Elemental Impurity Analysis: Technical Challenges and Overview of an Inter-laboratory Study PQRI Elemental Impurities Workshop March 31, 2015

Transcription

1 Elemental Impurity Analysis: Technical Challenges and Overview of an Inter-laboratory Study PQRI Elemental Impurities Workshop March 31, 2015 Donna Seibert on behalf of the Technical and Analytical Challenges Sub-team to the Coalition for Rational Implementation

2 Elemental Impurities Analytical Considerations Key challenge for industry and the regulatory community Analytical testing required to quantitate elemental impurities in drug products, APIs and excipients down to the levels listed in USP <232> and/or ICH Q3D guideline. Current work being done throughout the industry is showing that the analytical challenges may be greater than previously anticipated by regulators and pharmacopeias. Variety and complexity of pharmaceutical samples Many labs expanding capabilities Pharmaceutical labs adapting to ICP-MS analysis Existing spectroscopy labs adapting to the requirements of <233>

3 TECHNICAL/ANALYTICAL CHALLENGES PROJECT TEAM Sub-Team Objectives: Communicate the complexity of ICP/MS methodology (including, as appropriate, difficult and/or hazardous sample preparation studies) and portray a realistic picture of the technique for the benefit of industrial partners as they build expertise in elemental impurities analysis Compile, elucidate, and bring visibility to technical and analytical challenges that RM suppliers and FP manufacturers are facing during implementation of USP/ICH Elemental Impurities controls. Bring industrial, regulatory, and academic partners together to propose solutions to analytical challenges wherever possible. 3

4 TECHNICAL/ANALYTICAL CHALLENGES PROJECT TEAM Team Chartered June 2013 Membership (42 colleagues) Comprised of scientists from Coalition companies 5 pharmaceutical companies 7 raw material suppliers (API/excipient) 8 Contract laboratories 1 Government laboratory 1 University laboratory

5 Technical Challenges ICP/MS analysis requires specialized experience required around Maintenance/operation MS tuning Interference interpretation & remediation Chemical incompatibilities b/w analytes Use of internal standards, reaction gases, correction equations, etc. Sample preparation concerns Variety of pharmaceutical samples Aggressive microwave digestions Total metal extraction vs. acid leach methods High dissolved solids, acid concentrations, incomplete digestions, etc.

6 SAMPLE PREPARATION GAP Reference Materials Pharmaceutically Relevant Reference Materials Not established Not currently required by USP <233> Sample spiking is not indicative of extraction Pre-digestion spikes are not adequate for establishing recovery of elements from samples Existing certified materials found not relevant Sands, soils, vitamins Element identity and/or levels not suited to <232> purposes Currently, there exists no standardized way of assessing lab-to-lab consistency and validity of data 6

7 Inter-laboratory Collaborative Study Overview OBJECTIVES: Inter-laboratory data comparison for standardized samples Inter-laboratory evaluation of effectiveness of microwave digestion Compare bio-relevant leach/extraction techniques to total metal extraction Determine the correlation (good or bad) between the analysis of individual components vs. the formulated tablet (summation) 7

8 Inter-laboratory Collaborative Study Overview EVALUATION SAMPLES: Tableted evaluation samples prepared with and without spikes to show recovery via different methods of digestion Group A: Native levels provided by inclusion of magnesium aluminum silicate clay Group B: Enhanced levels added through inclusion of spiked lactose Material Percent mg/tablet Active (Replaced with Microcrystalline Cellulose) Magnesium Aluminum Silicate, NF Lactose, unspiked or spiked Pregelatinized Starch Stearic Acid Fumed Silicon Dioxide TOTALS Evaluation samples manufactured with commonly used excipients at Liverpool John Moores University Enhanced tablets spiked with ICH Class 1 & 2A elements As, Cd, Pb, Hg, Co, Mo Se, V (per ICH guidance at the time of tablet manufacture) Excipients used were analyzed to determine input values for summation

9 Inter-laboratory Collaborative Study Overview Identified testing labs: 13 member laboratories participated in the collaborative study Blinded participation was a pre-requisite for several laboratories Established standardized methods: Standard protocols established for all labs to use (Uniform method) Additional methods can also be performed for comparison (Non-uniform methods) Includes total dissolution and acid leach sample preparation methods Established uniform reporting template Distributed evaluation samples: Shipped to participating labs in April 2014 Final summary report: Project team to collect, review and assess results from each lab and summarize results for inter-lab consistency

10 Sample Preparation Variations EQUIPMENT Microwave models Individual Reaction Vessels Anton Paar MultiWave 3000 CEM Mars 5 and 6, Express CEM SPD Milestone Ethos Plus Single Reaction Chamber Shakers Heating Blocks Milestone UltraWave Milestone UltraClave III METHODS Uniform Method: Microwave Digestion 10:1 HNO 3 :HCl Ramp to 200 C over 20 min, hold 2 min Dilute to 50 ml final volume Non-Uniform Methods: Microwave Digestion Various acid combinations (HNO 3, HCl, HF, H 3 PO 4, HBF 4, HClO 4 ) Acid Leach Methods: Heating block or shaker Glycine/HCl 1 N HNO 3 HNO 3 + H 2 O 2 10

11 Instrumental Analysis Variations INSTRUMENTS ICP/MS Instrument Models Agilent 7500 Series (3) Agilent 8800 triple quad Bruker 820MS Perkin Elmer Nexion 300 Q Perkin Elmer Elan 6100 Perkin Elmer Elan DRC II Thermo Fisher Element II (high res) Thermo Fisher ICAP Q Thermo X-Series II ICP/OES Instrument Model Thermo icap Duo 6500 VARIABLE PARAMETERS Reaction gases H 2, He, 10% H 2 /He, O 2, NH 3, Internal standards Generally matched to analyte mass or ionization potential Across labs, 5-6 different internal standards used per element Correction equations Generally not used. Exceptions were Se and Co 11

12 Data Visualization Orientation Element Native Enhanced Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

13 Data Visualization Orientation Element Mean (ppm): STD: RSD (%): ND values: Native Enhanced Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

14 Data Visualization Orientation Element Mean (ppm): STD: RSD (%): ND values: Native Enhanced Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

15 Data Visualization Orientation Element ICH limit indicated when on scale Mean (ppm): STD: RSD (%): ND values: Native Enhanced Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

16 Data Visualization Orientation Element Mean (ppm): STD: RSD (%): ND values: Native Enhanced Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

17 Data Visualization Orientation Element Mean (ppm): STD: RSD (%): ND values: Native Enhanced Deviations from target spike are not critical. Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

18 Uniform Digestion Arsenic Native Enhanced Mean (ppm): STD: RSD (%): ND values: 12 0 Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

19 Uniform Digestion Cadmium Native Enhanced Mean (ppm): STD: RSD (%): ND values: 3 1 Native levels of Cd are insignificant Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

20 Uniform Digestion Lead Native Enhanced Mean (ppm): STD: RSD (%) ND values: 3 1 Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

21 Uniform Digestion Mercury One outlier excluded from plot: 8.31 ppm Native Enhanced Mean (ppm): STD: RSD (%): ND values: 9 6 Hg means significantly affected by outlying values Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

22 Uniform Digestion Cobalt Native Enhanced Mean (ppm): STD: RSD (%): ND values: 3 0 Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

23 Uniform Digestion Molybdenum Native Enhanced Mean (ppm): STD: RSD (%): Blue points are below LOQ, ND values: 13 0 Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

24 Uniform Digestion Selenium Native Enhanced Mean (ppm): STD: RSD (%): ND values: Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range Selenium below detection limits for many labs o Open circles indicate outliers

25 Uniform Digestion Vanadium Native Enhanced Mean (ppm): STD: RSD (%): ND values: 3 3 V means significantly affected by outlying values Blue points are below LOQ, Error bars set at 1.5 x Interquartile Range o Open circles indicate outliers

26 ppm Data Refinement Lead by Uniform Digestion Original 07-Jan-2015 Updated 05-Mar-2015 Orange data points indicate outliers o Open circles indicate outliers Error bars set at 1.5 x Interquartile Range

27 ppm Data Refinement Lead by Uniform Digestion Original 07-Jan-2015 Updated 05-Mar-2015 Orange data points indicate outliers o Open circles indicate outliers Error bars set at 1.5 x Interquartile Range Labs benefit from having comparator samples

28 ppm Non-Uniform Methods using HF Lead Results Native 6 HF methods vs. uniform digestions Limited ability to generate clear digestates (total digestion) No discernable increase in extraction of Pb from tablets with native levels 28

29 ppm Non-Uniform Methods using HF Lead Results Native 6 HF methods vs. uniform digestions Limited ability to generate clear digestates (total digestion) No discernable increase in extraction of Pb from tablets with native levels Use of HF does not guarantee clear digestate or total digestion 29

30 Acid Leach vs. Microwave Digestion For tablet analysis, some elements appear to show lower levels with shaker/acid leach methods. Caveats: Only native level tablets are relevant Many elements are near LOQ in these tablets Very few replicates within the data set (2 labs) More work is needed to assess differences between acid leach and microwave digestion methods 30

31 Summation vs. Tablet Testing (Native Levels) 31

32 Summation vs. Tablet Testing (Native Levels) Summation from raw material analysis reasonably approximates finished product testing for some elements 32

33 Data Overview Native Enhanced All elements across all labs Statistical analyses* indicate that the variability for each element across labs is significantly higher with the uniform method than the non-uniform method * Measurement Systems Analysis and Assessment of Variability Labs perform best when methodology is optimized for their instrumentation 33

34 Summary Statistics All Methods Uniform, Native Levels Element Mean (ppm) STD RSD (%) As Cd Pb Hg Co Mo Se V Non-uniform, Native Levels Element Mean (ppm) STD RSD (%) As Cd Pb Hg Co Mo Se V Uniform, Enhanced Levels Element Target (ppm) Mean (ppm) STD RSD (%) As Cd Pb Hg Co Mo Se V Non-uniform, Enhanced Levels Element Target (ppm) Mean (ppm) STD RSD (%) As Cd Pb Hg Co Mo Se V

35 Conclusions / Future Direction CONCLUSIONS Data for standardized samples show high variation across laboratories Labs benefit from access to standardized evaluation samples Tighter variation among non-uniform methods suggests need for flexibility in methodology for testing labs FURTHER WORK Data analysis is ongoing; further evaluation needed around Microwave type Total digestion vs. acid leach methods ICP/MS variations Possibility of a second round of testing? 35

36 Study Participants and Acknowledgements AQura Chemical Solutions ColorCon Dow Corning ELS JM Huber Liverpool John Moores Univ. Perrigo Pfizer P&G RTI Solvias US Geological Survey Cornel Venzago Francine Walker Gary Hayes Jason Sturm Brendan Murray Hal Garber, John Offidani Phil Riby James Bennett, Andy Kampfschulte, Josh Foote, Rob Lievense, Matt Schmitt, Saul Gylys Sam Powell, Nikki Clements Denise McClenathan, Roy Dobson, Andrei Shauchuk James Harrington Gisela Fontaine, Josephine Archinal Ruth Wolf and all other TAC Team members, Thank You!!! 36