Buffer Media Validation Issues

[chrisTRG]

Hi all,

First post, so please be patient with me

I have recently taken on the position of Validation Specialist in my company, and with that I have inherited a half-validated process which requires a total re-write as far as I can tell :mg: .

Basically, we have up-scaled a media production process whereby a number of various types of buffer solutions are made by the introduction of chemical compounds (e.g. NaCl, MgCl2, EDTA etc) to specification (e.g. 5M NaCl, 0.9% Saline) to purified water. These solutions are then pH neutralised, made up to final volume, prior to sterilisation. For obvious reasons I cannot be too specific here, sorry!

The small-scale system involved 1/2L of these buffers being made up in small glass bottles followed by autoclave sterilisation. No verification was performed in the past.

The up-scaled version uses 100/200L drums to make up these solutions followed by filter sterilization. Filters are then integrity tested and a sample is taken from the flush bag for dry-weight analysis.

The main issue we are having is that the small scale process was not verified using dry-weight analysis, so we have no way to compare these process'. This is made more difficult by the fact that autoclaving the small bottles also reduces the volume of buffer solutions, affecting the dry-weight reading, which we obviously don't see in filter sterilisation, hence large discrepancies in values.

Is there a more effective way to analyze these buffers? Does the process actually require validation? Should we be comparing new with old, or is the new process a whole new validation project of its own? How can a range of acceptable tolerances be defined for validation?

Thanks

 

[Bev D]

As you describe it, the two are different processes.
the buffer formulations may be the same %by volume or weight or whatever; but the use of a different sterilization method for small scale and large scales makes it by definition a different process. the use of two different mesurement systems that have not been assessed for critical differences (method comparison) only complicates the issue.

...so yes the large scale process must be validated separately.

 

[chrisTRG]

Thanks Bev.
Yes, I thought that was the case. Unfortunatly the inherited Validation Plan states that we will validate this "new" process as "like-for-like" which has caused numerous issues and a ton of unnecessary work.

That being said, the small scale production of media was not specifically validated. As such if we take on the up-scaled media validation as a new, separate validation project, how should I carry out Operational Qualification? Do I gather data from dry weight analysis of the upscaled product until I have enough to produce a standard deviation, from which I can use statistical significance to produce product parameters/operational windows?
Thanks again for the helpful advice!

 

[Bev D]

Well OQ is to validate that the output of the process meets requirements at the extremes of the allowed inputs so you create the variation deliberately and you don't really need lots of data. just replication. typically OQ is accomplished using a fractional factorial DoE design...