Accurate assessment of water quality for improved process control

By Michelle Stafford, Jaimie Russo, and Serge Ohresser, Millipore Corporation

Water is indisputably a critical element of many high-technology manufacturing processes. From rinsing devices and manufacturing lines to final product composition, pure water is a major component in the production of pharmaceuticals, beverages and microelectronics. As a result, bacterial contamination in purified water can compromise the quality of virtually any final product. The significant impact of water purity is further supported by the strict water standards established by numerous regulatory agencies including the Environmental Protection Agency, the United States Food and Drug Administration, the National Committee for Clinical Laboratory Standards and the American Society for Testing and Materials. Each regulation provides water quality standards and testing methodologies. Although all water systems are validated and routinely tested, significant amounts of time and money are still lost each year investigating contamination issues. Ultimately, the production downtime and the resulting financial impact can be minimized, if not eliminated, with effective monitoring practices and decontamination procedures. Accurate, regular assessment of the microbial quality of process water is essential for timely implementation of corrective actions.

Due to the significant economic ramifications of microbiological contamination, validation engineers and quality control specialists regularly test drops, tanks and baths of process-water loops to develop trending data as well as to identify as early as possible any spikes in bacterial levels. Currently, both the pharmaceutical and microelectronic industry-monitoring practices rely upon routine testing of small volumes, collecting 100 mL water samples from various points in the distribution system on a weekly rotation. These water samples are processed by membrane filtration and subsequently incubated on a validated culture media to detect the highest level of microorganisms in a specific sample. This accepted method produces results in three to five days due to the need for an extended incubation period, as the organisms typically found in high-purity water systems are stressed, slow-growing strains. To ensure the quality of pharmaceutical water, an action level for Water for Injection (WFI) and highly purified water is set at 10 CFU/100 mL by both the United States Pharmacopeia and the European Pharmacopeia (EP). The EP further stipulates that, for aseptic processing, stricter alert limits may need to be applied.

Low levels of bacteria and irregular contamination intervals can complicate the testing regime and compromise results. According to a recent report, the daily monitoring of a WFI system storage tank demonstrated a maximum total count of microorganisms to be 3 CFU/100 mL.1 Microbiologists will agree that this specification is often within the microbiological “margin of error.” Furthermore, it is estimated that 99 percent of bacteria in a high-purity water system are likely to exist as biofilm attached to internal surfaces.2 Biofilm is fundamentally a chronic microbial contamination that is difficult to control using conventional heat, mechanical or chemical treatment procedures. Today, little is known about the behavior of mature biofilm, including detachment and movement over solid surfaces. It is generally accepted that biofilm exists, yet the predicament is finding it early enough to eradicate.

Figure 1. This figure represents the number of P. aeruginosa colonies detected after sampling 1 L (light blue) or 100 mL (red) from the system.
Click here to enlarge image

A recent study was conducted to develop a test methodology to better detect low levels of microorganisms as well as a sporadic biofilm contamination in a purified water sample. In order to examine the contamination issues which plague purified water systems, a pilot-scale water system was contaminated with low levels (10 to 100 CFU/L) of planktonic bacteria. When sampling in accordance with water guideline recommendations, it was found that the standard, 100 mL test samples frequently did not detect the underlying contamination due to a limited sample volume (see Fig. 1). This study found that 100 mL samples failed to detect common water-system contaminants as much as 75 percent of the time. This could result in undetected contamination that could negatively impact process yields and final product quality. Given the fact that typical high-purity water system total-bacterial action and alert limits are generally less than 1 CFU/mL, statistical considerations dictate a minimum sample volume in the liter range versus milliliters. Therefore, the risk of false negative test results can be minimized by simply increasing the sample volume from 100 mL to one liter. An increased sample volume provides more accurate and statistically significant bacterial counts by improving the variability observed with low microbiological counts.

The simulated biofilm was also undetected with the lower volume sampling methodology. After four days of testing, each of the 100 mL samples provided negative results while the one liter samples clearly demonstrated a random release of bacteria from the biofilm (see Fig. 2). This study demonstrates that cells can be detached heterogeneously and randomly from a mature biofilm. Consequently, increasing both the sample volume and the frequency of sampling significantly increases the probability of detecting random detachment of bacteria from a biofilm.

Figure 2. Positive results observed after Day 4 can be attributed to biofilm detachment.
Click here to enlarge image

Data presented here from the two highlighted experiments strongly support increasing the volume of water from 100 mL to at least one liter as one improvement toward detection of biofilm and low-level microbial contamination. When membrane filtration is utilized, the additional sample volume does not impact final testing cost as one test device is utilized for both 100 mL and one-liter testing. This study clearly demonstrates that there are significant benefits to increasing the sample volume.

One could easily envision that another enhancement for testing purified water can be realized by employing rapid detection technologies along with existing test parameters. Microorganisms found in water systems are usually in a dormant metabolic state. When using traditional growth-based methods, this dormant metabolic state can result in a longer incubation period and, therefore, longer detection time. Furthermore, microorganisms present in water samples are commonly a mixed population where some organisms grow faster than others, limiting incubation time to the amount of time required by the slowest-growing organism. Rapid detection methods can reduce the wait for microbial results from days to hours when sampling mixed populations of stressed microorganisms.

Commercial rapid microbiology test systems that employ adenosine triphosphate (ATP) bioluminescence technology are available as a marker for cell viability. ATP is detected by a reaction that utilizes a luciferase enzyme and a luciferin substrate to produce light. Photons of light, emitted from bacterial cells, correlate to bacterial colony-forming units (CFUs) that are undetectable to the unaided eye. Using one such ATP bioluminescence-based rapid-detection instrument, a methodology was developed to rapidly detect mixed microbial populations of stressed microorganisms in high-purity water samples collected by using membrane filtration. Samples from an industrial water system were tested using rapid technology in conjunction with membrane filtration versus traditional membrane filtration alone. The membranes were plated on R2A medium at 30°C for both methods. By using membrane filtration in conjunction with the rapid system, detection and enumeration of the stressed microorganisms was reduced to 27.5 hours versus the 6 days required by the traditional method-a decrease in time by a factor of five. Furthermore, the faster-growing organisms did not interfere with the enumeration of the slower-growing organisms, as can commonly occur with traditional methods when fast growers overtake slow growers on a plate, thus obscuring smaller colonies and contributing to inaccurate microbial counts. Percent recovery of the microorganisms using the rapid method versus the traditional method was greater than 90 percent (see Table 1).

Table 1. Incubation time using rapid microbiology detection system at 30°C.
Click here to enlarge image

Significant gains in production efficiencies are obvious if water-test methodologies can be modified to include either increased sample volume or rapid microbial detection technology. The ability to capture low levels of contamination and biofilm improves along with the capacity to enumerate what was captured more quickly than traditional methods allow. Instead of six-day incubation delays, overnight results become possible. The capability to quickly and accurately detect and enumerate contamination in industrial water systems provides for better process control, higher product yields and faster time to market. Additionally, early knowledge of contamination can be essential in trend analysis and implementation of corrective actions, saving both time and money.

Michelle Stafford is the group technology manager for process monitoring tools for the Bioprocess division of Millipore Corporation in Billerica, Mass. She can be reached at [email protected]

Jaimie Russo is the rapid microbiology technology manager for the Bioprocess division of Millipore Corporation.

Serge Ohresser is the R&D application manager for the Bioprocess division of Millipore Corporation.


Millipore would like to thank Marie Pressel, R&D scientist, and Marilyn Romieux, R&D scientist, for their contribution to this research.


  1. Jahnke, M. “One-Way Distribution System for Water for Injection: Process Management, Microbiological Quality Control, and Meeting Regulatory Requirements.” PDA Journal of Pharmaceutical Science &Technology. Vol. 55 (1): 3-9. Jan/Feb 2001.
  2. VanHaecke, E., J-P Remon, M. Moors, F. Raes, D. DeRudder, and A. VanPeteghem. “Kinetics of Pseudomonas aeruginosa Adhesion to 304 and 316-L Stainless Steel: Role of Cell Surface Hydrophobicity.” Applied and Environmental Microbiology. Vol. 56(3), pp. 788-795. March 1990.
  3. “Detecting Low Concentrations of Microbiological Contamination in Pharmaceutical Grade Water Using Millipore’s MicropreSure® On-line Filtration Samplers.” Millipore Lit. No. AN0002EN00, 2005.


Easily post a comment below using your Linkedin, Twitter, Google or Facebook account. Comments won't automatically be posted to your social media accounts unless you select to share.