Method Robustness Considerations for Successful Product Commercialization

• Bristol-Myers Squibb, Co.

Analytical method robustness has gained significant attention in the pharmaceutical industry in recent years for many reasons, including regulatory environment change, globalization, and outsourcing trends.

Impact of Method Robustness to Pharmaceutical Development and Continued Process Verification

Change in the regulatory environment has been one of the important drivers to improve analytical method robustness. There has been significant shift in recent years from controlling the quality of the product by end process testing to building quality into the entire process. Per ICH Q10¹, the emphasis in pharmaceutical development has shifted to the control of critical quality attributes (CQAs) of a product via quality target product profile identification, risk analysis, and well-formulated control strategy per Quality by Design (QbD) principles. This has generated greater demands on analytical testing during product development. Fit-for-Purpose analytical methods must be able to produce consistent and reliable results to effectively support product development. Furthermore, when a product is commercialized, the process validation does not end with the successful manufacture of several batches of product on scale at the commercial site. Per FDA Guidance for Industry “Process Validation: General Principles and Practices”², a continued process verification is required. The drug’s sponsor is required to establish an ongoing program to collect and analyze product and process data to monitor process performance and make adjustments as necessary if there is a shift in that performance. This has led to a shift in expectations for method performance: from testing to verify a product meets pre-established specifications (e.g., 90-110% for assay) to being able to discern small differences in process performance, since the success of the continued process verification program hinges upon having robust analytical methods which are capable of generating meaningful data to truly reflect the variation of the process and the quality of the product.

Impact of Method Robustness to Global Technology Transfer

Pharmaceutical companies market products in many countries. The distribution of the product often requires manufacturing and packaging in multiple countries and locations. Some South American and Asian countries require in-country testing for marketed products. Additionally, to reduce overhead and equipment cost, many pharmaceutical companies outsource parts or all of their development and manufacturing to third party contract facilities. All of these manufacturing scenarios require transferring methods to multiple analytical laboratories, often located in different countries. To ensure successful technology transfer, the analytical methods must be robust enough to be adopted in many laboratories with variations of environment, equipment, reagent sources, analyst skills, operating practices, culture and training.

Factors to be Considered When Developing a Robust Method

Per ICH and USP definition^3,4, the robustness/ruggedness of an analytical procedure is a measure of its capacity to remain unaffected by small but deliberate variations in parameters listed in the procedure documentation and to provide an indication of its suitability during normal usage.

There are many factors, including basic parameters (e.g. solution stability), internal parameters (e.g. method parameter settings) and external parameters (e.g. environmental factors, instrument variation, source and quality of reagents, and technical skill of analysts), that must be considered when developing a robust analytical method. These considerations also vary depending on the product development stage^5,6,7. It is not unusual for analytical methods to undergo multiple iterations during the development life cycle of the pharmaceutical product. When a product is ready to be moved to commercialization, relevant analytical methods need to be finalized and validated for QC use. At this point, all parameters that potentially affect the robustness of the method should be carefully re-examined by performing appropriate risk analysis and conducting design of experiment (DoE) studies around the critical quality attributes (CQA) per QbD principles^8,9. It is especially important to focus on the hard to define and control parameters like the external parameters as discussed below. Through these studies, the method operating space will be better defined which will further improve method robustness.

Environment Concerns

Some analytical procedures are more susceptible to change by the laboratory environment than others. For example, a Karl Fischer method to determine water content may be sensitive to variation of the relative humidity of the laboratory. Results generated in the summer, in a laboratory located near the equator or near a rain forest may be significantly different from results generated in winter or dry/cold climate zones if the laboratory environment is not adequately controlled. Thus when setting method parameters, environmental factors need to be considered and experiments should be conducted to evaluate the robustness of the method relative to environmental factors. If it is shown that the method accuracy may be affected by the environmental conditions, further actions should be taken, such as modification of the method parameters, addition of a specific system suitability test, or requiring a controlled environment.

Instrument Concerns

Most types of analytical instruments are manufactured by multiple vendors. Although based on similar operating principles, differences in design may cause noticeable differences in performance. Even the same brand of instrument could have different models, minor configuration differences and/or age differences, leading to performance variation. For instance, the difference in HPLC system dwell volume and extra column dead volume could affect the peak shape, resolution and sensitivity. Before the method parameters are finalized, it is highly recommended to test the method on different brands/models of the instrument, especially those used in the target QC laboratory where the method will be transferred and routinely used to release commercial product.

Reagent Concerns

Most analytical methods involve use of various reagents, whether they are chemical or biological. When selecting reagents, the availability and potential variation in quality should be carefully considered. Multiple vendors exist for most reagents, thus the quality of the reagent could vary significantly and the susceptibility of the method to this variation should be considered. If necessary, experiments should be conducted to compare method performance using reagents of best quality to those of worse quality. If a significant difference is observed, the manufacturer and reagent grade need to be specified in the method, which may cause hardship at some QC laboratories or significantly increased testing cost. If possible, an alternative reagent should be considered.

Analyst Technical Skill Concerns

The term “QC friendly” is often used when a method involves only very commonly used techniques, is very easy to execute and is less prone to variation due to analyst skills and laboratory practices. When finalizing methods for product release testing, if multiple techniques are possible to determine certain CQA, the one that is less complex and analyst technique dependent should be given priority. Within a method, each of the operating steps should be carefully considered, too. Procedures that rely on analyst interpretation, observation, and individual judgment should be avoided whenever possible, for example “shake until dissolved” is far less descriptive of the desired outcome to “shake until the tablets are completely disintegrated and a fine powder is observed with no tablet fragments”. To ensure method robustness, the method should be challenged by having different analysts (preferably from different laboratories) test the same batch, following the procedure without giving any additional instructions.

Method Robustness Evaluation Using Potency Determination by HPLC as an Example

Since assay and impurities are two of the most important CQAs of a pharmaceutical product, an example method robustness evaluation is provided below for a solid dosage potency/impurities method by HPLC.

For a solid dosage potency/impurities method by HPLC, there are two aspects which should be considered: chromatographic conditions and sample preparation procedures. For a gradient HPLC method the chromatographic conditions to consider include: the type of gradient (e.g. single vs. multiple-stage gradient), slope of the gradient, isocratic hold, organic/pH modifier type, sample concentration and detection wavelength. In regard to sample preparation procedures, consider diluent selection, sample concentration, standard weight, dilution scheme and sample disintegration/extraction. Modeling studies for both aspects are highly desired.

HPLC Conditions

• Gradient type: During early development, API synthetic route and process conditions continue to change and the drug product formulation is not finalized, thus many impurities may need to be separated requiring a complicated gradient. When the project approaches commercialization, both the manufacturing processes of the API and product are locked or finalized and the physical/chemical stability of both are well understood, thus there are often fewer impurities/degradants monitored providing an opportunity to simplify the gradient. While this requires additional effort, a simple gradient provides fewer issues/challenges for the QC laboratory in the future.
• Mobile phase (MP): During development it is common to achieve the desired MP ratio by mixing two mobiles phases [one almost entirely aqueous (MP A) and the other almost all organic (MP B)] online, leveraging the instrument’s mixing capability. When finalizing a registrational method, premixing aqueous and organic to achieve the desired start and end ratios of mobile phase may be considered. Organic and pH modifier balance in both mobile phases should also be evaluated. Well studied and selected mobile phases can reduce baseline noise and baseline disturbances, and improve gradient repeatability.
• Isocratic hold: Incorporating an initial isocratic hold to a gradient program could be critical to the success of the method transfer from one instrument type to another. The various HPLC instrument models have different dwell volumes, and in some cases this variation could be significant. If a gradient is developed without an isocratic hold using a system with a very small dwell volume, the separation may not be able to be reproduced on a system with a significantly large dwell volume.
• Selection of detection wavelength: Wavelength selection should also be reconsidered at the registrational stage. It may not always be the best decision to select absorption maxima of the API as the HPLC detection wavelength. For example, if the compound has very strong absorption at the maximum wavelength, it may then require a relatively low working concentration to stay within the linear range of the UV detector, which in turn requires a larger sample volume or additional dilution steps for a given dosage strength. The example in Table 1 illustrates the balance between wavelength selection and sample volume and dilution steps. Simply changing the detection wavelength allows an increase in solution concentration (based on solubility), the sample solution volume can by reduced and secondary dilution can be eliminated. It is important to note that selecting a detection wavelength on a slope of the UV spectrum should be avoided whenever possible for obvious robustness concerns.

Table 1. Effect of Wavelength Selection on Required Sample Volume and Dilution

Sample Preparation Procedure

• Sample diluent: The final sample diluent composition should accommodate many potential variations considering the method will be used for years to come in multiple laboratories in different geographic areas with a variety of environment and reagent sources. A minor variation in product manufacturing process within the design space may also cause minor changes in product properties that may challenge the completeness of the active extraction/dissolution. The final composition of the diluent should be selected through a well designed DoE study, preferably using tablets manufactured at the edges of the process that might impact extraction (e.g., larger API particle size and high compression force). An example DoE study examining diluent variables shown as a surface plot is given in Figure 1. As can be seen in this example, both the acetonitrile (ACN) and trifloroacidic acid (TFA) concentrations affect the extraction efficiency of the drug product in the ranges studied. The final composition should be selected within the relative flat range of the surface plot to minimize the effect of slight variation of diluent composition to the extraction efficiency (percent label claim, %LC).

Figure 1. Surface plot of a tablet extraction DoE study. x-axis: %ACN from 15%-25%; z-axis: %TFA from 0.025% to 0.1%; y-axis: %Label Claim as a measure of extraction efficiency, from 97.5% to 101.0%.

• Standard weight: Based on the quality of the balance, variability of analyst’s skills, and the physical properties of the reference standard, such as hygroscopicity and static tendency, the defined amount of standard to be weighed should be carefully selected to further ensure method robustness. Low standard weights (e.g. <20 mg) or small solution volumes (e.g. 10 mL or less) would introduce unnecessary analytical error.
• Sample dilution scheme: The final method should strike a balance between the stock sample solution volume and the number of dilution steps. Too high of a solution volume (e.g. 1000 mL or larger) is difficult to handle and thus prone to insufficient mixing. In this case, further dilution steps should be considered. However too many dilution steps could also introduce errors. In addition, a low transfer volume (e.g. 1-2 mL) for further dilution would also increase the chance of analytical errors.
• Sample disintegration/extraction procedure: The most commonly used sample disintegration/extraction procedures for solid dosage forms are sonication, shaking, stirring or a combination of both. Sonicators, shakers and stirring plates are often not very well controlled and their effectiveness is difficult to quantify. Therefore, when developing a drug product sample disintegration/extraction procedure, these likely variations should be evaluated. A DoE study using different type/brand of instrument with different sonication/shaking/stirring time is highly recommended. Other technologies, such as using high shear forces to homogenize sample in diluent, could be good alternatives to these traditional techniques to improve method robustness.

Conclusion

Analytical method development is an essential part of pharmaceutical development and commercialization. Robustness of the analytical methods is very important throughout the development life cycle, but especially critical in late stage prior to commercialization. A robust analytical method will provide assurance for successful analytical technology transfers and implementation in quality control (QC) laboratories around the world and provide meaningful data to support continuous process verification upon product commercialization. All factors that potentially impact method robustness should be carefully considered and systematically studied using appropriate tools, such as risk analysis and DoE.

References

1. ICH Harmonised Tripartite Guideline: Pharmaceutical Quality System Q10.
2. FDA Guidance for Industry, Process Validation: General Principle and Practices.
3. ICH Harmonised Tripartite Guideline: Validation of Analytical Procedures: Text and Methodology, Q2(R1).
4. USP 37, Chapter 1225, Validation of Compendial Procedures.
5. Krull I., and Swartz M., “Method Validation and Robustness”, LCGC North America, May, 2006
6. Delaegher B. and Heyden Y., “Robustness Tests”, LCGC North America, Jul, 2006.
7. J. Ermer and P. W. Nethercote, “Method Validation in Pharmaceutical Analysis, A Guide to Best Practice”, (2nd Edition) by Wiley, 2014
8. Rozet E., et al, “Design Spaces for Analytical Methods”, Trends in Analytical Chemistry 42 (2013), 157-167.
9. Vogt F.G., and Kord A.S., “Development of quality-bydesign analytical methods”, J Pharm. Sci. 100 (2011), 797-812.

Xin Bu is a Principal Scientist working at Bristol- Myers Squibb Company for more than 17 years. She is responsible for CMC project management, analytical project management, technology transfer and plays an important role in regulatory filings. She received her PhD in analytical chemistry from the University of Washington.

John A. Castoro (Ph.D. in analytical chemistry at the University of California, Riverside) is an Associate Director at Bristol-Myers Squibb Company. His areas of focus include analytical method development as well as structure elucidation. He presently leads a group that provides analytical support for pharmaceutical development including CMC dossier submissions.