OR WAIT null SECS
James E. Akers is the president of Akers Kennedy & Associates, PO Box 22562, Kansas City, MO 64113, email@example.com.
James P. Agalloco is the president of Agalloco & Associates, P.O. Box 899, Belle Mead, NJ 08502, tel. 908.874.7558, firstname.lastname@example.org. He is also a member of Pharmaceutical Technology’s editorial advisory board.
Vaporous hydrogen peroxide, used for sterilization and decontamination, is highly potent but presents implementation challenges.
The use of hydrogen peroxide (H2O2) in the global healthcare industry and other industries that require high levels of contamination control has grown steadily. This growth is attributable to the chemical’s ability to kill spores and sterilize materials, which has been demonstrated in a variety of practical applications. Properly used, H2O2 is an effective sterilant capable of efficient and rapid elimination of contaminating microbes. Some difficulties have been associated with the implementation of H2O2 processes in the healthcare field although these issues appear to have been avoided in commercially sterile food and beverage manufacture. Specifically, persistent problems regarding the development of H2O2 processes and their subsequent validation have been reported. The author discusses the technical issues associated with achieving lethal concentrations of H2O2 delivered in vaporous form on decontamination targets, explores the core scientific principles behind H2O2‘s use in decontamination and sterilization, and provides experience-based solutions to frequently encountered operational issues.
Hydrogen peroxide (H2O2) is an extremely powerful oxidant that is capable of effectively killing resistant spore-forming bacteria over a wide range of concentrations; at concentrations of 3% or less, it is suitable for use as a topical antiseptic (1). H2O2 has been accepted by both FDA and the US Environmental Protection Agency (EPA) as a sterilizing agent for many years (2, 3). In the food industry, H2O2 is widely used to sterilize containers, closures, and aseptic chambers (i.e., isolators) used for manufacturing low-acid and dairy-based beverages as well as other applications (4).
The potency of H2O2 as a sterilant and its usefulness in a broad range of antimicrobial applications are beyond dispute. The problems associated with vaporized H2O2 processes in the healthcare industry lie in fundamental misunderstandings concerning physicochemical characteristics of H2O2 sterilization. These errors profoundly influence real-world H2O2applications.
What is Vapor?
There are three primary states of matter--solid, liquid, and gas. The term “vapor” is defined in several ways. Scientifically, a vapor is a gas at a temperature lower than its critical point; a vapor is a gas phase where the same substance can also exist as a liquid. An example is atmospheric water vapor. At temperatures above the dew point, water in the atmosphere is a gas. As the temperature is lowered through the dew point, the gaseous water condenses to form a fog or mist, or it can condense and form liquid water on a cold surface. Another definition of vapor is visible moisture in the air, as in fog or steam—a system in which a liquid is suspended in a gas.
Figure 1 shows water in various phases: the lake, the dense fog at the foot of the mountain, the wisps of cloud, and the blue sky above. The lake is certainly liquid water; the blue sky is just as clearly a gas which contains water in the gaseous state. The fog or cloud in the center is a mixture of a gas phase (comprised of nitrogen, oxygen, water, carbon dioxide, and trace amounts of inert gases) and a suspended liquid phase (small droplets of water). The density of the fog or cloud varies with its temperature. It is thickest (i.e., suspending the most liquid) near the base of the mountain where it is coldest. It is clearly less dense, with less suspended water droplets near the top of the image where the temperature is higher.
One of the major difficulties with hydrogen-peroxide (H2O2) processes is the use of a vapor for delivery of H2O2 and water (H2O) to the target chamber. It must be understood that a vapor is a mixture of air and liquid that is present within the chamber. In decontamination or sterilization using H2O2, the liquid phase is comprised of both H2O2 and H2O, and the concentration of each in the gas and suspended liquid state can vary across the system.
James P. Agalloco and James P. Akers
To fully understand the physical factors that affect the distribution of H2O2 in the vapor phase, one must consider the factors that affect vapors in general and the factors that allow them to exist in air, which is the medium in which H2O2 in the vapor phase is distributed within a decontamination target. Air contains varying, but small, amounts of water in the vapor phase, which is described using the term relative humidity (RH). An important factor in the distribution of a chemical is the dew point. The dew point is, in simplest terms, a function of both concentration and temperature. When the concentration of water exceeds the saturation point at a particular temperature, condensation occurs. The gaseous water converts to the liquid phase, and droplets of liquid water may appear. On the other hand, if the water concentration is below the saturation point, it will remain in the gas phase. When the temperature of the air is actively lowered (or simply drops as a function of thermodynamics) below the dew point, some portion of the water (H2O) present as a gas mixed with air condenses and forms liquid droplets. We observe this as clouds, dew, fog, or frost.
The typical H2O2 process
The process that most H2O2 generator and isolator manufacturers use for H2O2 introduction is one in which a hot air stream is used to introduce a heated H2O2/H2O gas into the target environment, which may be an aseptic chamber or isolator. Within the generator, the temperature of the air/H2O2/H2O mixture is sufficiently high that all three materials are in a gaseous state. The hot air is conventionally at temperatures in excess of 100 °C, which takes advantage of the respective boiling points of the pure components (i.e., H2O = 100 °C, H2O2 = 150.2 °C, and a 30-35% aqueous solution of H2O2 = approximately 108 °C). At these temperatures, both H2O2 and H2O are present as gases and are carried into the target vessel with the hot air. The H2O2/H2O is supplied as an aqueous solution of H2O2in varying percentages typically ranging from 31% to 50% H2O2. At typical room temperatures, each of these solutions is predominantly liquid, and the headspace air within the closed containers has a small amount of gas phase H2O2/H2O that is in equilibrium with the liquid.
If the concentration remains below the saturation point upon introduction into the target environment, then both the H2O2 and H2O will remain in the gas phase. When the hot and relatively humid gas mixture from a H2O2 generator is introduced to the target chamber, it will encounter colder air as well as ambient temperature surfaces of the chamber and materials inside it. As the hot gas mixture cools to the temperature of the chamber, it will fall below the dew-point temperature of both H2O2/H2O, and some portion of these materials will condense on the surfaces as liquids. In effect, the H2O2/H2O are returning to their initial equilibrium state of liquids in equilibrium with the adjacent gas, which they possessed before being converted to a gas in the generator.
Condensation that forms on the surfaces will tend to be nonuniform in concentration across the chamber for several reasons:
The extent of condensation that occurs depends upon the temperature (i.e., colder locations will have more condensation), the concentration or amount of H2O2/H2O introduced (and removed if a circulating process is used), the size of the enclosure (i.e., affects the surface/volume ratio), and the quantity of material within the chamber (i.e., adds to the surface area).
Phase states in the enclosure
In must be understood that the enclosure will contain a mixture of air/H2O2/H2O internally, with some of the H2O2/H2O in a liquid state on surfaces and the remainder in the gas phase. There is no simple means to establish how much H2O2/H2O is in each phase or where in the chamber a particular phase is present. Additionally one cannot know the percentage of H2O2 or H2O at any single location, and certainly not at every location within the enclosure. The Gibbs Phase rule makes it clear that conditions can vary across the system (see Equation 1).
F = C - P + 2 = 3 - 2 + 2 = 3 (Eq. 1)
where F = number of degrees of freedom (i.e., concentration, temperature, pressure), C = number of components in the system, and P = number of phases in the system.
Almost nothing is known with certainty with respect to concentration and location. There is, however, one constant in the process: H2O2 is lethal to microorganisms in both the gas and liquid phases. It is reasonable to assume that liquid-phase kill will be somewhat faster than the gas-phase kill for two important reasons as further outlined:
An older reference describes more rapid kill occurring with H2O2 in a gas-phase process compared to a liquid-phase process (6). This reference identifies a gas-phase process at 25 °C, with no mention of any liquid H2O2 present. At that temperature, however, H2O2 is a liquid, so there must be some liquid H2O2 in equilibrium with the gas. There is no means to establish that the kill in this “gas” process was actually accomplished in that phase. It is more likely that the cited kill was accomplished in the liquid phase. Misinterpreting what is actually “vapor” as a “gas” has led to the erroneous belief that gaseous-phase kill is more rapid than liquid-phase kill.
The expected microbial kill rates in the system might appear as shown in Figure 1, which visualizes H2O2 sterilization as a process that occurs within a band, bounded by the extremes of liquid and gas-phase kill. Figure 1 represents what is believed to occur and does not reflect any specific H2O2 process. The absolute slopes of the death curves are unknown. Given that the localized concentrations in both phases are variable due to temperature differences and proximity to the inlet with its heated air supply, it must be recognized that there will be different kill rates in different locations in both the liquid and gas phases. Figure 1 represents what might occur at a single point within the chamber; similar appearing death curves with differing slopes can be considered for other locations where the local conditions are different. These variations are the underlying cause of the variable performance experienced when using vapor-phase H2O2 as a lethal process.
Figure 1: Estimated relative kill rates in liquid and gas phases; the exact kill rates and their differences are unknown. (Figure courtesy of the authors.)
D-values for H2O2 decontamination
The death curves in Figure 1 seem to show that a D-value (or an approximation of one) could be established against a challenge microorganism for the combined processes. That assumption is faulty because there is no way of establishing what conditions (e.g., phase, concentration, or humidity) are present in the system at the point where the microorganism is killed. D-value determination requires knowledge of the specific lethal conditions to which a microorganism is exposed. In a single-phase sterilization process, gas or liquid, information on concentration of the agent, humidity (assumed at 100% for liquid processes), and temperature is readily determined. In the context of H2O2, this is easiest for liquids, and published D-values for Geobacillus stearothermophilus in various H2O2/H2O liquid solutions are available (1). These liquid phase D-values demonstrate extremely rapid kill (in seconds) at even modest H2O2 concentrations (7). At the estimated concentrations where condensation first occurs in vapor H2O2 processes, the D-values should be lower as the concentration will be substantially higher than that published in the literature. Unfortunately, no comparable data are available on H2O2, where a strictly gas-phase process is present. Thus, any labeled “D-values” for vapor H2O2 biological indicators must be considered nothing more than an approximation as the killing conditions are unknown. The conditions of kill may be consistent enough that they could be replicated in an independent study in the same test system. What cannot be established from these labeled “D-values“ is how that same biological indicator will respond in a different environment where the conditions are also unknown and most likely substantially different.
In the 20-plus years that this industry has been using H2O2 decontamination, a BIER (biological indicator evaluation resistometer) vessel for H2O2 has not been developed as a standard for compendial or routine use. The same conundrum faced with respect to variable and unknown biphasic conditions in a larger system has prevented the development of a H2O2 BIER. The absence of a BIER vessel and, thus, a fully useable “D-value” for H2O2 biological indicators has caused some difficulties. What can be established from the vendor “D-value” is the relative resistance of one lot to another from the same vendor. How any individual lot will perform under different conditions is something the user must determine for each application.
One suggested approach to get beyond this lack of a definitive D-value for a biological indicator is to establish a process or system “D-value” for a biological indicator within a large enclosure and rely upon that as the basis for destruction in the system rather than the vendor’s reported value. This approach presumes that the conditions used to establish the process/system “D-value” are representative of the entire system. That assumption is decidedly not the case, nor is it known whether the location(s) chosen for the process “D-value” determination are best case or worst case with respect to kill across the chamber. A number, which is not a D-value in the strict sense, can be calculated, but the utility of that number in any estimation kill rate across the chamber is essentially nil.
Reports of vapor-phase “D-value” variations as a consequence of different substrates must also be recognized as uncertain (8, 9). Because there is no objective biological indicator evaluation method available, published “D-values” are not standardized and thus of very limited use. Unless the concentration on the individual surfaces tested can be known and demonstrated to be constant, any hint that the substrate variations are meaningful must be viewed with some skepticism. There is also some published evidence that “D-values” may vary with spore concentration applied to the carrier material, which means kill may not be linear with concentration. That represents a serious flaw in the use of any biological indicator.
Is safety a concern with H2O2?
Given the rapid kill observed in the H2O2 liquid phase, the difficulties in attaining consistent kill with H2O2 vapor processes can only be explained by a lack of adequate condensation, for there is little doubt then when condensation does occur, kill will be quite rapid (10). Many of the newer generator designs, either freestanding or integrated into enclosures, rely on condensation to decontaminate/sterilize extremely rapidly.
Since the rapid kill provided by liquid H2O2 is well documented, why has industry been cautioned to avoid condensation in vapor H2O2 processes? The answer lies in the early teachings of AMSCO (now Steris) when the first H2O2 generator was introduced in the late 1980s. Caution was routinely raised regarding the potential hazards of high concentrations of liquid H2O2. (The H2O2 concentration in the gas phase at ambient temperature will always be substantially lower than its equilibrium concentration in the liquid phase.) The relevant safety issues with the use of H2O2 vapors are:
While there is a need for caution with respect to the use of vapor phase H2O2, undue concern is unwarranted. In more than 20 years of use in the global industry, there have been no reported incidents of personal injury or equipment damage associated with this process.
Claims that vapor-phase H2O2 processes do not result in condensation are speculative. The laws of physics and temperature within enclosures are such that some measure of condensation will always occur, and in many recent equipment and process designs the creation of condensation is intentional. Thus, within the context of real-world experience, the safety issues associated with vapor H2O2 systems where condensation is present appear to be adequately managed, assuming appropriate worker-safety precautions are maintained.
Limitations of multipoint process-control measurements
FDA’s Guideline on Sterile Drug Products Produced by Aseptic Processing recommends: “The uniform distribution of a defined concentration of decontaminating agent should also be evaluated as part of these studies” (15). This suggestion is made without reference to a specific methodology that could be employed. There is no technology that could address this expectation throughout a two-phase environment. Nor would the resulting data on concentration in the gas phase be useful in correlating to microbial kill on surfaces. When appropriate amounts of H2O2 are used for decontamination or sterilization, some of the available instruments, such as those that rely on near-infrared transmission, are unusable due to condensation on the lenses. Because accurate measurement is not possible, chemical indicators provide the only widely available means to confirm that H2O2 is, or was, present at a specific location.
Problems in an unsteady-state process
The introduction of H2O2 into a room-temperature enclosure uses vapor-process heating to convert the liquid solution into a gas for mixing and distribution in hot air. The temperatures in vaporizers are in the range of 105-150 °C. This high temperature results in some localized heating of the enclosure, primarily in locations close to the entry point of the heated materials. The effects of this heat input are multiple:
These phenomena are more problematic in those generators where H2O2 is fed and removed throughout the process. Systems that operate in a fill-and-soak mode may attain equilibrium conditions within the targeted volume.
The negative consequences of the unsteady-state nature of vapor-phase H2O2 processes are unavoidable in recirculating systems. The only means to establish a consistent process is to use enough H2O2 that even the warmest locations attain some measure of condensation. This solution is more easily accomplished in the non-circulating systems.
Penetration and adsorption by H2O2
Years of experience with vapor-phase H2O2 processes have shown how best to address the adverse impact of its adsorption as further explained:
The adverse consequences of decontamination and sterilization processes should be considered in the development and control of every process. Vapor-phaseH2O2 processes, because of their dual-phase nature, present new challenges. Were other gases to be used, similar, but different, concerns would present themselves and appropriate solutions would be identified. A more penetrating agent would only increase the penetration/aeration difficulties encountered, so while H2O2 penetration/absorption/desorption is a problem, the situation might be worse with alternative materials.
Biological indicator issues
Difficulties encountered in the destruction of biological indicators have been commonly reported and are so well known that there are some who doubt the efficacy of H2O2 as a sterilant. These problems are multifaceted but resolvable when the sterilization process is properly established.
First, H2O2decontamination and sterilization must be understood as a two-phase system. Considering it as a single, gas-phase process has caused more difficulties than anything else. The variability demonstrated in lethality is the direct result of applying process constraints that are suitable for a gas process but inadequate for two-phase H2O2 processes. Adapting process models and approaches from the most common gas sterilant, ethylene oxide (EO), to a vapor process created much of the problem. The largest flaw in this thinking is the deliberate avoidance of condensation in endeavoring to make what must be a two-phase vapor process into one that operates in a single phase. Some wrong assumptions are:
In the actual two-phase H2O2 process, none of these assumptions is correct or attainable at the present time. These assumptions led to the establishment of vapor processes that are inadequate for their intended purpose. They do not adequately induce condensation or use sufficient mixing and thus fail to deliver reasonably consistent conditions throughout the enclosure. The experienced difficulties are a consequence of poor cycle design and not problems with the lethality of H2O2.
Second, biological indicators must be specifically designed for the intended process. While there have been attempts at this design, what has been accomplished is largely empirical. The methods used for manufacturing H2O2 biological indicators may be identical to those used for other sterilization processes, but because correlation to actual process resistance is lacking, the process suggestions inferred from labeled resistance values are essentially unusable. In the absence of a BIER (and thus truly reproducible biological indicator resistance), the typical biological indicator process response can not be expected for vapor-phase H2O2 processes.
The most important attribute of any biological indicator is its reproducible resistance to the intended process. There is no established D-value method, which severely limits the certainty of process understanding and biological-indicator design and selection. Variable results with biological indicators could be attributable to either variations in the biological-indicator resistance or variation in the conditions resulting from poorly conceived controls for a complex process. Lacking a biological indicator whose response to the process is precise, vapor-phase decontamination and sterilization becomes a more challenging process to control.
Third, there is a demonstrated biological indicator concentration effect associated with the H2O2 processes unlike that seen in other sterilization processes. Biological indicators with a higher initial population have proven more difficult to kill with H2O2 than would be expected based upon the results of the same lot at a lower concentration (18). This phenomenon contradicts the core principle in all sterilization processes that microorganisms die at a constant logarithmic rate regardless of population. Occurrence of this phenomenon in H2O2 processes can be attributed to several possible causes:
All of these are correctable. Using a lower population biological indicator eliminates the first two of these difficulties. A hundred-fold reduction in spore population reduces the amount of debris present at the edge of the biological indicator drop and eliminates spore clumping significantly. This single change would result in more linear death curves than what has been evidenced. The third difficulty is a common mistake that is all too prevalent in the healthcare industry and has no basis in fact (22). The food industry has used H2O2 successfully for sterilization for many years and operates without this artificial and erroneous expectation. The last issue is an artifact of the limited process understanding still prevalent on many existing H2O2 processes. In cases for which condensation is actively promoted in the process, fewer problems with sterilization are encountered.
Much has been made recently of so-called “rogue” biological indicators. These rogues (i.e., outliers) are presumably biological indicators that failed to conform to the user’s expectations of their demise. There is little doubt that the production of spore crops, substrate selection, and the manufacture of biological indicators could result in clumping and encapsulation in contaminants that could result in a lack of uniform performance (23). Properly manufactured biological indicators should be largely free of outliers. Greater frequency of outliers detected in vapor H2O2 processes seems to be the result of poor understanding of vapor-phase H2O2 that results in marginally lethal processes and the creation of biofilms and clumps of spores on stainless steel at 106 concentrations, which result in what are effectively false-positive biological indicators that do not represent the elimination of normal flora at more diffuse concentrations.
Summary and recommendations
The successful use of any decontamination or sterilization process requires a thorough understanding of the underlying principles of the process with particular attention to those aspects that differentiate it from other methods because these represent potential new learning. The two-phase nature of the vapor-phase H2O2 process introduces complexities that, if not well understood, can prevent successful use. The healthcare industry has experienced considerable difficulty in the implementation of this process.
The greatest improvements in operating these processes can be obtained through the use of conditions that force some measure of condensation and by recognition that the desired log reduction of these processes need not be excessive given the end use of the enclosure. Only product contact parts must be sterilized, and shifting attention to those locales within the enclosure alone would result in substantial improvements in process outcomes.
About the Authors
James P. Agalloco is president of Agalloco & Associates, P.O. Box 899, Belle Mead, NJ 08502, tel. 908.874.7558, email@example.com. He is also a member of Pharmaceutical Technology’s editorial advisory board.
James E. Akers, PhD is president of Akers Kennedy and Associates.
Submitted: Feb. 1, 2013. Accepted: Mar. 1, 2013.