Implementation of QBD for Analytical Methods - Session Introduction -

Transcription

1 Implementation of QBD for Analytical Methods - Session Introduction - Sonja Sekulic January 24, 2014

2 Quality by Design (QbD) A Systematic Approach ICH Q8(R2) Product Profile Define quality target product profile Critical Quality Attributes Identify critical quality attributes (CQAs) Risk Assessment Link raw material attributes and process parameters to CQAs via experimentation and risk assessment Design Space Develop design space Control Strategy Establish comprehensive control strategy Continual Improvement Manage product lifecycle, including continual improvement 2

3 Analytical QBD Define Objectives (Method Design) Method Selection (Method Design) Develop Method Understanding Develop MODR & Control Strategy Knowledge Management Analytical Target Profile Develop MODR and Control Strategy (Risk Mitigation) Identify Quality Attributes Perform Experimental Strategy Method Understanding Quality Risk Assessment Identify and Prioritize MethodParameters Risk Assessment Prioritize Experiments ID Experiments Understand CQA = f(cpp) 3

4 Process QBD Process Robustness Ability of a process to tolerate variability of materials and changes of the process and equipment without negative impact on quality. [ICH Q8] 4

5 Method QBD Method Robustness "The robustness/ruggedness of an analytical procedure is a measure of its capacity to remain unaffected by small, but deliberate variations in method parameters and provides an indication of its reliability during normal usage" [1]. [1] ICH Harmonised Tripartite Guideline prepared within the Third International Conference on Harmonisation of Technical Requirements for the Registration of Pharmaceuticals for Human Use (ICH), Text on Validation of Analytical Procedures, 1994, ( 5

6 Process Analytics Community Generally require development of online methods such as NIR (obviously there are other technologies being used also) These NIR methods require reference methods (generally HPLC) The reference methods are also required for support of stability studies etc. A natural fit for AQBD/ATP discussion? 6

7 Drug Product Process and Control Strategy Cellulose Microcrystalline Croscarmellose Sodium Lactose Monohydrate API Magnesium Stearate First Blend Sieving Deagglomeration 1st Lubrication (Intragranular) RTRt Control Strategy API Particle Size Distribution Magnesium Stearate Dry Granulation and Milling Opadry II e.g. White Opadry II e.g. Blue Purified Water Preparation of Coating Suspension 2nd Lubrication (extragranular)/ Final Blending Tablet Compression Film Coating Packaging Assay/Content Uniformity (NIR) Identification (NIR) Performance Test (Disintegration) Purity (HPLC) Appearance Water (KF) 7 7 7

8 Analytical QbD for NIR methods Risk assessment conducted with involvement of the commercial site to ensure implementation continuity IR Tablets 8

9 Risk Assessment Qualitative or quantitative process to estimate risk associated with method parameters QRM Tools: - Cause & Effect Matrix - FMEA - Ishikawa Diagram Parameter Parent Parameter Name CU Assay Final Score A. Tablet Properties MCC Accuracy A. Tablet Properties Lactose Accuracy A. Tablet Properties API Distribution within Tablet A. Tablet Properties API Agglomeration A. Tablet Properties API Crystal Form B. Pre-processing Order of pre-processing steps B. Pre-processing Spectral Range for Model B. Pre-processing Spectral Range for SNV Smoothing window size for 2nd B. Pre-processing derivative A. Tablet Properties Water Content B. Weigh Tablets Balance Accuracy A. Tablet Properties API Salt Form B. Weigh Tablets Balance Drift B. Weigh Tablets Vibrations in room C. Model Suitability B. Weigh Tablets Diagnositic Algorithim Balance Precision C. Model Suitability Acceptance Value

10 Further Development of Robust Methods Risk Assessment led to experiments to investigate effects of high, medium risk factors Samples from process QbD study experiments were utilized to cover the process parameters in the design space Potential environmental variability (i.e. moisture content) was considered and built in the model Parameter Ranges Investigated Potency 70%-130% Hardness Low - High API particle size Specification range Process parameters Design space Excipients variability Multiple lots based on CoA space Manufacturing Scale 1.5kg 120kg Sample presentation Variability evaluated Moisture/environmental Specification Data preprocessing Various algorithms/parameters etc etc 10

11 Excipient Understanding 4 Lactose Scores Plot Multivariate PCA Scores of CoA Data MCC Scores PC: tps[2] Scores PC: tps[1] R2X[1] = R2X[2] = Ellipse: Hotelling T2PS (0.95) SIMCA-P+ 11-6/9/2010 4:04:26 PM Actively built-in excipients variation in the development for process understanding and model construction. 11

12 Variance Built-into the final NIR Model Factors Calibration Set Internal Validation External Validation Number of samples Potency % Intent by HPLC Hardness Low, Nominal, High Low, Nominal, High Nominal Thickness Low, Nominal, High Low, Nominal, High Nominal Process parameters Low, Nominal, High Low, Nominal, High Nominal Environmental/ Moisture Content by KF % % 4.0% MCC 2 lots 2 lots 3 lots Excipients Drug Substance Lactose 2 lots 2 lots 3 lots Croscarmellose sodium 1 lot 1 lot 2 lots MgSt 1 lot 1 lot 1 lot Lot 3 lots 3 lots 3 lots Particle size D(4,3) 8-39 micron 8-39 micron 8-32 micron Particle Size D micron micron micron Production Scale kg kg 120 kg Lot 15 lots 15 lots 3 lots (registration stability) It s the variance that matters. 12

13 Method System Suitability Considerations Calibration Production Hotelling T 2 (98.57%) Q Residuals (1.43%) x 10-8 Constructed a Spectral Quality Check(system suitability) test to ensure quality CU prediction is obtained from the NIR model. Provide the link to method lifecycle management - model maintenance

14 Method Positioning on Specification Ultimately filed the NIR and HPLC methods as alternative methods (potency, CU) NIR method was the primary for release Methods were shown to be statistically equivalent Question: does the ATP concept offer any other options for these types of applications? 14

15 How Much Robustness? No product is infinitely robust, nor should it be. It doesn t make financial sense. The product would be infinitely expensive and would take an infinite amount time to develop. But how much robustness is enough? An easier, and possibly more important, question to answer is how much is too little? Or, stated another way, what is the minimum level of product robustness? 15

16 6 Great Presentations & Discussion Panel at the end of the session 16