Quality-by-Design in Method Development. Mijo Stanic General Manager and Technical Director

Size: px

Start display at page:

Download "Quality-by-Design in Method Development. Mijo Stanic General Manager and Technical Director"

Adelia White
5 years ago
Views:

1 Quality-by-Design in Method Development Mijo Stanic General Manager and Technical Director

2 Agenda Chromicent who we are Traditional vs. Systematic strategies for method development QbD a new approach in development Case study : Quality by Design in analytical method development Case study : Fast UPLC method development and method transfer to HPLC (for business needs)

3 Who we are? Alexander H. Schmidt & Mijo Stanic, founder of Chromicent GmbH Together more than 40 years experience in pharm. industry in the development and validation of LC methods, Quality-by-Design, analytical instrument qualification, GMP, productivity and efficiency increase More than 60 peer-reviewed publications and presentations on international congresses Invited speaker at conferences (Concept Heidelberg, Informa) Guest lecturer at Molnár-Institute (DryLab), SMatrix (Fusion) Accociated professor at the Beuth University of applied Sciences, Berlin

4 Who we are? Company Pharmaceutical Service Provider offering Method Development in a Quality-by-Design framework in compliance with ICH Q8 Method Validation in compliance with ICH Q2 Forced degradation studies and impurity profiling in compliance with ICH Q1 Cleaning Validation Extractables and Leachables Method Transfer Consulting and Training Auditing (GMP)

local drug authority Allowance to handle controlled substances and

5 Who we are? Company Founded in 2013, based in Berlin Adlershof GMP approved by local drug authority Allowance to handle controlled substances and narcotics Reference customer and training facility for Waters Corp. Co-operation partner of WADA in development of methods

6 Who we are? Analytical dept. HPLC (Alliance) with PDA-, UV-, FL-, RI-, Conductivity, Electrochemical and Charged Aerosol Detection UPLC (Acquity classic, H-Class, I-Class) with PDA-, FL-, ELS-, single-ms and tandem-ms- Detection SFC (Acquity UPC² ) with PDA-, ELSD and tandem-ms-detection Ion chromatograph with Conductivity Detection Prep LC system with fraction collector

7 Traditional vs. Systematic strategies in method development

8 Traditional strategy for HPLC method development (I) Mainly by trial and error Varying one-factor-at-a-time (OFAT) Problems: - additional or missing peaks - changes in selectivity - decreasing resolution of the critical peak pair Trying to test quality into the method this is however the wrong way Results: - time consuming approach - no understanding of the influence of key factors

9 One factor at a time - OFAT Y X

10 One factor at a time - OFAT Y X

11 One factor at a time - OFAT Y X

12 Problems with non-robust methods System to system variations Day to day variation Column quality changes from batch to batch Peaks move with ph, temperature and %B Courtesy of Imre Molnár, Molnár-Institute, Berlin, Germany

13 Problems with non-robust methods System to system variations Day to day variation Column quality changes from batch to batch Peaks move with ph, temperature and %B Courtesy of Imre Molnár, Molnár-Institute, Berlin, Germany

14 Problems with non-robust methods System to system variations Day to day variation Column quality changes from batch to batch Peaks move with ph, temperature and %B Courtesy of Imre Molnár, Molnár-Institute, Berlin, Germany

15 Traditional strategy for HPLC method development (II) Chromatographers tried to adapt already validated methods published in pharmacopoeias (USP, EP, JP) or in literature

16 Example for adapting an old method Official HPLC method for ebastine published in E.P. Run time (unbelievable) 160 min Peak width for ebastine (API) = 8 min Retention factors k* = (recommended 2 20) Impurity Impurity A B Impurity D Impurity C Impurity F Impurity G Impurity E Ebastin Time (min)

17 Systematic LC method development I.) statistical software tools Using experimental design plans as an efficient and fast tool for method development. In a full or fractional factorial design a couple of experiments are carried out in which one or more factors are changed at the same time. Typical examples are Plackett-Burman design Software packages (e.g. Fusion AE, DesignExpert)

18 Systematic LC method development II.) modeling software tools A very smart and computer-assisted way of developing a chromatographic method is by using software modeling packages (e.g. DryLab, ChromSword, ACD/LC simulator).

19 Case study 1: Quality-by-Design in analytical method development

Initial situation Official HPLC method for ebastine published in E.P. Run time (unbelievable) 160 min Peak width for ebastine (API) = 8 min Retention factors k = 0.

20 Initial situation Official HPLC method for ebastine published in E.P. Run time (unbelievable) 160 min Peak width for ebastine (API) = 8 min Retention factors k = Who developed this method.??? Impurity A Impurity B Impurity A B Impurity Impurity D C Impurity F Impurity G Ebastin Impurity E Time (min)

21 Our method development strategy by using a Quality-by-Design approach Quality-by-Design Key components defined by ICHQ8 Analytical method development strategy Quality target product profile Define method goals / ATP Critical quality attributes Risk assessment Critical quality attributes Linking CQA to CPP Risk assessment Design of Experiments (DoE) Screening for stationary and mobile phase Optimization Design space Design space Select working point and verification Method validation and robustness testing Control strategy Control strategy system suitability test Continuous Improvements Continuous Improvements

22 Step 1: Define method goal / ATP Adequate baseline separation of all components (Rs > 2.0) Minimum analysis time (< 10 min). k*-values for ebastine and impurities should be between 2 and 10 Impurities: 0.05% lvl Visualize a design space, in which the method is robust

23 Step 2: Risk assessment Identified influencial parameters Column (chemistry) gradient time tg, start and end temperature T ternary composition of the eluent ph of the eluent

24 Step 3a: Design of experiments screening for the selection of column

25 Step 3a: Design of experiments screening for the selection of column

26 Design of screening experiments Four columns Linear gradients of 10 to 90% methanol acetonitrile 2-propanol

27 Column Acquity UPLC BEH C18 Acquity UPLC HSS T3 Acquity UPLC BEH Phenyl Acquity UPLC HSS C18 SB Mobile phase (Eluent B) Critical Resolution Rs (crit peak pair) Methanol 1.64 (A,D) Acetonitrile 1.94 (C,D) 2-propanol 1.88 (C,D) Methanol < 1.5 (C,D) Acetonitrile < 1.5 (C,D) 2-propanol 1.57 (C,D) Methanol < 1.5 (C,D) Acetonitrile < 1.5 (C,D) 2-propanol < 1.5 (C,D) Methanol < 1.5 (C,D) Acetonitrile < 1.5 (C,D, B ) 2-propanol < 1.5 (C,D, B)

28 Step 3b: Design of experiments Optimization phase

29 Run experiments on a WATERS Acquity UPLC H-class system WATERS Acquity UPLC H-class system Solvent Manager with SSV for up to 9 solvents Sample Manager FTN Column Manager for up to 4 columns and different temp. zones PDA-detector QDa single-ms-detector Empower 3

30 Step 4: Design space (Analyse and process data and build models) 2D-modell tg/t of 100%ACN Color code: Resolution Rs T [ C] tg [min]

31 2D-modell tg/t of 30% iproh in ACN Color code: Resolution Rs T [ C] tg [min]

32 2D-modell tg/t of 60% iproh in ACN Color code: Resolution Rs T [ C] tg [min]

33 3D-modell tg/t of 0-60% iproh in ACN

34 Select working point

35 Investigation of influence of ph Color code: Resolution Rs ph tg [min]

36 Verification of model with real experiment

37 Verification of model with real experiment 3,0 experimental retention time [min] 2,5 2,0 1,5 1,0 R² = 0,9996 0,5 0,0 0,0 0,5 1,0 1,5 2,0 2,5 3,0 DryLab predicted retention time [min]

38 Method robustness Variation of chromatographic parameters tg (3 min ± 0.3 min) T (60 C ± 6 C) tc (50% ± 5% ACN in PrOH ) flow rate (0.5 ml/min ± 0.05 ml/min) %start (30% ± 2%) %end (90% ± 2%) of the gradient Full factorial design: 3 levels (+1, 0, -1) = 3 6 = 729 experiments

39 Method robustness Variation of chromatographic parameters by using the Robustness Module of the DryLab 4.0 (in silico) tg (3 min ± 0.3 min) T (60 C ± 6 C) tc (50% ± 5% ACN in PrOH ) flow rate (0.5 ml/min ± 0.05 ml/min) %start (30% ± 2%) %end (90% ± 2%) of the gradient Full factorial design: 3 levels (+1, 0, -1) = 3 6 = 729 experiments

40 Method robustness 729 experiments with Rs > 2.0

41 Step 5: Control strategy Based on the validation data and the robustness of the method, the risk assessment indicates that there is extensive knowledge gained about the performance of the method, so that a suitable system suitability test may be the only control element needed in the method control strategy. Therefore, the resolution of the critical peak pair impurity C and D, which shows the lowest resolution of all impurity peaks was chosen as a system suitability test parameter and should be not less than 2.0.

42 Method validation Specificity Linearity, LoD, LoQ Coefficient of correlation > Accuracy and Precision (Repeatability) RSD < 5.0 %, Recovery rate between % Precision - Intermediate Precision RSD < 5.0 %, Mean-t-test must comply Precision - System Precision RSD < 2.0 %

43 Finale UPLC method for ebastine developed with DryLab Imp. C Imp. D Imp. A Imp. B Imp. F Imp. G Imp. E Ebastin 1,0 2,0 Time (min)

44 Comparison finale UPLC method vs. official HPLC method Imp. C Imp. D Imp. A Imp. B Imp. F Imp. G Imp. E Ebastin 1,0 2,0 Time (min) Impurity Impurity A B Impurity D Impurity C Impurity F Impurity G Impurity E Ebastin Time (min)

45 Conclusion All previous defined method goals were met: Baseline separation of the components of interest (Rs >>2.0) k*-values for ebastine and impurities are between 2.2 and 4.3 The design space an area in which the method is robust is defined and visualized Analysis time is only 4 min, which is a impressive 40-fold increase in productivity in comparison to the method published in the E.P. monograph and allowed purity testing of more than 360 samples per day.

46 Step 6: Continuous Improvement HPLC <-> UPLC <-> UPC²

47 Innovative QbD strategy presented at Pittcon 2012, Orlando Analytica 2012, Munich HPLC 2012, Los Angeles (nominated for best poster awards) Paper by Schmidt and Molnár published in J Pharm Biomed Anal (2013) 65-74

48 Case study 2: Fast UPLC method development and method transfer to HPLC (for business needs)

Case study Omeprazole Om Imp. A I 4 6 8 10 time (min) Imp. E Imp. F+G Imp. B Imp.

49 Case study Omeprazole Om Imp. A I time (min) Imp. E Imp. F+G Imp. B Imp. D Imp. H Imp. C Imp. E Imp. F+G Imp. B Imp. D Imp. Imp. I A time (min)

50 Development of a new UPLC method for omeprazole Design of Experiments: (Parameters and values) ph: 7.5, 8.0, 8.5, 9.0 Solvents: methanol, acetonitril Gradient time: 4, 10 min Temperature: 30, 60 C

51 Methanol as solvent Design space Robust region

52 Acetonitrile as solvent Design space Robust region

53 Conditions of the working point for new UPLC method Parameters and values ph: 8.5 Solvent: acetonitrile Gradient time: 4 min Temperature: 35 C Flow rate: 0.7ml/min Column: BEH C18, 50x2.1mm; 1.7µm Gradient: 10%-60% acetonitrile

54 New UPLC method for omeprazole predicted chromatogram Imp. A Imp. I Imp. E Imp. D Imp. B Omeprazol Imp. H Omepraz Imp. C Imp. F Imp. G time (min) experimental chromatogram Imp. A Imp. I Imp. E Imp. D Imp. B Imp. H Imp. C Imp. F Imp. G time (min)

Robustnes testing UPLC Frequency distribution of the R s,crit -values for all 729 experiments of the robustness study on the UPLC system. The six parameters t G (4 min ± 0.

55 Robustnes testing UPLC Frequency distribution of the R s,crit -values for all 729 experiments of the robustness study on the UPLC system. The six parameters t G (4 min ± 0.1 min), T (35 C ± 2 C), ph (8.75 ± 0.1), flow rate (0.7 ml/min ± 0.05 ml/min) and the %B start (10% ± 1%) and %B end (60% ± 1%) of the gradient were varied at 3 levels (+1, 0, -1).

56 Transfer from UPLC to HPLC for business reasons The transfer was calculated by DryLab by changing the values of the parameters UPLC HPLC Column: BEH C18, 50x2.1mm; 1.7µm 50x4.6mm; 2.5µm Dwell volume: 0.4 ml UPLC 1.0 ml HPLC

57 Transfered HPLC method predicted chromatogramm Imp. A Imp. I Imp. E Imp. D Imp. B Omeprazol Imp. H Imp. C Imp. F Imp. G time (min) ph: 8.5 Solvent: acetonitrile Gradient time: 7 min Temperature: 35 C Flow rate: 1.9 ml/min Column: BEH C18, 50x4.6mm; 2.5µm Gradient: 10%-60% acetonitrile Imp. A Imp. I Imp. E Imp. D Imp. B experimental chromatogramm Imp. H Imp. C Imp. F Imp. G time (min)

Robustness testing HPLC Frequency of the distribution of the resolution values R s,crit for all 729 experiments of the robustness study after the transfer to the HPLC system.

58 Robustness testing HPLC Frequency of the distribution of the resolution values R s,crit for all 729 experiments of the robustness study after the transfer to the HPLC system. The six parameters t G (7 min ± 0.1 min), T (35 C ± 2 C), ph (8.75 ± 0.1), flow rate (1.9 ml/min ± 0.1 ml/min) and the %B start (10% ± 1%) and %B end (60% ± 1%) of the gradient were varied at 3 levels (+1, 0, -1).

Verification of the UPLC method working point verification point 1 verification point 2 verification point 3 verification point 4 Flow rate [ml/min] 0.70 0.70 0.75 0.70 0.65 tg [min] 4.0 3.9 4.1 4.

59 Verification of the UPLC method working point verification point 1 verification point 2 verification point 3 verification point 4 Flow rate [ml/min] tg [min] Temp [ C] ph %start %end Retention time [min] Pred. Exp. Pred. Exp. Pred. Exp. Pred. Exp. Pred. Exp. Imp. A 1,06 1,14 1,13 1,18 1,04 1,09 0,96 1,08 1,09 1,15 Imp. I 1,45 1,48 1,50 1,52 1,37 1,41 1,30 1,32 1,43 1,46 Imp. E 1,71 1,74 1,75 1,77 1,65 1,68 1,57 1,59 1,69 1,73 Imp. D 1,97 2,00 2,01 2,02 1,91 1,93 1,79 1,83 1,90 1,91 Imp. B 2,17 2,21 2,20 2,21 2,11 2,14 2,06 2,08 2,15 2,18 Omeprazole 2,26 2,29 2,28 2,29 2,20 2,22 2,15 2,18 2,24 2,27 Imp. H 2,68 2,72 2,68 2,70 2,62 2,65 2,58 2,62 2,65 2,68 Imp. C 2,96 2,99 2,95 2,96 2,90 2,92 2,91 2,93 2,96 2,98 Imp. F 3,68 3,71 3,64 3,65 3,62 3,65 3,66 3,67 3,67 3,69 Imp. G 3,82 3,84 3,76 3,77 3,75 3,78 3,79 3,81 3,80 3,82 verification point 5 verification point 6 correlation between exp. vs. pred. RT Flow rate [ml/min] tg [min] Temp [ C] ph %start 11 9 %end Retention time [min] Pred. Exp. Pred. Exp. Imp. A 1,07 1,21 1,10 1,19 Imp. I 1,54 1,61 1,53 1,62 Imp. E 1,81 1,86 1,77 1,85 Imp. D 2,14 2,18 2,07 2,16 Imp. B 2,30 2,32 2,22 2,30 Omeprazole 2,38 2,40 2,30 2,38 Imp. H 2,84 2,85 2,72 2,80 Imp. C 3,11 3,10 2,96 3,04 Imp. F 3,88 3,84 3,66 3,71 Imp. G 4,02 3,97 3,79 3,84

Verification of the HPLC method working point verification point 1 verification point 2 verification point 3 verification point 4 Flow rate [ml/min] 1.9 1.9 2.0 1.9 1.8 tg [min] 7.0 6.

60 Verification of the HPLC method working point verification point 1 verification point 2 verification point 3 verification point 4 Flow rate [ml/min] tg [min] Temp [ C] ph %start %end Retention time [min] Pred. Exp. Pred. Exp. Pred. Exp. Pred. Exp. Pred. Exp. Imp. A 1,51 1,64 1,59 1,71 1,52 1,61 1,37 1,49 1,51 1,69 Imp. I 2,10 2,09 2,18 2,19 2,05 2,04 1,84 1,83 1,99 2,02 Imp. E 2,54 2,49 2,61 2,57 2,50 2,44 2,31 2,25 2,43 2,42 Imp. D 3,01 2,98 3,06 3,05 2,95 2,92 2,69 2,65 2,81 2,81 Imp. B 3,35 3,26 3,38 3,31 3,31 3,21 3,15 3,06 3,24 3,19 Omeprazole 3,50 3,40 3,52 3,44 3,45 3,35 3,31 3,21 3,39 3,32 Imp. H 4,24 4,13 4,23 4,14 4,20 4,09 4,06 3,94 4,10 4,03 Imp. C 4,73 4,59 4,69 4,58 4,70 4,55 4,64 4,51 4,65 4,55 Imp. F 6,00 5,84 5,90 5,77 5,97 5,81 5,95 5,77 5,89 5,78 Imp. G 6,23 6,08 6,12 5,99 6,20 6,04 6,18 6,03 6,10 6,00 verification point 5 verification point 6 Flow rate [ml/min] tg [min] Temp [ C] ph %start 11 9 %end correlation between exp. vs. pred. RT Retention time [min] Pred. Exp. Pred. Exp. Imp. A 1,52 1,56 1,59 1,67 Imp. I 2,21 2,19 2,26 2,26 Imp. E 2,66 2,55 2,66 2,58 Imp. D 3,24 3,20 3,18 3,17 Imp. B 3,51 3,36 3,43 3,31 Omeprazole 3,66 3,50 3,56 3,44 Imp. H 4,46 4,28 4,28 4,15 Imp. C 4,92 4,72 4,70 4,54 Imp. F 6,27 6,05 5,90 5,73 Imp. G 6,52 6,29 6,12 5,95

61 Implementation of both methods With this approach it is possible to switch between HPLC and UPLC instruments Chromatographic UPLC condition HPLC condition parameter Column ACQUITY BEH C18; 2.1 x 50 mm, 1.7 µm XBridge BEH C18; 4.6 x 50 mm, 2.5 µm Eluent A Eluent B 10mM ammoniumbicarbonat buffer ph 8.75 (± 0.1 ph units) acetonitrile Gradient linear increase from 10% (±1%) to 60% (±1%) of eluent B in 4.0 min (±0.05 min), followed by re-equilibration linear increase from 10% (±1%) to 60% (±1%) of eluent B in 7.0 min (±0.5 min), followed by re-equilibration Stop time 5 min 8 min Flow rate 0.70 ml/min (±0.05 ml/min) 1.90 ml/min (±0.05 ml/min) Column temp. 35 C (±2 C) Injection volume 2 µl 20 µl Detection 303 nm

62 UPLC Method Imp. A Imp. I Imp. E Imp. D Imp. B Omeprazol Imp. H Imp. C Imp. F Imp. G time (min) Imp. A Omeprazol Imp. I Imp. E Imp. D Imp. B HPLC Method Imp. H Imp. C Imp. F Imp. G time (min) Ph.Eu. Method Imp. E Imp. F+G Imp. Imp. D B Imp. H Imp. C Imp. A I time (min)

73 developed analytical methods: 58 methods for pharma 05 methods for chemistry 04 methods for food 04 methods for clinic / doping 02 methods for renewable energies comparison of 73 adapted methods

63 73 developed analytical methods: 58 methods for pharma 05 methods for chemistry 04 methods for food 04 methods for clinic / doping 02 methods for renewable energies comparison of 73 adapted methods vs. Chromicent developed methods 100,0% 100,0% 90,0% 80,0% purity testing in h 70,0% 60,0% 50,0% 40,0% 30,0% 20,0% 7,3% (14x) 3,3% (30x) 10,0% 0,0% Time & eluent adapted method time new method Eluent new method

64 Thank you very much for your interest!

[ APPLICATION NOTE ] INTRODUCTION APPLICATION BENEFITS WATERS SOLUTIONS KEYWORDS

[ APPLICATION NOTE ] INTRODUCTION APPLICATION BENEFITS WATERS SOLUTIONS KEYWORDS MS Identification of Trace level Impurities from a Non-MS Compatible Mobile Phase Using ACQUITY UPLC System with 2D Technology by Heart-cutting and Online Sample Concentration Bronsky Gopinadh, Dilshad