Finbar:
I think you are on the right track in considering ICH Q3A,
"Impurities in New Drug Substances."
If you know what impurities would be expected to carry over from
batch to batch, taking degradation due to the cleaning process into
account, you can put together a formal risk assessment for the
product carryover, based on the effect of the impurities on the
process, on the product, and on the patient. A maximum value for
carryover would probably be the ICH Q3A limit for identification of
impurities (0.10%).
Based on the 0.10% identification threshold for organic impurities, I
would base my Maximum Allowable Carryover (MACO) calculation on NMT
0.10% carry-over from batch-to-batch, using TOC or another suitable
assay for measurement. I would use the bioreactor protein production
in grams as the starting point. This is assuming a bioreactor or
fermenter process, where further downstream purification is expected
to occur. As stated by Dr. Kirsch, the actual MACO calculations can
be found in PDA TR29, or in Destin LeBlanc's book.
As an example, if you produced 100 grams of protein in a bioreactor
batch, your MACO into the next batch would be 0.1 gram of protein.
You would assume this 0.1 gram of protein would be distributed
evenly across the bioreactor surface. If you were to rinse the
bioreactor with 10 liter of WFI, assuming 100% recovery, your Maximum
Allowable Carryover would be 0.01 gram / liter, or 10 microgram / mL.
Obviously, you would have to factor in the actual rinse water
recovery, assay recovery, etc., to come up with your final limits.
Obviously, if your expected residues or degradants are expected to be
detrimental to the patient or process, you would take a look at
toxicity data such as the LD50 or NOEL limits, which could result in
limits tighter than the ICH Q3A Impurities limits.
Ed White
Senior Validation Scientist
Baxter Bioscience
Thousand Oaks, CA
ed_white@baxter.com
(805) 375-6779
Find purpose; the means will follow. -Gandhi
******************
Pharmaceutical Sci-Tech Discussion Group
To subscribe or unsubscribe go to:
http://www.pharmweb.net/pharmscitech.html
******************
|