CPV Slides SK11Aug16 - Part 1 only

1. Stephan O. Krause, PhD BioProcessing Summit, Boston MA Strictly Confidential 18-19 August 2016 CPV Acceptance Criteria and Conditions

2. Outline CPV/Commercial acceptance criteria - non-microbiological CQAs Control strategy d evelopment CPV conditions/rules “QA process” CPV/Commercial acceptance criteria - microbiological CQAs Not presented here but available via ShareSlide The content and views expressed by the author/presenter are not necessarily the views of the organization he represents.

3. 3 Typical CQA Development, CMC Changes, and Specifications From: Krause, S., WCBP, 30Jan13, Washington, DC.

4. Accelerated CQA Development, CMC Changes, and Specifications 4 From: Krause, S., CaSSS CMC Strategy Forum, 27Jan14, Washington, DC.

5. Accelerated CQA Development, CMC Changes, and Specifications 5 From: Krause, S., CaSSS CMC Strategy Forum, 27Jan14, Washington, DC.

6. 6 Typical CQA Development, CMC Changes, and Specifications From: Krause, S., WCBP, 30Jan13, Washington, DC. Formal CPV

7. S. Krause, PDA Annual Meeting - Las Vegas 17March15

8. CPV

9. CPV

10. CPV

11. CPV QA Process

12. Risk Assessment Process During Product Development S. Krause, PDA Annual Meeting - Las Vegas 17March15

13. Assessing Product Quality and/or Process Consistency Impact Critical Process Parameter (CPP): A process parameter whose variability has an impact on a critical quality attribute and therefore should be monitored or controlled to ensure the process produces the desired quality (ICH Q8 (R2)) 8 . Non-Critical Process Parameters are process parameters whose variability has no practically significant impact on critical quality attributes. Non-critical process parameters fall into two categories, Key Process Parameters and Non-Key Process Parameters. Key Process Parameter (KPP): A non-critical process parameter whose variability has a practically significant impact on process performance or process consistency. Non-Key Process Parameter (NKPP): A non-key process parameter is a non-critical process (control) parameter that has no practically significant impact on process performance or process consistency. Courtesy of Gisela Ferreira, MedImmune

14. Output Measurement Impact Type of Criterion / Limit In-Process Control (IPC) Determinant of product quality Acceptance Criteria: “Numerical limits, ranges, or other suitable measures for acceptance of the results of analytical procedures which the drug substance or drug product or materials at other stages of their manufacture should meet.” (ICH Q6B 9 ) Performance Attribute (PA) Used to indicate that the process performed as expected; may include quality measurements that do not directly determine final product quality Action Limit: “An internal (in-house) value used to assess the consistency of the process at less critical steps.” (ICH Q6B 9 ) Assessing Product Quality and/or Process Consistency Impact Courtesy of Gisela Ferreira, MedImmune

15. Considerations for CPV (Limits) Data Transformation of Non-Normal Distributions 15 Original-scaled data Log-scaled data Data are not symmetric around center:  mean and standard deviation not appropriate metrics . Courtesy of Steven Novick, MedImmune

16. Considerations for CPV (Limits) Data Transformation of Non-Normal Distributions 16 F rom : Mark DiMartino et al. on behalf of BioPhorum Operations Group, Responses to Signals from a Continued Process Verification System in the Biopharmaceutical Industry (to be submitted to an Autumn 2016 edition of the ISPE Journal of Pharmaceutical Engineering ) .

17. Considerations for CPV (Limits) Data Transformation of Non-Normal Distributions 17 How many lots before we move the center of CLs ? F rom : Mark DiMartino et al. on behalf of BioPhorum Operations Group, Responses to Signals from a Continued Process Verification System in the Biopharmaceutical Industry (to be submitted to an Autumn 2016 edition of the ISPE Journal of Pharmaceutical Engineering ) .

18. Understanding Campaign Differences and Batch Differences Within a Campaign A 95% upper confidence limit for a standard deviation is K * s , where K is given in the table and s = sample standard deviation. There is high uncertainty in the estimate of campaign-to-campaign variability when the data set contains only two campaigns; the true standard deviation might be 15.9x greater than the existing data . Courtesy of Steven Novick, MedImmune # of campaigns K 2 15.9 3 4.4 4 2.9 5 2.4 6 2.1 10 1.6 25 1.3 30 1.3 100 1.1

19. Example: Drug Substance Specifications and CPV Limit(s)

20. Example: Drug Substance Specifications and CPV Limit(s) Statistical Release Conditions

21. Example: Drug Substance Specifications and CPV Limit(s) “Practical” Alert/Action Conditions (based on non-stats DS release specs)

22. Example: Drug Substance Specifications and CPV Limit(s) OOT Release Conditions for Sequential Batches

23. Example: Drug Substance Specifications and CPV Limit(s) OOT Stability Conditions (OOT at 18M for 20-25 º C)

24. Example: Drug Substance Specifications and CPV Limit(s) OOT Stability Conditions (OOT for 2-8C at 18M)

25. PPQ Specifications , CPV Acceptance Criteria, and Non-Conformance C onditions

26. PPQ Specifications , CPV Acceptance Criteria, and Non-Conformance C onditions

27. PPQ Specifications , CPV Acceptance Criteria, and Non-Conformance C onditions

28. PPQ Specifications , CPV Acceptance Criteria, and Non-Conformance C onditions

29. PPQ Specifications , CPV Acceptance Criteria, and Non-Conformance C onditions

30. Reference CPV/Commercial acceptance criteria – microbiological CQAs Risk assessment process Alert and action level examples (Krause, S., “Alert , Action, and Specification Limits for Bioburden and Endotoxin”, PDA Annual Meeting, 2015)

31. 31