- Email: [email protected]
Development of an Extended-Release Formulation of Capecitabine Making Use of In Vitro–In Vivo Correlation Modelling
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Development of an Extended-Release Formulation of Capecitabine Making Use of In Vitro–In Vivo Correlation Modelling JELTE MEULENAAR,1,2 RON J. KEIZER,1,2 JOS H. BEIJNEN,1,3 JAN H. M. SCHELLENS,2,3 ALWIN D. R. HUITEMA,2 BASTIAAN NUIJEN1 1
Department of Pharmacy and Pharmacology, Slotervaart Hospital, Amsterdam, The Netherlands Department of Clinical Pharmacology, The Netherlands Cancer Institute, Amsterdam, The Netherlands 3 Department of Pharmaceutical Sciences, Faculty of Science, Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht University, Utrecht, The Netherlands 2
Received 6 June 2013; revised 18 September 2013; accepted 16 October 2013 Published online 5 December 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23779 ABSTRACT: An oral extended-release (ER) formulation of capecitabine was developed for twice daily dosing, theoretically providing a continuous exposure to capecitabine, thus avoiding the undesirable in-between dosing gap inherent to the dosing schedule of the marketed capecitabine immediate-release formulation (Xeloda ). The target 12-hour in vivo release profile was correlated to an in vitro dissolution profile using an in vitro–in vivo correlation model based on the pharmacokinetic (PK) and dissolution characteristics of Xeloda . Making use of the slow dissolution characteristics of amorphous capecitabine as reported previously and screening of a panel of ER excipients, an ER formulation was designed. Kollidon SR induced the most prominent ER. Moreover, it was shown that tablets prepared from CoSD capecitabine and Kollidon SR have an additional threefold delay in dissolution compared with tablets prepared from the same but only physically mixed components. Therefore, a prototype tablet formulation composed of co-spray-dried capecitabine and Kollidon SR (98/2%, w/w) mixed with colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w) was defined. This prototype shows similar dissolution characteristics as the modelled dissolution profile. Currently, the in vivo PK of our designed ER capecitabine formulations C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 103:478–484, is investigated in a clinical study. 2014 Keywords: capecitabine; amorphous; co-spray drying; extended release; IVIVC modelling R
R
R
R
R
INTRODUCTION Capecitabine (Fig. 1) is an orally administered chemotherapeutic agent used in the treatment of various tumour types including metastatic breast and colorectal cancer and is commercially available as an immediate-release tablet (Xeloda ).1 Capecitabine is a pre-prodrug that is enzymatically converted into 5-fluorouracil (5-FU) in the tumour, where it inhibits DNA synthesis and slows down growth of tumour tissue. To form 5-FU, capecitabine is activated by a three-step enzymatic pathway with two intermediary metabolites, 5 -deoxy5-fluorocytidine and 5 -deoxy-5-fluorouridine (Fig. 2). The absorption of capecitabine and its subsequent conversion into 5-FU is rapid (Cmax of 5-FU 60 min). Subsequently, 5-FU is cleared rapidly and follows the pharmacokinetic (PK) profile of capecitabine. 5-FU is undetectable in plasma after approximately 6 h.2–4 On the basis of this PK profile, it can be concluded that the approved twice daily oral dosing (morning– evening) schedule of capecitabine leaves an in-between dosing capecitabine-exposure gap of approximately 6 h, which is argued to be clinically undesirable.5,6 Indeed, several clinical studies have shown that, as a consequence of the prolonged exposure and reduction of the 5-FU peak plasma concentration, a continuous infusion instead of a bolus injection of 5-FU inR
Abbreviations used: 5-FU, 5-fluorouracil; 5 -DFCR, 5 -deoxy-5-fluorocytidine; 5 -DFUR, 5 -deoxy-5-fluorouridine; CoSD, co-spray-dried; Q50 the time to 50% dissolution of the active component. Correspondence to: Jelte Meulenaar (Telephone: +31-20-512-47-31; Fax: +3120-512-47-53; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 103, 478–484 (2014)
C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association
478
Figure 1. Chemical structure of capecitabine (Mw 359.35 g/mol).
creases the anti-tumour response rate and decreases toxicity in terms of a reduced occurrence and grade of adverse events.5,6 Therefore, to translate this concept to an oral dosing scheme, an oral extended-release (ER) dosage form of capecitabine is warranted. For most active pharmaceutical ingredients (APIs), the development of an ER formulation is straightforward. Various excipients are commercially available (e.g., polyvinyl acetate (PVAc)–polyvinyl pyrrolidone (PVP) co-polymer (Kollidon SR)7 and the methyl cellulose and ethylene oxide8 which, by simply mixing with the API in a certain ratio followed by, for example direct compression into tablets, results in ER of the compound of interest. However, an important factor is the amount of ER excipient required to obtain the desired release profile, in particular for high-dosed APIs as because of size little space per dose unit is available for additives. For example, Xeloda consists of oblong-shaped tablets containing 150 mg (5.4 × 11.5 mm2 ) or 500 mg (8.4 × 15.8 mm2 ) of capecitabine at a
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
R
R
479
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 2. (a) Schematic overview of the pathway of a DS from administration to the site of action (tumour). (b) IVIVC model for capecitabine. 5-DFCR: 5-deoxy-5-fluorocytidine; 5-DFUR: 5-deoxy-5-fluorouridine; 5-FU: 5-fluorouracil.
total weight of 187.5 and 625 mg, respectively, corresponding to 80% (w/w) API and only 20% (w/w) of excipients. Given the standard dosing regimen up to 1250 mg/m2 capecitabine twice daily, it is easily understood that from a patient compliance perspective this leaves limited space for additional excipients, which would result in either an increase in size or number of tablets that should be taken. Previously, we reported the anomalous dissolution behaviour of amorphous capecitabine which, in comparison with crystalline capecitabine used in Xeloda , displays unexpected slow and highly reproducible dissolution behaviour.9 Making use of this finding, we describe in this article the in vitro development of a prototype ER formulation for a twice-daily dosing (b.i.d.) schedule providing a continuous exposure to capecitabine. For this, simulations from a previously published human population PK model of capecitabine10 were used to obtain a theoretically optimal dissolution curve. This in vitro–in vivo correlation (IVIVC) model was subsequently used to screen a panel of ER excipients to identify the most efficient dissolution delaying excipient (i.e., extent of release in time in relation to the quantity of excipient used). R
MATERIALS AND METHODS Materials Capecitabine drug substance (DS), manufactured by Jiangsu ZW Pharmaceuticals Company (Changzhou, Jiangsu, China) was supplied by Dolder AG (Basel, Switzerland). For the screening of the sustained-release excipients, various polymer grades of methyl cellulose (MethocelTM K3 LV, K100 LV CR, K4M CR K15M CR, K100M CR), poly (ethylene oxide) (PolyoxTM 301 and coagulant) were kindly provided by Colorcon (Dartford, Kent, United Kingdom). PVAc–PVP copolymer (Kollidon SR) was kindly provided by BASF (Limburgerhof, Germany). Xeloda was purchased from Roche Laboratories Inc. (Nutley, New Jersey). All other excipients and solR
R
DOI 10.1002/jps.23779
vents (Ph.Eur./analytical grade) were originated from commercial suppliers. Methods
ER Capecitabine Formulation Screening Physical Mixtures. Physical mixtures (PMs) were prepared by mixing capecitabine (crystalline DS or amorphous) and polymers (80/20%, w/w) in a gallipot with a TURBULA –T 10 (Willy A. Bachofen AG – Maschinenfabrik, Muttenz, Switzerland) for 10 min. R
Spray-Drying and Co-Spray-Dried Formulations. The (co-) spray-dried (CoSD) powder of capecitabine DS was produced ¨ with a Buchi Mini Spray Dryer B-290, equipped with an Inert Loop B-295, Dehumidifier B-296, High Performance cyclone, 1.5 mm nozzle cap, and 0.7 mm nozzle tip (BUCHI Labortechnik AG, Flawil, Switzerland) using the following settings: feed 15%; N2-atomization flow 40 mm; aspirator flow 80%; inlet temperature 18◦ C; outlet temperature 18◦ C; inert loop temperature −20◦ C. Spray-drying solution consisted of capecitabine DS (300 mg/ mL) and Kollidon SR, dissolved (at 45◦ C) in 100% ethanol (HPLC grade). Kollidon SR was added at different ratios. R
R
Tableting Tablets (9 mm Ø, resistance to crushing 60 N) were pressed using an amount of formulation mixture equivalent to 224 mg capecitabine on a Korsch EK0 eccentric press (Korsch AG, Berlin, Germany). Tablet thickness was related to the volume and compaction properties of the tableting mixture. Crushing strength was determined using an Erweka TBH20 hardness tester (Erweka, Heusenstamm, Germany). For the prototype formulation, the CoSD powder (capecitabine Kollidon SR 98/2%, w/w) was mixed with 0.5% (w/w) colloidal silicon dioxide and 2.5% (w/w) magnesium stearate. Using this tableting R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
480
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
mixture, oblong tablets (8.4 × 15.8 mm2 , 100–140 N) containing 500 mg capecitabine were pressed.
Table 1.
Dissolution Testing. Dissolution testing was performed according to United States Pharmacopeia monograph using a type 2 (paddle) dissolution apparatus (Erweka) and water for Injections (B. Braun, Melsungen, Germany) at 37◦ C, stirred at 50 rpm, as medium.11 Samples collected at the various time points were filtrated using a 0.45-:m filter and subsequently analysed on a reversed-phase HPLC–UV (RP-HPLC–UV) system: column: Varian, Inertsil ODS-3, 15 × 4.6 mm2 ; 5 :m, Eluent methanol–water (7:3, v/v), flow: 1.0 mL/min, detection: 310 nm. The retention time of capecitabine in this RP-HPLC-UV system was about 3.5 min. The dissolution test was continued to 100% dissolution or until a constant dissolution percentage. Also, the dissolution profile of Xeloda (150–500 mg) was determined.
MethocelTM K3 LV MethocelTM K100 LVCR MethocelTM K4M CR MethocelTM K15M CR MethocelTM K100M CR PolyoxTM 301 PolyoxTM coagulant KollidonR SR KollidonR SR (CoSD) References Amorphous capecitabine XelodaR (Roche)a
R
IVIVC Model Development. Figure 2 shows a graphical overview of the IVIVC model used. The PK model previously published by Urien et al.10 developed for Xeloda was adapted to include an additional dissolution step prior to the absorption process. For this, the well-established Higuchi dissolution equation12,13 was converted to a differential equation describing the dissolution rate (Eqs. 1–3):
Time to 50% Dissolution (Q50 )
Polymer
Q50 (min) 53 271 638 701 838 666 739 1556 4966 45 11
Amorphous capecitabine (224 mg) tablets containing 20% (w/w) of different ER polymers (n = 3/formulation): methyl cellulose (MethocelTM K3 LV, K100 LV CR, K4M CR K15M CR, K100M CR); poly (ethylene oxide) (PolyoxTM 301 and coagulant); PVAc–PVP co-polymer (Kollidon SR). a Xeloda contains crystalline capecitabine.7 R
R
R
√ A(%) = a t + b
(1)
A(%) = at1/2 + b
(2)
1 dA(%) = at−1/2 = Rdis dt 2
(3)
where a is the slope of the found line, t is the time, Rdis is the dissolution equation and A is the percentage dissolved capecitabine as function of time. In the original model developed by Urien et al.,10 the absorption rate constant ka described both dissolution and absorption rate. To be able to introduce formulations with different dissolution characteristics, the true absorption rate constant kad was introduced, solely describing the absorption process. As original PK data were not available to estimate this parameter, simulations were used to approximate this parameter. For this, Eq. (3) was implemented combined with the determined Higuchi constant of 17.8 for Xeloda . Subsequently, the kad parameter was used to simulate different curves for various ER formulations with different Higuchi constants. The complete IVIVC model was described as a system of ordinary differential equations describing the mass transport between the different compartments as shown in Figure 2. This model was implemented in NONMEM (version VI 2.0; Icon Development Solutions, Ellicott City, Maryland) for simulation purposes.14 R
Modulated Temperature Differential Scanning Calorimetry. Modulated temperature differential scanning calorimetry (MTDSC) measurements were performed with a discovery differential scanning calorimeter (DSC) (TA Instruments, New Castle, Delaware). Temperature scale and heat flow were calibrated with indium. Samples of approximately 3 mg powder were weighed into Tzero aluminium pans (TA instruments), sealed and placed in the autosampler. Each sample was equilibrated at −30.00◦ C for 5 min, after which the sample was Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
heated to 125◦ C for capecitabine and 250◦ C for Kollidon SR at a speed of 2◦ C/min with a modulated temperature amplitude of 1◦ C per period of 60 s. R
X-ray Powder Diffraction. X-ray powder diffraction (XPRD) measurements were performed with an X’pert pro diffractometer equipped with an X-celerator (PANanalytical, Almelo, The Netherlands). Samples were placed in a 0.5-mm deep metal sample holder which was placed in the diffractometer. Samples were scanned at a current of 30–50 mA and a tension of 40 kV. The scanning range was 10◦ –60◦ 22, with a step size of 0.020◦ and a scanning speed of 0.002◦ /s.
RESULTS IVIVC Modelling As capecitabine is highly soluble over the whole physiological pH range as well as highly permeable, it is classified as a BCS class I drug.15–17 This means that no significant absorption delay as a consequence of solubility or permeability issues are to be expected after oral administration of capecitabine. Therefore, the release characteristics of the pharmaceutical dosage form will be the main variable determining capecitabine absorption during gastrointestinal transit. In addition to this, the availability of a well-described human PK model as well as the known dissolution characteristics of capecitabine from Xeloda served as an excellent starting point for the development of an IVIVC model. Using this model, the curve which approached the desired 12-h in vivo exposure profile for capecitabine best, proved to be an ER formulation with a Higuchi constant of 4.5. In Figure 3, the respective plasma concentration–time curve and corresponding dissolution profile are depicted. R
Screening of ER Polymers Next, several ER polymers were screened for their potential of obtaining this target capecitabine dissolution profile (Table 1). Spray-dried, amorphous capecitabine was used for the screening experiment, having already slow-release characteristics in itself and thus likely requiring less release-modifying excipient DOI 10.1002/jps.23779
481
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 3. Capecitabine plasma concentration–time and corresponding dissolution curves , capecitabine (XelodaR ) concentration–time curve [Urien model (10)]; , XelodaR (500 mg) dissolution curve (experimental data); , simulated 12-h capecitabine exposure concentration–time curve; , simulated 12-h capecitabine dissolution curve (Higuchi constant of 4.5).
◦
•
to meet the target release profile, favourable in view of the final size and feasibility of the high-content capecitabine dosage form.9 Polymers from structurally different chemical groups and suitable for direct compression into monolithic matrices were selected, mixed with amorphous capecitabine, pressed into tablets and subjected to dissolution testing. From Table 1, it can be seen that PVAc–PVP co-polymer (Kollidon SR) induced the most prominent ER effect with an approximately 141.5-fold and 34.5-fold delay in dissolution compared with Xeloda and amorphous capecitabine, respectively. On the basis of these results, it was investigated whether capecitabine could be CoSD with Kollidon SR. This might have potential advantages in terms of a reduction of processing (mixing) steps, which is favourable in view of upscaling. Also, the addition of Kollidon SR (having mixture Tg ’s of 47◦ C and 161◦ C determined by MTDSC on a single batch purchased Kollidon SR, data not shown), because of its various components gives a, although limited, Tg elevation (CoSD Tg of 32◦ C, data not shown) as compared with 100% amorphous capecitabine (Tg of 19◦ C 9 . Theoretically, this Tg elevation may result/be reflected in an increased physical stability. Moreover, and unexpectedly, it was shown that tablets prepared from CoSD capecitabine and Kollidon SR have an additional threefold delay in dissolution compared with tablets prepared from the same but only physically mixed components (Table 1). R
R
R
R
R
R
Optimization of Percentage Kollidon SR in the CoSD Mixture R
On the basis of these results, Kollidon SR, when CoSD with capecitabine, was identified as the most promising ER polymer to reach the target release profile using as less excipient as possible. To examine this further, in vitro dissolution as a function of the percentage Kollidon SR in CoSD capecitabine Kollidon SR mixtures was assessed. For comparison, dissoR
R
R
DOI 10.1002/jps.23779
lution curves of physically mixed crystalline (DS) and amorphous capecitabine Kollidon SR mixtures were recorded as well. Figure 4 gives the results of these experiments, and clearly shows the significant dissolution delay of the CoSD material as compared with the PMs of both crystalline and amorphous capecitabine using the same percentages (w/w) of Kollidon SR. Moreover, it was shown that the target Higuchi constant of 4.5 can be reached by adding only a minimal quantity of Kollidon SR to the CoSD mixture. The percentages of Kollidon SR required to obtain a dissolution curve with a Higuchi constant of 4.5 were 1.2%, 7.3%, and 19.2% (w/w) for the CoSD, amorphous capecitabine PM, and crystalline PM, respectively (Fig. 4). R
R
R
R
Final Prototype Formulation As only a minimal quantity of Kollidon SR was needed in the CoSD capecitabine Kollidon SR mixture, this was selected for further development into a prototype formulation. Aiming for tablets as the preferred dosage form, at least a glidant and lubricant need to be added to the CoSD capecitabine Kollidon SR mixture. The CoSD capecitabine Kollidon SR mixture by itself was found to be of high density and good compressibility compared with the crystalline capecitabine Kollidon SR PM (data not shown). Therefore, experiments were performed to investigate whether the addition of colloidal silicon dioxide and/or magnesium stearate to the CoSD powder affects the release characteristics of capecitabine. Colloidal silicon dioxide did not show any release-modifying effects (data not shown) but magnesium stearate, however, significantly accelerated capecitabine release. Adding a standard percentage of 2.5% (w/w) of magnesium stearate to a tableting mixture composed of CoSD capecitabine Kollidon SR (99/1% w/w) and colloidal silicon dioxide 0.5% (w/w), the time to 50% dissolution (Q50 ) of capecitabine from the tablets dropped from 4.5 to 2 h. R
R
R
R
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
482
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 4. Dissolution as a function of the percentage KollidonR SR in CoSD capecitabine KollidonR SR. Dissolution curves of 224 mg capecitabine tablets (n = 3) were recorded and expressed as their Higuchi constants. , Physically mixed crystalline capecitabine DS and KollidonR SR; , physically mixed amorphous capecitabine and KollidonR SR; , CoSD capecitabine KollidonR SR. The insert shows the full dissolution curves of tablets containing 10% (w/w) KollidonR SR of the respective formulations.
◦
•
Figure 5. Dissolution of the ER capecitabine prototype formulation as compared with the target dissolution curve with a Higuchi constant of 4.5. ◦, Dissolution curve with Higuchi dissolution constant 4.5 (dotted lines show the f2 interval for similarity of dissolution 16 ; , In vitro dissolution curve (n = 3) of the ER capecitabine prototype formulation with 2.0% KollidonR SR. NB: the error bars are showing the variability in dissolution performance for the tested tablets. The insert shows the dissolution curves of the tablets containing increasing amounts (0%–2.5%) of KollidonR SR. , 0%; , 0.5%; , 1.0%; , 1.5%; , 2.0%; , 2.5%.
•
•
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
DOI 10.1002/jps.23779
483
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 6. X-ray diffraction patterns of PMs of crystalline capecitabine (1) amorphous capecitabine (2) and KollidonR SR, CoSD capecitabine and KollidonR SR (3) and a powdered prototype tablet after storage over 11 months at −20◦ C, consisting of CoSD capecitabine and KollidonR SR (4). Note: The small peaks at 43◦ –45◦ are caused by the scattering of the metal sample holder.
To match the release characteristics to the Higuchi constant of 4.5 that we aimed for, dissolution curves were recorded of CoSD capecitabine Kollidon SR mixtures (range of 0.5%–2.5%, w/w, Kollidon SR) in the presence of 0.5% colloidal silicon dioxide and 2.5% (w/w) magnesium stearate (insert Fig. 5). From these experiments, a percentage of 2.0% (w/w) Kollidon SR in the CoSD mixture was found to potentially approach the target release profile. Figure 5 shows the theoretical dissolution profile with a Higuchi constant of 4.5 and the experimental dissolution curve of the formulation containing 2.0% (w/w) Kollidon SR in the CoSD mixture. The similarity factor (f2 ) for this dissolution curve was found to be 52 and therefore considered similar to the theoretical curve (specification: 50 < f2 < 100).18 On the basis of this result, the prototype formulation selected for further development is composed of CoSD capecitabine Kollidon SR (98/2%, w/w), colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w). R
R
R
R
R
R
R
R
R
R
R
DISCUSSION As pharmaceutical development mainly involves in vitro experiments, it is highly useful to have some guidance on the correlation with in vivo characteristics when developing an ER dosage form for a specific compound. For capecitabine, its BCS class 1 classification, the availability of a PK model as well as the known dissolution characteristics of Xeloda offered the opportunity for developing an IVIVC model. This was performed by a relatively simple combination of two well-established models: the Higuchi dissolution model and a compartmental population PK model. Simulations using this model resulted in a to-develop dosage form having release characteristics with a Higuchi constant of 4.5 (Fig. 3). R
DOI 10.1002/jps.23779
From the screening experiments, Kollidon SR induced the most prominent ER effect. This slow release of capecitabine can be attributed to the insoluble PVAc component of Kollidon SR that forms an inert hydrophobic matrix after hydration. The increasing length of the diffusion pathway over time is responsible for the ER of capecitabine. In contrast to Kollidon SR, the methyl cellulose and poly(ethylene oxide) are gel-forming hydrophilic colloids. Because of the erosion of the gel formed upon hydration, the diffusion pathway is limited and dissolution is mainly governed by the viscosity of the polymer (e.g., dissolution rate from MethocelTM K100M CR < MethocelTM K4M CR). Moreover, co-spray drying of capecitabine and Kollidon SR appeared to result in an additional, significantly delayed dissolution as compared with the physically mixed individual components. Apparently, the Kollidon SR matrix responsible for the add-on ER characteristics of amorphous capecitabine is even more pronounced when being part of a homogeneous, amorphous structure (Fig. 6). Indeed, only a minor amount (1.2%, w/w) of Kollidon SR was required to obtain the target release profile. The addition of magnesium stearate, however, resulted in an unexpected and significant acceleration of the dissolution rate of CoSD capecitabine Kollidon SR tablets. As reported previously, the slow-release characteristics of amorphous capecitabine are attributed to an instant gelling (collapse) upon contact with water.9 It is likely that, by covering the capecitabine Kollidon SR particles with a thin, hydrophobic film, magnesium stearate interferes with this gelling property. We speculate that in the presence of magnesium stearate the resulting gel surface is not as tight and leaves space for capecitabine to dissolve more quickly. When adding a slightly higher concentration of matrix former (2%, w/w, Kollidon SR), however, this issue was resolved. R
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
484
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
The prototype capecitabine ER formulation comprises only a very limited amount of slow-release excipient (2%, w/w). For comparison, the excipient monograph of the manufacturer advises quantities of Kollidon SR from 40% to 55% (w/w) to be added to a freely soluble API to obtain ER drug profiles.7 Indeed, the low amount of Kollidon SR in the prototype ER formulation is highly attractive in terms of clinical feasibility of the final dosage form. The currently marketed immediaterelease dosage form of capecitabine (Xeloda ) is composed of 80% (w/w) of API and 20% (w/w) of excipients. As the prototype ER formulation consists of 95% (w/w) of API and 5% (w/w) of excipients, it is clear that no limitations are to be expected with respect to size of the final dosage form. However, one issue which must be addressed is the stability of the final dosage form that shows the low Tg (32◦ C, data not shown) of the CoSD capecitabine Kollidon SR. Previously, the reluctance of amorphous capecitabine to crystallize was shown with a 48h-isothermal differential scanning calorimetric analysis at a temperature (80◦ C) far exceeding the Tg of amorphous capecitabine.9 Further, Figure 6 shows that the XPRD pattern of the powdered prototype tablet (11 months stored at −20◦ C) is equal to the freshly prepared CoSD capecitabine and thus an amorphous system is present. Additional, the DSC analysis ran over −25◦ C to 125◦ C 10◦ C/min of that powdered prototype tablet showed a Tg at 32◦ C but lacks a melting peak (data not shown). The absence of a melting peak in this DSC run is an extra and further indication that the required amorphous system for the defined release characteristics of the prototype formulation is fully intact. Taking these results into account, the statement can be made that the preliminary stability results show that the tablet formulation is chemically (HPLC, data not shown) and physically stable for at least 11 months when stored at −20◦ C.
istics and the applicability of the IVIVC model used, the in vivo PK of this formulation is investigated in a clinical study.
R
R
R
R
CONCLUSIONS An oral ER formulation of capecitabine was developed for twice daily dosing, theoretically providing a continuous exposure to capecitabine and 5-FU thus avoiding the undesirable in-between dosing gap inherent to the dosing schedule of the marketed capecitabine immediate-release formulation (Xeloda ). To achieve this aim, the target 12-h in vivo release profile was correlated to an in vitro dissolution profile using an IVIVC model based on the PK and dissolution characteristics of Xeloda . Making use of the slow dissolution characteristics of amorphous capecitabine as reported previously 9 and screening of a panel of ER excipients, a prototype tablet formulation composed of CoSD capecitabine and Kollidon SR (98/2%, w/w) mixed with colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w) was defined. To confirm its ER characterR
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
REFERENCES 1. Roche Laboratories Inc. 2011. Product information XelodaR . Accessed April 29, 2013, at: www.roche-australia.com/fmfiles/re7229005/ downloads/oncology/xeloda-pi.pdf 2. Walko CM, Lindley C. 2005. Capecitabine: A review. Clin Ther 27:23– 44. 3. Pentheroudakis G, Twelves C. 2002. Capecitabine (Xeloda): From the laboratory to the patient’s home. Clin Colorectal Cancer 2:16–23. 4. Schellens JHM. 2007. Capecitabine. Oncologist 12:152–155. 5. Seifert P, Baker LH, Reed ML, Vaitkevicius VK. 1975. Comparison of continuously infused 5-fluorouracil with bolus injection in treatment of patients with colorectal adenocarcinoma. Cancer 36:123–128. 6. Meta-analysis Group In Cancer. 1998. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol 16:301–308. 7. BASF SE Care Chemicals Division Pharma Ingredients & Services. 2008. Technical information KollidonR SR (BASF). EMP 030728e-07:1– 12. 8. Colorcon. Methocel product application data sheet. Accessed August 8, 2013, at: http://www.colorcon.com/resources/lit/Products. 9. Meulenaar J, Beijnen JH, Schellens JHM, Nuijen B. 2013. Slow dissolution behaviour of amorphous capecitabine. Int J Pharm 441:213– 217. 10. Urien S, Reza´ı K, Lokiec F. 2005. Pharmacokinetic modelling of 5-FU production from capecitabine—A population study in 40 adult patients with metastatic cancer. J Pharmacokinet Pharmacodyn 32:817– 833. 11. Marques MRC, Mao F. 2011. USP34–NF29 capecitabine tablets. The United States Pharmacopeia–National Formulary 2148–2148. 12. Paul DR. 2011. Elaborations on the Higuchi model for drug delivery. Int J Pharm 418:13–17. 13. Higuchi T. 1963. Mechanism of sustained-action medication. Theoretical analysis of rate of release of solid drugs dispersed in solid matrices. J Pharm Sci 52:1145–1149. 14. Beal SL, Sheiner LB. 1989. NONMEM users guides. Ellicott City, Maryland: Icon Development Solutions. ¨ H, Shah V, Crison J. 1995. A theoretical 15. Amidon G, Lennernas basis for a biopharmaceutic drug classification: The correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm Res 12:413–420. 16. Reigner B, Verweij J, Dirix L, Cassidy J, Twelves C, Allman D, Weidekamm E, Roos B, Banken L, Utoh M, Osterwalder B. 1998. Effect of food on the pharmacokinetics of capecitabine and its metabolites following oral administration in cancer patients. Clin Cancer Res 4:941–948. 17. Blesch KS, Gieschke R, Tsukamoto Y, Reigner BG, Burger HU, Steimer JL. 2003. Clinical pharmacokinetic/pharmacodynamic and physiologically based pharmacokinetic modeling in new drug development: The capecitabine experience. Invest New Drugs 21:195–223. 18. EMEA Committee for Medicinal Products for Human Use. 2008. Guideline on the investigation of bioequivalence. Guideline CPMP/EWP/QWP/1401/98 Rev. 1:1–29.
DOI 10.1002/jps.23779
Development of an Extended-Release Formulation of Capecitabine Making Use of In Vitro–In Vivo Correlation Modelling JELTE MEULENAAR,1,2 RON J. KEIZER,1,2 JOS H. BEIJNEN,1,3 JAN H. M. SCHELLENS,2,3 ALWIN D. R. HUITEMA,2 BASTIAAN NUIJEN1 1
Department of Pharmacy and Pharmacology, Slotervaart Hospital, Amsterdam, The Netherlands Department of Clinical Pharmacology, The Netherlands Cancer Institute, Amsterdam, The Netherlands 3 Department of Pharmaceutical Sciences, Faculty of Science, Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht University, Utrecht, The Netherlands 2
Received 6 June 2013; revised 18 September 2013; accepted 16 October 2013 Published online 5 December 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23779 ABSTRACT: An oral extended-release (ER) formulation of capecitabine was developed for twice daily dosing, theoretically providing a continuous exposure to capecitabine, thus avoiding the undesirable in-between dosing gap inherent to the dosing schedule of the marketed capecitabine immediate-release formulation (Xeloda ). The target 12-hour in vivo release profile was correlated to an in vitro dissolution profile using an in vitro–in vivo correlation model based on the pharmacokinetic (PK) and dissolution characteristics of Xeloda . Making use of the slow dissolution characteristics of amorphous capecitabine as reported previously and screening of a panel of ER excipients, an ER formulation was designed. Kollidon SR induced the most prominent ER. Moreover, it was shown that tablets prepared from CoSD capecitabine and Kollidon SR have an additional threefold delay in dissolution compared with tablets prepared from the same but only physically mixed components. Therefore, a prototype tablet formulation composed of co-spray-dried capecitabine and Kollidon SR (98/2%, w/w) mixed with colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w) was defined. This prototype shows similar dissolution characteristics as the modelled dissolution profile. Currently, the in vivo PK of our designed ER capecitabine formulations C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 103:478–484, is investigated in a clinical study. 2014 Keywords: capecitabine; amorphous; co-spray drying; extended release; IVIVC modelling R
R
R
R
R
INTRODUCTION Capecitabine (Fig. 1) is an orally administered chemotherapeutic agent used in the treatment of various tumour types including metastatic breast and colorectal cancer and is commercially available as an immediate-release tablet (Xeloda ).1 Capecitabine is a pre-prodrug that is enzymatically converted into 5-fluorouracil (5-FU) in the tumour, where it inhibits DNA synthesis and slows down growth of tumour tissue. To form 5-FU, capecitabine is activated by a three-step enzymatic pathway with two intermediary metabolites, 5 -deoxy5-fluorocytidine and 5 -deoxy-5-fluorouridine (Fig. 2). The absorption of capecitabine and its subsequent conversion into 5-FU is rapid (Cmax of 5-FU 60 min). Subsequently, 5-FU is cleared rapidly and follows the pharmacokinetic (PK) profile of capecitabine. 5-FU is undetectable in plasma after approximately 6 h.2–4 On the basis of this PK profile, it can be concluded that the approved twice daily oral dosing (morning– evening) schedule of capecitabine leaves an in-between dosing capecitabine-exposure gap of approximately 6 h, which is argued to be clinically undesirable.5,6 Indeed, several clinical studies have shown that, as a consequence of the prolonged exposure and reduction of the 5-FU peak plasma concentration, a continuous infusion instead of a bolus injection of 5-FU inR
Abbreviations used: 5-FU, 5-fluorouracil; 5 -DFCR, 5 -deoxy-5-fluorocytidine; 5 -DFUR, 5 -deoxy-5-fluorouridine; CoSD, co-spray-dried; Q50 the time to 50% dissolution of the active component. Correspondence to: Jelte Meulenaar (Telephone: +31-20-512-47-31; Fax: +3120-512-47-53; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 103, 478–484 (2014)
C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association
478
Figure 1. Chemical structure of capecitabine (Mw 359.35 g/mol).
creases the anti-tumour response rate and decreases toxicity in terms of a reduced occurrence and grade of adverse events.5,6 Therefore, to translate this concept to an oral dosing scheme, an oral extended-release (ER) dosage form of capecitabine is warranted. For most active pharmaceutical ingredients (APIs), the development of an ER formulation is straightforward. Various excipients are commercially available (e.g., polyvinyl acetate (PVAc)–polyvinyl pyrrolidone (PVP) co-polymer (Kollidon SR)7 and the methyl cellulose and ethylene oxide8 which, by simply mixing with the API in a certain ratio followed by, for example direct compression into tablets, results in ER of the compound of interest. However, an important factor is the amount of ER excipient required to obtain the desired release profile, in particular for high-dosed APIs as because of size little space per dose unit is available for additives. For example, Xeloda consists of oblong-shaped tablets containing 150 mg (5.4 × 11.5 mm2 ) or 500 mg (8.4 × 15.8 mm2 ) of capecitabine at a
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
R
R
479
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 2. (a) Schematic overview of the pathway of a DS from administration to the site of action (tumour). (b) IVIVC model for capecitabine. 5-DFCR: 5-deoxy-5-fluorocytidine; 5-DFUR: 5-deoxy-5-fluorouridine; 5-FU: 5-fluorouracil.
total weight of 187.5 and 625 mg, respectively, corresponding to 80% (w/w) API and only 20% (w/w) of excipients. Given the standard dosing regimen up to 1250 mg/m2 capecitabine twice daily, it is easily understood that from a patient compliance perspective this leaves limited space for additional excipients, which would result in either an increase in size or number of tablets that should be taken. Previously, we reported the anomalous dissolution behaviour of amorphous capecitabine which, in comparison with crystalline capecitabine used in Xeloda , displays unexpected slow and highly reproducible dissolution behaviour.9 Making use of this finding, we describe in this article the in vitro development of a prototype ER formulation for a twice-daily dosing (b.i.d.) schedule providing a continuous exposure to capecitabine. For this, simulations from a previously published human population PK model of capecitabine10 were used to obtain a theoretically optimal dissolution curve. This in vitro–in vivo correlation (IVIVC) model was subsequently used to screen a panel of ER excipients to identify the most efficient dissolution delaying excipient (i.e., extent of release in time in relation to the quantity of excipient used). R
MATERIALS AND METHODS Materials Capecitabine drug substance (DS), manufactured by Jiangsu ZW Pharmaceuticals Company (Changzhou, Jiangsu, China) was supplied by Dolder AG (Basel, Switzerland). For the screening of the sustained-release excipients, various polymer grades of methyl cellulose (MethocelTM K3 LV, K100 LV CR, K4M CR K15M CR, K100M CR), poly (ethylene oxide) (PolyoxTM 301 and coagulant) were kindly provided by Colorcon (Dartford, Kent, United Kingdom). PVAc–PVP copolymer (Kollidon SR) was kindly provided by BASF (Limburgerhof, Germany). Xeloda was purchased from Roche Laboratories Inc. (Nutley, New Jersey). All other excipients and solR
R
DOI 10.1002/jps.23779
vents (Ph.Eur./analytical grade) were originated from commercial suppliers. Methods
ER Capecitabine Formulation Screening Physical Mixtures. Physical mixtures (PMs) were prepared by mixing capecitabine (crystalline DS or amorphous) and polymers (80/20%, w/w) in a gallipot with a TURBULA –T 10 (Willy A. Bachofen AG – Maschinenfabrik, Muttenz, Switzerland) for 10 min. R
Spray-Drying and Co-Spray-Dried Formulations. The (co-) spray-dried (CoSD) powder of capecitabine DS was produced ¨ with a Buchi Mini Spray Dryer B-290, equipped with an Inert Loop B-295, Dehumidifier B-296, High Performance cyclone, 1.5 mm nozzle cap, and 0.7 mm nozzle tip (BUCHI Labortechnik AG, Flawil, Switzerland) using the following settings: feed 15%; N2-atomization flow 40 mm; aspirator flow 80%; inlet temperature 18◦ C; outlet temperature 18◦ C; inert loop temperature −20◦ C. Spray-drying solution consisted of capecitabine DS (300 mg/ mL) and Kollidon SR, dissolved (at 45◦ C) in 100% ethanol (HPLC grade). Kollidon SR was added at different ratios. R
R
Tableting Tablets (9 mm Ø, resistance to crushing 60 N) were pressed using an amount of formulation mixture equivalent to 224 mg capecitabine on a Korsch EK0 eccentric press (Korsch AG, Berlin, Germany). Tablet thickness was related to the volume and compaction properties of the tableting mixture. Crushing strength was determined using an Erweka TBH20 hardness tester (Erweka, Heusenstamm, Germany). For the prototype formulation, the CoSD powder (capecitabine Kollidon SR 98/2%, w/w) was mixed with 0.5% (w/w) colloidal silicon dioxide and 2.5% (w/w) magnesium stearate. Using this tableting R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
480
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
mixture, oblong tablets (8.4 × 15.8 mm2 , 100–140 N) containing 500 mg capecitabine were pressed.
Table 1.
Dissolution Testing. Dissolution testing was performed according to United States Pharmacopeia monograph using a type 2 (paddle) dissolution apparatus (Erweka) and water for Injections (B. Braun, Melsungen, Germany) at 37◦ C, stirred at 50 rpm, as medium.11 Samples collected at the various time points were filtrated using a 0.45-:m filter and subsequently analysed on a reversed-phase HPLC–UV (RP-HPLC–UV) system: column: Varian, Inertsil ODS-3, 15 × 4.6 mm2 ; 5 :m, Eluent methanol–water (7:3, v/v), flow: 1.0 mL/min, detection: 310 nm. The retention time of capecitabine in this RP-HPLC-UV system was about 3.5 min. The dissolution test was continued to 100% dissolution or until a constant dissolution percentage. Also, the dissolution profile of Xeloda (150–500 mg) was determined.
MethocelTM K3 LV MethocelTM K100 LVCR MethocelTM K4M CR MethocelTM K15M CR MethocelTM K100M CR PolyoxTM 301 PolyoxTM coagulant KollidonR SR KollidonR SR (CoSD) References Amorphous capecitabine XelodaR (Roche)a
R
IVIVC Model Development. Figure 2 shows a graphical overview of the IVIVC model used. The PK model previously published by Urien et al.10 developed for Xeloda was adapted to include an additional dissolution step prior to the absorption process. For this, the well-established Higuchi dissolution equation12,13 was converted to a differential equation describing the dissolution rate (Eqs. 1–3):
Time to 50% Dissolution (Q50 )
Polymer
Q50 (min) 53 271 638 701 838 666 739 1556 4966 45 11
Amorphous capecitabine (224 mg) tablets containing 20% (w/w) of different ER polymers (n = 3/formulation): methyl cellulose (MethocelTM K3 LV, K100 LV CR, K4M CR K15M CR, K100M CR); poly (ethylene oxide) (PolyoxTM 301 and coagulant); PVAc–PVP co-polymer (Kollidon SR). a Xeloda contains crystalline capecitabine.7 R
R
R
√ A(%) = a t + b
(1)
A(%) = at1/2 + b
(2)
1 dA(%) = at−1/2 = Rdis dt 2
(3)
where a is the slope of the found line, t is the time, Rdis is the dissolution equation and A is the percentage dissolved capecitabine as function of time. In the original model developed by Urien et al.,10 the absorption rate constant ka described both dissolution and absorption rate. To be able to introduce formulations with different dissolution characteristics, the true absorption rate constant kad was introduced, solely describing the absorption process. As original PK data were not available to estimate this parameter, simulations were used to approximate this parameter. For this, Eq. (3) was implemented combined with the determined Higuchi constant of 17.8 for Xeloda . Subsequently, the kad parameter was used to simulate different curves for various ER formulations with different Higuchi constants. The complete IVIVC model was described as a system of ordinary differential equations describing the mass transport between the different compartments as shown in Figure 2. This model was implemented in NONMEM (version VI 2.0; Icon Development Solutions, Ellicott City, Maryland) for simulation purposes.14 R
Modulated Temperature Differential Scanning Calorimetry. Modulated temperature differential scanning calorimetry (MTDSC) measurements were performed with a discovery differential scanning calorimeter (DSC) (TA Instruments, New Castle, Delaware). Temperature scale and heat flow were calibrated with indium. Samples of approximately 3 mg powder were weighed into Tzero aluminium pans (TA instruments), sealed and placed in the autosampler. Each sample was equilibrated at −30.00◦ C for 5 min, after which the sample was Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
heated to 125◦ C for capecitabine and 250◦ C for Kollidon SR at a speed of 2◦ C/min with a modulated temperature amplitude of 1◦ C per period of 60 s. R
X-ray Powder Diffraction. X-ray powder diffraction (XPRD) measurements were performed with an X’pert pro diffractometer equipped with an X-celerator (PANanalytical, Almelo, The Netherlands). Samples were placed in a 0.5-mm deep metal sample holder which was placed in the diffractometer. Samples were scanned at a current of 30–50 mA and a tension of 40 kV. The scanning range was 10◦ –60◦ 22, with a step size of 0.020◦ and a scanning speed of 0.002◦ /s.
RESULTS IVIVC Modelling As capecitabine is highly soluble over the whole physiological pH range as well as highly permeable, it is classified as a BCS class I drug.15–17 This means that no significant absorption delay as a consequence of solubility or permeability issues are to be expected after oral administration of capecitabine. Therefore, the release characteristics of the pharmaceutical dosage form will be the main variable determining capecitabine absorption during gastrointestinal transit. In addition to this, the availability of a well-described human PK model as well as the known dissolution characteristics of capecitabine from Xeloda served as an excellent starting point for the development of an IVIVC model. Using this model, the curve which approached the desired 12-h in vivo exposure profile for capecitabine best, proved to be an ER formulation with a Higuchi constant of 4.5. In Figure 3, the respective plasma concentration–time curve and corresponding dissolution profile are depicted. R
Screening of ER Polymers Next, several ER polymers were screened for their potential of obtaining this target capecitabine dissolution profile (Table 1). Spray-dried, amorphous capecitabine was used for the screening experiment, having already slow-release characteristics in itself and thus likely requiring less release-modifying excipient DOI 10.1002/jps.23779
481
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 3. Capecitabine plasma concentration–time and corresponding dissolution curves , capecitabine (XelodaR ) concentration–time curve [Urien model (10)]; , XelodaR (500 mg) dissolution curve (experimental data); , simulated 12-h capecitabine exposure concentration–time curve; , simulated 12-h capecitabine dissolution curve (Higuchi constant of 4.5).
◦
•
to meet the target release profile, favourable in view of the final size and feasibility of the high-content capecitabine dosage form.9 Polymers from structurally different chemical groups and suitable for direct compression into monolithic matrices were selected, mixed with amorphous capecitabine, pressed into tablets and subjected to dissolution testing. From Table 1, it can be seen that PVAc–PVP co-polymer (Kollidon SR) induced the most prominent ER effect with an approximately 141.5-fold and 34.5-fold delay in dissolution compared with Xeloda and amorphous capecitabine, respectively. On the basis of these results, it was investigated whether capecitabine could be CoSD with Kollidon SR. This might have potential advantages in terms of a reduction of processing (mixing) steps, which is favourable in view of upscaling. Also, the addition of Kollidon SR (having mixture Tg ’s of 47◦ C and 161◦ C determined by MTDSC on a single batch purchased Kollidon SR, data not shown), because of its various components gives a, although limited, Tg elevation (CoSD Tg of 32◦ C, data not shown) as compared with 100% amorphous capecitabine (Tg of 19◦ C 9 . Theoretically, this Tg elevation may result/be reflected in an increased physical stability. Moreover, and unexpectedly, it was shown that tablets prepared from CoSD capecitabine and Kollidon SR have an additional threefold delay in dissolution compared with tablets prepared from the same but only physically mixed components (Table 1). R
R
R
R
R
R
Optimization of Percentage Kollidon SR in the CoSD Mixture R
On the basis of these results, Kollidon SR, when CoSD with capecitabine, was identified as the most promising ER polymer to reach the target release profile using as less excipient as possible. To examine this further, in vitro dissolution as a function of the percentage Kollidon SR in CoSD capecitabine Kollidon SR mixtures was assessed. For comparison, dissoR
R
R
DOI 10.1002/jps.23779
lution curves of physically mixed crystalline (DS) and amorphous capecitabine Kollidon SR mixtures were recorded as well. Figure 4 gives the results of these experiments, and clearly shows the significant dissolution delay of the CoSD material as compared with the PMs of both crystalline and amorphous capecitabine using the same percentages (w/w) of Kollidon SR. Moreover, it was shown that the target Higuchi constant of 4.5 can be reached by adding only a minimal quantity of Kollidon SR to the CoSD mixture. The percentages of Kollidon SR required to obtain a dissolution curve with a Higuchi constant of 4.5 were 1.2%, 7.3%, and 19.2% (w/w) for the CoSD, amorphous capecitabine PM, and crystalline PM, respectively (Fig. 4). R
R
R
R
Final Prototype Formulation As only a minimal quantity of Kollidon SR was needed in the CoSD capecitabine Kollidon SR mixture, this was selected for further development into a prototype formulation. Aiming for tablets as the preferred dosage form, at least a glidant and lubricant need to be added to the CoSD capecitabine Kollidon SR mixture. The CoSD capecitabine Kollidon SR mixture by itself was found to be of high density and good compressibility compared with the crystalline capecitabine Kollidon SR PM (data not shown). Therefore, experiments were performed to investigate whether the addition of colloidal silicon dioxide and/or magnesium stearate to the CoSD powder affects the release characteristics of capecitabine. Colloidal silicon dioxide did not show any release-modifying effects (data not shown) but magnesium stearate, however, significantly accelerated capecitabine release. Adding a standard percentage of 2.5% (w/w) of magnesium stearate to a tableting mixture composed of CoSD capecitabine Kollidon SR (99/1% w/w) and colloidal silicon dioxide 0.5% (w/w), the time to 50% dissolution (Q50 ) of capecitabine from the tablets dropped from 4.5 to 2 h. R
R
R
R
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
482
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 4. Dissolution as a function of the percentage KollidonR SR in CoSD capecitabine KollidonR SR. Dissolution curves of 224 mg capecitabine tablets (n = 3) were recorded and expressed as their Higuchi constants. , Physically mixed crystalline capecitabine DS and KollidonR SR; , physically mixed amorphous capecitabine and KollidonR SR; , CoSD capecitabine KollidonR SR. The insert shows the full dissolution curves of tablets containing 10% (w/w) KollidonR SR of the respective formulations.
◦
•
Figure 5. Dissolution of the ER capecitabine prototype formulation as compared with the target dissolution curve with a Higuchi constant of 4.5. ◦, Dissolution curve with Higuchi dissolution constant 4.5 (dotted lines show the f2 interval for similarity of dissolution 16 ; , In vitro dissolution curve (n = 3) of the ER capecitabine prototype formulation with 2.0% KollidonR SR. NB: the error bars are showing the variability in dissolution performance for the tested tablets. The insert shows the dissolution curves of the tablets containing increasing amounts (0%–2.5%) of KollidonR SR. , 0%; , 0.5%; , 1.0%; , 1.5%; , 2.0%; , 2.5%.
•
•
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
DOI 10.1002/jps.23779
483
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Figure 6. X-ray diffraction patterns of PMs of crystalline capecitabine (1) amorphous capecitabine (2) and KollidonR SR, CoSD capecitabine and KollidonR SR (3) and a powdered prototype tablet after storage over 11 months at −20◦ C, consisting of CoSD capecitabine and KollidonR SR (4). Note: The small peaks at 43◦ –45◦ are caused by the scattering of the metal sample holder.
To match the release characteristics to the Higuchi constant of 4.5 that we aimed for, dissolution curves were recorded of CoSD capecitabine Kollidon SR mixtures (range of 0.5%–2.5%, w/w, Kollidon SR) in the presence of 0.5% colloidal silicon dioxide and 2.5% (w/w) magnesium stearate (insert Fig. 5). From these experiments, a percentage of 2.0% (w/w) Kollidon SR in the CoSD mixture was found to potentially approach the target release profile. Figure 5 shows the theoretical dissolution profile with a Higuchi constant of 4.5 and the experimental dissolution curve of the formulation containing 2.0% (w/w) Kollidon SR in the CoSD mixture. The similarity factor (f2 ) for this dissolution curve was found to be 52 and therefore considered similar to the theoretical curve (specification: 50 < f2 < 100).18 On the basis of this result, the prototype formulation selected for further development is composed of CoSD capecitabine Kollidon SR (98/2%, w/w), colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w). R
R
R
R
R
R
R
R
R
R
R
DISCUSSION As pharmaceutical development mainly involves in vitro experiments, it is highly useful to have some guidance on the correlation with in vivo characteristics when developing an ER dosage form for a specific compound. For capecitabine, its BCS class 1 classification, the availability of a PK model as well as the known dissolution characteristics of Xeloda offered the opportunity for developing an IVIVC model. This was performed by a relatively simple combination of two well-established models: the Higuchi dissolution model and a compartmental population PK model. Simulations using this model resulted in a to-develop dosage form having release characteristics with a Higuchi constant of 4.5 (Fig. 3). R
DOI 10.1002/jps.23779
From the screening experiments, Kollidon SR induced the most prominent ER effect. This slow release of capecitabine can be attributed to the insoluble PVAc component of Kollidon SR that forms an inert hydrophobic matrix after hydration. The increasing length of the diffusion pathway over time is responsible for the ER of capecitabine. In contrast to Kollidon SR, the methyl cellulose and poly(ethylene oxide) are gel-forming hydrophilic colloids. Because of the erosion of the gel formed upon hydration, the diffusion pathway is limited and dissolution is mainly governed by the viscosity of the polymer (e.g., dissolution rate from MethocelTM K100M CR < MethocelTM K4M CR). Moreover, co-spray drying of capecitabine and Kollidon SR appeared to result in an additional, significantly delayed dissolution as compared with the physically mixed individual components. Apparently, the Kollidon SR matrix responsible for the add-on ER characteristics of amorphous capecitabine is even more pronounced when being part of a homogeneous, amorphous structure (Fig. 6). Indeed, only a minor amount (1.2%, w/w) of Kollidon SR was required to obtain the target release profile. The addition of magnesium stearate, however, resulted in an unexpected and significant acceleration of the dissolution rate of CoSD capecitabine Kollidon SR tablets. As reported previously, the slow-release characteristics of amorphous capecitabine are attributed to an instant gelling (collapse) upon contact with water.9 It is likely that, by covering the capecitabine Kollidon SR particles with a thin, hydrophobic film, magnesium stearate interferes with this gelling property. We speculate that in the presence of magnesium stearate the resulting gel surface is not as tight and leaves space for capecitabine to dissolve more quickly. When adding a slightly higher concentration of matrix former (2%, w/w, Kollidon SR), however, this issue was resolved. R
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
484
RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
The prototype capecitabine ER formulation comprises only a very limited amount of slow-release excipient (2%, w/w). For comparison, the excipient monograph of the manufacturer advises quantities of Kollidon SR from 40% to 55% (w/w) to be added to a freely soluble API to obtain ER drug profiles.7 Indeed, the low amount of Kollidon SR in the prototype ER formulation is highly attractive in terms of clinical feasibility of the final dosage form. The currently marketed immediaterelease dosage form of capecitabine (Xeloda ) is composed of 80% (w/w) of API and 20% (w/w) of excipients. As the prototype ER formulation consists of 95% (w/w) of API and 5% (w/w) of excipients, it is clear that no limitations are to be expected with respect to size of the final dosage form. However, one issue which must be addressed is the stability of the final dosage form that shows the low Tg (32◦ C, data not shown) of the CoSD capecitabine Kollidon SR. Previously, the reluctance of amorphous capecitabine to crystallize was shown with a 48h-isothermal differential scanning calorimetric analysis at a temperature (80◦ C) far exceeding the Tg of amorphous capecitabine.9 Further, Figure 6 shows that the XPRD pattern of the powdered prototype tablet (11 months stored at −20◦ C) is equal to the freshly prepared CoSD capecitabine and thus an amorphous system is present. Additional, the DSC analysis ran over −25◦ C to 125◦ C 10◦ C/min of that powdered prototype tablet showed a Tg at 32◦ C but lacks a melting peak (data not shown). The absence of a melting peak in this DSC run is an extra and further indication that the required amorphous system for the defined release characteristics of the prototype formulation is fully intact. Taking these results into account, the statement can be made that the preliminary stability results show that the tablet formulation is chemically (HPLC, data not shown) and physically stable for at least 11 months when stored at −20◦ C.
istics and the applicability of the IVIVC model used, the in vivo PK of this formulation is investigated in a clinical study.
R
R
R
R
CONCLUSIONS An oral ER formulation of capecitabine was developed for twice daily dosing, theoretically providing a continuous exposure to capecitabine and 5-FU thus avoiding the undesirable in-between dosing gap inherent to the dosing schedule of the marketed capecitabine immediate-release formulation (Xeloda ). To achieve this aim, the target 12-h in vivo release profile was correlated to an in vitro dissolution profile using an IVIVC model based on the PK and dissolution characteristics of Xeloda . Making use of the slow dissolution characteristics of amorphous capecitabine as reported previously 9 and screening of a panel of ER excipients, a prototype tablet formulation composed of CoSD capecitabine and Kollidon SR (98/2%, w/w) mixed with colloidal silicon dioxide (0.5%, w/w) and magnesium stearate (2.5%, w/w) was defined. To confirm its ER characterR
R
R
Meulenaar et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:478–484, 2014
REFERENCES 1. Roche Laboratories Inc. 2011. Product information XelodaR . Accessed April 29, 2013, at: www.roche-australia.com/fmfiles/re7229005/ downloads/oncology/xeloda-pi.pdf 2. Walko CM, Lindley C. 2005. Capecitabine: A review. Clin Ther 27:23– 44. 3. Pentheroudakis G, Twelves C. 2002. Capecitabine (Xeloda): From the laboratory to the patient’s home. Clin Colorectal Cancer 2:16–23. 4. Schellens JHM. 2007. Capecitabine. Oncologist 12:152–155. 5. Seifert P, Baker LH, Reed ML, Vaitkevicius VK. 1975. Comparison of continuously infused 5-fluorouracil with bolus injection in treatment of patients with colorectal adenocarcinoma. Cancer 36:123–128. 6. Meta-analysis Group In Cancer. 1998. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol 16:301–308. 7. BASF SE Care Chemicals Division Pharma Ingredients & Services. 2008. Technical information KollidonR SR (BASF). EMP 030728e-07:1– 12. 8. Colorcon. Methocel product application data sheet. Accessed August 8, 2013, at: http://www.colorcon.com/resources/lit/Products. 9. Meulenaar J, Beijnen JH, Schellens JHM, Nuijen B. 2013. Slow dissolution behaviour of amorphous capecitabine. Int J Pharm 441:213– 217. 10. Urien S, Reza´ı K, Lokiec F. 2005. Pharmacokinetic modelling of 5-FU production from capecitabine—A population study in 40 adult patients with metastatic cancer. J Pharmacokinet Pharmacodyn 32:817– 833. 11. Marques MRC, Mao F. 2011. USP34–NF29 capecitabine tablets. The United States Pharmacopeia–National Formulary 2148–2148. 12. Paul DR. 2011. Elaborations on the Higuchi model for drug delivery. Int J Pharm 418:13–17. 13. Higuchi T. 1963. Mechanism of sustained-action medication. Theoretical analysis of rate of release of solid drugs dispersed in solid matrices. J Pharm Sci 52:1145–1149. 14. Beal SL, Sheiner LB. 1989. NONMEM users guides. Ellicott City, Maryland: Icon Development Solutions. ¨ H, Shah V, Crison J. 1995. A theoretical 15. Amidon G, Lennernas basis for a biopharmaceutic drug classification: The correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm Res 12:413–420. 16. Reigner B, Verweij J, Dirix L, Cassidy J, Twelves C, Allman D, Weidekamm E, Roos B, Banken L, Utoh M, Osterwalder B. 1998. Effect of food on the pharmacokinetics of capecitabine and its metabolites following oral administration in cancer patients. Clin Cancer Res 4:941–948. 17. Blesch KS, Gieschke R, Tsukamoto Y, Reigner BG, Burger HU, Steimer JL. 2003. Clinical pharmacokinetic/pharmacodynamic and physiologically based pharmacokinetic modeling in new drug development: The capecitabine experience. Invest New Drugs 21:195–223. 18. EMEA Committee for Medicinal Products for Human Use. 2008. Guideline on the investigation of bioequivalence. Guideline CPMP/EWP/QWP/1401/98 Rev. 1:1–29.
DOI 10.1002/jps.23779