Key points from the guides
Figure 1: Product realization using a QbD approach. DoE is design of experiments. CPPs is critical process parameters. (ALL
FIGURES ARE COURTESY OF THE AUTHORS)
|
The workflow given in the ISPE Guide series is based on a schematic developed by an EFPIA team (see Figure 1) (14). Although
Figure 1 suggests the workflow is linear, there is much discussion and exemplification to indicate that the work is more likely
to be iterative, particularly the process development step of sequential and iterative application of quality risk management
and design of experiments (DoE) studies. There are many detailed examples of quality risk assessment exercises linked to DoE
studies in Part 2, with the output of a DoE study cycling back into another risk-assessment exercise to demonstrate reduction
and eventual acceptance of risk.
Figure 2: Process parameter continuum of criticality. CQA is critical quality attribute. PP is process parameter. RPN is risk
priority number.
|
How to assign criticality is discussed, the concept of a continuum of criticality based on risk assessment is suggested with
examples. For CQAs, criticality is shown based solely on severity of harm (i.e., safety and efficacy) as linked to the patient
through the quality target product profile as indicated in ICH Points to Consider document (10). For CPPs (and CQAs), the
flowchart shown in Figure 2 indicates how quality risk-assessment steps can lead to a continuum of "criticality" based on
severity, probability, and detectability. Many examples are given, such as types of risk assessment criteria with detailed
discussion and explanation.
Figure 3: Representation of design space as inputs and outputs.
|
Design space is represented pictorially in Figure 3. Multifactorial relationships in the middle box in Figure 3 can be based
on mechanistic understanding or first principles, or empirical, or hybrid models as represented in Figure 4.
Figure 4: Commonly used model types.
|
There are discussion and examples of derivation of different types of design space, some not requiring DoE experimentation
when they are based on first principles, in the ISPE series. With this relatively new concept of design space, the ISPE guides
describe how to use design space in operation of flexible processes, how to consider the impact of scale and link of design
space to control strategy, and how to update and continually improve a design space.
Because development and operation of a control strategy is mandatory, significant attention is given in both parts of the
series to development, choice, implementation in manufacturing, link to process validation, and continual improvement of control
strategy. In Part 2, particularly, linkage of drug-substance control-strategy elements (e.g., particle size to drug product
CQAs impacting patients), dissolution and uniformity-of-dosage-units tests are shown, including how to develop an algorithm
in which drug substance particle size determined by in-line measurement in a high shear wet milling unit operation is a major,
but not only contributor to predict dissolution values. Knowledge of drug-substance particle size can be used in a feed-forward
manner to determine some relevant process parameters (e.g., lubrication time to assure the dssolution CQA acceptance criteria
are achieved). Prediction of dissolution values is one element of a complex control strategy, the elements of which are derived
from the enhanced, QbD approach discussed. Required CGMP elements also contribute to the control strategy and examples are
given.
Figure 5: Summary of some PAT elements of a real-time release testing scheme.
|
A summary of some elements of a control strategy leading to real time release is given in Figure 5 in comparison with a traditional
approach. Figure 5 shows that RTRT usually involves more analysis (in-line, online, or at-line) compared with a traditional
end-product testing approach coupled with, for example, in the application of fixed process parameter ranges.
|
