Real-World Vapor Phase Hydrogen Peroxide Decontamination

sciencephoto/stock.adobe.com

The introduction of isolation technology in the pharmaceutical industry by the French firm La Calhene in the 1980s required a means for the reliable microbial decontamination of the isolator interior. This was initially performed using a mist of peracetic acid/water; however, this was considered undesirable for a variety of reasons, with objections over its use primarily aimed at the resultant corrosion of surfaces, wet surfaces, and lengthy aeration times.  In the late 1980s, AMSCO (now Steris) offered the first widely available alternative with their VHP-1000 generator, which delivered vapor phase hydrogen peroxide (VHP). With the introduction of the VHP-1000, this technology and its derivatives became the dominant means for isolator decontamination. 

Initially, the performance expectations for the decontamination process varied according to the end user’s protocol requirements. The process target (decontamination or sterilization) and the means to establish them (the biological indicator [BI] population to use and the selection of cycle duration) varied widely. Firms with near identical systems and practices considered them differently.  

Regulatory influences

The first definition of process expectations was provided in United States Pharmacopeia (USP) 28 <1208>, Sterility Testing-Validation of Isolator Systems (2000), with an expectation for sterilization, “The sterilization methods used to treat isolators, test articles, and sterility testing supplies are capable of reproducibly yielding a six log kill against an appropriate, highly resistant biological indicator” (see Figure 1) (1).

Figure 1. Log-kill and log reduction.

The inclusion of a “six-log kill” requirement is problematic in that this terminology is uncommon in sterilization; there being no generally accepted definition for it in sterilization practice. It could be interpreted as either a six-log reduction or complete kill of a six-log population (a nine-log reduction, see Figure 1). Stating that and expecting “sterilization” in the first version of USP<1208> was clearly problematic, as sterilization is more widely accepted as a minimum 12-log reduction of a BI.

The Parenteral Drug Association issued its publication on isolators in 2001. It included several different means to establish its decontamination requirement, “for the purposes of isolator decontamination a Total Kill Analysis study of a suitable bioindicator with a population of 10⁵ or greater is considered an overkill cycle. Such a cycle indicates a spore log reduction value of >7 logs” (2). 

FDA formally addressed isolator decontamination first in 2002, with the initial draft of its revised aseptic processing guidance (3). The stated expectation was, “For most production applications, demonstration of a six-log reduction of the challenge BI is recommended.” The draft was also explicit in stating decontamination, as opposed to sterilization, of the interior was expected. Sterilization was correctly identified as the requirement for product contact surfaces.

The European Medicines Agency, through the assembly of inspectors with the Pharmaceutical Inspection Co-Operation Scheme (PIC/S), issued its first isolator-related guidance in 2004 and aligned closely with FDA, “…, but a target of six log reductions is often applied” (4). This document was explicit in defining a six-log reduction, as not requiring the complete destruction of all microorganisms on a 10⁶ population BI. This document adds confusion of a different kind by referring to the decontamination processes primarily as a gaseous process, which is a serious error (see below).

The final version of the FDA guidance published in September 2004 relaxed the log reduction position; “Normally, a four- to six-log reduction can be justified depending on the application” (5). This guidance in conjunction with the PIC/S six-log definition has largely shaped global industry decontamination practice to this date.

In 2008, USP revised chapter <1208> to align more closely with the regulatory and industry guidance documents that had been issued, “The ability of the process to reproducibly deliver a greater than three-log kill is confirmed in three consecutive validation studies” (6). In this revision, the process is now termed decontamination consistent with the other standards.

USP significantly revised its sterilization content in 2013–2018 and added vapor phase sterilization in those changes. The new content includes two related subchapters that directly impact VHP processes (7,8). The USP  content changes expectations in many ways. First, vapors are considered two-phase systems, the presence of which substantially complicates process design and execution. Second, because of the dual-phase nature of the process, D-value estimation is not possible because the conditions of microbial death are indeterminate. Lastly, if the D-value is indeterminate, the log reduction expectations may be inadequate. The USP’s chapters disrupt many of the commonly held misconceptions regarding VHP validation.

A regulatory perspective was provided by the United Kingdom’s Medicines and Healthcare products Regulatory Agency (MHRA) in an April 2018 blog on their website. “VHP, when well controlled and validated, is a useful method for the decontamination of the surrounding workspace, e.g., an isolator environment. However, given the above concerns, our current stance is that VHP cannot be used to sterilize critical items” (9). This position creates substantial difficulties for many current installations. The in-situ sterilization of stopper bowl, feed tracks, and many other surfaces in direct contact with sterilized items (and presumably considered “critical”) is called into question by this MHRA position. Industry response has been one of alarm, as it potentially invalidates many systems current in use. 

Inside the real world 

A shift in perspective is essential to succeed in the changing landscape. The existing log reduction requirements are poorly suited to the emerging perspectives of decontamination using a vapor treatment. Increasing usage of vaporized hydrogen peroxide has resulted in numerous publications describing applications of various types, from small pass-through chambers to entire suites of rooms. Unfortunately, past mistakes and misstatements have been incorporated which have adversely influenced industry practices. This document endeavors to clarify the process relying on the core science underlying it. Many of the problematic statements regarding vapor decontamination are provided in this paper followed by the scientific reality.

Common misconception: Vapor phase hydrogen peroxide decontamination is a single phase gas process.

Scientific reality: There are varying definitions of vapor, some of which suggest that vapors are equivalent to single phase gas mixtures. However, at the ambient temperatures used for VHP processes condensation of hydrogen peroxide (H₂O₂), and, to a lesser extent, water (H₂O), is unavoidable (10,11). Vapor decontamination processes operate below the boiling point of both materials, and, while some of each will remain in the gas phase due to their vapor pressure, they are both liquids at ambient temperature. The amount of H₂O₂and H₂O maintained in the gas phase decreases as the temperature decreases below that used to introduce them into the system. Thus, VHP must be understood to be a two-phase process in which both gas and liquid are present. Depending on the amount of liquid phase present, it may or may not be visible. The original characterization of VHP as a completely gas phase process originates in the originating patent; however, the inventors did not consider or determine what phase was actually present at the point of kill (12). 

Common misconception: The kill of resistant BIs by H₂O₂is more rapid in the gas phase than in the liquid phase.

Scientific reality: The experiments that support this claim can be interpreted differently to assert that the exact opposite is true (13). No determination of phase or concentration measurement at the point of kill was performed in the experiments. Condensation is highest at the lowest operating temperature, and these demonstrated the fastest kill rate. Published data on H₂O₂solutions indicate that spores are killed extremely rapidly (14–16). This suggests that liquid-phase kill will be more rapid than gas-phase kill because the concentration of H₂O₂in the liquid will always be higher than in the gas phase (17). This has been independently confirmed in challenge studies conducted using various decontamination systems (18–20).

Common misconception: VHP decontamination is a single phase gas process that resembles ethylene oxide (EtO) and other sterilizing gas processes.

Scientific reality: Considering gases and vapors as identical goes to the very origins of hydrogen peroxide decontamination (12,21). The confusion in terminology may originate in the means to introduce a 35% aqueous solution into a hot air stream where high heat rapidly converts the H₂O₂aqueous solution into a multi-component gas at elevated temperature. Above their boiling points, both liquids are converted to gases. The hot gas mixture is introduced into the ambient temperature target system where it loses heat to its surrounding surfaces. As this occurs, the H₂O₂and H₂O concentrations in the gas phase are above their saturation vapor pressure and condensation must occur (22). The presence of two phases in an ambient-temperature vapor hydrogen peroxide process is unavoidable. Hydrogen peroxide having a higher boiling point and lower vapor pressure than water condenses first and concentrations of H₂O₂in the liquid phase will be higher than the original solution percentage. 

Common misconception: Because VHP decontamination is a gas phase process, temperature is not an important influence on process lethality.

Scientific reality: VHP is a multi-phase process with both liquid and gas phases present (23). The amount of condensation on the surface is directly related to the local temperature. As the H₂O₂/H₂O solution is first vaporized and then introduced with hot air, there will be temperature differences across the system. The hottest areas, typically closest to the vapor inlet and any operating equipment, are potentially “worst case” because of reduced condensation. Changes in temperature during the process duration are somewhat unavoidable, especially in smaller chambers, which will alter the amount of condensation and, thus, process lethality. It is important to maintain near-constant room conditions to minimize process variability during the individual VHP process and between multiple VHP cycles over time. 

Common misconception: Measurement of H₂O₂gas concentration can be used to reliably control VHP processes. 

Scientific reality: The gas- and liquid-phase concentrations at any location is dependent upon the temperature of the system at that location. Gas-phase measurement cannot be used to estimate concentrations on the surface unless the system reaches equilibrium. Variations in temperature across location and process duration are unavoidable with most VHP processes, which minimizes the utility of concentration measurements taken in the gas phase. 

Common misconception: Condensation during VHP decontamination is to be avoided, as it slows kill and extends aeration times. Thus, dehumidification prior to processing helps prevent condensation.

Scientific reality: Condensation is unavoidable, and, while not always visible, it actually supports more rapid kill (see previous scientific explanations), condensation being necessary to assure rapid kill is actually desirable. Dehumidification prior to the introduction of H₂O₂is unnecessary because it delays condensation, thus, increasing process time and system cost. Kill is quicker in condensing systems such that the process dwell can be shortened substantially. 

Common misconception: Condensed H₂O₂on surfaces is a corrosion and explosion hazard and is to be avoided.

Scientfic reality: Corrosion is not a concern when appropriate materials are used for construction of the system. Condensation of H₂O₂has always been present, even if not always readily visible. Nevertheless, in more than 30 years of vapor phase hydrogen peroxide use, there has never been a reported explosion. This includes the many newer systems where H₂O₂is intentionally condensed to expedite the process.

Common misconception: Condensation of H₂O₂on surfaces can result in greater adsorption of H₂O₂than a gas phase process without condensation. This can result in lengthy aeration.

Scientific reality: Gases can permeate Tyvek wrapping and be readily absorbed by polymers, while condensed liquids cannot permeate the hydrophobic material and remain on the exterior surface. It might seem counter intuitive, but a “wet” process can result in less adsorption than a “dry” process. Note that the distinction between “wet” and “dry” processes is an artificial one. As mentioned, all VHP processes have some condensation present, so the “wet” vs. “dry” distinction refers to the differing beliefs of whether kill occurs best when the BI is contacted with a liquid or a gas. With rapid kill, diffusive adsorption of H₂O₂from the gas phase is reduced and overall process duration may be shorter in “wet” processes than in “dry” processes. The longest segment of many VHP cycles can be the aeration period, which is limited by slow desorption of H₂O₂from polymeric materials and permeable packaging.

Common misconception: Labeled “D-values” for VHP BIs are definitive and accurate (author’s note: throughout the document, the term “D-values” in quotes denotes instances where it is believed that the values are improperly identified as such, and the term D-values without quotes is used elsewhere).

Scientific reality: A D-value can only be established when the conditions (concentration, relative humidity, and temperature) to which the microorganisms are exposed are known (8). The presence of two phases in VHP systems makes D-value determination impossible. BIs labeled for VHP processes do not state the conditions of kill, but typically cite the H₂O₂injection rate in the BI manufacturers test system. The injection rate cannot be correlated to precise and reproducible destruction of microorganisms in a different system. None of the BI labels report accurate “D-values” because the conditions at the point of kill are unknown and, thus, the labels themselves are misleading. 

Common misconception: Vendors selling VHP BIs provide reliable “D-values” on their certificates of analysis.

Scientific reality: BI vendors have responded to customer expectations and the “D-value” information they provide does assure consistent resistance for their products, but only in the BI vendors test system (24,25). The performance of those same indicators in users’ systems cannot be predicted from the certificates provided because of the differing conditions in the users’ systems.

Common misconception: Publications showing the effect of varying materials on the “D-values” of BIs in VHP processes are useful because the BIs used in those studies are positioned on the surface (26–29).

Scientific reality: As stated earlier, D-value can only be established when the conditions (concentration, relative humidity, and temperature) to which the microorganisms are exposed are known (8). The various studies do not provide that information, however. Further, because the internal conditions inside the various systems must be understood to vary with location and time, the relative resistance associated with different materials and finishes cannot be established.

Common misconception: BIs for VHP can demonstrate anomalous behavior (30,31). These are often termed “rogues”, and their survival is an indication of defective BIs, not flaws in the process being evaluated. 

Scientific reality: BIs are produced for a variety of sterilization processes using well-established and consistent methods. There is no comparable problem with “rogue” BIs associated with any sterilization or decontamination process other than VHP. The “rogue” BI is more likely the result of poorly defined decontamination processes where the lethality delivered is insufficient to kill all of the BI microorganisms present.

Common misconception: Because there are “rogue” BIs, multiple BIs should be used with all VHP decontamination processes.

Scientific reality: Properly developed VHP processes that use sufficient H₂O₂to create modest amounts of condensation have been proven reliable without resorting to multiple BIs in an attempt to compensate for either a minimally lethal or excessively variable decontamination process.

Common misconception: A “system D-value” can be used to establish reliable VHP processes; “D-value of a BI measured in a specified gas generator/separative enclosure combined with a defined sporicidal vapor-phase (decontamination cycle)” (31).

Scientific reality: As indicated previously, D-value determination requires maintenance of specified conditions (concentration, relative humidity, and temperature). This is not currently possible for two-phase systems using a BIER type design (5,8). Suggestions that it could be achievable and useful in the operation of a large system with potentially greater variation in the critical parameters are poorly founded. 

Summary recommendations

Putting aside the unusual MHRA blog position, there are principles and practices to follow with respect to VHP decontamination that include: 

• The BI should have a population of 10⁴ colony-forming unit (CFU)/unit where decontamination is the process expectation.

• BI positioned on product contact surfaces should have a population of 10⁶ CFU/unit where sterilization must be demonstrated.

• All BIs should be fully inactivated. The use of multiple indicators at each location is unnecessary, especially if the intent of multiple BIs is to explain away positive results.

• VHP is most effective when provided using a “wet” process, and this may entail the use of greater quantities of H₂O₂.

• Labeled “D-values” for VHP processes are inaccurate and should never be used to determine cycle duration. They may be useful in comparing different BI lots from the same manufacturer.

• Log-kill times cannot be estimated from “D-values” and can only be established using BIs in the system being evaluated.

Conclusion

The preceding could be misinterpreted as denigrating the utility and efficacy of VHP as a decontaminating or sterilizing agent. There are certainly difficulties with its use; however, properly delivered, its many advantages outweigh the negative aspects and assure its continued and expanded usage (22). It lacks the simplicity of a gas process, and adaptation to the unique requirements of its two-phase nature is essential to success. Reliable decontamination/sterilization cycles and systems require the following:

• Acknowledgement that condensation is a necessary element of reliable kill

• Awareness that BIs provide prima facie evidence of lethality

• Abandonment of gas concentration as a useable metric  

• Acceptance of empirical evidence based on BI destruction as the definitive indicator of process suitability.

In spite of it being misrepresented as a single-phase gas process, hydrogen peroxide usage has increased steadily since its introduction. The cited difficulties have been overcome by many practitioners. Implementing vapor H₂O₂processes following sound scientific principles of chemistry, physics, and engineering can establish designs for the delivery of a complex, but lethal, chemical agent in the most effective manner. Numerous vendors and end users have realized success with vapor hydrogen peroxide, and, despite the difficulties, some have experienced, increased future use is near certain.

References

1. USP, <1208>, “Sterility Testing-Validation of Isolator Systems” USP 28 (US Pharmacopeial Convention, Rockville, MD, 2000).
2. PDA, TR 34: Design and Validation of Isolator Systems for the Manufacturing and Testing of Health Care Products (Bethesda, MD, 2001).
3. FDA, Draft Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing, September (Rockville, MD, 2002).
4. Pharmaceutical Inspection Co-Operation Scheme, Recommendation on Isolators Used for Aseptic Processing and Sterility Testing, PI-014-2, July 2004.
5. FDA, Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing (Rockville, MD, September 2004).
6. USP, <1208> “Sterility Testing-Validation of Isolator Systems” USP 30 (US Pharmacopeial Convention, Rockville, MD, 2008).
7. USP, Supplement 1, <1229.11>, “Vapor Phase Sterilization” USP 38 (US Pharmacopeial Convention, Rockville, MD, 2015).
8. USP, Supplement 1, <1229.5>, “Biological Indicators for Sterilization”USP 39 (US Pharmacopeial Convention, Rockville, MD, 2016).
9. MHRA, “VHP (Vapour Hydrogen Peroxide) Fragility,” MHRAInspectorate.blog.gov.uk, April 20, 2018.
10. W.M. Haynes, Ed., CRC Handbook of Chemistry and Physics (CRC Press LLC, Boca Raton, FL, 95th ed., 2014) pp. 4–67.
11. T.E. Daubert, and R.P. Danner, Physical and Thermodynamic Properties of Pure Chemicals Data Compilation(Taylor and Francis, Washington, DC, 1989).
12. USPTO, “Hydrogen Peroxide Vapor Sterilization Method,” US Patent 4,169,123, Sept. 25, 1979. 
13. D. Eddington, et al, “Chemical and Biological Aspects of Hydrogen Peroxide Vapor,” Presentation at ISPE Barrier Isolation Technology Conference (Washington, DC, 2000).
14. R. Toledo, et al., Applied Microbiology, 26 (4) 592–597 (1973).
15. S.S. Block, “Peroxygen Compounds,” in Disinfection, Sterilization and Preservation(Lippincott Williams and Wilkins, Philadelphia, PA, 5th ed., 2001), pp. 185–204.
16. M. Ahlgren, and Z. Bjorkland, “Robustness of a Biological Indicator Resistance Test Rig-a Study of Different Parameters, Methodologies, and Their Impact on the Measured D-value of the Biological Indicator,” Lund University, LTH School of Engineering, 2005.
17. C. Hultman, et al., Pharmaceutical Engineering 27, 22–32 (2007).
18. J.E. Akers, et al., “A New Approach to VPHP Decontamination,” presentation to ISPE, 2006.
19. D. Watling, “Is H₂O₂a wet or dry process?” presentation at PDA Isolator Technology Conference, April 29-30, 2002.
20. B. Unger-Bimczok, et al., J. Pharm. Innovation 3, 123–133 (2008).
21. G.S. Graham, et al., “Sterilization of Isolators and Lyophilizers with Hydrogen Peroxide in the Vapor Phase,” proceedings from the International Congress of the Parenteral Drug Association (Basel, Switzerland, 1992), pp. 32–51.
22. J. Agalloco and J. Akers, Pharmaceutical Technology 37 (9) 46-56 (2013).
23. D. Watling and M. Parks, Pharmaceutical Technology Europe, 16 (3) (2004).
24. G. Krushefski, Spore News 9 (4) (2009).
25. J. Drinkwater, et al., EJPPS 14(1) 5–11 (2009).
26. V. Sigwarth and C. Moirandat, PDA J Pharm Sci Technol 54 (4) 286–304 (2000).
27. V. Sigwarth and A. Stark, PDA J Pharm Sci Technol 57 (1) 3–11 (2003).
28. T. Mau, et al., PDA J Pharm Sci Technol 58 (3) 130–146 (2004).
29. B. Unger, et al., PDA J Pharm Sci Technol 61 (4) 255–275 (2007).
30. P. Templeton and J. Hillebrand, “Case Study: Isolator Sanitisation Cycle Development, Validation and Revalidation,” presentation at ISPE Isolation Barrier Forums (Washington, DC, and Prague, 2005).
31. PDA, TR 51: Biological Indicators for Gas and Vapor-Phase Decontamination Processes: Specification, Manufacture, Control and Use (Bethesda, MD, 2010). 

About the author

James Agalloco, jagalloco@aol.com, is principal of Agalloco & Associates.

Article Details

Pharmaceutical Technology
Vol. 44, No. 1
January 2020
Pages: 53–57

Citation

When referring to this article, please cite it as J. Agalloco, “Real-World Vapor Phase Hydrogen Peroxide Decontamination,” Pharmaceutical Technology 44 (1) 2020.