Scenario - WFI Loop OOS
Engineering drops you an email to say their continuous conductivity measurement is on the rise and the microbiology department calls to report a high Total Viable Count on your water loop.
As ever, an approach….
First thought
Both sets of “signals” indicate that the water loop has something in it that shouldn't be present. We are no longer making/using WFI and we probably need to stop ASAP.
An aside
This is not really a call your QP should need to make, If you employed on a sterile manufacturing site (assumption here as WFI is being mentioned) then anyone who is trained in all but the introductory roles in Production, Microbiology, QC, Engineering, Validation and QA at a minimum should understand the importance of a clean WFI system remaining in control. This is a fundamental part of your Contamination Control Strategy being effective. If this isn’t the case and people don’t either 1) know how to spot a problem or 2) feel empowered to say “hang on a second” on their own, you might need to look at technical training and or improving your quality culture.
While this aside alone may seem contrived, its important that you understand that underlying culture is fundamental to everything else discussed below. It is also the hardest thing to fix with a CAPA in your eQMS system.
In my own viva my first answer to one, quite detailed technical question around the CAPA expected to address significant substandard behaviour of a supplier was “I would stop working with them”. This visibly made my examiners relax and while alone it’s not demonstrating knowledge, a key thing you need to demonstrate is when you know something isn’t right on a level beyond the technical problem presented.
Back to the question, quit waffling. How would you manage this?
So let’s start with process. The key QMS activities would be an OOS and likely a deviation record to:
Confirm that the water system is out of control. It is still possible that two systems have shown inaccurate readings but given that these are independent means of measuring water quality, that’s a remote possibility. You still need to check the micro tester is trained and if the conductivity monitor was recently calibrated – coincidences do occur.
Document what the organism is.
Develop a recent history around the WFI systems.
When did the conductivity reading start to rise?
Have there been other micro results out of specification.
When has the system been sampled? Whats the frequency? How many points?
Where is the conductivity sensor relative to the point sampled by microbiology?
What does conductivity say now? Is it trending up still?
What are the other sampling systems telling me – is there a rise in endotoxin levels? (is our method capable of quantifying endotoxin levels below an OOS result is a great little micro weekend read if you need a diversion). What are the TOC levels.
Who has used WFI recently? What has it been used for? Is it used for API manufacture, Finished Goods manufacture, or cleaning?
When was maintenance last performed on the system and test equipment?
When was the system last sterilised or sanitised?
What are the recent temperature trends for the loop?
Document your (fairly immediate) CAPAs
As soon as possible the loop should be sanitised to minimise the chance that a biofilm could form, and the contamination becomes more entrenched.
Quarantine impacted material.
Document a formal risk assessment.
In general you will need to define a point of “When the water was last classifiable as WFI” and risk assess the suitability of everything that has not used WFI.
In this respect the type of organism is important as is consideration of downstream controls. Is your WFI drawn off to make a product that is subject to further sterile filtration or terminal sterilisation. That all sounds relatively low risk until micro reports it’s a gram negative bacteria that produces endotoxins.
Document any CAPAs that arise based on your investigation.
In general, WFI system are well controlled and root cause will be tied to a special circumstance. If a system is overall not suitable, such as design deficiency or engineering defect – it is generally not possible to validate.
Note:
Until the recent update to EU guidance the only way to make WFI was using distillation which is a fairly absolute control mechanism as bacterial contamination is not going to survive a steam-born transfer in a still. More recently means of high-grade filtration can be applied. Whilst this enables some smaller and cheaper WFI systems to be available it does leave open a new failure mode – when a filter develops a crack or split – whatever it has been keeping out will get in. (My inner WFI purist prefers distillation).
Investigative approach.
As water systems benefit from continuous monitoring of some parameters and are generally cared for by multiple departments (when was the last time your utilities engineer set foot in a biological safety cabinet) a shared timeline view is incredibly helpful in both risk assessment and root cause investigation. Get everyone in a room and a long whiteboard or use the boardroom table.
As usual there would also be an expectation that other parts of the business should be notified as they may be impacted such as the sales or planning team. It can take while to get back to normal.