WhatsApp
Beta Pramesti Asia
betapramestiasia

Pharma’s water wars: distillation’s heat vs. ultrafiltration’s finesse

  • beta-pramesti-asia
  • industry-pharmaceutical
  • process-purified-water-pw-dan-water-for

Pharma’s water wars: distillation’s heat vs. ultrafiltration’s finesse

Endotoxins are the tiny, heat-stable fragments that drug makers fear most in their water. Distillation has long kept them at bay, but ultrafiltration is muscling in as a final polish with multi‑log removal and lower energy.

Industry: Pharmaceutical | Process: Purified_Water_(PW)_&_Water_for_Injection_(WFI)_Generation

Pharmaceutical water quality is a zero‑defect game. Water for Injection (WFI, the highest‑purity grade used for parenterals) must hit ≤0.25 EU/mL (endotoxin units per milliliter), and endotoxins—lipopolysaccharide (LPS) fragments from bacterial cell walls—are nonvolatile, heat‑tolerant troublemakers. In this environment, distillation has been the “gold standard” because boiling cleanly separates nonvolatile impurities and endotoxins from the steam that ultimately condenses into WFI.

That thermal separation is more than folklore. Distillation “ensures that bacterial endotoxins, bacteria, and other nonvolatile impurities are left behind with the water, ultimately removed from the evaporator… by blowdown” (Pharmaceutical Technology). Historically, surveys found distillation as the final purification step in more than 99% of WFI systems across the US, EU, and Japan (studylib.net), and as recently as 2017, the EU still permitted only distillation as the compendial route (A3P).

In Indonesia, Good Manufacturing Practices are explicit: WFI must be “diproduksi melalui cara penyulingan atau cara lain yang akan menghasilkan mutu yang sama” (“produced by distillation or other means giving the same quality”), and WFI must be stored and circulated in clean, inert systems at >70 °C (CPOB; CPOB).

Thermal separation and real‑world limits

Run properly, distillation consistently yields WFI that meets the endotoxin spec (for example, <0.25 EU/mL). Industry surveys reported that more than 90% of WFI loops routinely showed <10 CFU/100 mL microbial counts (A3P), implying endotoxin control is well in hand.

But stills are not infinite sinks. Regulators have warned that “the log reduction of endotoxin that stills can provide is limited…; in many failures an insufficient pretreatment system produced an endotoxin feed…greater than [the still] could remove” (Pharmaceutical Technology). Biofilm risk adds another constraint: “biofilm formation can occur in WFI stills unless the system is kept hot (≥65–70 °C) continuously” (Pharmaceutical Technology). That’s why Indonesian CPOB calls for >70 °C circulation during storage and distribution (CPOB).

Energy and cost considerations

Multi‑effect and vapor‑compression stills are energy‑intensive. One analysis puts WFI from multi‑effect distillation (with PW feed) at roughly €30/m³, with site‑dependent ranges around 12–40 €/m³, plus about €8/m³ for the PW pretreatment (A3P). By contrast, one estimate places ultrafiltered WFI at ~$0.25/L versus ~$1.00/L by distillation (University of Arkansas Inquiry).

Ultrafiltration as endotoxin barrier

Ultrafiltration (UF) uses membranes with very tight sieving characteristics, defined by a molecular‑weight cutoff (MWCO). The USP notes UF is “most often employed in pharmaceutical water systems for removing endotoxins” (USP <1231> via studylib.net), with typical devices in the 10,000–20,000 Dalton range (USP <1231> via studylib.net). In practice, endotoxin removal is multi‑log: endotoxin‑retentive filters in dialysis report LRVs (log reduction values, i.e., orders of magnitude reduction) of ≥6 for endotoxin (Membranes journal, 2025), and a 10 kDa polyethersulfone UF membrane reduced endotoxin below the USP limit of 0.25 EU/mL in controlled trials (University of Arkansas Inquiry).

Membrane‑based WFI systems—now permitted by USP and, post‑2017, recognized by the EP—typically finish with an ultrafilter as a final polish. In “cold” designs, double‑pass RO plus continuous microbial control capped by UF is common; a 2019 WHO draft guideline explicitly describes WFI via RO and deionization with an ultrafiltration step (ECA/WHO draft). Field data show these “non‑distilled” loops can meet strict specs: over 90% reported <10 CFU/100 mL and met endotoxin limits (A3P).

As a final barrier in PW or WFI, UF is a targeted choice; many facilities deploy dedicated ultrafiltration units as the last unit operation to capture residual LPS that may slough from biofilms. The USP chapter <1231> explicitly cites UF as an “intermediate or final purification step” that “may be appropriate” for endotoxin removal (USP <1231> via studylib.net; USP <1231> via studylib.net).

Reverse osmosis and membrane trains

In membrane trains, reverse osmosis (RO) does the heavy lifting on ions and small organics, but ordinary RO membranes (MWCO ~300 Da) have negligible endotoxin rejection; double‑pass RO alone cannot be relied upon for endotoxin control (Pharmaceutical Technology). System designers increasingly combine RO with UF in modern membrane systems, and some adopt double‑pass RO prior to UF when targeting WFI‑grade quality. When double‑pass is selected, facilities may standardize on robust RO skids before the final ultrafilter stage.

Other endotoxin removal methods

Some unit operations help; others do not. Activated carbon beds are excellent for chlorine and organics, but typically ineffective against endotoxins, so they cannot be counted on for LPS removal (USP <1231> via studylib.net). In practice, carbon is still valuable upstream for organics; many PW trains include activated carbon stages for that purpose.

Ion‑exchange “organic scavenger” resins—especially weakly basic, macroreticular anion exchangers—can adsorb negatively charged endotoxins; the USP notes these resins “remove organic material and endotoxins from the water” and can be regenerated with caustic brine (USP <1231> via studylib.net). Operating concerns include fouling and regeneration frequency (USP <1231> via studylib.net). Where adsorption is part of the strategy, plants source dedicated ion‑exchange resins designed for pharmaceutical service.

Sterilizing‑grade microfilters (commonly 0.22 µm) remove bacteria but not dissolved endotoxins, which are far smaller and can pass such pores. For bioburden control prior to UF or at points‑of‑use, however, facilities often standardize on validated cartridge filters while recognizing that endotoxin removal is UF’s domain (USP <1231> via studylib.net).

Performance, regulations, and trends

Both distillation and UF can meet endotoxin specs, with different operating economics and failure modes. Distillation inherently strips nearly all endotoxin (often >99.99% if feed levels are moderate) and “virtually assures” pyrogen removal under controlled conditions (Pharmaceutical Technology), while UF routinely achieves multi‑log (LRV 3–6 and higher) endotoxin reduction (Membranes journal, 2025).

Industry surveys once showed >99% reliance on thermal distillation for WFI (studylib.net). Since 2004, however, the USP has permitted “distillation or a purification process that is equivalent or superior…in the removal of chemicals and microorganisms,” opening doors to validated membrane routes (studylib.net). The EP’s 2017 revision likewise allowed membranes (A3P), and WHO’s 2019 draft described RO‑ and deionization‑based WFI with UF (ECA/WHO draft).

In Indonesia, CPOB still emphasizes distillation or equivalent quality and >70 °C circulation (CPOB; CPOB). Globally, the trend is toward membrane adoption with robust validation and hot‑loop or equivalent microbial control.

Design choices and combined strategies

Distillation plus hygienic design and 70–80 °C distribution generally produces extremely low endotoxin levels; a PW train with double‑pass RO and a high‑grade 10–20 kDa UF can likewise achieve sub‑spec endotoxin. Quantitatively, both routes meet the 0.25 EU/mL target, and UF often achieves LRV 3–6 or higher (Membranes journal, 2025). Combined strategies are feasible—adding a point‑of‑use ultrafilter after a traditional still provides a final safety net.

Ultimately, selection hinges on (a) local regulations (for example, Indonesia’s distillation‑first stance: CPOB), (b) system design (hot loops versus ambient “cold” systems), and (c) cost/energy profile—balanced against the need to consistently meet endotoxin release assays and ensure patient safety (Pharmaceutical Technology; A3P).

References and source links