Advertisement
Articles
Advertisement

New CDS Functions to Meet New Regulatory Requirements?

Thu, 06/05/2014 - 4:17pm
R.D. McDowall

Do chromatography data systems (CDS) have adequate functions to help analytical scientists meet requirements when the pharmacopoeia is updated?Ensuring efficient method development and routine operation into the future

A recent United States Pharmacopoeia (USP) stimulus to the revision process paper1 has taken a life cycle approach to the development, validation and use of analytical procedures. Do chromatography data systems (CDS) have adequate functions to help analytical scientists meet these requirements when the pharmacopoeia is updated?

The USP paper, using principles from ICH Q8(R2),2 outlines an integrated life cycle for the development, validation and operational use of analytical procedures. Traditionally, regulatory guidance and regulations have focused only on the validation of an analytical method and the development phase was only mentioned in passing.3-5 In contrast, the USP article1 uses life cycle management which includes six main requirements:

  • establish the Analytical Target Profile (ATP) for the procedure
  • determination of an analytical procedure’s design space during method development
  • define the analytical control strategy
  • procedure performance qualification (PPQ) or validation of the procedure to ensure fitness for purpose
  • procedure performance verification (PPV) in the testing laboratory and on-going trending to demonstrate acceptable performance of a method during use
  • determining if a procedure modification is a change or an adjustment

This paper will look at these areas and consider what functions should be available in a CDS to help analytical scientists in the future. I will focus on the chromatographic elements of a procedure and only mention the preparation in passing. This is a personal view that is intended to stimulate discussion and, hopefully, develop these functions in future CDS. Within the space allowed, I can only outline the broad principles of what I think should be available in future CDS.

Figure 1: Life cycle of an analytical procedure from method development to applicationExperimental Design in Method Development
The start of method development (stage 1 in Figure 1) begins with defining an analytical target profile (ATP) which is the specification of the procedure1 that is independent of the analytical technique. From the ATP requirements, the method is designed. Assuming that we are using HPLC, a procedure could consist of the following elements:

  • sampling plan for the procedure and transport to the laboratory
  • number of aliquots of the sample per test and number of replicates per aliquot for the procedure
  • sample preparation requirements
  • sample presentation to the chromatograph
  • HPLC instrument and conditions e.g. vial storage conditions prior to injection, sample injection volume(s), isocratic or gradient separation, column make and type, mobile phase composition, column temperature, detector type with appropriate settings
  • data processing and calculation requirements to generate the reportable value

The key to method development is the understanding of how key variables in the procedure impact the quality of the separation and robustness of the method. From the perspective of a CDS, the system needs to automate the design, conduct and evaluation of experiments. Some CDS have been integrated with experimental design software and are able to control chromatographs so that results of individual experiments can be fed back into the design software for evaluation. The aim of these robustness experiments is to determine a design space within which the procedure performs as required against the ATP,1 as illustrated in Figure 2. Note that Figure 2 is only an illustration of a design space, there may be more variables involved than are possible to draw in three dimensions.

Figure 2: Design space for an analytical procedureTo ensure that the design space is understood, a risk-based analytical control strategy is devised. One of the risk assessment methodologies discussed in the USP paper1 is CNX (Control, Noise, Experimental). Each variable in the analytical procedure is ranked as a factor to be controlled (C), those that are noise (N) and the ones that need to be examined experimentally (X). This is shown in Table 1 only for some of the HPLC method variables. Further experiments are conducted to understand how these variables e.g. temperature, percentage of organic modifier in the mobile phase, impact the procedure. The CNX assessment can be updated after this work and enables the developer to define the analytical control strategy that specifies the controls required to ensure the procedure operates within the design space. To help provide evidence that the procedure is under control during routine use, the system suitability test (SST) parameters need to be defined and verified at this stage. The SST parameters now become key to demonstrating that the operational method is controlled.

Returning to the theme of this article, it is important that a CDS has the functions to perform with the risk assessment and that it remains within the analytical control strategy that defines a procedure’s design space. Whenever the procedure is run later, the CDS can verify that the chromatographic conditions used are within the analytical control strategy parameters and, hence, the procedure is operated in a state of control.

Procedure Performance Qualification (PPQ)
Procedure performance qualification (stage 2 in Figure 1) is the new term for method validation.1 At this point in the life cycle, the CDS needs functions for the users to define the experiments for determining the various PPQ parameters together with predefined acceptance criteria for an individual procedure. On completion of the work, the calculated results can be interpreted by the CDS against the acceptance criteria and secure result tables created for inclusion in the method validation report that will be prepared outside of the data system.

Table 1: Outline Control Strategy for Some of the Variables of an HPLC SeparationProcedure Performance Verification (PPV)
When a procedure is established in a new laboratory, then transfer or procedure performance verification (stage 3 in Figure 1) experiments can be designed and executed. When these are complete, the results can be compared with the original studies to confirm whether the laboratory has set up the method correctly. The CDS also can be used to ensure that the analytical control strategy parameters are enforced to ensure that the method is within the design space of the procedure. Therefore, the functions required in the future for any CDS for PPQ and PPV should be:

  • PPQ experiments, such as intermediate precision and accuracy/uncertainty measurement, limits of quantification/determination, linearity over the measurement range, stability of analyte(s) in vials etcetera.
  • PPV experiments to demonstrate that a testing laboratory can establish a procedure and the results generated from the procedure are equivalent to the originating laboratory. If laboratories are using different CDS systems, there needs to be a mechanism for transferring the analytical control strategy from one CDS to another.

Operational Use of the Method and Data Trending
The requirements of ICH Q106 and EU GMP Chapter 6.97 for trending of QC data have been applied by the USP stimulus paper to control of methods during routine use. Therefore, a CDS needs to have the ability to trend data, such as the individual and mean results, along with the key SST parameters defined by users. These data can be presented as, say, Shewhart plots with action and warning limits or Cusum plots with the aim of proactively identifying trends before a procedure produces out-of-specification results. The CDS should then allow a user to drill down from the plot to look at, say, the instrument(s) and column(s) used to see if there are any issues around a specific instrument or column.

Managing and Controlling Method Changes
When the CDS contains the design space of a method, any attempted changes can be checked to see if they are within the analytical control strategy. Under the USP stimulus paper,1 this would be an adjustment with no revalidation impact. This would demonstrate that an analytical method was under control and no unauthorized changes were allowed. If a change was proposed that was outside of the analytical control strategy, then the CDS could force the user to the validation module where incremental (the ATP is unchanged) or method development and full validation (the ATP is changed) would take place to show that the updated method was validated.

Summary
I have suggested changes for future CDS functions to cope with the new approaches to method development, procedure performance qualification and verification to allow efficient routine operation. Currently, no CDS has all of these functions. So, it is either for users to demand more from their CDS suppliers or for suppliers to be proactive and develop these features for their current and future users.

References
1. G.P.Martin et al , Lifecycle Management of Analytical Procedures: Method Development, Procedure Performance Qualification, and Procedure Performance Verification, Pharmacopoeial Forum 39 (5) September – October 2013 (Note there are no page numbers as this is only available on line at www.usp.org)
2. International Conference on Harmonisation (ICH) Q8(R2) Pharmaceutical Development, 2009
3. FDA draft Guidance for Industry, Analytical Procedures and Methods Validation for Drugs and Biologics, February 2014
4. International Conference on Harmonisation (ICH) Q2(R1) Validation of Analytical Procedures: Text and Methodology, 2005
5. United States Pharmacopoeia, <1225> Validation of Analytical Procedures
6. International Conference on Harmonisation (ICH) Q8(R2) Pharmaceutical Development, 2009
7. EU GMP Chapter 6 Quality Control 2006 (note this chapter is currently under review and the public comment period has now closed)

R.D. McDowall is Principal, McDowall Consulting. He may be contacted at editor@ScientificComputing.com.

Topics

Advertisement

Share this Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading