High Purity Water Systems: FDA Inspection Guide Explained

In short: The US FDA’s “Guide to Inspections of High Purity Water Systems” (July 1993) is still the document many inspectors think of when they walk into a water plant. Its limits and pharmacopoeial references are old, though. This article explains what the guide is, then summarises what current USP, Ph. Eur., EMA and WHO texts expect from Purified Water (PW) and Water for Injection (WFI) systems. It is for QA, QC, microbiology and engineering staff. Always take limits from the current official texts, not from this page.

What the FDA guide is

The guide is an FDA inspection reference titled “High Purity Water System (7/93)”. FDA still hosts it on its website under inspection guides. It was written for FDA investigators, and it says plainly that it does not bind FDA or give rights to anyone. It is not a regulation and not a guidance for industry.

It looks mainly at the microbiology of water systems used to make drug products and drug substances, and it also covers system design and problems that inspectors had seen in the field. Its sections deal with system design and validation, microbial limits, WFI systems (stills, heat exchangers, holding tanks), pumps and piping, reverse osmosis, purified water systems and an inspection strategy.

FDA also has a separate Inspection Technical Guide, “Water for Pharmaceutical Use” (ITG no. 46), which lists the water types used in the industry. It is useful background, and it too is old.

Why inspectors still use it

The guide explains practical failure modes in plain language. Several ideas in it have not aged:

• A system should be described by a drawing showing all equipment from feed water to the points of use, with every sampling point marked. A water system without a current drawing is a common observation.
• Microbial limits are not simple pass/fail numbers. When a limit is exceeded, the firm should investigate, fix the cause and assess the effect on product.
• Organisms attach to pipe and tank walls (biofilm) and are released unevenly. Results from one sample point can therefore swing widely from one day to the next.
• Conductivity meters tell you about ionic purity, not about microbial quality.
• Circulating loops are preferred to one-way systems, and hot systems are described as self-sanitising.
• Validation should run in phases and cover a full year of routine data.

What is out of date

The guide refers to USP XXII and to early-1990s drinking water regulations, and it predates today’s USP chapters on conductivity and TOC. It also treats distillation and reverse osmosis as the USP routes for WFI, which is a narrower view than current texts. Use it to understand inspection thinking, not as a specification.

Water grades in plain terms

The European Pharmacopoeia also had a “Water, highly purified” monograph, but EMA’s 2020 guideline notes that it has been suppressed. If an older SOP or specification still names highly purified water, review it against your current dossier and pharmacopoeia.

Current official texts to work from

USP chapters (United States)

• USP <1231> Water for Pharmaceutical Purposes. An informational chapter on water types, system design, operation, validation, sampling and microbial control. It is where the well-known microbial action levels are discussed.
• USP <645> Water Conductivity. The method and limits for conductivity as a measure of ionic impurities in PW and WFI, including the staged approach for offline and online measurement.
• USP <643> Total Organic Carbon. The method and limit for TOC in PW and WFI.
• The Purified Water and Water for Injection monographs in USP–NF carry the official specification, including the bacterial endotoxin requirement for WFI.

Do not copy numbers for conductivity or TOC from blogs, old SOPs or this article. Take them from the current USP–NF chapters, which are updated from time to time. For orientation only: <645> Stage 1 conductivity for bulk water is 1.3 µS/cm at 25 °C (the limit is temperature-dependent, read from the chapter’s table), and the <643> TOC limit for bulk Purified Water and WFI is 0.50 mg/L (500 ppb) as carbon.

The commonly quoted microbial and endotoxin numbers

USP’s own public FAQ on pharmaceutical water states that the <1231> action levels, 100 CFU/mL for Purified Water and 10 CFU/100 mL for Water for Injection, are generally considered levels above which the water is unfit for use. The same FAQ says these are process-control tools rather than release specifications. It recommends that in-house alert and action levels be set no higher than, and preferably lower than, these values, based on your own trending. These two numbers are also the ones in the old FDA guide.

The WFI endotoxin limit is less than 0.25 USP Endotoxin Units per mL (USP Water for Injection monograph; Ph. Eur. 0169 gives less than 0.25 IU/mL). FDA’s 1993 guide only mentions this value in examples.

In practice, the figures above are action levels and the monograph limit. A site’s alert level is normally lower, and an alert should trigger a look at the trend before anything fails. Which numbers apply to your product also depends on your market, your marketing authorisation and your own justified specification.

Ph. Eur. and the 2017 change for WFI

Until April 2017, the European Pharmacopoeia allowed WFI to be made by distillation only. According to EMA’s guideline, the revised WFI monograph (0169) was published in Ph. Eur. Supplement 9.1 and took effect on 1 April 2017. It now allows a purification process equivalent to distillation, for example reverse osmosis (single or double pass) combined with other suitable techniques such as electrodeionisation, ultrafiltration or nanofiltration. EMA notes this brings Ph. Eur. closer to the USP and the Japanese Pharmacopoeia. The Purified Water monograph is 0008. The text of the Ph. Eur. monographs themselves is not freely quoted here.

EMA guideline on the quality of water for pharmaceutical use

EMA/CHMP/CVMP/QWP/496873/2018, dated 20 July 2020, came into effect on 1 February 2021. It replaces the 2002 Note for Guidance (CPMP/QWP/158/01) and the CPMP position statement on water for vaccines. Key points:

• It sets the minimum water quality for different products and manufacturing steps, in tables. For example, WFI is the minimum for parenteral products, biologics (including vaccines and ATMPs) and for haemofiltration, haemodiafiltration, peritoneal dialysis and irrigation solutions. Purified Water is the minimum for oral, rectal/vaginal, and most nasal/ear and cutaneous preparations, unless a risk assessment says otherwise.
• Potable water may be used during the manufacture of active substances and in the early stages of cleaning equipment, unless technical or quality needs require better water.
• Water grade is a risk-based choice and should be discussed in the marketing authorisation application.
• It is meant to be read with the EU GMP/GDP Inspectors Working Group Q&A on producing WFI by non-distillation methods, reverse osmosis and biofilms (EMA/INS/GMP/443117/2017, Questions and answers on production of water for injections by non-distillation methods – reverse osmosis and biofilms and control strategies). The guideline also says that when reverse osmosis is introduced at a site for WFI, notice should be given to the GMP supervisory authority before implementation.

WHO TRS 1033, Annex 3 (2021)

“Good manufacturing practices: water for pharmaceutical use”, WHO Technical Report Series No. 1033, Annex 3, 2021, revises the earlier WHO text (TRS 970, Annex 2, 2011). Its background section explains that water, unlike most ingredients, is drawn on demand and is not tested and released by batch before use, so control of the system matters more than end testing. WHO revised the text so that WFI may be produced by technologies other than distillation, and cross-refers to its separate 2020 guideline on that topic (TRS 1025, Annex 3).

The WHO text covers drinking water, bulk purified water and bulk WFI, general design considerations, sanitisation, storage and distribution, qualification, continuous monitoring, maintenance, periodic review and inspection. For microbial and chemical purity it points to the pharmacopoeia and requires alert and action limits based on system knowledge and trend data.

Design: what good systems have in common

Numbers below are typical or guidance values taken from the sources named, not universal requirements. Your own design, risk assessment and validation decide.

Other design points worth checking on a walk-through are the quality of materials (usually 316L stainless steel, orbital welds with records), the positions of UV units and filters, how sample valves and points of use are configured, and how the system is sanitised (steam, hot water, ozone or chemical agents) with proof that the sanitising agent is removed.

Validation phases

Both the FDA guide and WHO describe a phased approach. WHO sets out three phases covering at least one year of operation across seasons: Phase 1 of at least two weeks with intensive sampling and no use of water for manufacturing, Phase 2 of at least a further two weeks, and Phase 3 to complete at least 12 months, with sampling that can be reduced according to a routine plan. FDA’s older guide describes an initial phase of two to four weeks, a confirmation phase and then a year of routine data.

Source water testing should be part of the programme and continue in routine monitoring.

Sampling and trending

A good sampling plan answers four questions: where, how often, how, and what do we do with the results.

• Where. Feed water, after key purification steps, loop return and representative points of use. USP’s FAQ says analysis must represent the water as used, and that sample-port results can look better than the point of use.
• How. Use the same method that production uses (hose, flushing, draining) so that the sample reflects what the product sees. WHO makes the same point.
• Volume. FDA’s old guide states that WFI samples of 100 to 300 mL are expected and samples under 100 mL are unacceptable. Check the current compendial method for your test.
• Trend. WHO suggests statistical review at defined intervals (monthly, quarterly, annually) with control limits such as 3 sigma, and alert and action levels based on historical data.
• React. Investigate adverse trends and out-of-limit results for root cause, and when microbial contamination of bulk WFI occurs, identify the organism (WHO).

Read this together with our SOP example on sampling and testing of purified water in microbiology and the water purification system operation SOP. Both are illustrative templates to be adapted under your own quality system.

A short inspection checklist

1. Current P&ID or system drawing with sample points matching the field.
2. Qualification and validation reports, including the full-year data and seasonal effect.
3. Alert and action levels, with the rationale and the trending history.
4. Excursion investigations and CAPA, including any organism identification.
5. Sanitisation records and proof of removal of agents.
6. Change control for modifications, and records of preventive maintenance and calibration of online instruments.
7. Periodic system review.

Frequently asked questions

Is the 1993 FDA guide still valid?

It is still published by FDA as an inspection reference, but it is old and non-binding. Use it for inspection logic and for failure examples. For specifications, use the current pharmacopoeia.

Is 100 CFU/mL a limit for Purified Water?

It is the action level discussed in USP <1231>, which USP describes as a process-control tool and not a release specification. Set your own alert and action levels at or below it based on trending, and follow the requirement of your market and product.

Can WFI be made by reverse osmosis?

Under the revised Ph. Eur. monograph 0169 (effective 1 April 2017) and the USP, a purification process equivalent to distillation is accepted, provided it is validated and controlled. WHO’s 2021 guideline also allows it. Check your product’s regulatory filings before changing a method.

Why is conductivity not enough?

Conductivity and TOC show chemical quality in real time. They do not show microbial contamination or endotoxin, so microbial and endotoxin testing remain necessary.

Do I need a 12-month validation for a new system?

WHO describes a three-phase approach covering at least one year of operation. Your regulator and company procedure decide what is required; document any justified reduction for changes to existing systems through risk assessment.

References

1. US FDA, Guide to Inspections of High Purity Water Systems (7/93), FDA Inspection Guides (ORA).
2. US FDA, Inspection Technical Guide No. 46, Water for Pharmaceutical Use (issued 1986; web page current as of 2014).
3. USP–NF, General Chapters <1231> Water for Pharmaceutical Purposes, <645> Water Conductivity and <643> Total Organic Carbon; monographs Purified Water and Water for Injection (current official USP–NF).
4. European Pharmacopoeia monographs Water for Injections (0169) and Purified Water (0008), (current edition; WFI may be produced by non-distillation methods since the 2017 revision of monograph 0169).
5. EMA/CHMP/CVMP/QWP/496873/2018, Guideline on the quality of water for pharmaceutical use, 20 July 2020 (effective 1 February 2021).
6. EU GMP/GDP Inspectors Working Group, Questions and answers on production of water for injections by non-distillation methods, reverse osmosis and biofilms and control strategies, EMA/INS/GMP/443117/2017.
7. WHO, Good manufacturing practices: water for pharmaceutical use, WHO Technical Report Series No. 1033, Annex 3, 2021.
8. WHO, Production of water for injection by means other than distillation, WHO TRS No. 1025, Annex 3, 2020.