FAQs: Water for Pharmaceutical and Analytical Purposes

Because of the various uses of these waters, microbial requirements are not included in these monographs since this would unnecessarily burden users with meaningless and/or inconsequential or inappropriate requirements, e.g., water used for many laboratory analyses. Microbial guidelines are provided under the informational chapter Water for Pharmaceutical Purposes <1231> where it states that the user should establish in-house specifications or fitness for use microbial levels above which the water is unsuitable for use.

Alert and Action Levels are process control terms and should be established at levels indicative of the water system trending outside of its normal microbial control range. These levels should be established at levels no higher than, and preferably lower than, those listed in Water for Pharmaceutical Purposes <1231> based on the normal microbial performance trends in your water system. The purpose of Alert and Action Levels is to trigger additional proactive, rather than routine, microbial control measures. These additional control measures should prevent objectionable levels and types of microorganisms from being present in the water, based on the water’s use.

USP is silent on a specific answer to this question. It is understood that some laboratories may not be able to test immediately, or they may have their analyses performed by external laboratories – which may take several days or longer. For this reason, there is no time limit.

In general, you can wait as long as you have qualified the storage time and conditions. It is advised to test as soon as practical for the following reasons: 1) when stored, the water purity only degrades over time due to environmental, ambient, or container factors; and 2) water is typically not produced in batches, but rather it is usually purified, produced, and consumed continuously. The water may have had direct product impact or contact before any lab analysis is executed. Delays in testing only increase the amount of potential product impact – in the event of a failed test.

For lab analyses, samples should be stored in containers that do not adversely impact the test results. This is to prevent false positives and unnecessary investigations. For example, storage of water in a glass container for a few hours is usually good, but storage for a longer time will result in a modest increase in the sample conductivity. This is due to the leaching of sodium silicate from the glass, raising the pH and the water conductivity, and threatening to fail Water Conductivity <645>. In general, clean plastic containers are a better choice for long term storage of samples for Water Conductivity <645> testing. For Total Organic Carbon <643>, there is a similar rationale - many types of non-shedding plastics or glass suffice.

In general, storage at ambient or refrigerated temperatures is best for these chemical tests. Cleanliness of any container is most critical. Due to the very high purity of these waters, fingerprints, soaps, and other residues must be avoided. False positives can result.

Yes. These two chapters specifically state that these tests can be performed off-line or on-line. There are benefits and challenges for each approach, and they are described in more detail in these chapters and in Water for Pharmaceutical Purposes <1231>. In general, on-line testing avoids the risk of contamination of off-line samples by humans, containers, or the environment, and it provides immediate analysis and direct opportunities for real-time control, decision and intervention. For example, with on-line testing you can continuously test and accept the water as passing (for these chemical attributes). With this real time analysis, you can prevent the utilization of the water in the event of a failed test result. Whether using on-line testing or laboratory analyses of grab samples, the tested water must be representative of the water used in production.

For compendial compliance, the water analysis must be representative of the water used in production. Typically, the online instruments are located downstream of the last outlet, and not after any of the purification unit operations within the distribution system. If there are multiple loops in a distribution system, the impact from each loop and sub-loop should be evaluated for compliance.

For process control, on-line instruments could also be connected in additional locations, for example, where water from the purification system enters the holding tank as well as at the beginning of distribution loops, each location with advantages and disadvantages in immediacy of contaminant entry awareness.

No.  The USP general chapters <645> and <643> allow online and offline water testing. On-line analyses can be very useful for real-time process control in addition to compliance with <645> and <643>.  Where multiple water systems and/or distribution system complexity could require many on-line instruments, grab sample testing on a centralized laboratory instrument may offer a more economical choice, but with the process control sacrifice of not having continuous real-time results.

There is a "target limit response" of 500 µg of Carbon/L. The true limit is the response of the TOC measurement system to a 500 µg Carbon/L (prepared from sucrose) solution, Rs, corrected for the response to reagent water, Rw. This limit is equal to Rs – Rw. The actual number will vary based upon your reference standard solution, your equipment, background carbon, etc. USP Reference Standards are required for use.

USP General Chapter <643> intentionally says nothing about how often the system suitability test (SST) should be run. The reasoning is that this frequency depends on the stability of the Total Organic Carbon (TOC) instrument response and other factors associated with the water quality and risk. If the TOC of a quality water system is very low, say <20 ppb, then many opt to reduce the frequency of testing due to less risk. The stability of different TOC measurement technologies may vary over extended periods of time. The instrument manufacturer can advise the user on this matter and user experience can also be valuable in determining a suitable frequency. Another factor is the risk of a non-conforming system suitability test result since the Rs-Rw result used in this calculation is the limit response for the instrument, the crucial pass/fail value for the TOC test. If a non-conforming system suitability test is obtained, it implicates the inaccuracy of all TOC test results since the previous successful system suitability test. For this reason, many users choose to perform the system suitability test more frequently than the stability of the TOC instrument response might suggest, just to minimize the impact of a possibly non-conforming result. This is why a typically low TOC water system is at less risk, even with a failed SST. If the SST fails, some remediation is necessary to re-adjust the instrument, replace a lamp, or some other means of instrument improvement. But even a 50% error will have little impact on the past TOC readings (since the readings, even with this error, are so low relative to the Limit). On a high TOC water system, the failure of the SST is possibly more critical. This is up to the risk the user is willing to assume, knowing the historic stability of their instrument and other factors. Therefore, the Total Organic Carbon <643> is silent on the frequency of performing the system suitability test because it is up to the user to decide what is appropriate.

Where USP is silent on storage conditions and the stability of prepared Total Organic Carbon (TOC) reference standard solutions, the solutions should be 1) prepared fresh or 2) used within the expiry if procured from 3rd party supplier or 3) used within a timeframe determined by stability studies. In all cases, USP Reference Material is specified. Several factors can influence the stability of the reference standard solutions. These include temperature, light, oxygen, microbial decomposition, and adsorption to the container surface. The developments of turbidity, additional color, or performance variability relative to freshly prepared solutions are indicators of instability. Most of the suppliers of solutions specify expiry dates. But as a practical matter, concentrated reference standard solutions of Sucrose last 3-6 months, and analogous solutions of 1, 4 Benzoquinone (pBQ) last about 2 months, assuming they are stored at appropriate temperatures in appropriate containers and protected from light (for pBQ). It is recommended to use refrigeration since this slows down solution degradation, and reduces microbial growth, particularly in the sucrose solution.

In Stage 3, a neutral electrolyte (KCl) is added to increase the ionic strength and accurately measure the pH of the solution. If the ionic strength of the solution is not increased, the pH measurement will be highly unstable and inaccurate. So KCl is added to make a valid pH measurement as a part of the Water Conductivity <645> - Stage 3 test. The increase in the ionic strength is needed so that there is minimal concentration gradient across the pH electrode diaphragm/junction. A large concentration gradient results in a lack of equilibrium and unstable pH response.

There is no need to perform stages 1 and 2 in order. You can go directly to Stage 2 if offline testing in preferred - you do not have to fail stage 1 first.

The cell constant accuracy must be ±2% of the certified value, not the nominal value.

In general, any material that does not impact the conductivity in any appreciable way is suitable. Many plastic containers including PTFE, HDPE, LDPE and some polycarbonates are appropriate. Glass containers for immediate testing are appropriate. Regardless of the material, they have to be clean and free of any cleaning reagents such as soaps. Soaps are very conductive.

Yes, this is correct. There has never been a test for nitrates for USP waters. The heavy metals test on USP waters was deleted in 1996. The pH test was deleted in 1998. [Note - There is a pH measurement (not a test) as a part of Stage 3 test for Water Conductivity <645>, but this is still a conductivity limit test]. Note that you cannot fail the former pH specifications of water (pH 5.0 – 7.0) if you pass the conductivity specifications. You also cannot fail the heavy metals test or the nitrate test if you pass conductivity and your water system starts with water compliant with the requirements for one of the drinking waters specified in the monographs (for the US, EU, Japan, or WHO). In some cases, these tests may be required by other pharmacopoeia.

You may do so, but only under certain circumstances. The microbial quality of the water within the system, as reflected by water from that sample port, may be better than the quality that is delivered to the point of use (POU) during manufacturing use. This is because of microbial contamination of the system water that can occur as it is transferred from the system outlets to the POU. It is the quality of water DELIVERED from the system to the POU that affects products and other uses.

If you have good water use practices such that the microbial count from a sample port is essentially the same as at a POU when delivered by the manufacturing use practice, then the risk of the sample port microbial counts falsely reflecting the quality of the delivered water is low.

Generally, water release for use should be based on a POU sample reflecting manufacturing’s water use practices and not on sample port data.

No. The destination of that water where it will be used for product formulation or cleaning or where it enters a manufacturing process is the true point of use. The quality of water at the true point of use, as delivered by manufacturing (or by a sampling process identical to the manufacturing water delivery process) must be known at all points of use receiving water from the system. The water quality at the true point of use is where the water must be “fit for use”, i.e. pass your water specifications.

Yes. Action Levels in USP <1231> (100cfu/mL for Purified Water and 10cfu/100mL for Water for Injection) are generally considered to represent a level above which the water is unfit for use. That is why an OOS investigation must be undertaken if those Action Levels are exceeded.

So whether you declare microbial specifications or not, they are assumed to be those “compendia action level” values contained in General Chapter <1231>. To avoid ever exceeding a water microbial specification, trend-based Alert and Action Levels should be used to monitor and control the water system so it always produces water that is fit for use.

1. One common problem is where there is a cold WFI sub-loop off of a heated system with a large shell and tube heat exchangers used for cooling in that sub-loop. When the sub-loop is hot water sanitized, not enough contact time is allowed for the cooling heat exchangers (and their trapped chilled water) to get them thoroughly hot and sanitized. When incompletely sanitized, any surviving biofilm will immediately reinoculate the cold sub-loop after resumed cold operation and be present as detectable micro counts.
2. Other common problems with cold WFI systems are dead legs, sometimes temporary ones that are created by open hard-piped connections to equipment that is not in use and not drawing water. The hot water during sanitization doesn’t mix well with the trapped water in that dead leg, so the dead leg never gets sanitized. If there was any contamination that got into that side leg during previous use, it will grow unabated in the unsanitized dead leg and continuously contaminate the loop water.
3. Another common problem is overwhelming the distillation purification process with a high level of endotoxin in the water going to the still (100+ EU/mL). This can happen with poor maintenance of pretreatment unit ops such as carbon beds, and also when coincident with high endotoxin levels in the city water when they switch over to straight chlorine from chloramine for a part of a year.

It would not be surprising if substantial biofilm were allowed to be present from infrequently used chemical sanitants. However, if hot water is used for sanitization, it would denature the nuclease enzymes, so this phenomenon might not occur with hot water sanitized systems.

Yes. A temperature of 80˚C is very “forgiving” of cooler locations which can still be sanitized even with a 10-15˚C temperature loss as it penetrates throughout the system by convection and conduction, so it is very effective. Cooler temperatures (down to 65˚C) can also be used but is “unforgiving” of yet cooler locations such as outlet valves off of the main loop.  So such cooler locations must be flushed with this slightly cooler hot water in order to assure that all surfaces reach sanitizing temperatures greater than 60˚C. Unless systems are specifically designed for this, temperatures hotter than 80˚C can impact the longevity of system materials (e.g. gaskets and diaphragms). A temperature of 80˚C is well hot enough to kill the most heat resistant biofilm organisms that will colonize a water system (D value of about 5 milliseconds).

1. If the sampling is for QC “release” of the water for manufacturing use, then the outlet used by manufacturing must be sampled in EXACTLY the same fashion as it is used by manufacturing – same outlet sanitization (if any), same manufacturing hose (no matter how grungy or poorly maintained), same pre-flushing (if any), same everything. The purpose of the sample data is to duplicate the same quality of water that manufacturing is using, so you have to duplicate in sample collection how the water is drawn from the system for use. Those procedures of water use can significantly contaminate pristine water within a water system when it exits, so that “nasty” water is delivered to a manufacturing operation. If you sample the water differently (better) than it is used by manufacturing, you will get lower (better) micro counts that are not representative of the water quality that is actually be used. Sampling like manufacturing water use for QC release is required by FDA to be identical. If it is not, this could earn you an FDA483 observation or worse.
2. If the water is being sampled for process control (PC) for the purpose of water system monitoring and systemic microbial control, it might be done through sampling ports that are not used by manufacturing. Since we know that the outlets themselves can contribute to the bioburden of the collected water, extreme efforts can be used to assure that the outlet does not add to the microbial content of the water as it exits the system (using extreme outlet sanitization, very vigorous and thorough flushing, sterile hoses, etc.). For PC, you are interested in the quality of the water within the system behind the valve and do not want contamination in a sampling port to bias the interpretation of the data.
3. However, water collected from sampling ports (rather than manufacturing use outlets) usually cannot be used for final release (QC) of water since it is not collected in the manner it is actually used. Manufacturing does not generally use water drawn from sampling ports.

1. Endotoxin levels are typically a concern only for WFI systems. Most WFI systems are sanitized by elevated temperatures (hot water is better than steam since no special engineering is needed for hot water sanitization and it is plenty adequate), though more may employ ozone in the coming years as ambient non-distillation purification technologies become more widespread with EP’s relaxation of the methods of preparing WFI in their WFI monograph. Since thermal or ozone sanitization of WFI systems is typically no less frequent than weekly, that is not enough time for biofilm (with its endotoxin) to develop in the system and be released by periodic sanitization.  If the systems are much less frequently sanitized, there is a chance that developing biofilm could release detectable endotoxin when killed by periodic sanitization.
2. If chemical sanitizers other than ozone are used (this would be very atypical for a WFI system or an endotoxin-controlled Purified Water system), the sanitizer would have to be rinsed out, which would also rinse out any released endotoxin.

a. If you do not have a balance to accurately weigh the low mass, prepare a solution with a higher concentration that provides the correct solution accuracy. Then perform an analytical dilution to the desired concentration for executing the test method.

b. If preparing a concentrated solution to dilute, be aware of the solubility of the reference standard to ensure that the solid will completely dissolve in the concentrated solution.

a. Preparation of the reference solutions must be performed to achieve the accuracy as indicated by the significant digits in the test method, that is, 0.50 mg/L of carbon.

b. The solution should be prepared to an accuracy of +/- 0.005 mg/L of carbon

c. Use a combination of appropriate analytical balances and volumetric glassware to achieve the solution accuracy.

Organic extractable components from the packaging that contribute to the TOC profile of the sterile packaged water should have been identified, quantified, and evaluated for safety/toxicity during packaging development activities in order for the packaging to have been approved by FDA. The methodologies used at that time could be the basis for any identification, quantification, and safety/toxicity studies needed for showing current compliance of sterile water product batches with their respective USP monographs. Additional guidance is available in General Chapters <661> and its sub-chapters, <1663> and <1664>.

If unknown organic impurities are also present, these will have to be evaluated using analytical methods most suitable for the determination.

Note Some factors that may cause high TOC can include the packaging system components and packaging process controls. The composition/concentrations of the specific packaging components used in the packaging of the water under investigation may be subtly different than what was originally qualified (e.g., resin differences from alternate suppliers). Also, the controls associated with the packaging/fabrication process (e.g., adhesives and mold release compounds) may have changed since original packaging approval resulting in unknown organic impurities.

The “Nominal Container Volume,” as stated in <643>, refers to the intended that the container is intended to hold.

If a container is intended to hold 5 mL and has an actual fill of 3 mL or a fill of 5.2 mL (to accommodate extractable volume), the Nominal Container Volume is considered 5 mL, and L1 is 32…

If a container is intended to hold 10 mL and has an actual fill of 5 mL, the Nominal Container Volume is considered 10 mL, and L1 is 24

Purified Water and Water for Injection monographs state that the waters “contain no added substance”. In <1231> Water for Pharmaceutical Purposes Section 3, “no added substances” is intended to mean “no added substances that aren't sufficiently removed”.  Companies are required to reduce the ozone level below a limit of detection prior to use. Every company may perform its individual risk assessment on the use, reduction, and detection of the ozonation process and take action if deemed appropriate.