1、Product Development:Case Study Overview, ICH, November 2010,Disclaimer,The information within this presentation is based on the ICH Q-IWG members expertise and experience, and represents the views of the ICH Q-IWG members for the purposes of a training workshop., ICH, November 2010,Outline of Presen
2、tation,Key Steps for Quality by DesignCase Study OrganizationIntroducing API and Drug Product Discussion of concepts of Quality Target Product Profile, processes, compositionDescription of API & Drug Product process development Discussion of illustrative examples of detailed approaches from the case
3、 studyBatch release, ICH, November 2010,Key Steps for a product under Quality by Design (QbD),Product/Process Development,Pharmaceutical Development,PQS & GMP,Local Environment,Commercial Manufacturing,Quality Unit (QP,.) level support by PQS,Manage product lifecycle, including continual improvement
4、,Design Space (DS), RTR testing,Link raw material attributes and process parameters to CQAs and perform Risk Assessment Methodology,Potential CQA (Critical Quality Attribute) identified & CPP (Critical Process Parameters) determined,QTPP : Definition of intended use & product,Quality TargetProduct P
5、rofile,CPP : CriticalProcess Parameter,CQA : CriticalQuality Attribute,Risk Management,Opportunities,Design to meet CQA using Risk Management & experimental studies (e.g. DOE),DOE : Design of Experiment,Control Strategy,Technology Transfer,Batch ReleaseStrategy,Prior Knowledge (science, GMP, regulat
6、ions, .),Continualimprovement,Product/Process Understanding,QRM principle apply at any stage,Marketing Authorisation,Quality System PQS, ICH, November 2010,Purpose of Case Study,Illustrative exampleCovers the concepts and integrated implementation of ICH Q8, 9 and 10Not the complete content for a re
7、gulatory filingNote: this example is not intended to represent the preferred or required approach., ICH, November 2010,Case Study Organization, ICH, November 2010,Basis for Development Information,Fictional active pharmaceutical ingredient (API) Drug product information is based on the Sakura Tablet
8、 case studyFull Sakura case study can be found at http:/www.nihs.go.jp/drug/DrugDiv-E.html Alignment between API and drug productAPI Particle size and drug product dissolutionHydrolytic degradation and dry granulation /direct compression, ICH, November 2010,Organization of Content,Quality Target Pro
9、duct Profile (QTPP)API properties and assumptionsProcess and Drug product composition overviewInitial risk assessment of unit operationsQuality by Design assessment of selected unit operations, ICH, November 2010,Quality attribute focus,Technical Examples,APIDrug Product,Compression,Real Time Releas
10、e testing(Assay, CU, Dissolution),Blending,APICrystallization,- Final crystallization step,- Blending- Direct compression,- Particle size control,- Assay and content uniformity - Dissolution,Process focus, ICH, November 2010,Process Step Analysis,For each exampleRisk assessmentDesign of experimentsE
11、xperimental planning, execution & data analysisDesign space definitionControl strategyBatch release,Design ofExperiments,Design Space,Control Strategy,Batch Release,QRM, ICH, November 2010,QbD Story per Unit Operation,Process Variables,Design ofExperiments,QualityRisk Management,Illustrative Example
12、s of Unit Operations:,QTPP & CQAs,Design Space,Control Strategy,Batch Release,Compression,Real Time Release testing(Assay, CU, Dissolution),Blending,APICrystallization, ICH, November 2010,Introducing API and Drug Product, ICH, November 2010,Assumptions,API is designated as AmokinolSingle, neutral po
13、lymorphBiopharmaceutical Classification System (BCS) class II low solubility & high permeabilityAPI solubility (dissolution) affected by particle sizeDegrades by hydrolytic mechanismIn vitro-in vivo correlation (IVIVC) established allows dissolution to be used as surrogate for clinical performanceDr
14、ug product is oral immediate release tablet, ICH, November 2010,Assumptions & Prior Knowledge,API is designated as AmokinolSingle, neutral polymorphBiopharmaceutical Classification System (BCS) class II low solubility & high permeabilityAPI solubility (dissolution) affected by particle sizeCrystalli
15、zation step impacts particle sizeDegrades by hydrolytic mechanismHigher water levels and elevated temperatures will increase degradationDegradates are water soluble, so last processing removal point is the aqueous extraction stepDegradates are not rejected in the crystallization stepIn vitro-in vivo
16、 correlation (IVIVC) established allows dissolution to be used as surrogate for clinical performanceDrug product is oral immediate release tablet, ICH, November 2010,Quality Target Product Profile (QTPP)Safety and Efficacy Requirements,QTPP may evolve during lifecycle during development and commerci
17、al manufacture - as new knowledge is gained e.g. new patient needs are identified, new technical information is obtained about the product etc., ICH, November 2010,API Unit Operations,Coupling Reaction,Aqueous Extractions,Distillative Solvent Switch,Semi ContinuousCrystallization,Centrifugal Filtrat
18、ion,Rotary Drying,Coupling of API Starting Materials,Removes water, prepares API for crystallization step,Addition of API in solution and anti-solvent to a seed slurry,Filtration and washing of API,Drying off crystallization solvents,Removes unreacted materials. Done cold to minimize risk of degrada
19、tion,Understand formation & removal of impurities,Example from Case Study, ICH, November 2010,Tablet Formulation, ICH, November 2010,Drug Product Process,Blending,Lubrication,Compression,Film coating,API and ExcipientsAmokinolD-mannitolCalcium hydrogen phosphate hydrateSodium starch glycolate,Lubric
20、antMagnesium Stearate,CoatingHPMC，Macrogol 6000titanium oxideiron sesquioxide, ICH, November 2010,Overview of API and Drug Product Case Study ElementsRepresentative Examples from the full Case Study, ICH, November 2010,Overall Risk Assessment for Process,Process Steps,CQA,Example from Case Study, IC
21、H, November 2010,Overall Risk Assessment for Process,Process Steps,CQA, ICH, November 2010,API Semi-Continuous Crystallization,Designed to minimize hydrolytic degradation (degradate below qualified levels)Univariate experimentation exampleFMEA of crystallization process parametersHigh risk for tempe
22、rature, feed time, water levelTest upper end of parameter ranges (represents worst case) with variation in water content only and monitor degradationProven acceptable upper limits defined for above parametersNote that in this case study, the distillative solvent switch prior to crystallization and c
23、rystallization itself are conducted at lower temperatures and no degradation occurs in these steps, ICH, November 2010,API Semi-Continuous Crystallization,Designed to control particle sizeMultivariate DOE example leading to predictive modelFMEA of parameters using prior knowledgeHigh risk for additi
24、on time, % seed, temperature, agitationDOE: half fraction factorial using experimental ranges based on QTPP, operational flexibility & prior knowledgeDesign space based on predictive model obtained by statistical analysis of DOE dataParticle size distribution (PSD) qualified in formulation DOE and d
25、issolution studies, ICH, November 2010,Risk Assessment: Particle Size Distribution (PSD) Control,To be investigatedin DOE, ICH, November 2010,Options for Depicting a Design Space,Large square represents the ranges tested in the DOE.Red area represents points of failureGreen area represents points of
26、 success.,Oval = full design space represented by equation Rectangle represent rangesSimple, but a portion of the design space is not utilizedCould use other rectangles within ovalExact choice of above options can be driven by business factors,For purposes of this case study, an acceptable design sp
27、ace based on ranges was chosen,Seed wt%, ICH, November 2010,Options for Expanding a Design Space,Why expand a Design Space?Business drivers can change, resulting in a different optimum operating spaceWhen is DS Expansion possible?Case A: When the original design space was artificially constrained fo
28、r simplicityCase B: When some edges of the design space are the same as edges of the knowledge space, ICH, November 2010,API Crystallization: Design Space & Control Strategy,Control Strategy should address:Parameter controls Distillative solvent switch achieves target water contentCrystallization pa
29、rameters are within the design space TestingAPI feed solution tested for water contentFinal API will be tested for hydrolysis degradate Using the predictive model, PSD does not need to be routinely tested since it is consistently controlled by the process parameters, ICH, November 2010,Design Space
30、/ Control StrategyParameter controls & Testing,Particle size will be tested in this example, since the result is includedin the mathematical model used for dissolution.,Example from Case Study, ICH, November 2010,Drug Product,Immediate release tablet containing 30 mg AmokinolRationale for formulatio
31、n composition and process selection providedIn vitro-in vivo correlation (IVIVC) determinationCorrelation shown between pharmacokinetic data and dissolution resultsRobust dissolution measurement neededFor a low solubility drug, close monitoring is important, ICH, November 2010,Drug Product Direct Co
32、mpression Manufacturing Process,Focus of Story,Example from Case Study,Lubrication, ICH, November 2010,Initial Quality Risk Assessment,Impact of Formulation and Process unit operations on Tablet CQAs assessed using prior knowledgeAlso consider the impact of excipient characteristics on the CQAs,Exam
33、ple from Case Study, ICH, November 2010,Drug Product CQA Dissolution Summary,Quality risk assessmentHigh impact risk for API particle size, filler, lubrication and compressionFillers selected based on experimental work to confirm compatibility with Amokinol and acceptable compression and product dis
34、solution characteristicsAPI particle size affects both bioavailability & dissolutionMultivariate DOE to determine factors that affect dissolution and extent of their impactPredictive mathematical model generatedConfirmed by comparison of results from model vs. actual dissolution testingPossible grap
35、hical representations of this design space, ICH, November 2010,Predictive Model for DissolutionA mathematical representation of the design space,Factors include: API PSD, lubricant (magnesium stearate) specific surface area, lubrication time, tablet hardness (via compression force),Confirmation of m
36、odel,Example from Case Study,Continue model verification with dissolution testing of production material, as needed, ICH, November 2010,Dissolution: Control Strategy,Controls of input material CQAsAPI particle sizeControl of crystallisation stepMagnesium stearate specific surface areaSpecification f
37、or incoming materialControls of process parameter CPPsLubrication step blending time within design spaceCompression force (set for tablet hardness) within design spaceTablet press force-feedback control systemPrediction mathematical modelUse in place of dissolution testing of finished drug productPo
38、tentially allows process to be adjusted for variation (e.g. in API particle size) and still assure dissolution performance, ICH, November 2010,Drug Product CQA -Assay & Content Uniformity Summary,Quality risk assessmentPotential impact for API particle size, moisture control, blending, and lubricati
39、onMoisture will be controlled in manufacturing environmentConsider possible control strategy approachesExperimental plan to develop design space using input material and process factorsIn-process monitoringAssay assured by weight control of tablets made from uniform powder blend with acceptable API
40、content by HPLCBlend homogeneity by on-line NIR to determine blending endpoint, includes feedback loopAPI assay in blend tested by HPLCTablet weight by automatic weight control with feedback loop, ICH, November 2010,Blending Process Control Options,Decision on conventional vs. RTR testing,Example fr
41、om Case Study, ICH, November 2010,Process Control Option 2 Blend uniformity monitored using a process analyser,On-line NIR spectrometer used to confirm scale up of blendingBlending operation complete when mean spectral std. dev. reaches plateau regionPlateau may be detected using statistical test or
42、 rulesFeedback control to turn off blenderCompany verifies blend does not segregate downstreamAssays tablets to confirm uniformityConducts studies to try to segregate API,Data analysis model will be providedPlan for updating of model available,Acknowledgement: adapted from ISPE PQLI Team,Example fro
43、m Case Study, ICH, November 2010,Conventional automated control of Tablet Weight using feedback loop:Sample weights fed into weight control equipment which sends signal to filling mechanism on tablet machine to adjust fill volume and therefore tablet weight.,Tablet Weight Control in Compression Oper
44、ation, ICH, November 2010,Batch Release Strategy,Finished product not tested for assay, CU and dissolution Input materials meet specifications and are testedAPI particle size distributionMagnesium stearate specific surface areaAssay calculationVerify (API assay of blend by HPLC) X (tablet weight)Tab
45、let weight by automatic weight control (feedback loop), %RSD of 10 tabletsContent UniformityOn-line NIR criteria met for end of blending (blend homogeneity)Tablet weight control results checkedDissolutionPredictive model using input and process parameters calculates for each batch that dissolution m
46、eets acceptance criteriaInput and process parameters used are within the filed design spaceCompression force is monitored for tablet hardnessWater content NMT 3% in finished product (not covered in this case study), ICH, November 2010,Drug Product Specifications,Use for stability, regulatory testing
47、, site change, whenever RTR testing is not possibleInput materials meet specifications and are testedAPI PSDMagnesium stearate specific surface areaAssay calculation (drug product acceptance criteria 95-105% by HPLC)Verify (API assay of blend by HPLC) X (tablet weight)Tablet weight by automatic weig
48、ht control (feedback loop)For 10 tablets per sampling point, 2% RSD for weightsContent Uniformity (drug product acceptance criteria meets compendia)On-line NIR criteria met for end of blending (blend homogeneity)Tablet weight control results checkedDissolution (drug product acceptance criteria min 8
49、5% in 30 minutes)Predictive model using input and process parameters for each batch calculates whether dissolution meets acceptance criteriaInput and process parameters are all within the filed design spaceCompression force is controlled for tablet hardnessWater content (drug product acceptance criteria NMT 3 wt% by KF),
