Statistical Statistical Evaluation Using Design of Evaluation Using Design of Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis (TOC) to Combine Cleaning Agent Method with (TOC) to Combine Cleaning Agent Method with Drug Product Methods Drug Product Methods Pt W l h Nik jV M i N it h Pt W l h Nik jV M i N it h Peter Walsh, Nikunj Vasoya, Mariann Neverovitch, Peter Walsh, Nikunj Vasoya, Mariann Neverovitch, and Antonio Fernandez and Antonio Fernandez Analytical and Analytical and Bioanalytical Bioanalytical Development, Development, Bristol Bristol-Myers Squibb Co., New Brunswick NJ Myers Squibb Co., New Brunswick NJ 19 19 th th November, 2013 November, 2013 1
12
Embed
Lean Clean - Total Organic Carbon Analysis for Cleaning Validation in Pharmaceutical Manufacturing.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
StatisticalStatistical Evaluation Using Design of Evaluation Using Design of Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis Experiments of Total Organic Carbon Analysis (TOC) to Combine Cleaning Agent Method with (TOC) to Combine Cleaning Agent Method with
Drug Product MethodsDrug Product Methods
P t W l h Nik j V M i N it hP t W l h Nik j V M i N it h
gg
Peter Walsh, Nikunj Vasoya, Mariann Neverovitch, Peter Walsh, Nikunj Vasoya, Mariann Neverovitch, and Antonio Fernandezand Antonio Fernandez
Analytical and Analytical and BioanalyticalBioanalytical Development, Development, BristolBristol--Myers Squibb Co., New Brunswick NJ Myers Squibb Co., New Brunswick NJ
1919thth November, 2013November, 2013
1
OutlineOutline
Cleaning BackgroundCleaning Background Current Practice & Project GoalCurrent Practice & Project Goaljj Design of Experiments (DoE)Design of Experiments (DoE) Data Analysis & ResultsData Analysis & Results Data Analysis & ResultsData Analysis & Results Risk AnalysisRisk Analysis ConclusionConclusion
2
Why Do We Clean?Why Do We Clean?yy
Cleaning is importantCleaning is important EquipmentEquipmentCleaning is importantCleaning is importantEvery time we make a batch of Drug Every time we make a batch of Drug Product we have a chance of Product we have a chance of contamination from the previous batchcontamination from the previous batchcontamination from the previous batchcontamination from the previous batchWe avoid this by cleaning inWe avoid this by cleaning in--between between runs with detergent. Then in order to runs with detergent. Then in order to verify it’s clean we take a swab and test itverify it’s clean we take a swab and test itverify it s clean we take a swab and test it verify it s clean we take a swab and test it for API/detergent.for API/detergent.
3
What Is Considered Clean?What Is Considered Clean?
Cleaned EquipmentCleaned EquipmentCleaned EquipmentCleaned EquipmentThe equipment is considered clean if the swab sample is below The equipment is considered clean if the swab sample is below its worst case Permitted Residue Limit (PRL). This is calculated its worst case Permitted Residue Limit (PRL). This is calculated from the contaminant’s Subject Exposure Limit (SEL) provided from the contaminant’s Subject Exposure Limit (SEL) provided by a toxicologist.by a toxicologist.
Cleaning Agent Limit:Cleaning Agent Limit:
PRL Limit: 13.5 parts per million Carbon (ppmC) PRL Limit: 13.5 parts per million Carbon (ppmC) Cleaning Limit: 1.00 ppmC (Alert Cleaning Limit)Cleaning Limit: 1.00 ppmC (Alert Cleaning Limit)
4
Total Organic Carbon AnalysisTotal Organic Carbon Analysisg yg y
The TOC method is nonThe TOC method is non--specific in that it analyzes specific in that it analyzes how much organic carbon how much organic carbon i i l tii i l tiis in a solution.is in a solution.
This means ANY and ALLThis means ANY and ALLThis means ANY and ALL This means ANY and ALL contamination will be contamination will be accounted for in one accounted for in one analysis.analysis.
5
Current Project And Project GoalCurrent Project And Project Goalj jj jBlender (Equipment to clean)Blender (Equipment to clean)
Current PracticeCurrent Practice Project GoalProject Goal
Drug Drug ProductProduct
Cleaning Cleaning AgentAgent Drug Product + Cleaning Drug Product + Cleaning
AgentAgentExample Settings:Example Settings:
•• 20 mL20 mL•• pHpH--22•• 2.0 µL/min2.0 µL/min
Example Settings:Example Settings:
•• 40 mL40 mL•• pHpH--77•• 0.0 µL/min0.0 µL/min
AgentAgent
Our Goal: To analyze only one sample for Our Goal: To analyze only one sample for both Drug Product and Cleaning Agent.both Drug Product and Cleaning Agent.
2.0 µL/min2.0 µL/min 0.0 µL/min0.0 µL/min
Drug Product and Cleaning Agents Drug Product and Cleaning Agents can be swabbed separately due to can be swabbed separately due to
different optimal solubility different optimal solubility parameters.parameters.
To do this we needed to prove that no To do this we needed to prove that no variation occurs when changing the variation occurs when changing the parameters for the Cleaning Agent.parameters for the Cleaning Agent.
6
Design of Experiment (DoE)Design of Experiment (DoE)g p ( )g p ( )Full Factorial DoE was conducted with 3 factors, 2 levels, Full Factorial DoE was conducted with 3 factors, 2 levels,
and 5 replicates in the statistical program Minitaband 5 replicates in the statistical program Minitab
Main Effects PlotsMain Effects PlotsDoE results for Cleaning AgentDoE results for Cleaning Agent
Little to NO Little to NO slope!slope!
Very slight Very slight Variation forVariation forVariation for Variation for
OxidizerOxidizer
Lowest Lowest Reading Reading
Observed was Observed was 0.15 ppmC 0.15 ppmC less than less than
8
theoreticaltheoretical
Interaction Plots Interaction Plots DoE results for Cleaning AgentDoE results for Cleaning Agent
-- No angles!No angles!
-- Lines are Lines are almost almost
completely completely p yp yoverlapping!overlapping!
9
Quantifying Oxidizer Variation Quantifying Oxidizer Variation Q y gQ y gOxidizer was individually varied for Cleaning AgentOxidizer was individually varied for Cleaning Agent
Therefore the Therefore the detergent detergent gg
contamination will contamination will ALWAYS be well ALWAYS be well
within the PRL limit.within the PRL limit.
NO RISK!NO RISK!
11
ConclusionConclusion
Based on the data gathered we conclude Based on the data gathered we conclude there is NO significant variation when there is NO significant variation when changing parameters for both detergentschanging parameters for both detergentschanging parameters for both detergents.changing parameters for both detergents.
Th f bi l i fTh f bi l i fTherefore, we can combine analysis of Therefore, we can combine analysis of Detergent with API.Detergent with API.