Design of drug-product stability studies for early development

Batch selection for stability studies. As noted earlier, early stage clinical supplies often are simple oral formulations or powder-in-bottle or powder-in-capsule. Stability studies at this early stage need to support the use of the product in the clinical studies and to facilitate further development. Unless a specific formulation is required for early phase studies (e.g., due to poor solubility, poor bioavailability, the need for once-a-day formulations, or the need for a sterile injectable formulation), a simple oral solid-dosage form is typically employed. For the DS in capsule or bottle, stability of the DS will support early stage development, however, the clinical supply may be placed on stability (e.g., retained samples) and tested, if necessary. The study would continue through completion of the clinical study and could also be monitored at later time points to increase product knowledge, if appropriate.

For formulated oral solid-dosage products, preformulation and process development studies can be performed to define the initial formulation. These studies will also provide initial stability information on the DS and its compatibility with excipients, moisture, processing conditions, and so forth. At this stage, a development batch (non-GMP) may be manufactured and placed on stability. If this batch is representative of the clinical supply (same or closely similar composition, process, and packaging) it can support the IND/IMPD filing and the initial use date for the product. If more than one strength is required, a bracketing stability design can be used. Changes to the clinical drug product need to be evaluated and additional batches placed on stability, as necessary. Using the same tools and thought processes that were mentioned in the DS portion of this paper, the first step is to determine what DP attributes have an effect on stability. If the new DP changes one or more of the stability-related quality attributes, the new batch should be placed on stability. Identification of these stability-related quality attributes can be facilitated through paper-based risk assessments, prior knowledge, or through short-term stability challenges. Changes that could impact product stability include major DS changes (e.g., solid form), formulation changes (e.g., drug to excipient ratio, different excipients), significant process changes (e.g., direct compression to wet granulation), and packaging changes (e.g., contact materials, blister versus bottle, change in size/headspace). Changes unlikely to affect stability for simple oral dosage forms include changes in scale, equipment, and site of manufacture.

Number of batches. Typically, there are limited batches in early development. Because these clinical studies are small, there is often a need for only one batch to be produced. For clinical materials expected to be used only through Phases 1 and 2a, the authors believe long-term data collected on the clinical batch or a batch representative of the clinical batch may be sufficient to ensure stability of the clinical product in early phase development. For stable products, additional batches of the same formulation do not need to be placed on stability.

Study duration and time points. Studies should generate data that give assurance that the clinical supplies retain their quality within acceptable limits at least through the use of the supplies in the clinic. Because early stage clinical studies can be relatively short, stability data need only support short-use periods. However, in many cases, extra samples are in reserve to extend studies for longer duration as a matter of practicality; for example, if the supplies are needed for longer duration than initial plans due to delays in the clinical program.

As noted, the ICH Q1A guideline on stability testing includes recommendations for stability studies for market application requirements but these are not meant to support early clinical trials. However, many of the ICH constructs are still useful such as the storage conditions and spacing of the time points. Spacing of the time points should be designed to capture the overall stability trend, noting that having a few earlier time points during screening studies or more formal stability studies to establish a use date by the time of filing can be desirable.

For DP, the early phase formulation usually will not be used in later clinical trials. As such, there is little value to carry the stability study beyond the time in the clinic.

The selected analytical tests for stability studies should cover the quality attributes that may change with time. Typically, these tests would include assay, impurities (i.e., degradation products), drug release (e.g., disintegration or dissolution), and description. If chiral conversion has been shown to be an issue in DS then it should be monitored in DP. It may not be necessary to perform all tests at every time point depending on the purpose of the test.

Data fromHT/HH stress studies may be useful in impact assessment for temperature excursions and to design efficient long-term studies. Another approach is the stress studies coupled with modeling, such as ASAP as discussed above. ASAP in some instances can give a good use period estimate in a shorter time than the traditional long-term stability approach. The use date derived from stress-extrapolation can be later verified by placing the first clinical batch into real-time stability studies under ICH accelerated and long term conditions.

Strategy for use dating and use-date extensions. Extrapolation of use periods is a commonly accepted practice and nearly a necessity in terms of getting clinical product packaged, labeled, released, and shipped to clinical sites with some use time remaining for storage at clinical sites. There has been much debate on how extrapolation of existing stability knowledge should be used to set use periods of clinical supplies. As was discussed at the beginning of this article, many factors will determine the amount of extrapolation that is defendable. Health authorities may be reluctant to accept extrapolations of more than 12 months past the existing long-term stability data, but it is possible to justify more or less extrapolation based on scientific grounds (see section on DS stability data collection). A one-size-fits-all approach is difficult to arrive at in terms of an exact number for allowed extrapolations, but an approach based on science and risk management is recommended.

With sufficient DS and drug product stability knowledge and a stable clinical formulation, the risk to the quality of the material is low when allowing for stability retest/use period extensions 12 months beyond available real-time stability data. This, however, would not be appropriate for an unstable drug product or one for which there is insufficient stability knowledge. This risk-based approach is based on three principles:

1) Stability attributes that limit shelf life are determined from early development studies and appropriate packaging and storage conditions selected.

2) Use-date assignments are monitored and confirmed though traditional (real time and accelerated) stability studies and appropriate actions taken if required.

3) Companies have internal procedures describing processes for setting and extending initial use dates. Internal procedures for updating use-date assignments normally involve coordination of clinical relabeling and in some cases CMC filings.

It is worth stressing the point that stability data first included in regulatory filings are often just the beginning of a program to monitor the stability of the clinical supplies and continued monitoring to confirm storage claims and initial trends can be very useful for products that continue development.

The authors support the following recommendation from A. Kane: Companies often struggle with the strategy for updating retest/use period assignments throughout development and the consequent updates to the CMC submissions for IMPD and IMPD type dossiers. "Arguably the preference of many applicants would be to update assignments as more stability data are generated without the need for submission of a substantial amendment. To avoid the need for subsequent substantial amendments, the IMPD should contain the proposed specification and provide a clear explanation of how extrapolation is/will be applied to assign the retest/use period assignments" (14). It is the authors' belief that with a well-planned monitoring program, extrapolation with limited initial data is justified.