One of the myths is that provided the pressure does not become negative, the considered opinion is that the possibility of microbial intrusion is very low and with some opinion it is seen as impossible. However, experimental data drawn from various industries challenges this logic and demonstrates that a considerable risk of microbial contamination exists during low pressure events.
Hence, periodic leak assessments and verification of pressures of pipes, tubing and containers both represent critical quality attributes.
This article assesses this research and unpicks the key factors that need to be considered when undertaking an assessment.
An important objective with pharmaceutical processing is to exclude environmental microorganisms from the process. There are different areas where microbial contamination can arise and be transferred into the product stream, ranging from raw materials with high bio burdens and specific pathogens to operator interventions in relation to open product. Another important area is through leaks to pipes, vessels, tubing, and connectors, either for holding or transferring product, or as part of the supply of pharmaceutical grade water.
Even if a leak occurs, as through a faulty valve, a connector that does not seal, or structural damage to pipes or tubing, under conditions of a sufficiently high pressure microorganisms cannot move against the pressure gradient. Under conditions of negative pressure, this causes, in effect, a suction effect and a contamination risk must be assumed. However, the risk is not a binary one of positive and negative pressure. Contrary to some perspectives, this does not require negative pressure to occur, for low-pressure conditions can also pose a risk and therefore a risk-based decision is required (and in the case of any sterile product, post-final filter, the outcome of such a risk assessment may well be batch rejection).
Defining low pressure is not straightforward. There will be a point when a system becomes vulnerable to microbial incursion. What represents low is based on the system design. For example, with a water distribution system, low is around 12 kPa (0.1 ‘Bar’). For U.S. drinking water supplies (Virginia Department of Health), low pressure and the risk of contamination is set at not less than 13 kPai. Below these levels, fluids will always flow from high to low pressures, thus when the pressure within a pipe drops below the surrounding pressure, a potential driving force for contaminant ingress exists.
In terms of more specific factors, should microbial contamination be present, then the factors that can affect intrusion include:
- Pressure changes around the crack area.
- The surrounding environmental conditions.
- The re-intrusion of any liquid that was lost during the leak process.
- The nature of the flow through the orifice (crack) and its driving force.
- The overall flow rate through the pipe or tube (risks are higher under conditions of a low Reynolds numberii).
- The type of microorganism.
The physical factors are discussed in more detail in the section below.
With microbial types, those bacteria that are capable of motility are more likely to be able to intrude than microorganisms that are non-motile. Motile bacteria use appendages for propulsion. The appendage could be a relatively stiff helix that is rotated by a motor embedded in the cell wall, as in the case of Escherichia coli, or it could be a flexible filament undergoing whip-like motions due to the action of molecular motors distributed along the length of the filamentiii. Examples here are peritrichous flagella found all around the periphery of the bacterial cell, as with many members of the Enterobacteriaceae family.
The concern with motile bacteria is that they can move against a slightly imbalanced pressure gradient provided that the flow is not too high (that is the flow will, at some point, be too great for the swimming speed of the bacteria to overcome). However, some bacteria can move against flow velocities as high as 13-fold above the cell swimming speed, as was observed with Pseudomonas aeruginosa in one studyiv where the bacterial cell swimming speed was recorded as 45 µm s−1 against a flow of 600 µm s−1.
The ability of bacterial cells to swim against a given flow also varies according to bacterial elongation and swimming traits (speed and tumbling rate, both which are phenotypic characteristicsv). In addition, motile bacteria become trapped close to flat surfaces, inducing a strong concentration of cells close to the area of the leak, with vells more likely orientated in a leeward direction. This is also influenced by the curvature of the surface. Where local flow is disrupted, the interplay between local flow and bacterial motility affects both the attachment rate and the attachment site of bacteria, due to the deflection of the trajectories of swimming bacteria by the flow, and that this effect strongly depends on the magnitude of the flowvi.
The ingress of organisms is aided by the presence of liquid coming from the area of damage. It has also been shown experimentally that there is a direct correlation between having the liquid in the defect pathway and occurrence of microbial growthvii. Here, the liquid seems to act as a transport mechanism for the bacteria to enter into apparatus (like a single-use system). Hence, the location of the defect in terms of liquid inside the system also influences the likelihood or otherwise of bacterial intrusion. A related mechanism that can lead to contamination is where liquid leaves the system, mixes with the contaminant and then the solution is absorbed back into the system, should the pressure fluctuate between higher and lower pressures.
Factors influencing microbial incursion
Arguably the most deterministic factor affecting the likelihood of microbial intrusion, assuming a defect is present, is pressure, for there will be a theoretical pressure for any system which is positive (and sufficiently high) to make a microbial incursion impossible. It also stands that any negative pressure will draw in any liquids and microorganisms in the immediate area of the defect. In-between these two poles are variations during which intrusion may or may not occur. Studies of drinking water distribution systems indicate that pressure does not need to be negative in order for microorganisms to enter the fluid pathway. Therefore, during low-pressure events, microorganisms can enter liquid through pipeline leaksviii.
Pressure in water distribution systems is seldom steady for hydraulic pressure transients to occur(meaning circumstance in which pressure is changing with time)ix. At times these may oscillate quite rapidlyx. A pressure gradient is a physical quantity that describes the direction and rate that a pressure increases most rapidly around a particular location (expressed as Pascals per metre). This effect typically occurs in the form of a wave, rising and decreasing over time. During transient phases, the pressure will drop sometimes to levels that would make microbial intrusion more likely should a hole or other damage be present.
For water systems, the main determinant of pressure transients is alterations to flow. Any change in flow in a pipe (such as due to valve closure, pipe fracture, or pump stoppage) will result in an exchange of energy between flow and pressure and pressure variation. Kirmeyer and colleagues showed that for every 0.3 m/sec of velocity forced to a sudden stop, water pressures increase by 345 to 414 kPa, depending on the pipe materials, topography, and so on. Conversely, for a sudden velocity increase, there is an instantaneous low or negative pressurexi. For example, if a pipe is under static pressure and a valve is suddenly opened, a depressurization wave propagates through the system. In certain areas of a pipe network (a dead end) transient pressure waves superimpose. This can lead to momentary depressurisation and negative pressuresxii . As a second example, when a valve is closed instantaneously, water will decelerate, and a transient pressure wave will travel upstream and downstream from the valve and through the pipe or tubing. The wave continues until the kinetic energy is dissipated by friction. Therefore, the more frequently that valves are opened and closed, the greater the transient pressure.
As well as velocity, the magnitude of the pressure change is also influenced by the materials of construction and pipe characteristics. The roughness of the inside of the pipe is also influential on pressure. Rougher pipes have higher resistance factors (roughness coefficients), which means that the water flowing through the rough pipes loses energy more quickly than it would if flowing through a smooth pipe.
For pharmaceutical water systems, where power surges, excessive opening and closing of valves, and power cuts occur, these increase system vulnerabilities. However, the most significant risks emerge following repairs and maintenance.
Where the pressure is unknown the size of the leak may provide clues since leakage is lower under conditions of a lower pressure and, conversely, greater under conditions of higher pressure. However, this is only of relevance if the leak is visible. At the same time, lower pressure presents a greater chance of intrusion therefore the presence of a ‘small leak’ means that the system is potentially at a greater risk (with other factors to be considered)xiii.
Size of defect to tubing or pipework
The larger the hole, the greater the risk of ingress. Experiments have assessed this at a steady pressure (30 kPa). The experimental outcomes have been expressed in terms of probability, with a significant shift occurring between 2 and 3 micronsxiv:
- 0 to 2 µm = 5% probability of ingress
- 3 to 5 µm = 20% probability of ingress
- 6 to 10 µm = 60% probability of ingress
- 11 to 15µm = 80% probability of ingress
- 16 to 20µm = 90% probability of ingress
- >21µm = 100% probability of ingress
The studies used an aerosolization challenge containing 106 CFU/cm2 of Bacillus atrophaeus, a far greater challenge than would be found in any pharmaceutical environment. It is important to note that this organism is non-motile. In terms of each of the hole sizes presented in the above data, the smallest size the human eye can actually see ranges anywhere from 50 to 400µm. Hence, under optimal conditions any leak area that presents a risk is not detectable with unaided vision.
Ingress of bacteria was also assessed by pressuring the test assembly at other ranges, from 5 to 100 kPa. The risk for 5 kPa was a factor of two greater than for 30 KPa. This demonstrates that the risk of ingress is both size and pressure dependent.
The shape of the hole has some influence on the time taken for microbial intrusion (albeit that any time required is very short). The shape also influences the expansion rate under pressure with round holes showing the smallest expansion with pressure, followed by circumferential cracks and then longitudinal cracksxv. The material type will also influence this, with the concerns being greater with certain plastics material strain (such as pressurized visco elastic pipes).
Time is an important factor, as with the minimum time required for the any contaminant to begin entering the system, although it is not as important as the defect size or pressure since if contamination is present the time needed for intrusion is very short (some studies infer ingress is immediate; other studies show an increase in time relative to rising pressure albeit at no more than 3 seconds per 100 kPaxvi). While based, experimentally, on industrial systems what is interesting about these pressure-time relationships is that the pressures are similar to the pressure parameters used for sterile filtration activities using single-use systemsxvii xviii.
Hence, although time and pressure are proportional, in that faster initial intrusion times occur under low operating pressure(and where the volume of intrusion is connected to the duration and magnitude of the low- or negative pressure transient)xix, if microbial numbers are high any event that creates the physical conditions for microbial intrusion presents an immediate risk in the pharmaceutical context for a single-use system (there will be some variation of risk for a large water system).
A further noteworthy factor of time is that minimum time required for the contaminant to intrude into the system is not affected by how long the system was running before the leak or low pressure event occurs, provided that the pressure is sufficiently low for intrusion to occurxx.
How likely are defects to occur?
The discussion so far is predicated by defects being present. The larger the system and the more maintenance required then the risk of defects is probably greater on an operational basis, which puts water systems at greater risk. However, single-use systems, if they have not been subject to quality-by-design principles during their development and rigorous testing, can also suffer with leaks.
Water systems are more likely to see defects along pipework, from pin holes, cracks, and corrosion clusters; and from and joint/connection weaknessesxxi. Whether intrusion occurs is based on the factors presented above. It may be that a leak flushes contaminants away from the leakxxii or it may be that short duration, oscillating (but extreme) transient pressure events result in contaminant ingressxxiii. Once in the system, progress of the contaminant within the network depends mainly on the velocity of flow, making parts of the pharmaceutical process with higher water consumption more exposed to the organismsxxiv.
One of the factors that causes pressure transients is from the rapid closure of an upstream valvexxv xxvi. To this can be added the sudden shutdown of pumps, power outages, and start-up testsxxvii
With single-use systems many vulnerabilities occur around connections (such as joining two devices together for fluid flow or adding a sampling or mixing bag) or where moulding has occurred, linking separately manufactured components together. Tearing and punctures can occur through the poor manufacturing and the poor handling of systems.
Manufacturers of single-use systems will have assessed maximum allowable leakage limits by subjecting systems to various integrity testsxxviii. However, these may not have captured every factor of concern or provide the level of sensitivity required to assess a barrier against microbial ingress or liquid loss.
For water systems, preventative maintenance is essential. As well as maintaining a water system for water flow (velocity and direction), engineers should also seek to detect leaks and to predict where leaks may occur and undertake actions to address these. To assess the potential for pharmaceutical pipework to leak, a series of calculations based on the size of the pipe, the speed of the pressure wave, and the distance it has to travel, are requiredxxix. Where a leak is suspected or observed, to determine where precisely a leak has occurred requires assessment, relating to identifying the effect of the leak on the measured pressure response. To aid leak detection, traditional acoustic methods (such as listening rods and noise correlators) can accurately pinpoint a leak. However, such tools take time and patient to deploy. To overcome these limitations, deep learning can help to strength a transient frequency response based leak identification framework.xxx xxxi
Actions can also be taken to ensure leaks do not occur. The likelihood of a leak from a water distribution system is partly the product of choice of water pipe material and the specific material grade (such as PVC or stainless steel; with plastics, the time-history of the stresses in the pipe is also a determinant). Material choices can greatly affect the potential for pipe deterioration and resistance to stresses which consequently influences the potential for cracks and leaks. Related to this are the quality of welds and the possibility of joint/connection failures.
For single use systems, testing can be performed before and after application. The most commonly used non-destructive test methods for checking a SUS for integrity are:
• Pressure decay: The system is inflated with gas at a given pressure and the drop of the pressure is measured after a given time.
• Flow measurement: The system is inflated with gas at a given pressure. The flow required to keep this pressure constant is measured over time.
• Gas tracer method: The system is placed into a vacuum chamber and connected to the gas tracer (typically helium for SUS) line. The gas leaking out of the system is detected and quantified.
Should a leak be observed in operation, the significance will depend on the process stage and whether opportunities for microbial removal exist. For aseptic processing, post-filtration, there are unlikely to be any circumstances where a leak can be justified since the risk of microbial intrusion is too great unless a very high pressure was maintained and even here, in the absence of any conclusive microbial test, caution is advised. Pre-filter, an assessment could be made with considerable caution. This could only be attempted with a thorough understanding of the factors discussed below.
The information presented in this review suggests that the probability of bacterial ingress is a product of some or all of the following:
- A pathway: leakage or breakage:
o Defect size, as expressed in micrometres (risk becomes high for 2 µm and greater),
o Fluid flows within the system (the risk decreases >600 µm s−1).
- A driving force: low or negative pressure within the system:
o Pressure relative to the surrounding environment (risks exist from 50 to 70 kPa). A significantly high pressure is somewhere above >100 kPa,
o The differential pressure across the defect,
o Small leaks may be signals of low pressure and therefore greater risks than larger leaks.
- Presence of liquid from the defect:
o Liquid from the defect providing a bridge for organisms to cross,
o External liquid sucked back into the system due to oscillating pressures,
o For conditions typical of prefinal filtration, bacterial dell intrusion will occur within three seconds under favourable pressure.
- The availability of a microbial contaminant source nearby:
o Size and shape of organisms,
o Whether organisms are capable of motility,
o Time that organisms will have spent within the fluid path and the potential for growth.
In addition, for single-use systems, process time needs to be considered. Should microorganisms have entered in low numbers pre-filtration, should conditions exist that would enable the microorganisms to increase in number then time to complete the process is also a determinant. Given the doubling rates of many bacterial species, a time of more than two hours presents a far higher risk of an excessive bacterial challenge to the filter.
This review paper has looked at the risk of microbial intrusion into tubes, pipework, vessels and other systems used for the transportation or holding or liquids – be that pharmaceutical grade water or product in a single-use system. With most areas of pharmaceutical microbiology, risks can only be presented in terms of probability and due to the probabilistic nature of the microorganisms not every liquid leak leads to microbial ingress. However, there are a set of additive risk factors that will increase the likelihood of microbial ingress occurringxxxii.
The regular assessment of pharmaceutical water systems for maintenance of pressure and likelihood of leaks is important, such as the use of electronic pressure devices situation in key locations. Such data should be reacted to should the pressure drop below the threshold where microbial ingress could occur. For single-use systems, integrity testing is of great value supported by any detection by operators. The risks depend on the process stage, with the risk level for aseptic processing in relation to final filtration being greatest (pre-filter) and unacceptable under any conceivable scenario post-filter.
iAnon. Pressure in the Distribution System. At: https://water.mecc.edu/concepts/pressuredistr.html
iiSecchi, E., Vitale, A., Miño, G.L. et al. The effect of flow on swimming bacteria controls the initial colonization of curved surfaces. Nat Commun 11, 2851 (2020). https://doi.org/10.1038/s41467-020-16620-y
iiiBerg H C 2009 Bacterial flagellar motor Curr. Biol. 18 R689–91
ivShen, Y., Siryaporn, A., Lecuyer, S., Gitai, Z. & Stone, H. A. Flow directs surface-attached bacteria to twitch upstream. Biophys. J. 103, 146–151 (2012)
vRusconi, R., Guasto, J. S. & Stocker, R. Bacterial transport suppressed by fluid shear. Nat. Phys. 10, 2–7 (2014)
viTaylor, R., Cronin, A., Pedley, S., Barker, J. & Atkinson, T. The implications of groundwater velocity variations on microbial transport and wellhead protection—Review of field evidence. FEMS Microbiol. Ecol. 49, 17–26 (2004).
viiAliaskarisohi, S.; Kumar, C.; Hogreve, M.; Montenay, N.; Cutting, J.; Mundrigi, A.; Paramathma, A. Single-Use System Integrity II: Characterization of Liquid Leakage Mechanisms. PDA Journal of Pharmaceutical Science and Technology, May 2021, 75 (3), 273-288
viiiPotential for pathogen intrusion during pressure transients, Journal of the American Water Works Association, Mohammad R. Karim, MortezaAbbaszadegan, Mark Lechevallier. First published: 01 May 2003. https://doi.org/10.1002/j.1551-8833.2003.tb10368.x
ix Savalan Pour Akbarkhiavi, MonzurAlamImteaz, A novel ‘pressure index’ for predicting number of pipe bursts in water distribution system, Proceedings of the Institution of Civil Engineers – Water Management, 10.1680/jwama.20.00076, 174, 6, (278-290), (2021).
xFox, S.; Shepherd, W.; Collins, R.; Boxall, J. Experimental quantification of contaminant ingress into a buried leaking pipe during transient events. J. Hydraul. Eng. 2016, 142, 04015036
xiKirmeyer, G. J., Friedman, M., Martel, K., Howie, D., LeChevallier, M., Abbaszadegan, M., Karim, M., Funk, J. & Harbour, J. 2001. Pathogen Intrusion into the Distribution System. AWWA. Research Foundation and the American Water Works Association, Denver, CO
xiiIlaria Butera, J. Jaime Gómez-Hernández, Silvia Nicotra, Contaminant-Source Detection in a Water Distribution System Using the Ensemble Kalman Filter, Journal of Water Resources Planning and Management, 10.1061/(ASCE)WR.1943-5452.0001383, 147, 7, (2021).
xiiiPressure vs. Flow Relationship for Pipe Leaks,May 2009. DOI: 10.1061/41036(342)10. Conference: World Environmental and Water Resources Congress 2009.
xivAliaskarisohi, S.; Hogreve, M.; Langlois, C.; Cutting, J.; Barbaroux, M.; Cappia, J. M.; Menier, M. C. Single-Use System Integrity I: Using a Microbial Ingress Test. PDA Journal of Pharmaceutical Science and Technology, September 2019, 73 (5), 459- 469
xvA. M. Cassa, J. E. van Zyl & R. F. Laubscher (2010) A numerical investigation into the effect of pressure on holes and cracks in water supply pipes, Urban Water Journal, 7:2, 109-120, DOI: 10.1080/15730620903447613
xviMahmoud, Mohammed, Ashraf Farahat, Mohamed A. Hamouda, Muhammad Al-Zahrani, Muhammad Nadeem Sharif, and Rehan Sadiq. 2019. “Experimental Investigations and CFD Modeling of Contaminant Intrusion in a Water Network during Transient Events” Water 11, no. 7: 1425. https://doi.org/10.3390/w11071425
xviiAllmendinger A, Mueller R, Huwyler J, Mahler HC, Fischer S. Sterile Filtration of Highly Concentrated Protein Formulations: Impact of Protein Concentration, Formulation Composition, and Filter Material. J Pharm Sci. 2015 Oct;104(10):3319-29. doi: 10.1002/jps.24561. Epub 2015 Jul 6. PMID: 2614974
xviiiMaynard, J., Meltzer, T. and Jornitz, M.(2005) Sterilizing Filters: Right Flow, Right Size Critical, Pharmaceutical Manufacturing: https://www.pharmamanufacturing.com/articles/2005/341/
xixJones, S.; Shepherd, W.; Collins, R.; Boxall, J. Experimental quantification of intrusion due to transients in distribution systems. Procedia Eng. 2014, 89, 1306–1313
xxLeChevallier, M.W.; Gullick, R.W.; Karim, M.R.; Friedman, M.; Funk, J.E. The potential for health risks from intrusion of contmainants into the distribution system from pressure transients. J. Water Health 2003, 1, 3–14
xxiClayton, C., and van Zyl, J. (2007). “The effect of pressure on leakage in water distribution systems.” Proc. ICE Water Manage., 160(2),109–114
xxiiS. Fox, W. Shepherd, R. Collins, J. Boxall, Experimental Proof of Contaminant Ingress into a Leaking Pipe During a Transient Event, Procedia Engineering, Volume 70,2014,Pages 668-677, ISSN 1877-7058, https://doi.org/10.1016/j.proeng.2014.02.073
xxiiiImproved Model for Contaminant Intrusion Induced by Negative Pressure Events in Water Distribution Systems.Yan Yang, Ph.D.; David Z. Zhu, M.ASCE; Tuqiao Zhang, Ph.D.; Weichao Liu, Ph.D.; and Shuai Guo, Ph.D.
xxivIslam, N.; Farahat, A.; Al-Zahrani, M.A.M.; Rodriguez, M.J.; Sadiq, R. Contaminant intrusion in water distribution networks: Review and proposal of an integrated model for decision making. Environ. Rev. 2015, 23, 337–352
xxvEbacher, G., Besner, M. C., Clément, B., and Prévost, M. (2012). “Sensitivity analysis of some critical factors affecting simulated intrusion volumes during a low pressure transient event in a full-scale water distribution system.” Water Res., 46(13), 4017–4030
xxviFox, S., Shepherd, W., Collins, R. et al. (1 more author) (2015) Experimental quantification of contaminant ingress into a buried leaking pipe during transient events. Journal of Hydraulic Engineering. ISSN 1943-7900
xxviiCollins, R., Fox, S., Beck, S., Saul, A., and Boxall, J. (2012b). “Intrusion and leakage through cracks and slits in plastic (MDPE) pipes.” Proc., 14th Water Distribution Systems Analysis Conf., Engineers Australia, Barton, ACT, Australia.
xxviiiUSP<1207> Package Integrity Evaluation: Sterile Products. United States Pharmacopeial Convention, Inc. Rockville, MD, 1 May 2018
xxixOn site leak location in a pipe network by cepstrum analysis of pressure transients, J.D Shucksmith, J.B.Boxall, W.J.Staszewski, A.Seth and S.B.M.Beck, Journal – American Water Works Association https://doi.org/10.5942/jawwa.2012.104.0108
xxxZiyuan Liao, Hexiang Yan, Zhenheng Tang, Xiaowen Chu, Tao Tao,Deep learning identifies leak in water pipeline system using transient frequency response, Process Safety and Environmental Protection,Volume 155, 2021,Pages 355-365,
xxxiYong Deng, Wen Jiang, Rehan Sadiq,Modeling contaminant intrusion in water distribution networks: A new similarity-based DST method, Expert Systems with Applications,Volume 38, Issue 1,2011,Pages 571-578,ISSN 0957-4174,
xxxiiNilufar Islam, Ashraf Farahat, Mohammad Abdullah M. Al-Zahrani, Manuel J. Rodriguez, and Rehan Sadiq. Contaminant intrusion in water distribution networks: review and proposal of an integrated model for decision making. Environmental Reviews. 23(3): 337-352. https://doi.org/10.1139/er-2014-0069
Dr. Tim Sandle is the author of the book Digital Transformation and Regulatory
Considerations for Biopharmaceutical and Healthcare Manufacturers, Volume 1: Digital Technologies for Automation and Process Improvement, available via the PDA Bookstore: https://www.pda.org/bookstore/product-detail/5897-digitaltransformation-volume-1.