2021 Volume 23 Issue 2 Pages 23-44
It is common to release compendial waters (Purified Water and Water for Injection) continuously based on the online testing for its chemical attributes (TOC and conductivity).
However, the release of this water would be “at risk” since its microbial attributes would not have been evaluated due to the time required to grow and enumerate microbial colonies.
As an example, during the monitoring of the microbial attributes of our water system by traditional growth-based methods, we observed inconsistent and sometimes failing results from week-to-week.
Although the water system in question is a product development tool, not a water system for use with a pharmaceutical product, the maintenance of the chemical and microbial control is paramount.
Instead of increasing the frequency of the traditional sampling with more samples per week, we employed an at-line, real-time, fluorescence-based microbial detection technology.
These periodic test failures, which later turned out to be the result of purification process failures, were occurring every 24-30 hours, and traditional growth-based microbial detection methods could not have easily detected these failures.
By examining the continuous data, we correlated the auto-fluorescent count to a tank-filling process, then correlated the tank filling process to an exhaustion of mixed-bed ion-exchange resin.
The observation was a sudden increase in microbial counts followed by a subsequent decay over hours as the recirculation loop reduced the microbial counts.
By measuring continuously (every second), we observed periodic and serious purification failures.
We corrected those failures, thereby producing consistent and compliant water quality, and now have a predictive tool to identify maintenance needs of key purification processes, i.e., processes that are consumed over time such as carbon bed and ion-exchange resin.
