Does Purified Water have to be tested for Endotoxins?

When planning new purified water (PW) systems, the question arises as part of the risk analysis or later, when drawing up the qualification and sampling plans as to whether endotoxins should be tested for.

According to the specifications of the pharmacopoeias (e.g. USP or Ph.Eur.), endotoxin testing is not required for Purified Water. The endotoxin content is not a test parameter in the valid pharmacopoeia monographs for purified water.

However, if purified water is used as feed water for distillation plants and pure steam generators for the production of WFI (water for injection) or pure steam, it may be useful to test for endotoxins as part of the qualification process. This is due to the fact that distillation systems and pure steam generators - depending on the technology and manufacturer - can only achieve a reduction of 3 to 6 log levels of endotoxins.

For the validation of the water system, the performance of the purification process must be proven. This applies to all quality parameters, including the endotoxin content in the WFI. However, such validation is only possible if the initial concentration of a contaminant is known.

Many operations with PW and WFI systems do not measure endotoxins in PW as part of routine sampling. Instead, endotoxins are measured in the WFI, where there are usually values below the detection limit (<0.06 IU/ml). However, some companies consider these random sample measurements in the WFI to be insufficient. If endotoxin values above the detection limit (0.06 IU/ml) but below the pharmacopoeia limit for WFI (0.25 IU/ml) are occasionally found, it is necessary to analyse the cause as part of the trend evaluation. For this reason, some sites have included the measurement of endotoxins in their sampling plans for PW for information purposes.

PW systems with membrane technology generally have endotoxin levels below the pharmacopoeia limit for WFI (0.25 IU/ml). However, PW systems without membrane technology, especially older, poorly flowing deionized water systems (with anion and cation exchangers), can have values significantly above the pharmacopoeia limit.

There are also applications of purified water in the biotechnology sector in which endotoxins can play a role, for example in the fermentation process or during purification. In such cases, purified water should also be tested for endotoxins.

A 'mini-risk analysis' is recommended here to clarify two questions:

• What is the purified water used for? 
• In which cases could endotoxins be problematic?

Based on these answers, the necessity of testing for endotoxins can then be concluded and, if necessary, alarm and limit values as well as specifications for the test interval can be defined.