Pharmaceutical water system validation: the three phases explained

Water is the most used raw material in a pharma plant and a common source of Gram-negative contamination. It is used in formulations, for cleaning equipment, for washing components, and as the base for many laboratory reagents. A water system is validated in three phases so that you see how it behaves over weeks of intensive testing and then through every season of a full year.

This article explains each phase, the limits you validate against, how to sample and test properly, and what to keep watching once the system is in routine use.

Types of pharmaceutical water

• Purified Water (PW): used for non-sterile products, cleaning and as feed for WFI generation. Usually produced by reverse osmosis, often with electrodeionisation, from potable water.
• Water for Injection (WFI): used for parenteral products and final rinsing of product-contact equipment in sterile manufacturing. Traditionally produced by distillation; Ph. Eur. and USP now also accept membrane-based processes if equivalent quality is shown.
• Highly purified and sterile waters have their own monographs and uses but are less common in routine production.

Limits you are validating against

Attribute	Purified Water (PW)	Water for Injection (WFI)
Microbial action level (typical)	100 CFU/mL	10 CFU/100 mL
Bacterial endotoxins	Not specified	Less than 0.25 EU/mL
TOC	Not more than 500 ppb	Not more than 500 ppb
Conductivity	Per USP <645> / Ph. Eur.	Per USP <645> / Ph. Eur.

Microbial levels for PW and WFI are action levels, not pass/fail specifications, in USP <1231>. Your own alert levels should come from trend data. Check results with the water count checker.

Before validation: qualification

Validation starts after the system has passed design, installation and operational qualification:

• DQ: the design meets user requirements, including materials, sanitisation method, flow velocity and dead legs.
• IQ: the system was built as designed: materials, welds, slopes, instruments and documentation.
• OQ: each part works as intended: pumps, heat exchangers, UV lamps, alarms, sanitisation cycles.

Performance qualification is the three-phase monitoring programme described below.

Phase 1: 2–4 weeks of intensive testing

• Sample every point of use and every critical point in generation, storage and distribution, every day.
• Develop and confirm operating, cleaning, sanitisation and maintenance SOPs.
• Water is not used for production during this phase.

The aim is to understand the system and fix problems before routine use. Expect some results to be high at first, particularly from new outlets or after construction work. Each one should be investigated and the fix documented.

Phase 2: another 2–4 weeks with final SOPs

• Continue intensive sampling with the SOPs from Phase 1.
• Show the system consistently produces water of the required quality when run by its procedures.
• At the end, water can usually be released for production, subject to your protocol.

If the SOPs change during Phase 2, the clock usually restarts, because the purpose is to prove that the final procedures work.

Phase 3: one year of routine monitoring

• Sample at the routine frequency and locations defined in the protocol.
• Capture seasonal variation in the feed water and its effect on the system.
• At the end, the system is considered validated, and routine trending continues for its life.

In India, the monsoon can change feed water quality sharply. Phase 3 should show that pretreatment and the main system cope with that change.

Typical sampling plan

Location	Phase 1 and 2	Phase 3 and routine
Feed water (potable)	Daily	Weekly or monthly
After each treatment step	Daily	Weekly
Storage tank	Daily	Weekly
Points of use, PW	Daily	Each outlet at least weekly, on rotation
Points of use, WFI	Daily	Each outlet daily or weekly, depending on use and risk
Return loop	Daily	Daily or weekly

Adjust the plan to your system design and risk assessment, and write the rationale into the protocol.

Sampling technique

Bad sampling causes more water excursions than bad water. Follow a few rules:

• Sample the way production uses the water: from the same outlet, with the same hose, after the same flush time. If production uses a hose, sample through the hose.
• Use sterile, depyrogenated containers for endotoxin samples, and sterile containers for micro samples.
• Do not sanitise the outlet with alcohol just before sampling, unless production does the same, because it can kill organisms and give a false low result.
• Test promptly. Analyse within 2 hours or refrigerate at 2–8 °C and test within the time your validation supports, usually 12–24 hours.

Microbiological testing tips

• Use a low-nutrient medium such as R2A agar with longer incubation (for example 5–7 days at 30–35 °C, as justified) to recover stressed waterborne bacteria. Plate count agar with shorter incubation can miss them.
• Filter at least 100 mL for WFI by membrane filtration. For PW, 1 mL by pour plate is the minimum, but filtering a larger volume gives better sensitivity.
• Run a negative control for every session and growth promotion for every medium lot.
• Identify recurring isolates; Pseudomonas, Ralstonia, Burkholderia, Sphingomonas and other Gram-negative rods point to biofilm.

Understanding biofilm

Biofilm is a layer of bacteria embedded in a slime they produce, attached to pipe walls. Once established, it is very hard to remove, and it releases bacteria and endotoxin into the water intermittently. Signs of biofilm include counts that creep up between sanitisations, the same organism appearing at several outlets, and endotoxin spikes without high counts.

Prevention is easier than cure:

• Keep water moving with turbulent flow (typically above 1 m/s in distribution).
• Avoid dead legs; the usual target is a length less than three times the pipe diameter.
• Keep hot loops above 65 °C, or sanitise cold loops regularly with heat or ozone.
• Use smooth, electropolished stainless steel and sanitary fittings.

What to trend

• Counts by sampling point over time, against alert and action levels
• Organism identities and new species
• Sanitisation dates against counts (do counts creep up between sanitisations?)
• Conductivity, TOC and temperature alongside microbial data
• Endotoxin results for WFI, even when well within limits

Present trends monthly for routine review and summarise them in an annual product or system quality review. A rising trend inside limits is a warning; act on it before it becomes an excursion.

Handling an excursion

1. Confirm the sampling and testing were correct.
2. Check the point of use, recent maintenance and sanitisation records.
3. Resample the outlet and nearby points, including the return loop.
4. Identify the organism and compare it with previous isolates.
5. Assess the impact on any batch or equipment that used water from that outlet since the last good result.
6. Decide on corrective action, such as extra sanitisation or replacing a valve or hose.

Common audit findings

• Sampling through a different outlet or hose than production uses
• Excursions closed as "sampling error" without evidence
• No trending, or trend reports with no conclusions
• Alert levels not based on data
• Dead legs and poor slopes found during walk-downs
• No justification for incubation conditions or media

A well-designed, well-run water system is quiet: low counts, stable trends and few surprises. The validation phases are your first chance to prove that, and routine trending keeps proving it.