Biomedical Engineering Reference
In-Depth Information
development many different lots may be used at different times after manufacture.
The use of multivariate statistical analysis for evaluating product attributes and clinical
findings may be informative [28].
There are many potential strategies for understanding the input of complex product
attributes on clinical performance. Attributes may interact and a multivariate attribute
space may be more useful than univariate ranges for such attributes. However, complex
products may have many thousands of possible attribute states. Even using all these
strategies, this large number of attributes and interactions cannot be fully evaluated. Thus,
a risk assessment, utilizing prior knowledge in conjunctionwith other approaches, will be
needed. An important consideration for complex products is that the more the attributes
dismissed based on assumptions, the greater the uncertainty that no critical attribute or
interaction was missed. This greater uncertainty needs to be considered in defining
attributes as unimportant. For complex products, mechanism of action and biological
characterization may contribute to understanding the importance of product attributes.
We have discussed assigning risk and importance to many attributes of complex
products. For regulatory purposes, clear ways to distinguish attributes will be needed.
Draft guidance [11] has defined a critical quality attribute (CQA) as a physical, chemical,
biological, or microbiological property or characteristic that should be within an
appropriate limit, range, or distribution to ensure the desired product quality. This CQA
definition is process independent and does not consider the control system; a CQA is only
defined as a property or characteristic of the product. Other suggestions have included an
intermediate level of attribute importance, analogous to intermediate levels of parameter
importance or key parameters [29]. For complex products with greater uncertainties
regarding attribute importance, intermediate attribute categories, between noncritical
and critical, may be of value. Linkage of potential attribute risk to the appropriate control
strategy may be best achieved with a variety of attribute risk categories. Attribute
importance can be assigned through a risk assessment, as described above. CQAs do not
necessarily need to be controlled by end-product testing (classical specifications), but the
sensitivity of biotechnology product attributes to environmental conditions should be
considered in any upstream approaches to CQA control.
Assessment of product attributes using the above methods may require significant
efforts. However, the knowledge gained from linking product attributes to clinical
performance can be leveraged into other products and may facilitate discovery and
design of new products. This information does not need to be complete at the time of
marketing application submission. As product knowledge grows, CQAs may evolve. Not
all sponsors may invest in an extensive risk assessment of all their products before
approval. Historically, sponsors have used the ranges of attributes in clinical study lots to
assure clinical performance. This approach may still be appropriate in some cases but
could limit flexibility.
2.4 MANUFACTURING PROCESS
In Fig. 2.3, a schematic for a typical biotechnology manufacturing process is described.
The upstream process begins with expansion of the cell substrate that produces the
Search WWH ::




Custom Search