Biomedical Engineering Reference
In-Depth Information
The next three practices, data collection and formal experimental design, the use of
multivariate tools, and in-line, online, or at-line process analyzers are intended to
optimize the quality and quantity of information about the process. They enable the
development of sophisticated models based on detailed understanding of multiple
interactions and enable the final practice, process control. When all critical quality
attributes have been accomplished the process has reached its end point.
Because of their complexity, biotechnology processes typically have a greater
degree of characterization and control than small molecule processes, and in many cases
there are already a significant amount of relatively simple PAT controls around both
upstream and downstream processes. Incoming raw materials have been studied to see
how any variability may impact process and product quality and to develop suitable
screening control initiatives and feed strategies [31]. Upstream raw materials used for
cell culture or fermentation processes are a major source of variability, either because
they are inherently complex (as in the case of hydrolysates) or, in the case of chemically
defined media, because they contain a wide array of energy sources, growth factors, and
trace nutrients. Multivariate tools are proving useful in determining the interactions
between materials and process conditions [32], but because of the complexity of
microbial or mammalian cell culture, it may be harder to claim higher level understand-
ing of every detail of the process. PAT tools may be as useful in establishingwhat does not
need to be measured and controlled in manufacturing as for what does.
Recovery and purification operations, on the contrary, are relatively simpler, and
there are well-developed theoretical models for many of the typical unit operations for
centrifugation, filtration, and chromatography. The problem in development is to
establish the clinical significance of process- and product-derived contaminants, which
may be monitored and controlled but whose significance will not be fully understood.
13.5.4 Manufacturing
If QbD, design space, and product knowledge are the instruments and objectives of
process development, PAT analyzers, process understanding and controls, and continu-
ous improvement are the tools and goals of the manufacturing engineer.
PAT controls are already in place for many biotech processes. Although QbD seeks
to define a multidimensional design space, controls may be univariate or multivariate,
depending on what is important for the process (e.g., monitoring to an end point, or
more interactive control of a continuous process based on multiple parameters).
PAT seeks to assure quality either directly, by monitoring and controlling product
quality attributes during the process, or indirectly, by reducing variability and thereby
providing a higher assurance of quality. PAT can also be used to provide continued
assurance of fitness for purpose of process equipment and facilities, see Fig. 13.4. PAT
can be applied at multiple levels, and just as QbD can be applied without use of PAT, so
PAT can be applied to situations that are not specifically for the design space established
by QbD.
The focus of QbD is largely on the design and development of the product to ensure
efficacy and patient safety, and considers the product in terms of CQAs. Some parts of
product design are fixed and cannot be impacted by PAT. Instead, PAT focuses on the
Search WWH ::




Custom Search