1. An Industry Perspective on the Potential for Emerging Process Analytical Technologies Norman Winskill Steve Hammond Pfizer Global Manufacturing Services

2. Agenda l History of PAT and vision for the future l Some specific, current PAT applications of interest l Possible implementation scenarios for these and other applications - shaping the future l Proposals to ensure that implementation proceeds as we ALL would like u “the win-win scenario”

3. Evolution of NIR/PAT Within Pfizer l Mid ‘80s u Control of fermentations, hazardous reactions, solvents, raw materials, packaging materials, drug product process troubleshooting l Early ‘90s u Dedicated group formed (‘90), libraries for APIs and RMs developed, automated methods for sample presentation, some DP applications. Still largely NIR based l Late ‘90s u Other techniques emerge (Raman, vision systems, acoustic, LIF etc. Increasingly used in DP processes

4. PAT Applications at DP Sites l RM Testing (warehouse based) l Packaging Components l Blending (at-line or on-line) l Drying l Tableting (potency and CU) l Encapsulation (potency and CU) l Tablet Coating (coating thickness) l Packaged product l Equipment cleaning (on line monitoring of CIP) l Equipment cleaning (surface monitoring) Note - Less than 15% of applications at US sites

5. What Does the Future Hold? l Significant increase in number of applications l Broadening of type of analysis e.g. mid-IR, Raman, LIF, acoustic, vision l Availability of “off the shelf” solutions, including analyzers offered as options by equipment vendors l Integration of applications. Used for “Continuous Quality Verification” as opposed to control of discrete unit operations

6. Shift the Manufacturing Paradigm

7. Changing the Manufacturing Paradigm - Challenges l Considerable progress in some areas u chemometrics, fiber optics, “industrialization” of some instruments (e.g. NIR) l Opportunities still exist in other areas u smaller, faster, cheaper instruments (e.g. Raman) u sample interfaces l Real or perceived regulatory hurdles may effect the way PAT is implemented Technical Regulatory

8. On-Line analysis of blending l Driven by safety issues with new potent API’s l Uses battery power and radio communication l Small fast diode array instrument l Mounted on the moving blender l Controlled outside of containment area l Results appear outside of containment area

9. Layout of Containment Blending Room Corridor Blender Room 2nd Containment Room 1st Containment Room Store Room PC with RF Link Blender with NIR & RF Link ~25 feet

10. On- line NIR bin blender Reading Head NIR Instrument Battery RF Module Fibre optic

11. Corona Sensor Head Light Lid of blender & window Focal Point just inside bin 30mm Fibre Optic Collector Fibre Optic to Detector Contributing sample weight is ~300mg

12. Saccharin Specific Absorption Saccharin Avicel Lactose Saccharin increasing absorbance with blending

13. Absorbance of Blend Components (Step 1)

14. Blend Uniformity All ingredients(Step 1)

15. Magnesium Stearate Uniformity (Step 2)

16. On-line blending - benefits l No operator contact l No sampling errors - no thief l real-time information l Multi-ingredient uniformity l Process understanding l Process finger- printing for scale up l Right first time l Fast release of the blend - reduced cycle times

17. Tablet analysis by NIR transmission l Started at-line with plant operators analysing 200 - 400 cores per lot l With containment issues needs to be non attended l Development of a fully automated test station sited at the press l NIR chemical analysis - weight thickness hardness by normal module l Integrated into one unit.

18. NIR in Production

19. NIR Tablet Transmission Device

20. Tablet core potency - blend segregation in the bin

21. Automated tablet analysis NIR Identity Potency Uniformity Bruker Weight Thickness Hardness Dr S Pharmatron

22. Absorbance of active tablets Vs placebos

23. NIR transmission Vs conc in tablets 0.1% - 2.0%

24. Microscopic View of Dosage Form

25. NIR Microscopy mapping process

26. Distribution of Lubricant in Blend Spectra obtained every 15  m, 0.6mm X 0.6mm area Good Flow BlendBad Flow Blend

27. DiluentBinderActiveDisintegrantLubricant Maps of sticking and non-sticking tablets Non sticking Sticking

28. MatrixNIR - Spectral Dimensions

29. Changing the Manufacturing Paradigm - Challenges l Considerable progress in some areas u chemometrics, fiber optics, “industrialization” of some instruments (e.g. NIR) l Opportunities still exist in other areas u smaller, faster, cheaper instruments (e.g. Raman) u sample interfaces l Real or perceived regulatory hurdles may effect the way PAT is implemented Technical Regulatory

30. The “Don’t Use” Scenario l What: u Modern PAT methods not used/developed during product development so not used for routine process control l Why: u Fear of regulatory delays u Wasteful of resources to duplicate method development - “current methods work OK” u Concern of “raising the bar” unnecessarily: information generated for one process may be expected from all l Issues: u Loss of benefit of PAT - improved process information and control

31. Example of “Don’t Use” - Antifungal Polymorph Conformation l Powder XRD l Established technique l Not common in IPC/QO labs l Did not exist at site of manufacture l Samples sent to another site 3,500 miles away for release (LT 1 week plus) l NIR l New technique (for polymorph) l Common in IPC/QO labs l Existed at site of manufacture l All assays could be done on site (LT minutes if necessary) Method A Method B

32. The “Don’t Tell” Scenario l What: u PAT methods not registered but used in parallel with registered (conventional) methods to gain greater process insight and control l Why: u Concern over delays in regulatory approval u Concern that data may be interpreted inappropriately by regulators. More data will lead to more deviations from “norms” - need to be able to determine which are relevant and which are not l Issues: u Duplication, inefficiency, environment of “mistrust”

33. Control of Antibiotic Fermentation

34. Benefits of Improved Control l Conventional methods give product which is fit for intended use, but… l Advanced control gives u better batch to batch consistency u better quality (less impurities) u fewer reworks/rejects u better productivity u lower cost u improved process understanding u faster response times

35. The “Win - Win” Scenario l What: u Modern PAT methods used to gain greater understanding of processes during development, are registered and used as in-process control (and release?) methods u PAT methods accepted as alternatives to traditional lab based methods - but not required l Why: u Methodology understood and accepted by regulators and industry alike l Issues: u This is where we should all want to get to. Making progress, but we are not there yet……….

36. How can we create a “Win-Win” Environment? l Sponsor joint forums to promote discussion and enhance understanding of the issues and opportunities offered by PAT l Develop an effective process for the evaluation of new PATs l Develop appropriate guidelines for the development, validation and implementation of new PATs u lab based extraction/chromatography rules don’t apply u participate in “dummy run” submissions l Ensure consistent approach to PAT by Review and Inspection Dealing with the real or perceived regulatory hurdles: