Understanding of DoE and its advantages in Pharmaceutical development as per QbD Approach
Yogita M. Kolekar
Shivaji University, Kolhapur, Maharashtra, India
*Corresponding Author E-mail: kolekaryogita13@gmail.com
ABSTRACT:
Design of experiments as per QbD approach is systematic statistical technique for planning & execution of experiments in terms of screening, advanced screening & optimization. Also, to draw meaningful inferences from the data along with predefined objective. [4,10] Quality of pharmaceutical product is affected by variety of factors to overcome from this problem with the appropriate linkage of critical material attributes and critical process parameters to critical quality attributes is must. DoE examines how the input factors (CMA’s/ CPP’s) affect on an output (Responses/CQA’s) of the pharmaceutical product at a same time. This paper illustrates broad theoretical as well as practical view of screening, advanced screening & optimization designs. As well as to provide an overview and workflow to DoE and tasks to be executed in developmental activities for drug product formulation. In addition to the statistical concepts Analysis of variance (ANOVA), regression analysis, parato chart, residual diagnosis, p-value, curvature, lack-of-fit, coefficient of determination (R2), design space and multiple response prediction.
KEYWORDS: Design of experiments, Quality by design, Formula DoE, Process DoE, Screening & optimization designs.
INTRODUCTION:
Why/How DoE as per QbD? OR What’s need of DoE/QbD in pharmaceutical development?
Sir Ronald Fisher had given great contribution to the applying statistical theory in the planning and execution stage of the pharmaceutical product development rather than end in terms of Design of Experiments as per quality by design. It checks cause & effect relationship between independent factors/variables (CMAs/CPPs) [11] and dependent factors/variables (Responses/CQAs) Also; it enables to build quality into the product.
DOE/QbD enable to reduce cost, to save time, to get reliable quality and provide robust formulation.
To achieve Quality by Design, you need to develop your product and processes to ensure predefined product quality, safety and efficacy. This requires you to know how your product makeup and processes influence quality, safety and efficacy. [7]
According to ICH (International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use) Q8 QbD is defined as “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”. [7]
Manufacturing without QbD
Manufacturing with QbD
Figure 1: Manufacturing with or without QbD
Experimental Approaches: [1,3]
1. Trial and Error 2. OFAT 3. Design of Experiments as per QbD |
|
Predominant use of trial and error, OFAT and extreme trials
|
Software’s Used:
· Minitab
· Design-Expert
· Unscrambler
· Fusion-Pro
· JMP
Key Steps in DoE:
The experimental design shall be part of development for Screening, Advanced Screening/Characterization and Optimization of multiple variables and effect on responses. The DoE application for development helps in identifying appropriate factors, their impact and identifying the design space and control strategy.
The DoE study consist of following key steps [1]
Figure 2: Key Steps for DoE Study
1. Pre-requisites for applying DoE:
In order to formulate clear objective for DoE, scientist must ensure availability of information derived from following enlisted fundamental steps before planning DoE study. [7]
a) Literature Review
b) Reference Listed Drug Characterization (Including Reverse Engineering)
c) Dissolution development & In Vitro - In Vivo Relation (IVIVR)
d) Defining QTPP, CMAs, CPPs, CQAs and Risk assessment
e) Feasibility Trials
2. Defining objectives:
Above enlisted pre-requisites and knowledge gained from previous experience, problem statement (s) is/are to be listed down by respective scientist. Make a plan for design of experiment study with specific focus on identifying levels/ranges for medium and high-risk variables to be included in DoE. i.e. to Screen or Optimize. Identify feasibility of CMAs, CPPs and variables’ ranges and role of specification.
Consider the equipment/Area/Process availability & feasibility at respective plant location with the proper scale-up factor calculation.
i.e. Lab scale Pilot scale Commercial
Scale (As per DoE/QbD philosophy)
Formulate clear, concise, testable and measurable objective (s) considering multivariate requirements. (DoE/QbD)
3. Brainstorming:
Essential session while designing DoE, including respective project scientist (s), statistician and cross functional teams (CFTs). Make a proper plan/map along with objectives for further planning and execution of design of experiments study (DoE).
4. Selection of suitable design:
After framing of clear objective for design of experiments (DoE), Statistician shall take a call for selection of best suitable study design. Choose study design which fits the problem and the resources.
Stages of DoE -
a. Screening
b. Advanced Screening/Characterization
c. Optimization
d. Confirmation
e. Verification
Along with the three Basic principles of Design of Experiments – [1,2,3,5]
a. Randomization - To avoid bias during trial execution
b. Replication - It reduces known but irrelevant sources of variation between groups
c. Blocking - It allows the estimation of pure experimental error
Selection of suitable design along with objective based on three stages of Design of experiments (DoE) as below,
Figure 3: Stages of DoE
Stages of DoE |
Particulars for design selection |
Screening (Plackett - Burman Design) Resolution III class Number of factors ≥ 6 |
· Goal of screening designs is to check only significant effects among maximum number of factors. · Screening design allows for the simultaneous and individual study of CMAs and CPPs on CQAs based on risk assessment. · It allows performing an experiment, varying the levels of the factors (CMAs and CPPs) simultaneously rather than one at a time. (In terms of time and cost) · Screening design can be used for Quantitative and Qualitative factors. · Feasibility trials can be executed with the help of screening designs. · Screen out main effects only. |
Advanced Screening/Characterization (Full factorial design, Fractional factorial design, General factorial design and Taguchi design) Full factorial design - Min. two factors Fractional factorial design - Min. four factors General factorial design - Min. three factors Taguchi design - Min. three factors |
· Goal of advanced Screening designs is to identify all possible significant effects (i.e. Main effects and interaction effects). · Advanced screening designs allow screening; Individual effects as well as interaction effects within predefined levels for study of significant factors identified from screening stage. · CMAs and CPPs need to design separately in screening stage for better understanding of CQAs. · Advanced screening design can be used for Quantitative and Qualitative factors. |
Optimization (Central composite design, Box-Behnken design and Mixture design) Central composite design - Min. two factors Box-Behnken design - Min. three factors Mixture design - Min. three factors |
· Goal of optimization designs is to identify the factor combination at which the desired response profile fulfilled i.e., Either maximize or minimize response. · Optimization design required when curvature effect is observed I.e. significant from screening/advanced screening designs. · Optimization designs allow to perform trials with predefined ranges as well as it includes all possible ranges to capture data trend. · The optimum combination of factors can yield desired CQAs near the optimum. · It allows how a specific CQA is affected by changes in the level of the factors over the specified level of interest. |
Select appropriate design and sufficient number of runs such that, it would be easy to test for curvature, Lack-of-Fit and estimation of pure error. Also consider minimum two center points to understand reproducibility and estimate pure error in design matrix. As well as consider basic principles of design of experiments such as randomization, replication and blocking for eliminating bias, estimation of variability and estimation of pure error respectively. [1]
Why addition of center points:
Figure 4
The addition of center points to factorial designs provides a test to determine if the use of a response surface design is needed or not. (i.e. to identify curvature effect)
Replicating the center point gives an estimate of pure error.
5. Execution of design matrix:
Planning and execution of experiments as per design matrix received from statistician. Identify and manage unplanned and unexpected sources of variability, particularly those involving test operations and measurements e.g., Noise factor (Temperature, humidity, machine variability, man to man variability etc.) should be mitigated at the time of execution of trials and generating data from trials.
In case of qualitative CQAs or intermittent CQAs under monitoring, response shall be recorded while execution of experiments and to be shared with statistician for analysis.
6. Statistical analysis and interpretation:
First verify fundamental assumptions such as residual diagnosis (Normal probability plot, Histogram, Versus fit, Versus order plot). In residual diagnosis, residual values must be lies between (-2, +2).
Identifying significant effect of CMAs/CPPs on CQAs, perform analysis of variance (ANOVA) and draw inference with the help of p-value (Significant or Non-significant) Also, pictorially observe parato chart/ half normal plot for checking significant or non-significant effects.
Subsequently check main effect and interaction effect plots for visual effect of CMAs/CPPs on CQAs.
Identify curvature effect and lack-of-fit further inference. Validation of model, using coefficient of determination (R2) for fitting of best fitted model or not.
Generation of overlaid contour plot for identifying design space (as per ICH Q8 guideline). Include most significant CMAs/ CPPs in contour plot generation which would be identified from statistical analysis. [5]
Perform robust testing whether drawing inference from analysis is correct or not. For this use point prediction technique for validating further future CQAs/responses outcome. Simultaneously observe whether predicted responses falling in design space or not. [6,13]
Design Space:
It is the multidimensional combination and interaction of input variables that have been demonstrated to provide assurance of quality. It is proposed by the applicant and subject to regulatory assessment and approval. [1,3,5,10,11]
It developed at lab or pilot scale can be proposed for commercial scale, but needs to be verified at production scale for scale dependent parameters. It may be scale and equipment dependent. Therefore, the design space determined at laboratory scale may need to be justified for use at commercial scale.
As per ICH Q8 guideline, working within design space (multidimensional region) not generally considered as a change. Movement out of design space is a change → regulatory post approval change process. [4,5,8,11]
Figure 5: Design Space
REFERENCES:
1. D. C. Montgomery, Design and Analysis of Experiments, 5th edition, John Wiley and Sons, Inc. New York, 2001., 8 ed., John Wiley & Sons, Inc, 2013.
2. D. C. Montgomery, Statistical Quality Control, John Wiley, 2009.
3. D. C. Montgomery and G. C. Runger, Applied Statistics and Probability for Engineers, Willey, 2005.
4. Role of Statistics in Pharmaceutical Development Using Quality-by-Design Approach – an FDA Perspective by Chi-wan Chen, PhD and Christine Moore, Ph.D. office of New Drug Quality Assessment CDER/FDA.
5. ICH Q8 Pharmaceutical development guidance.
6. Statistical Optimization of Pharmaceutical Formulations by P.K. Shiromani President Shirman Pharmaceutical Consulting.
7. Hardik P, Shardda P and Bhavna P. A Comprehensive review on quality by design (QbD) in pharmaceuticals. International Journal of Pharmaceutical Sciences Review and Research. 2013; 21:223-23
8. Isa Martins Fukuda1, Camila Francini Fidelis Pinto1, Camila dos Santos Moreira1, Alessandro Morais Saviano1, Felipe Rebello Lourenço1*, Design of Experiments (DoE) applied to Pharmaceutical and Analytical Quality by Design (QbD) in Brazilian Journal of Pharmaceutical Sciences 2018;54(Special): e01006
9. Sangshetti JN, Deshpande M, Zaherr Z, Shinde DB, Arote R. Quality by design approach: regulatory need. Arab J Chem. 2017;10(Suppl 2): S3412-S3425.
10. Zhang L, Mao S. Application of quality by design in the current drug development. Asian J Pharm Sci. 2017;12(1):1-8.
11. K. P. R. Chowdary*, K. Ravi Shankar and P. Suneel Kumar, Recent Research on QbD Approach in Formulation Development Volume 4-Issue1, Jan 2014. pp.282-292
12. Khade MM and Sunil JP. Quality by design (QBD)-A Quality improvement perspective for pharmaceutical development”, International Journal of Pharmaceutical Research and Bio- Science. 2013; 2:144-166.
13. Sumit K and Shikha T. A Quantitative Approach for Pharmaceutical Quality by design patterns. Invention Rapid: Pharm Analysis & Quality Assurance. 2012; 4:1-8.
Received on 02.08.2019 Accepted on 16.10.2019
© Asian Pharma Press All Right Reserved
Asian J. Pharm. Tech. 2019; 9 (4):271-275.
DOI: 10.5958/2231-5713.2019.00045.X