Chemistry Reference
In-Depth Information
degradant of the API will generally also be a CQA. Other properties may or may not be
critical because they may or may not impact safety or ef
cacy, which needs to be
evaluated on a case-by-case basis. For example, particle size would be a CQA if it affects
dissolution to a degree that impacts the drug
'
s exposure, but may not be critical
otherwise.
Importantly, the risk assessment that determines criticality of an attribute is not a
once-and-for-all process; instead, risk levels are re-evaluated during development as the
material under consideration is better understood. In fact, the initial risk assessment aims
primarily to identify areas of concern, where ongoing experimentation will be necessary.
This determination may be based on prior familiarity with the process or product. To
carry out this and any subsequent assessment, the development team may assign a risk
priority number. This is a semiqualitative determination that multiplies the magnitude of
potential harm, the likelihood of the harm occurring, and the likelihood that the harm
would be detected. This is similar to failure modes and effects analysis (FMEA)
commonly used in process development. Clearly, the process of assigning a risk priority
number is not precise, but it does allow the team to classify a CQA broadly as low,
medium, or high risk.
A number of parameters may have some effect on a CQA. The effect becomes an
impact
, in regulatory terms, when it causes the CQA to cross a
criticality threshold
.This
threshold is crossed when the change to the CQA is likely to cause changes to safety or
ef
-
cantly) likely to change (or impact) a CQA beyond its acceptable limits (the criticality
threshold), that parameter needs to be controlled during development and manufacture.
Regardless of the risk level, once a parameter is de
cacy beyond accepted limits. To summarize, when a parameter is (statistically signifi-
ned as a CQA, it is subject to
quite different handling from other measurements of the drug. As far as regulators are
concerned, an attribute cannot be
a little bit
critical
either it is critical or it is not. While
the target range of a noncritical parameter could be changed without any pre-approval
from regulatory authorities, such a change would require an update in an annual report.
Conversely, changes to CQAs that expand the design space are major maneuvers and
require a lengthy process of approval and review.
We propose utilizing a conceptual approach to design space development as
illustrated in Figure 11.3. In the
—
first stage (step 1), preliminary design of experiments
(DoEs) should be done on individual unit operations (i.e., solution preparation, spray
drying, secondary drying, and so on for the spray drying process; and blending,
compression, and so on for the tablet process). These experiments will provide
information on both viable operating ranges for the individual unit operations and
the effect of material and process parameters for the given operation on output
parameters from that unit operation (e.g., the effect of spray drying process parameters
on SDD properties such as particle size and bulk density).
In the second stage (step 2), the interaction between outputs from one particular
operation on downstream unit operations should be investigated by performing DoEs
incorporating more than one unit operation. For example, the effect of spray drying on
final tablet dissolution should be investigated, as a function of incoming SDD properties
(such as particle size and bulk density), by performing DoEs incorporating the spray
drying and tablet compression steps, in a discontinuous mode of operation.
