Quality by Design (QbD)-Based Crystallization Process Development

May 3, 2011 - Synopsis. The thermodynamically stable Form II of the antidiabetic drug tolbutamide exhibits a thin fiber needle shape which renders it ...
2 downloads 0 Views 3MB Size
Download PDF
ARTICLE pubs.acs.org/crystal

Quality by Design (QbD)-Based Crystallization Process Development for the Polymorphic Drug Tolbutamide Satyanarayana Thirunahari,*,†,‡ Pui Shan Chow,‡ and Reginald B. H. Tan*,†,‡ † ‡

Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576 Institute of Chemical and Engineering Sciences, A*STAR (Agency for Science Technology and Research), 1 Pesek Road, Jurong Island, Singapore 627833

bS Supporting Information ABSTRACT: The thermodynamically stable Form II of the antidiabetic drug tolbutamide exhibits a thin fiber needle shape which renders it intractable for isolation and downstream processing. This work implements two in situ process analytical technology (PAT) methods, namely, attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) with orthogonal partial least-squares-principal component analysis (OPLSPCA) for monitoring solute concentration, and Raman spectroscopy with dynamic PCA based multivariate statistical process monitoring (MSPM) for detection of solid form purity, to derive the robust design space for cooling crystallization of the desired Form IL.

1. INTRODUCTION One of the challenging issues in the development of crystallization processes for active pharmaceutical ingredients (APIs) is to obtain the desired polymorphic form of the product. Though chemically identical, polymorphs are different crystal forms resulting from variation in molecular arrangement in the crystal lattice. Consequently, they exhibit different physicochemical properties such as morphology, stability, solubility, bioavailability, etc. which can impact both the processability and performance of the drug. Therefore, in pharmaceutical production, it is vital to have a robust crystallization process for the consistent production of the desired crystal form. Although there is a vast literature available dealing with the phenomenon of polymorphism, only a few address the issue of how to control polymorphism in crystallization. Some important ones are using tailor-made additives to stabilize the desired form,1,2 use of solvents to promote the nucleation of the desired polymorph,35 sonication,6,7 use of supercritical medium,8 etc. However, these methods are system specific and some require good understanding in molecular aspects of crystallization where our knowledge is extremely poor. A generic approach which has currently been practiced by industry is based on Quality by Design (QbD) which is motivated by the new recent initiatives from the Food and Drug Administration (FDA) to adopt modern technology to develop scientific processes.9 A few authors10,11 have discussed the concept of QbD, its benefits and application to develop robust crystallization processes with several examples. Essentially, QbD involves experimental screening of input and process parameters to determine a design r 2011 American Chemical Society

space (operating region where the crystallization of the desired polymorph is guaranteed) and with appropriate controls installed to operate within the control space (upper and/or lower limits of operating parameters between which the process is controlled) to produce the desired form. To this aim, in situ analytical technologies, usually referred to as process analytical technologies (PATs) play a crucial role in process characterization, analysis, and control.12,13 The FDA also strongly recommends the liberal use of PAT to assist the implementation of QbD. Some of the potential benefits of PATs are improved process understanding,14,15 faster and focused process development,1621 online assessment of polymorphic quality,2225 and implementation of advanced control strategies.2628 Tolbutamide (TB) is a sulfonylurea drug for type II diabetic patients to control the blood sugar levels. It is known to exist in five polymorphic forms (Forms IL, IH, and IIIV). Forms IL and IH are kinetically reversible with a transition temperature of 38 °C.29 Therefore, at any given temperature, only one form is experimentally accessible. In our previous work, crystal polymorphism of TB was reinvestigated to establish the structural and thermodynamic features of TB polymorphs.30 Crystal structures of four polymorphic forms were solved and a schematic energytemperature diagram was constructed to elucidate their stability relationships. Recently, Nath et al.31

Received: March 10, 2011 Revised: April 26, 2011 Published: May 03, 2011 3027

dx.doi.org/10.1021/cg2003029 | Cryst. Growth Des. 2011, 11, 3027–3038

Crystal Growth & Design

ARTICLE

reported a novel sixth form (Form V) which was obtained in an attempt to cocrystallize TB. The main objective of the present work is to develop a robust cooling crystallization process for the production of the desired polymorph of TB in a QbD paradigm. The PAT framework used in this work mainly consists of two widely used PATs in crystallization: attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) for monitoring solute concentration and Raman spectroscopy for detection of polymorphic purity. ATR-FTIR was coupled with orthogonal partial least-squares-principal component analysis (OPLSPCA),32,33 a calibration based robust chemometric method, which is a recent modification to partial least-squares (PLS). In recent times, Raman spectroscopy has emerged as the most suitable technique for in situ monitoring of polymorphs. However, calibration based Raman monitoring is associated with many challenges. First of all, Raman calibration is a laborious and time-consuming task.34,35 The other issue is the presence of unrelated variation in Raman data due to change in particle size, suspension density, etc. which can complicate calibration models leading to robustness issues.36 In the present work, a dynamic PCA-based multivariate statistical process monitoring (MSPM) has been applied with Raman for polymorph monitoring. An advantage of MSPM is that it is completely data-based; therefore, no sophisticated calibration is required to apply these techniques.

2. EXPERIMENTAL SECTION 2.1. Materials. TB (g98% purity) was purchased from Junda Pharmaceutical Co. Ltd. (Jiangsu, PR China). The powder X-ray diffraction (PXRD) data of TB confirms that it belongs to Form IL and used as such without further purification. The solvents (ethanol and deionized water) used were of analytical reagent grade. A mixed solvent ethanol/water (60:40 w/w) was prepared and used for all the experiments. To prepare Form II, Form IL crystals were suspended in excess in solution at 30 °C and stirred for 45 h to allow solution mediated polymorphic transformation (SMPT). The obtained Form II crystals were filtered and dried in a vacuum oven at room temperature (RT). Polymorphic purity was verified using PXRD. 2.2. Mathematical Methods. In the following text, matrices are denoted by bold uppercase letters. Vectors are assumed to be column vectors and denoted by bold lowercase letters. Scalars are denoted by italicized lowercase letters. X(n  m) denotes a set of mean centered spectral data matrix with n samples and m variables (intensities at wavenumber) and y(n  1) denotes the solute concentration vector. Principal Component Analysis (PCA). PCA models single space X by finding the latent variables called principal components (PCs) that explain the maximum variance in X.37 PCA model is given by eq 1. X ¼ t1 pT1 þ t2 pT2 þ 3 3 3 þ tK pTK þ E ¼ TPT þ E

ð1Þ

K is the number of PCs. T(n  K) is the matrix of X scores, P(m  K) is the matrix of X loadings, and E is the residual matrix, respectively. ti and pi are the vectors of T and P corresponding to ith PC, respectively. To derive an optimum PCA model, a good rule of thumb is to include PCs until the subsequent PC captures γ(IL,IH) and AIII > A(IL,IH). On the other hand, slower cooling rates might have promoted the nucleation of Form (IL, IH) by favoring J(IL,IH) > JIII with conditions γ(IL,IH) > γIII and A(IL,IH) > AIII. When intermediate cooling rates were used, the balance of these contributions to nucleation rates of Forms IL and III might have favored the condition JIL = JIII thus resulting in concomitant nucleation of both forms. Nevertheless, this analysis clearly suggests that the design space for the isolation of desired form of TB is D2 where primary nucleation of Form (IL, IH) is guaranteed. 3.6. Design and Operation of Selective Crystallization. For selective crystallization, a cooling crystallization process was designed to operate within the design space. However, one possibility

ARTICLE

Figure 15. Desupersaturation profile measured using ATR-FTIR during batch crystallization while operating in the design space.

Figure 16. PVM image of Form IL crystals obtained during batch crystallization operation.

which cannot be ignored while operating in the design space is the primary nucleation of more stable Form II via SMPT. To prevent any possibility of undesired SMPT to Form II, crystallization should be designed such that the total crystallization time (excluding the induction time) should be within the induction period for the nucleation of Form II (∼200 min). Keeping this in mind, a process was designed with a saturated solution corresponding to a temperature of 50 °C which is cooled at 0.08 °C/min until 25 °C. Figure 15 shows the desupersaturation profile measured using ATR-FTIR during the crystallization operation. A hump in concentration profile in the region of 3941 °C which is due to the solid state transformation42 of Form IH f IL clearly indicates that the desired form has been achieved. Figure 16 shows the PVM image of crystals obtained. The prismatic morphology typical of Form (IL, IH) further confirms the identity of the desired form. Hence, a robust crystallization process for the production of the desired form of TB has been successfully demonstrated.

4. CONCLUSIONS A crystallization design space for the isolation of the desired polymorph Form (IL, IH) of TB has been determined and a 3036

dx.doi.org/10.1021/cg2003029 |Cryst. Growth Des. 2011, 11, 3027–3038

Crystal Growth & Design crystallization batch process was operated within the design space to achieve the desired form. A QbD based strategy not only assisted in developing a robust process but also provided a good understanding about the crystallization behavior of TB. Characterization of polymorphic transformation of TB polymorphs Form IL f Form II suggests that the primary nucleation of the stable Form II is the controlling step for the transformation. MZW measurements indicate that the cooling rate has a strong influence on the nucleation energy barrier and can be used as a manipulator for achieving different TB polymorphs. This work also demonstrates the application of two PATs, ATR-FTIR combined with OPLS-PCA and Raman with MSPM in crystallization process characterization and monitoring. OPLS-PCA ensures the robustness of the calibration models which is particularly important in industrial environments where process disturbances are not uncommon. MSPM can be combined with Raman and used as a PAT tool for the characterization of polymorphic transformations and nucleation experiments with an additional advantage of in situ detection of the polymorph nucleated. The most appealing part of MSPM is that it is calibration free, and therefore it can reduce the experimental effort and speed up the process development.

’ ASSOCIATED CONTENT

bS

Supporting Information. Loading plots of PC1 and PC2 for PCA reference models built for monitoring SMPT and MZW experiment (109 mg/g concentration, 0.1 °C/min cooling rate) and PXRD patterns of crystals obtained in MZW experiments A, B, and C. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*(R.B.H.T.) Tel.: þ65 6796 3841. Fax: þ65 6316 6183. E-mail: [email protected]. (S.T.) Tel.: þ65 6796 3844. Fax: þ65 6316 6183. E-mail: [email protected].

’ ACKNOWLEDGMENT Financial support provided by the Science and Engineering Research Council of A*STAR (Agency for Science, Technology and Research), Singapore, is gratefully acknowledged. We thank Dr. Yu Zaiqun and Dr. Martin Wijaya Hermanto of ICES (Institute of Chemical and Engineering Sciences) for useful discussions and Ms. Agnes Nicole Phua for technical support. ’ REFERENCES (1) Davey, R. J.; Blagden, N.; Potts, G. D.; Docherty, R. J. Am. Chem. Soc. 1997, 119 (7), 1767–1772. (2) He, X. R.; Stowell, J. G.; Morris, K. R.; Pfeiffer, R. R.; Li, H.; Stahly, G. P.; Byrn, S. R. Cryst. Growth Des. 2001, 1 (4), 305–312. (3) Blagden, N.; Davey, R. J. Cryst. Growth Des. 2003, 3 (6), 873–885. (4) Weissbuch, I.; Torbeev, V. Y.; Leiserowitz, L.; Lahav, M. Angew. Chem., Int. Ed. 2005, 44 (21), 3226–3229. (5) Kitamura, M.; Hara, T.; Takimoto-Kamimura, M. Cryst. Growth Des. 2006, 6 (8), 1945–1950. (6) Kurotani, M.; Hirasawa, I., Chem. Eng. Res. Des. 88, (9A), 12721278.

ARTICLE

(7) Gracin, S.; Uusi-Penttila, M.; Rasmuson, A. C. Cryst. Growth Des. 2005, 5 (5), 1787–1794. (8) Kordikowski, A.; Shekunov, T.; York, P. Pharm. Res. 2001, 18 (5), 682–688. (9) Yu, L. X. Pharm. Res. 2008, 25 (4), 781–791. (10) Bohlin, M.; Jones, H.; Black, S. Am. Pharm. Rev. 2009, (November/December), 5257. (11) Wu, H. Q.; White, M.; Khan, M. A. Int. J. Pharm. 2011, 405 (12), 63–78. (12) Yu, L. X.; Lionberger, R. A.; Raw, A. S.; D0 Costa, R.; Wu, H. Q.; Hussain, A. S. Adv. Drug. Delivery Rev. 2004, 56 (3), 349–369. (13) Variankaval, N.; Cote, A. S.; Doherty, M. F. AIChE J. 2008, 54 (7), 1682–1688. (14) Chen, J.; Patience, D. B.; Falco, N.; Pachera, R. Am. Pharm. Rev. 2009, April. (15) Howard, K. S.; Nagy, Z. K.; Saha, B.; Robertson, A. L.; Steele, G.; Martin, D. Cryst. Growth Des. 2009, 9 (9), 3964–3975. (16) Zhou, G.; Moment, A.; Cote, A.; Hu, T. Am. Pharm. Rev. 2009, (July/August), 6268. (17) Birch, M.; Fussell, S. J.; Higginson, P. D.; McDowall, N.; Marziano, I. Org. Process Res. Dev. 2005, 9 (3), 360–364. (18) Saranteas, K.; Bakale, R.; Hong, Y. P.; Luong, H.; Foroughi, R.; Wald, S. Org. Process Res. Dev. 2005, 9 (6), 911–922. (19) Starbuck, C.; Spartalis, A.; Wai, L.; Wang, J.; Fernandez, P.; Lindemann, C. M.; Zhou, G. X.; Ge, Z. H. Cryst. Growth Des. 2002, 2 (6), 515–522. (20) Muller, M.; Meier, U.; Wieckhusen, D.; Beck, R.; Pfeffer-Hennig, S.; Schneeberger, R. Cryst. Growth Des. 2006, 6 (4), 946–954. (21) Desikan, S.; Parsons, R. L.; Davis, W. P.; Ward, J. E.; Marshall, W. J.; Toma, P. H. Org. Process. Res. Dev. 2005, 9 (6), 933–942. (22) Ono, T.; ter Horst, J. H.; Jansens, P. J. Cryst. Growth Des. 2004, 4 (3), 465–469. (23) Caillet, A.; Rivoire, A.; Galvan, J. M.; Puel, F.; Fevotte, G. Cryst. Growth Des. 2007, 7, 2080–2087. (24) Klimakow, M.; Leiterer, J.; Kneipp, J.; Rossler, E.; Panne, U.; Rademann, K.; Emmerling, F. Langmuir 2010, 26 (13), 11233–11237. (25) Scholl, J.; Bonalumi, D.; Vicum, L.; Mazzotti, M.; Muller, M. Cryst. Growth Des. 2006, 6 (4), 881–891. (26) Kee, N. C. S.; Arendt, P. D.; Tan, R. B. H.; Braatz, R. D. Cryst. Growth Des. 2009, 9 (7), 3052–3061. (27) Kee, N. C. S.; Tan, R. B. H.; Braatz, R. D. Cryst. Growth Des. 2009, 9 (7), 3044–3051. (28) Abu Bakar, M. R.; Nagy, Z. K.; Rielly, C. D. Org. Process Res. Dev. 2009, 13 (6), 1343–1356. (29) Hasegawa, G.; Komasaka, T.; Bando, R.; Yoshihashi, Y.; Yonemochi, E.; Fujii, K.; Uekusa, H.; Terada, K. Int. J. Pharm. 2009, 369 (12), 12–18. (30) Thirunahari, S.; Aitipamula, S.; Chow, P. S.; Tan, R. B. H. J. Pharm. Sci. 2010, 99 (7), 2975–2990. (31) Nath, N.; Nangia, A. CrystEngComm 2011, 13, 47–51. (32) Trygg, J.; Wold, S. J. Chemometr. 2002, 16 (3), 119–128. (33) Svensson, O.; Kourti, T.; MacGregor, J. F. J. Chemometr. 2002, 16 (4), 176–188. (34) Fevotte, G. Chem. Eng. Res. Des. 2007, 85 (A7), 906–920. (35) Cornel, J.; Lindenberg, C.; Mazzotti, M. Ind. Eng. Chem. Res. 2008, 47 (14), 4870–4882. (36) Chen, Z. P.; Fevotte, G.; Caillet, A.; Littlejohn, D.; Morris, J. Anal. Chem. 2008, 80 (17), 6658–6665. (37) Jackson, J. A User0 s Guide to Principal Components; John Wiley & Sons, Inc: New York, 2003. (38) Kourti, T. Int. J. Adaptive Control Signal. Process. 2005, 19 (4), 213–246. (39) Gurden, S. P.; Westerhuis, J. A.; Smilde, A. K. AIChE J. 2002, 48 (10), 2283–2297. (40) Pollanen, K.; Hakkinen, A.; Reinikainen, S. P.; Rantanen, J.; Minkkinen, P. Chemom. Intell. Lab. Syst. 2006, 84 (12), 126–133. (41) Thirunahari, S.; Chow, P. S.; Tan, R. B. H. 2011, manuscript in preparation. 3037

dx.doi.org/10.1021/cg2003029 |Cryst. Growth Des. 2011, 11, 3027–3038

Crystal Growth & Design

ARTICLE

(42) Kawakami, K. J. Pharm. Sci. 2007, 96 (5), 982–989. (43) Davey, R. J.; Cardew, P. T.; McEwan, D.; Sadler, D. E. J. Cryst. Growth 1986, 79 (13), 648–653. (44) Simon, L. L.; Oucherif, K. A.; Nagy, Z. K.; Hungerbuhler, K. Ind. Eng. Chem. Res. 2010, 49 (20), 9932–9944. (45) Simon, L. L.; Oucherif, K. A.; Nagy, Z. K.; Hungerbuhler, K. Chem. Eng. Sci. 2010, 65 (17), 4983–4995. (46) Hu, Y. R.; Liang, J. K.; Myerson, A. S.; Taylor, L. S. Ind. Eng. Chem. Res. 2005, 44 (5), 1233–1240. (47) Volmer, M. Kinetics Der Phasenbildung; Steinkopff: Leipzig, 1939.

3038

dx.doi.org/10.1021/cg2003029 |Cryst. Growth Des. 2011, 11, 3027–3038