“

”
Mohamad Mukharir, S.Si
 Goals of analytical procedures
 Analytical Procedures Development (APD): Traditional x Enhanced
approaches
 Enhanced approach: Analytical Quality by Design (AQbD)
 Regulatory perspectives for APD and AQbD: ICH Q14 - Analytical
Procedures Development
 Analytical Procedure Development
 The goal and purpose of the method should reflect the phase of drug development

PRECLINICAL
EARLY SAFETY STUDIES
TECHNOLOGY MANUFACTURING POST MARKETING
DEVELOPMENT EARLY CLINICAL TRIALS
TRANSFER PROCESS SURVEILLANCE
DRUG DISCOVERY FORMULATION
DEVELOPMENT

Analytical methods
Analytical Methods support are intended to Methods are developed to support drug testing against
preclinical safety evaluations, establish the identity, specifications during manufacturing and quality release
pre-formulation studies, and purity, physical operations, as well as during long-term stability studies
prototype product stability characteristics,
studies. potency of drug
products and
evaluate drug
performance

Method development is a continuous process that progresses in parallel with the evolution of the drug
product, and may be refined/expanded, based on increased API and drug product knowledge.
 AMD

 Method related activities are iterative,

particularly during early drug
development phases.
 Changes encountered during drug
development may require modifications to
existing analytic methods.
 AMD Approach

 APPROACHES USED FOR ANALYTICAL PROCEDURES DEVELOPMENT

 “Traditional approach”
One-factor-at time experiments (OFAT) testing of factors and their effects one
at a time instead of multiple factors simultaneously

 “Enhanced approach”
Analytical Quality by Design (AQbD) systematic approach which studies
multiple factors simultaneously to evaluate the impact on method
performance
Traditional Development and Transfer
of Methods
Common finding during method transfer
(Traditional Approach )
Traditional Approach
vs
QbD Approach
Enhanced approach - Analytical
Quality by Design (AQbD)

 What can we do with the regression models?

 Variables selection •
 Identify analytical conditions which impact significantly on the analytical responses •
 The interaction between factors can be estimated systematically. –

 Analytical response prediction •

 Access to experimental information in a larger region of the factor space. •
 Identify analytical conditions for analytical response optimization
 QbDApproach for Method Development

 Analytical QbD begins by defining goals (Analytical Target Profile, ATP) and
identifying potential method variables and responses that affect method
quality
 Statistical „Design of Experiments“ (DOE) has been applied to the selected
method variables leading to process and method understanding and to
create a list of critical method parameters (CMPs- flow rate, temperature
etc.) and critical method attributes (CMAs- resolution, peak tailing)
 The experimentally measured responses were modeled to determine the
Design Space (defined in ICH Q8 (R2))
 A verified and validated method can be modified within the Design Space
to compensate any unforseen variables yet delivering consistent results
 AQbD: Regulatory perspective

 ICH Guideline Q8: Pharmaceutical Development (2004) Quality by Design

concept (QbD) “A systematic approach to development that begins with
predefined objectives and emphasizes product and process understanding
and process control, based on sound science and quality risk
management”

 ICHQ8(R2) doesn’t explicitly discuss analytical method development.

However, concepts apply to Analytical Procedure Development!
 Analytical QbD Workflow
•QUALITY TARGET PRODUCT PROFILE (QTPP) Predefined objective that stipulates the performance
1 •ANALYTICAL TARGET PROFILE (ATP) requirements for the analytical procedure

METHOD
DESIGN
•CRITICAL QUALITY ATTRIBUTE(CQA)
Analytical responses which represent the quality of the
•CRITICAL METHOD ATTRIBUTE (CMA) method and reflect method suitable performance
2
LIFECYCLE/KNOWLEDGE

•CRITICAL PROCESS PARAMETER (CPP) Analytical conditions (input factors) which impact
•CRITICAL METHOD PARAMETER (CMP) significantly on method performance - Prior

METHOD UNDERSTANDING
3
knowledge/initial risk assessment
MANAGEMENT

•DESIGN OF EXPERIMENT (DOE) Select suitable experimental design and acquire data,
4 •KNOWLDGE SPACE & DESIGN SPACE Regression analysis/Statistical analysis Design Space=Method
Operable Design Region (MODR) Risk assessment
•WORKING POINTS (Method Operable Design Region (MODR) Select points for experimental
•ROBUSTNESS TEST
5 verification and design space
validation
•CONTROL STRATEGY Assign system suitability test
6
 Analytical Target Profile (ATP)
 General ATP for analytical procedures is as follows:

 Target analytes selection (API and impurities) •

 ICH Q3 and all other regulatory guidance explained the consideration of impurities in the API synthetic
route –

 Technique selection (HPLC, GC, HPTLC, Ion Chromatography, chiral HPLC, etc.)
 Analytical test item and purpose of test are also important for selecting the technique –

 Method requirements selection (assay or impurity profile or residual solvents)

 Method requirements can vary from one method to another. The common ATPs for impurity profile by
HPLC method
 Critical Quality Attributes (CQA)
CQA for analytical methods includes method attributes and method parameters. Each
analytical technique has different CQA.

 HPLC (UV or RID) CQA are mobile phase buffer, pH, diluent, column selection, organic
modifier, and elution method.
 GC methods CQA are gas flow, oven temperature and program, injection temperature,
sample diluent, and concentration.
 HPTLC method CQA are TLC plate, mobile phase, injection concentration and volume, plate
development time, color development reagent, and detection method

Note : Nature of impurities and DS can define the CQA for analytical method development
such as solubility, pH value, polarity, charged functional groups, boiling point, and solution
stability
 DOE
 Enhanced approach - Analytical Quality by Design (AQbD) Input factors will be
varied all at the same time using Design of Experiments (DOE).
 DOE is defined as a branch of applied statistics that deals with:
 Planning, conducting, analyzing,
 Interpreting controlled tests to evaluate the impact of the factors on the process
parameters.
 DOE is a powerful data collection and analysis tool that can be used in a variety
of experimental situations.
 Types of experimental design: Screening: Full factorial design, fractional
factorial design, placket-burmann, mixtures design, optimal designs
Optimization: Central composite design, Box Benken, Doehlert etc.
 Process/method understanding acquisition!
 Screening, Optimization and Selection of DoE tools
Design Number of variables and Advantage Disadvantage
usage
Full factorial Optimization/2–5 variables Identifying the main and Experimental runs increase
design interaction effect without any with increase in number of
 Design of Experiments (DoE) confounding variables

Fractional factorial  Screening,

Optimization/and OptimizationRequiring
screening and Selection ofofDoE tools
lower number Resolving confounding effects
design or Taguchi variables experimental runs of interactions is a difficult job
methods
Plackett-Burman Screening/or identifying vital few Requiring very few runs for It does not reveal interaction
method factors from large number of large number of variables effect
variables
Pseudo-Monte Carlo Quantitative risk Behaviour and changes to the For nonconvex design spaces,
sampling analysis/optimization model can be investigated this method of sampling can
(pseudorandom with great ease and speed. be more difficult to employ.
sampling) method This is preferred where exact Random numbers that can be
calculation is possible produced from a random
number generating algorithm

Full factorial Optimization/ 2–5 variables Identifying the main and Experimental runs increase
design interaction effect without any with increase in number of
confounding variables
QUALITATIVE RISK ASSESSMENT
 Qualitative Risk Assessment Criteria
 Risk estimation helps to identify what to study as a part of analytical method
development
 Evaluation of qualitative risk is ultimately linked back to potential harm to
the patient
Qualitative Risk Assessment :
Prioritization Matrix
Qualitative Risk Assessment :
Prioritization Matrix
 QUANTITATIVE RISK ASSESSMENT
 To study the critical factors, the team conducted a risk assessment using a
FMECA.
 Output from the risk assessment study was based on risk score which was
used to identify the critical factors required for the study
 Risk priority scores included an estimate for detectability, severity and
probability
Quantitative Risk Assessment: FMECA
Quantitative Risk Assessment: FMECA
Quantitative Risk Assessment: FMECA
Response ( CQA )
Experimental Factors
Constant Factors
Design Selection Matrix
Screening Design : Resolution
Residual Analysis : Sanity check of
Experimental Trials
Initial results For Resolution: ANOVA
Initial Result For Retention Time: ANOVA
 Response Survace Methodology (RSM)

• RSM selected as it better to fit non linear data

• Most surfaces are flatter further away from optimal settings.
• Use linear models when we are far from the optimums.
• Use quadratics to approximate the surfaces near the peaks
Prioritized Terms for Resolution
Prioritized Terms for Retention time
 Significant factors
 Out of the five factors selected for the screening design, the factors that are significant are:

 After discussion and analysis, following factors will be selected for optimization design:
 Column length (Experimental): Axial Low (150), Axial High (250)
 Column Temperature: Axial Low (25), Axial High (45)
 pH: Axial Low (2), Axial High (7)
 %ACN: Axial Low (100), Axial High (100)
 Particle Size (Constant): 5 micron
 A 30 trial Central Composite design has been suggested for optimization
Optimized RSM Design
Final Model for Resolution : ANOVA
 Resolution : Model Interpretation

 As per the ANOVA table above, it can be seen that:

 The squared term pH is significant for resolution

 There is an interaction between Column temp & pH, and Column length & pH

 The R-Sq values are 88.83% which determines that the model is good for prediction of
the response
Final Model for Retention Time : ANOVA
 Retention Time : Model Interpretation

 As seen from the ANOVA tables above, it can be concluded that:

 There is a squared effect of Column Temperature and % Acetonitrile on the response

 Interaction exists between Column Temperature & %ACN, Column Temp & Column
Length, and %ACN & Column Length
 Also, the R-Sq value of the reduced model is 99.95% which is in indicator that the
predictability of the model will be very good
 Optimum Settings For Validation
 Parameters Goal Lower Target Upper
 Resolution Maximum 4 6 6
 Retention Time Minimum 55 55 75
 Optimum settings
 Col Temp = 45
 %ACN = 100
 pH = 6.6
 Column Length = 150
 Predicted Responses
 Resolution = 5.0123
 Retention Time = 52.209
 Design Space
 The desired profile for both the responses are:
 Parameters Goal Lower Target Upper
 Resolution Maximum 4 6 6
 Retention Time Minimum 55 55 75
 Conclusion

 Method development may follow traditional OFAT aproach as well as

Enhanced approach AQbD
 Traditional approach produce fixed parameter while AQbD produce
flexible design space
 AQbD QbD is a state of the art approach to remove variability of analytical
methods
 AQbD focused on robustness while traditional more on reproducibility
 AQbD requires the right ATP and Risk Assessment and usage of right tools
and performing the appropriate quantity of work within proper timelines.
AQbD workflow and available software
Why different separation conditions?