Biomedical Engineering Reference
In-Depth Information
study. However, the term “worst case” is often applied to, for example, high protein
load, but it is not clear that a high load is always the worst case. The same is true for
other independent variables (e.g., conductivity and pH) and the interactions between
independent variables.
In a QbD approach, the manufacturer would run a series of virus clearance studies
containing a strong element of statistical design of experiments (DOE), which would
allow the quantification of main effects and the more likely two-way interactions. The
outcome would be a statistical and perhaps mechanistic process understanding that,
along with the product CQA requirements and the knowledge base of viral clearance
technologies, forms the basis of a design space. In this context, Table 8.4 provides an
overview of the typical significance of several common independent variables on both
inactivation and removal.
The concept of “bracketed generic clearance” was described in 2003. In a controlled
study, it was observed that an LRVof
4.6 log
10
of rodent type C retrovirus was achieved,
TABL E 8.4. Significance of Various Product, Process, and Testing Variables on Virus
Inactivation and Removal
Removal
Chromatography
Inactivation
Virus Filtration
Product
pH
þþ
þþ
þ
þþ
þþ
þ
Conductivity/ionic strength
Temperature
þþ
þþ
þ
Buffer composition
þþ
þþ
þ
Feedstream purity
þþ
þþ
þ
Aggregate concentration
þ
þ
þ
Protein concentration
þþ
þþ
þþ
Type of product
þ
þ
þ
Process
Unit operation time
þþ
þ
þ
Flux or flow rate
NA
þþ
þþ
Pressure
NA
þ
Cleaning
NA
þþ
NA
Reuse
NA
þþ
NA
Temperature gradient/heat transfer
þ
NA
NA
Mixing efficiency
þ
NA
NA
Virus testing
Virus spike purity
þþ
þþ
þþ
Virus spike viability
þþ
þ
þ
Virus spike volume
þþ
þþ
þþ
Virus titer
þþ
þþ
þþ
Availability of scale-down model
þþ
þþ
þþ
Virus class
þþ
þþ
þ
Virus size
þ
þ
þþ
þþ
, very significant in most cases;
þ
, significant in some cases; and NA, not applicable.
