Quality by Design (QbD) has become popular within the pharmaceutical industry and the FDA has cited a risk-based approach to drug development as desirable for the near future. More complete information and transparency relating to risk assessment will be made available for new drug applications, speeding the approval process and preventing late-stage failures.
Several principles lie at the heart of QbD, all of which lead toward the goal of building in quality from the earliest stages of drug discovery, including increased experimental scope, modelling robustness and retaining experimental knowledge. To be successful, QbD must be applied at every stage in the development and manufacturing process, though one can consider processes within drug discovery on an individual basis when applying QbD principles.
One such process is the development of chromatographic methods for impurities and degradant studies. Ensuring both robustness and optimization from an efficiency standpoint is time-consuming and difficult. Generally, method development is carried out using a trial-and-error approach and requires a large amount of manual data interpretation, and QbD principles can be directly applied to this process to achieve better, more robust separations.
Increasing experimental design space
The variety of parameters involved in chromatographic separation means that the combinations and permutations of possible parameters for investigation quickly become unfeasible to consider by manual experimentation. To cover a sufficiently large experimental space, the matrix of possible column types, temperatures, gradients and buffer concentrations would require thousands of experiments, requiring far more time and resources than are typically available to chromatographers in an industrial R&D lab.
To work within limitations, the chromatographer has to choose a certain subset of parameters and begin varying one at a time—single variable optimization. The result is that localized optima are established before moving on to the next parameter. The true optimum will often be missed using this approach as it fails to take into account the influence that individual parameters have on each other. The result of the trial-and-error approach commonly used today is a compromise between the best use of available resources and experimental rigour.
To satisfy QbD, a compromise is not sufficient. While it may not be necessary or even feasible to optimize all parameters in parallel, some multivariate optimization is beneficial to the final product—the chromatographic method. ACD/Labs offers method development software that increases the feasibility of more thorough, systematic studies through intelligent selection of starting points, simulation of experimental results, recycling of previous experiments and project management, including automatic peak tracking and information overview, and instrument control for some of the industry’s most popular chromatographic systems.
Starting point selection
By intelligently choosing a good starting point, combinations of parameters that are clearly inappropriate are eliminated from the scope of experimentation in favour of those that are most likely to produce results. But how is a good starting point selected? Chromatographers use their experience and expertise, while software is capable of different approaches, such as: (1) Comparing the chemical structures of compounds to be separated against a database of successful methods used for compounds with similar structural features, (2) Using predictive algorithms to model chromatograms and predict retention times.
In the first approach, software like ACD/ChromGenius is used to identify existing methods from an applications database that catalogues the separation of compounds similar in structure to the compound of interest. Optimization for the new compound then proceeds, and any obtained experimental data is added back to the database to improve the software's performance in future development.
The second approach uses software algorithms present in ACD/LC Simulator to predict physicochemical properties of the compound of interest, leading to estimates of retention times under particular system conditions. Such an approach is appropriate when there is a lack of experimental data upon which to build a database or when the compound of interest belongs to a novel chemical class not represented in the application database.
Modeling robustness
A key measure of the quality of a chromatographic method is its robustness or the tendency of the method to perform within specification even with slight changes to method conditions. With traditional approaches, robustness is measured after method development is complete, usually as part of the validation process. This approach—in which a non-robust method may have to be abandoned or dramatically redesigned after validation has begun—is clearly in conflict with QbD principles. Instead, optimization tools incorporate the concept of robustness into the optimization procedure.
Visual examination of resolution maps can help pinpoint areas of low/high robustness. Modern tools can facilitate this process and reject areas of insufficient robustness automatically. Additionally, the same tool set can be used to measure robustness across an entire series of variables.
The figure illustrates this approach as an overlay on a standard resolution map—the range of viable values for the parameters is unshaded.
Knowledge retention
An important aspect of QbD is maintaining complete knowledge of the process and having the ability to retrieve and share this information in the future. Manually collecting the data produced in a stability study is untenable when one considers the vast amount of data that is generated. However, building an applications database using ACD/ChromGenius or ACD/AutoChrom MDS makes collecting, organizing and storing this information in an electronic database infinitely more manageable and provides means of securing, reporting and accessing the data for audit by colleagues or regulators.
A particular benefit of knowledge retention for impurity studies is the ability to retain chemical structure information and search for it again later. Impurities that reappear can be quickly identified by searching the accumulated project data, eliminating the task of re-characterizing the impurity again.
Summary
QbD principles are directly applicable to LC method development in several ways, including expansion of the experimental design space, modelling robustness throughout method development and retention of project information. ACD/Labs chromatographic method development tools make this possible by guiding the process of choosing a better starting point, using predictive power to guide optimization and allowing chromatographers to utilize the wealth of chemical knowledge already accumulated by applying it to current situations. Quality is built into the development of the method itself, resulting in better, more robust separations, into the stability study process by producing well-developed, traceable, error-free results, and finally into the drug product by bringing novel products to market faster and more cost-effectively.
(For more information about ACD/ Labs software for chromatographic method development, see website: www.acdlabs.com/qbd)
The author is Director of Asia & Pacific Operations,
Advanced Chemistry Development Inc. (ACD/Labs)