4 minute read
assessment tools
Critical quality attributes
processes. Although the complex physiology of mammalian cells (e.g., CHO) has been investigated in many studies, little is known about how this information can be coupled to product quality. Therefore, physiological parameters such as specific rates have to be defined to support the extraction of scalable and science-based information on physiology–product quality interactions in bioprocess development. The identification of such physiological parameters in early-stage development has to be facilitated by structured risk assessment approaches. Ishikawa diagrams and risk questions support the proper understanding of linkages targeted by risk assessments. With the help of these tools, interdisciplinary team members can come to an agreement on the scope of risk assessments. Due to the high complexity of targeted bioprocesses, the approach depicts a promising tool for more efficient process development, especially in its early stages.
Advertisement
Risk assessment tools convert subject knowledge into quantitative information in order to assess criticality. The outcome is the risk number (RN), which is calculated by multiplying two or more factors. Although the ICH Q9 guideline lists a variety of risk assessment tools, it does not provide a clear definition as how to select the most appropriate one for the specific purpose. Some studies have reported the use of these tools during biopharmaceutical product development, but they did not provide extensive information on the reasons for selection. As already discussed, the factor severity is always included in the risk assessment tool to express the potential harm to pharmaceutical quality as the basis for the determination of criticality. Additional factors are used to improve the risk assessment tool by breaking up the risk into multiple components. For example, uncertainty is often used as a second factor besides severity to include the quality of input data as a possible source of risk. This is especially relevant in early-stage process development, where scientific knowledge is often lacking to fully understand the linkage between product- and process-related parameters. The two factors severity and uncertainty are included in both CQA and CPP risk assessment tools within this study. If additional information is available that can increase the selectivity of the risk assessment, the tool has to be appended to process all the information at hand. An example is the original product’s quality profile for biosimilars. Biosimilar guidelines in the EU and the United States put emphasis on analytical comparability with the original product. Consequently, the quality profile of biosimilars is highly determined by the originator product. In order to involve this additional information in biosimilar development, a third factor called deviation was added to the here-described risk assessment tool for CQA selection (Table 6.1). This factor incorporates the extent of acceptable deviations from the 251
Biosimilarity: The FDA Perspective
Table 6.1 Overview of the Risk Assessment Tools for CQA and CPP Selection
CQA Risk Assessment CPP Risk Assessment
Linkage
CQAs–QTPP specification Risk question How critical is the effect of a possible deviation from the innovator’s quality profile with respect to safety and efficacy?
Risk assessment tool RN = severity × uncertainty × deviation
Scores for the third factor Deviation (from the quality profile of the reference material): 1. No deviation in the quality profile 2. The low deviation in the quality profile or robust purification method deviation, limited purification efficiency 3. The severe deviation in the quality profile limited purification efficiency 4. The severe deviation in the quality profile, a variant cannot be purified CPPs–CQA ranges How critical is the effect of the process parameter or process variable on CQAs?
RN = severity × uncertainty × complexity
Complexity (of the mechanism responsible for the CPP–CQA effect): 1. The mechanism described by physical law 2. Simple mechanism with well-known characteristics 3. Complex mechanism with previously reported quantitative interactions 4. Complex mechanism without quantified characteristics 5. Very complex mechanism
252 originator product’s quality profile. Quality attributes with minor importance would receive a low deviation score, indicating a higher acceptable deviation. Thus, this factor helps to prioritize the quality attributes for product development based on their effect on biosimilarity. However, as communicated by regulatory bodies, the effect of a deviation from the originator in attributes with low relevance has to be justified as well in biological assays. The factor deviation can also contain information about the purification capacity of downstream process steps if the risk assessment is conducted for the determination of CQAs in upstream process development (see Table 6.1). Another example to incorporate additional information into the risk assessment in this study was, considering the complexity of the mechanisms, how process parameters can affect the investigated quality attributes. Accordingly, besides the factors severity and uncertainty, a third factor called complexity was added to the risk assessment tool of CPP selection. This factor quantifies as to which extent the mechanism of the CPP–CQA interaction can be described by a scientifically developed formula (Table 6.1). The higher the score, the more complex the mechanism and the less information available for its quantification. As the lack of reliable information on CPP–CQA interactions raises the uncertainty score of almost each process parameters and variables in early-stage process development, including complexity as a third factor, helped to differentiate CPP candidates based on scientific considerations. Introducing this factor also emphasizes the scope of CPP risk assessment at this stage of process development, which is not to select critical parameters for a finalized manufacturing process but rather to rank parameters in order to prioritize experiments for process development. These considerations justify the development of novel tools as described earlier instead of
