Inside Pharma’s Most Audited Utility: How to Validate a Purified Water and WFI System

In pharma, water is a product-contact utility that never gets a free pass. Regulators expect a documented Validation Master Plan (VMP) showing that Purified Water (PW) and Water for Injection (WFI) consistently hit pharmacopeial limits—think TOC (total organic carbon) ≤0.5 mg/L, endotoxin ≤0.25 EU/mL, and total microbial count ≤100 CFU/mL—citing Indonesian Pharmacopeia (Farmakope VI) as an example that aligns with USP and Ph. Eur. (pdfcoffee.com) (pharmaknowledgeforum.com) (www.usp.org).

For conductivity, the headline number is ≤1.3 µS/cm at 25 °C; for WFI, the practical microbiological expectation is even tougher at <10 CFU/100 mL, with endotoxin ≤0.25 EU/mL (pharmaknowledgeforum.com) (www.usp.org). The VMP must define roles, scope, risk assessment, and a qualification path (IQ/OQ/PQ) (gmpua.com), and, unlike many utilities, “extensive PQ is indispensable” for water systems (gmpua.com).

A site‑specific Master Plan is recommended to enumerate equipment, documentation, schedules, and IQ/OQ/PQ acceptance criteria for all critical components (gmpua.com) (gmpua.com).

Validation master plan structure

Governance rests on GMP expectations (e.g., CPOB in Indonesia; USP/Ph. Eur./WHO internationally). The VMP for PW/WFI defines how Installation Qualification (IQ), Operational Qualification (OQ), and Performance Qualification (PQ) will be executed, by whom, and with what pass/fail criteria (gmpua.com) (gmpua.com).

Pharmacopeial anchors include PW limits for conductivity (≤1.3 µS/cm at 25 °C), TOC (≤500 ppb), microbial count (≤100 CFU/mL), and WFI expectations of <10 CFU/100 mL and endotoxin ≤0.25 EU/mL (pharmaknowledgeforum.com) (www.usp.org). Indonesian Pharmacopeia (Farmakope VI) cites TOC ≤0.5 mg/L (500 ppb), endotoxin ≤0.25 EU/mL, and microbial ≤100 CFU/mL (pdfcoffee.com).

Installation qualification (IQ) checks

IQ confirms the system is installed per design and GMP: vendor docs and P&IDs; materials of construction (e.g., 316 L stainless steel, USP Class VI surfaces); weld quality; aseptic fittings and flange assemblies; calibration certificates for instruments; correct piping slopes; and proper installation of pumps, tanks, and valves (gmpua.com). Test points include “Material proof,” weld integrity, flange design, electrical/control installations, and documentation completeness (gmpua.com).

Binary acceptance applies: installations must match design (e.g., all welds pass dye‑penetrant test; instruments calibrated to NIST standards), with deviations corrected and re‑inspected. IQ procedures, checklists, and calibration certificates are controlled records, and results are reviewed before OQ (www.pharmtech.com).

Example: verify hot and cold distribution loops are sloped 0.5–1% toward drain; measure random segments; confirm sanitary hangers and insulation; match instruments and tag numbers; ensure labels/manuals and SOPs are present. Where 316L components are specified, plants commonly use 316L stainless steel housings for pharmaceutical and food grade applications such as stainless steel cartridge housings, aligning with the “materials of construction” requirement.

Operational qualification (OQ) simulations

OQ demonstrates the system works to spec, including worst‑case and system‑limit conditions. Instrumentation is challenged: inject a known salt solution to confirm conductivity response; pass a potassium hydrogen phthalate standard to check the TOC analyzer; verify accuracy ±3% or better; and test setpoints so high conductivity (e.g., 1.1 µS/cm alarm) diverts to drain.

Flow/pressure ranges are exercised: recirculation pump flow at min/max speeds meets design (±5%); backwashing or sanitization sequences fire on time; pressure sensors read plausible values. For sanitation, heat the loop to ≥80–90 °C per design and verify temperature uniformity; if UV is used, confirm sanitizer function by monitoring units (www.pharmtech.com). Plants that opt for non-chemical disinfection often rely on ultraviolet systems for a 99.99% pathogen kill rate without chemicals and low operating cost.

Controls and interlocks are stressed: simulate raw‑water loss to check pump shutdown/alarm; trigger high‑temperature or low‑pressure conditions; verify drain‑on‑high‑conductivity logic; and confirm data logging and alarm recording, including any remote monitoring. As one guidance frames it, “The OQ will verify the operation of the system… acceptance criteria, particularly for conductivity and temperature… Which conductivity…and what are the actual limits?” with each step, expected result, and data capture predefined (www.pharmtech.com) (www.pharmtech.com).

Acceptance means instruments read within calibration; setpoints trigger correct actions; no dead‑legs or stagnant zones are detected; alarms activate under fault conditions (www.pharmtech.com) (gmpua.com). Example: ramp to ≥85 °C (±2 °C) and hold for the required time; ensure each conductivity cell reads within 0.05 µS of a calibrated standard at 25 °C.

Performance qualification (PQ) monitoring

PQ collects extended, routine‑mode data to show consistent quality over time. Robust programs span multiple weeks and yield 20–30 or more measurements per parameter, concentrating on conductivity, TOC, microbes, and, for WFI, endotoxins (www.pharmtech.com). Alert and action levels are set from compendia and historical performance; USP’s informational chapter treats 100 CFU/mL for PW and 10 CFU/100 mL for WFI as Action Levels that render water unfit for use if exceeded (www.usp.org).

Sampling targets representative points, typically the loop end near use points. Conductivity and TOC are often monitored continuously online, with daily offline checks at 25 °C (post‑polish loop). Microbial count is sampled once or twice weekly by membrane filtration (~100–200 mL) or pour‑plate (1–10 mL) on R2A agar, with incubation at 30–35 °C for ≥5 days; for WFI, USP/Ph. Eur. practice filters 100–200 mL on R2A at 30–35 °C for ≥5 days (www.drugfuture.com). Endotoxin (WFI) is tested with LAL (Limulus amoebocyte lysate) gel‑clot or chromogenic methods per USP <85>.

Acceptance criteria by attribute

Conductivity: ≤1.3 µS/cm at 25 °C for both PW and WFI (equivalent to ≥18 MΩ·cm); on‑line meters are qualified against 0.055 M KCl standards, and daily values are logged, often with tighter in‑house alarms near 1.0–1.1 µS (pharmaknowledgeforum.com).

TOC: ≤0.5 mg/L (500 ppb) for PW and WFI; analyzers are calibrated with potassium hydrogen phthalate, and samples measured within minutes to avoid CO₂ ingress (pdfcoffee.com).

Microbial count: ≤100 CFU/mL for PW; for WFI a practical operational target of <10 CFU/100 mL (0.1 CFU/mL) is imposed in practice; tests use validated incubation conditions (e.g., R2A or Plate Count agar at 30–35 °C, 48–72+ hours). The Farmakope example states “Uji batas mikroba: total plate count ≤100 koloni/mL” (www.usp.org) (pdfcoffee.com).

Endotoxins (WFI): ≤0.25 EU/mL (also cited for PW in the FI example to discourage parenteral use); each batch of tests runs LAL kit QC standards (pdfcoffee.com).

Test protocols and compendial chapters

Each PQ protocol cites USP general chapters—TOC <643>, conductivity <645>, microbial enumeration <1226>, endotoxin <85>—and includes blanks and standards. Example: “Conductivity test: Use a calibrated cell, measure at 25 °C±0.5 °C per USP<645>. Acceptance: ≤1.3 µS/cm” (pharmaknowledgeforum.com).

Data are trended over time (e.g., control charts). Alert levels (e.g., 80% of spec) prompt preemptive actions such as sanitization; Action Levels—USP notes 100 CFU/mL for PW and 10 CFU/100 mL for WFI—trigger investigation (www.usp.org). For example, if conductivity drifts above 1.0 µS/cm, the system is flagged for resin regeneration—an operational reality in systems using ion exchange, including ion-exchange resins used for demineralization. FDA/WHO‑aligned reasoning expects “controlled” operation with microbial attributes monitored against alert/action levels (www.drugfuture.com) (www.usp.org).

Closeout and example outcomes

PQ is not judged on a single datapoint but on population control across the study window. An example end‑of‑study summary: conductivity = 1.10 ± 0.05 µS/cm (25 °C), TOC ≤450 ppb in all samples, microbial 30‑day median = 0 CFU/mL (monthly max <10 CFU), endotoxin <0.1 EU/mL in all samples—showing all episodes met the strictest compendial criteria and none exceeded USP Action Levels (pdfcoffee.com) (www.usp.org).

References cited here—USP/Ph. Eur., PIC/S‑GMP, WHO—underscore the practical focus on conductivity, TOC, microbes, endotoxin (www.pharmtech.com) and the use of alert/action levels for trending (www.usp.org). Indonesian Pharmacopeia (Farmakope VI) explicitly sets values—TOC ≤0.5 mg/L; microbial ≤100 CFU/mL; endotoxin ≤0.25 EU/mL for PW—consistent with USP/Ph. Eur. standards; WFI also has these plus sterility requirement (pdfcoffee.com). All test methods and acceptance limits should be documented in the IQ/OQ/PQ protocols, with analyses performed per validated procedures to meet these criteria in 100% of Qualification samples.