmatrices.

Acknowledgments

The work was supported by the Polish Ministry of Science andHigher Education grant NN518 407438. The project was co-financed by the European Regional Development Fund under theInfrastructure and Environment Program UDA-POIS.13.01-023/09-00.

G.-M. Rappen and D. Haznar-Garbacz would like to thank to theGerman Federal Ministry of Education and Research for thefinancial support (BMBF FKZ 03IPT612C).

We gratefully acknowledge to Jolanta Klaja (Oil and GasInstitute, National Research Institute, Kraków, Poland) for assis-tance with micro-CT, to Marta Zakrzewska for technical assistancein preparation of tablets and to Marco LH Gruwel (NRC-CNRCAquatic and Crop Resource Development, Winnipeg, Canada) forcritical reading of the manuscript.

References

Bettini, R., Catellani, P., Santi, P., Massimo, G., Peppas, N., Colombo, P., 2001.Translocation of drug particles in HPMC matrix gel layer: effect of drugsolubility and influence on release rate. J. Control. Release 70, 383–391.

Cheer, S.M., Wagstaff, A.J., 2004. Quetiapine: a review of its use in the managementof schizophrenia. CNS Drugs 18, 173–199.

Chen, M.L., Shah, V.P., Ganes, D., Midha, K.K., Caro, J., Nambiar, P., Rocci Jr., M.L.,Thombre, A.G., Abrahamsson, B., Conner, D., Davit, B., Fackler, P., Farrell, C.,Gupta, S., Katz, R., Mehta, M., Preskorn, S.H., Sanderink, G., Stavchansky, S.,Temple, R., Wang, Y., Winkle, H., Yu, L., 2010. Challenges and opportunities inestablishing scientific and regulatory standards for determining therapeuticequivalence of modified-release products: workshop summary report. Clin.Ther. 32, 1704–1712.

Chen, C., Gladden, L.F., Mantle, M.D., 2014. Direct visualization of in vitro drugmobilization from Lescol XL tablets using two-dimensional 19F and 1Hmagneticresonance imaging. Mol. Pharm. 11, 630–637.

Colombo, P., Bettini, R., Peppas, N.A., 1999. Observation of swelling process anddiffusion front position during swelling in hydroxypropyl methyl cellulose(HPMC) matrices containing a soluble drug. J. Control. Release 61, 83–91.

DeVane, C.L., Nemeroff, C., 2001. Clinical pharmacokinetics of quetiapine. Clin.Pharm. 40, 509–522.

244 P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245

Dickinson, P.A., Lee, W.W., Stott, P.W., Townsend, A.I., Smart, J.P., Ghahramani, P.,Hammett, T., Billett, L., Behn, S., Gibb, R.C., Abrahamsson, B., 2008. Clinicalrelevance of dissolution testing in quality by design. AAPS J. 10, 380–390.

Doro _zy�nski, P.P., Kulinowski, P., Mendyk, A., Młynarczyk, A., Jachowicz, R., 2010.Novel application of MRI technique combined with flow-through celldissolution apparatus as supportive discriminatory test for evaluation ofcontrolled release formulations. AAPS PharmSciTech 11, 588–597.

Doro _zy�nski, P., Kulinowski, P., Mendyk, A., Jachowicz, R., 2011. Gastroretentive drugdelivery systems with L-dopa based on carrageenans andhydroxypropylmethylcellulose. Int. J. Pharm. 404, 169–175.

Doro _zy�nski, P.P., Kulinowski, P., Młynarczyk, A., Stanisz, G.J., 2012. MRI as a tool forevaluation of oral controlled release dosage forms. Drug Discov. Today 17,110–123.

Doro _zy�nski, P., Kulinowski, P., Jamróz, W., Juszczyk, E., 2014. Geometry of modifiedrelease formulations during dissolution – influence on performance of dosageforms with diclofenac sodium. Int. J. Pharm. 477, 57–63.

EMEA, 2010. Guideline on the Investigation of Bioequivalence. In: EM Agency (Ed.),European Medicines Agency, London.

Garbacz, G., Klein, S., 2012. Dissolution testing of oral modified-release dosageforms. J. Pharm. Pharmacol. 64, 944–968.

Garbacz, G., Wedemeyer, R.S., Nagel, S., Giessmann, T., Monnikes, H., Wilson, C.G.,Siegmund,W.,Weitschies,W., 2008. Irregular absorptionprofiles observed fromdiclofenac extended release tablets can be predicted using a dissolution testapparatus that mimics in vivo physical stresses. Eur. J. Pharm. Biopharm. 70,421–428.

Garbacz, G., Klein, S., Weitschies, W., 2010. A biorelevant dissolution stress testdevice – background and experiences. Expert Opin. Drug Deliv. 7, 1251–1261.

Garbacz, G., Kandzi, A., Koziolek, M., Mazgalski, J., Weitschies, W., 2014. Releasecharacteristics of quetiapine fumarate extended release tablets underbiorelevant stress test conditions. AAPS PharmSciTech 15, 230–236.

Hinz, D.C., 2006. Process analytical technologies in the pharmaceutical industry: theFDA’s PAT initiative. Anal. Bioanal. Chem. 384, 1036–1042.

ICH, 2009. ICH Harmonised Tripartite Guideline, Pharmaceutical Development.International Conference on Harmonisation of Technical Requirements forRegistration of Pharmaceuticals for Human Use ICH Q8 (R2).

Kamba,M., Seta, Y., Kusai, A., Ikeda,M., Nishimura, K., 2000. Aunique dosage form toevaluate the mechanical destructive force in the gastrointestinal tract. Int. J.Pharm. 208, 61–70.

Karakosta, E., Jenneson, P.M., Sear, R.P., McDonald, P.J., 2006. Observations ofcoarsening of air voids in a polymer-highly-soluble crystalline matrix duringdissolution. Phys. Rev. E 74, 11504.

Kulinowski, P., Doro _zy�nski, P., Jachowicz, R., Weglarz, W.P., 2008. An integratedsystem for dissolution studies and magnetic resonance imaging of controlledrelease, polymer-based dosage forms – a tool for quantitative assessment ofhydrogel formation processes. J. Pharm. Biomed. Anal. 48, 685–693.

Kulinowski, P., Doro _zy�nski, P., Młynarczyk, A., Weglarz, W.P., 2011. Magneticresonance imaging and image analysis for assessment of HPMC matrix tabletsstructural evolution in USP apparatus 4. Pharm. Res. 28, 1065–1073.

Kulinowski, P., Młynarczyk, A., Doro _zy�nski, P., Jasi�nski, K., Gruwel, M.L., Tomanek, B.,Weglarz, W.P., 2012. Magnetic resonance microscopy for assessment ofmorphological changes in hydrating hydroxypropylmethyl cellulose matrixtablets in situ. Pharm. Res. 29, 3420–3433.

Kulinowski, P., Młynarczyk, A., Jasi�nski, K., Talik, P., Gruwel, M.L., Tomanek, B.,Weglarz, W.P., Doro _zy�nski, P., 2014. Magnetic resonance microscopy forassessment of morphological changes in hydratinghydroxypropylmethylcellulose matrix tablets in situ – is it possible to detectphenomena related to drug dissolution within the hydrated matrices? Pharm.Res. 31, 2383–2392.

Kuo, B., McCallum, R.W., Koch, K.L., Sitrin, M.D., Wo, J.M., Chey, W.D., Hasler, W.L.,Lackner, J.M., Katz, L.A., Semler, J.R., Wilding, G.E., Parkman, H.P., 2008.Comparison of gastric emptying of a nondigestible capsule to a radio-labelledmeal in healthy and gastroparetic subjects. Aliment. Pharm. Ther. 27, 186–196.

Laity, P.R., Cameron, R.E., 2010. Synchrotron X-raymicrotomographic study of tabletswelling. Eur. J. Pharm. Biopharm. 75, 263–276.

Laity, P.R., Mantle, M.D., Gladden, L.F., Cameron, R.E., 2010. Magnetic resonanceimaging and X-ray microtomography studies of a gel-forming tabletformulation. Eur. J. Pharm. Biopharm. 74, 109–119.

Mantle, M.D., 2011. Quantitative magnetic resonance micro-imaging methods forpharmaceutical research. Int. J. Pharm. 417, 173–195.

Mantle, M.D., 2013. NMR and MRI studies of drug delivery systems. Curr. Opin.Colloid Interface Sci. 18, 214–227.

PAT, 2010. Process Analytical Technology: Spectroscopic Tools and ImplementationStrategies for the Chemical and Pharmaceutical Industries, 2nd ed. John Wileyand Sons.

Rägo, B.S., 2008. Drug regulation: history, present and future, In: van Boxtel, C.J.,Santoso, B., Edwards, I.R. (Eds.), Drug Benefits and Risks: International Textbookof Clinical Pharmacology, Revised International Textbook of ClinicalPharmacology. 2nd ed. IOS Press, pp. 65–77.

Sauri, J., Millan, D., Sune-Negre, J.M., Colom, H., Tico, J.R., Minarro, M., Perez-Lozano,P., Garcia-Montoya, E., 2014. Quality by design approach to understand thephysicochemical phenomena involved in controlled release of captopril SRmatrix tablets. Int. J. Pharm. 477, 431–441.

Schiller, C., Frohlich, C.P., Giessmann, T., Siegmund, W., Monnikes, H., Hosten, N.,Weitschies, W., 2005. Intestinal fluid volumes and transit of dosage forms asassessed by magnetic resonance imaging. Aliment. Pharmacol. Ther. 22,971–979.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T.,Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.Y., White, D.J.,Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: an open-sourceplatform for biological-image analysis. Nature Methods 9, 676–682.

Shah, V.P., Tsong, Y., Sathe, P., Liu, J.P., 1998. In vitro dissolution profile comparison –

statistics and analysis of the similarity factor, f2. Pharm. Res. 15, 889–896.Weitschies, W., Kosch, O., Monnikes, H., Trahms, L., 2005. Magnetic marker

monitoring: an application of biomagnetic measurement instrumentation andprinciples for the determination of the gastrointestinal behavior ofmagnetically marked solid dosage forms. Adv. Drug Deliv. Rev. 57, 1210–1222.

Yin, X., Li, H., Guo, Z., Wu, L., Chen, F., de Matas, M., Shao, Q., Xiao, T., York, P., He, Y.,Zhang, J., 2013. Quantification of swelling and erosion in the controlled releaseof a poorly water-soluble drug using synchrotron X-ray computedmicrotomography. AAPS J. 15, 1025–1034.

Zhang, Q., Gladden, L., Avalle, P., Mantle, M., 2011. In vitro quantitative 1H and 19Fnuclear magnetic resonance spectroscopy and imaging studies of fluvastatin inLescol(R) XL tablets in a USP-IV dissolution cell. J. Control. Release 156, 345–354.

P. Kulinowski et al. / International Journal of Pharmaceutics 484 (2015) 235–245 245