In recent years, acute awareness has been focused on ethylene oxide health impacts by community stakeholders, regulators, and attorneys across the United States. Following several investigations in the 2010s, ethylene oxide received compounding scrutiny from a number of regulatory agencies and nongovernmental organizations (NGOs) on the potential health hazards the compound presents, from both occupational and environmental exposures. This newfound exposure recognition—driven mainly by fugitive community exposures—has led to evolving assessments and methods to quantify community human health risk. Subsequent to these investigations have been the claims that ambient ethylene oxide exposures have caused cancer. The impacts that these exposures have on human health are heightened by many in the press, by numerous communities, and at the federal level. An understanding of where, how, and why ethylene oxide exposures are occurring is necessary to contextualize current human health exposure and risk assessment efforts.
Ethylene oxide is a colorless, sweet-smelling flammable gas at room temperature that has a variety of uses in everyday life. At some point, we have used products made with ethylene oxide, whether we know it or not. In fact, more than 97% of ethylene oxide is used as a chemical intermediate in the production of other industrial products, including surfactants, adhesives, fiberglass, perfumes, textiles, lithium-ion batteries (e.g., electric vehicles), personal care items (e.g., cosmetics, shampoo, etc.), and even the antifreeze used to cool internal combustion vehicles. In addition to serving as a manufacturing intermediate, ethylene oxide is a registered U.S. Environmental Protection Agency (EPA) active ingredient for use as a sterilizing antimicrobial agent in a number of different healthcare scenarios as identified by the U.S. Food and Drug Administration (FDA). Ethylene oxide is available for spice and herb sterilization (e.g., licorice, sesame seeds, spearmint, etc.) with established residue thresholds. Ethylene oxide is also found in tobacco smoke and is a byproduct of fuel combustion, both of which serve as additional environmental exposure sources.
Historically, ethylene oxide exposure concerns centered mainly around occupational exposure scenarios. Health-based occupational exposure thresholds are established to protect those who are using the material daily as part of the work process. Manufacturing systems utilizing ethylene oxide are typically closed systems (e.g., sealed) because of the chemical’s physical characteristics during use. The use of ethylene oxide as a sterilant may create more unique exposure scenarios than manufacturing because the sterilization systems may have exposure points that cannot be easily engineered out of the process.
Ethylene oxide has established occupational exposure thresholds, both enforceable (i.e., Occupational Safety and Health Administration permissible exposure limit (OSHA PEL)) and recommended (i.e., National Institute for Occupational Safety and Health recommended exposure limit (NIOSH REL) and American Conference of Governmental Industrial Hygienists threshold limit value (ACGIH TLV)). Because ethylene oxide is used as a registered antimicrobial pesticide (i.e., sterilant), the EPA uses a safety paradigm to evaluate use and prescribes safe use conditions on the product label, including personal protective equipment (PPE) in occupational settings to protect below the established safety thresholds.
In recent years, the release of ethylene oxide from sterilization operations into the ambient environment in various communities across the United States alerted regulators to the potential for community exposures from the sterilization process. Both commercial facilities and operational facilities (e.g., hospitals, dental clinics, etc.) routinely use ethylene oxide for the sterilization of heat-sensitive and other types of medical devices. Depending on the facility type, there may be separate chambers for fumigation (i.e., sterilization of the equipment) and aeration (i.e., desorption of ethylene oxide from the sterilized surface), whereas other facilities combine the processes in one chamber. The recognition of fugitive emission exposures from this process caught the attention of regulators and NGOs to highlight any potential human health risks from these facilities.
It is important to understand how environmental and community releases are occurring to gain a better picture of ambient ethylene oxide concentrations. The levels that are observed from ambient air monitoring by the EPA and others are presumed to originate from medical device sterilization facilities. The putative ways that ethylene oxide releases occur include but are not limited to ineffective industrial hygiene control mechanisms, incidental releases during sterilization, leakage of material due to deferred maintenance or other mechanical issues, or simply gross accidental releases. However, other known operations involving ethylene oxide should be investigated to determine how they contribute to ambient levels. Understanding the different ways ethylene oxide is released into the environment is crucial in addressing the issue. By identifying and investigating various operations that involve ethylene oxide generation, a comprehensive approach to determining its impact on ambient levels can be developed.
To dissect how ethylene oxide may increase someone’s health risk, the hazard and risk must first be understood along with the potential health endpoints. Generally, a hazard is a potential source of harm from a substance, whereas risk is the probability of potential adverse health effects resulting from exposure to a hazard. The degree to which risk exists is the driving factor when evaluating a hazard, including the contribution of exposure factors. For this reason, there have been a few attempts to assess ethylene oxide human health risk from ambient exposures.
The major focus on ambient environmental ethylene oxide exposure has been on chronic inhalation health outcomes. Inhalation is the primary pathway through which ethylene oxide can enter the body. While acute ethylene oxide exposure can cause health effects (i.e., neurotoxicity), those effects are thought to occur at higher levels than the cancer-causing effects. For this reason, regulators focus on the chronic exposure outcomes for many of the current ambient exposure situations.
In the mid-2010s, the EPA (via the National Air Toxics Assessment) identified ethylene oxide monitoring data in communities surrounding sterilization facilities (singled out via census tracts) and ultimately determined there was elevated carcinogenic risk in those areas. Along with updated scientific evaluations, the EPA concluded that the risk assessments at certain facilities showed an unacceptable level of risk based on the model input variables. Since the initiation of recent EPA risk assessments, there have been assessments performed by other regulatory agencies in an attempt to understand the carcinogenic risk presented by ethylene oxide from the ambient environment.
In 2012, the International Agency for Research on Cancer (IARC) identified that ethylene oxide is a group one carcinogen (carcinogenic to humans) based on its hazard assessment (the first step in determining if a substance is a human health risk driver). Findings by organizations like IARC typically initiate reviews and/or re-reviews by other organizations to perform their own hazard and risk assessments. Based on additional assessment, a number of different domestic organizations (e.g., EPA, National Toxicology Program (NTP)) have identified ethylene oxide as a human carcinogen based on regulatory review of the data and underlying studies (e.g., observational epidemiology, mechanistic data, and animal models). However, other organizations like the Texas Commission on Environmental Quality (TCEQ) have evaluated the data and determined that ethylene oxide is “likely to be carcinogenic to humans” (as opposed to “carcinogenic to humans” as the EPA and NTP have determined). This nuanced distinction highlights the different assessment processes that have been performed when striving to understand the carcinogenic potential. There are currently disagreements on the weight of evidence as to which cancer outcome is associated with ethylene oxide overexposure (i.e., lymphohematopoietic cancers versus breast cancer).
An evaluation of how the regulatory information meshes with the exposure and risk assessment data is also a complicated process. There have been differing opinions on the most accurate risk level because of the different methodologies used by the various agencies when assessing risk. Many recent risk assessments have focused on the nonoccupational “bystander” exposure scenarios (i.e., ambient environmental exposures). Some of these recent risk assessments further divide this into residential and nonresidential bystander site receptors (i.e., the exposed population). Quantitative risk assessments typically employ exposure modeling to understand how the site receptors (e.g., bystanders) would be exposed and to identify any elevated risk. Risk assessment models typically use the most sensitive endpoint for characterizing the cancer risk.
Ethylene oxide risk assessments have been evolving over the past decade. In a recent draft risk assessment, the EPA selected a certain linear risk assessment model to quantify risk; but in a 2023 risk assessment addendum, the EPA defaulted to its 2016 Integrated Risk Information System (IRIS) cancer characterization (TCEQ used a different model than the EPA and, thus, identified different risk levels). The variations in the methodology make pinpointing the exact risk in the population more difficult when using these conservative assessment methods.
The exposure parameters used in the risk assessment models are also important to determine individual exposures. The EPA based original modeling on census tract information for locations surrounding sterilization facilities. When performing the assessment, the default assumption for evaluating residential risk is to assume essentially continuous exposure to a carcinogenic substance (i.e., 24 hours a day for 365 days per year, which is then averaged over 70 years). These assessments are conservative to account for the upper bound of the possibility of cancer development. However, for a more realistic picture of an individual’s exposure and risk, the exposure parameters of that individual’s exposure must be identified. For example, if the individual travels outside of their residential setting during work and/or weekends, does the model take into consideration the reduced exposure duration? The standard risk assessment models also assume that there is continuous exposure to ambient outdoor air, even when not outdoors. The exposure model should account for any attenuation offered by residential conditions (e.g., home fresh air intakes, natural versus mechanical ventilation, etc.) for a more realistic exposure picture.
Along with realistic exposure parameters, data quality from ambient air sampling should be well understood along with the laboratory analytical methodologies. If relying on ambient monitoring versus personal monitoring (the type typically performed in occupational settings), are the data comparable? Discrepancies in data may complicate the modeling process. If there are established ambient limits, the user will need to ensure that the laboratory has a reliable method sensitive enough to detect down to that low level, or the data may be uninterpretable.
In addition to exposure parameters, the dose is also important to consider when assessing a causal link between exposure and health outcome (think Bradford Hill’s causal criteria). Generally, exposure to a chemical is the opportunity to come into contact with a substance and the opportunity to internalize a dose. Dose is the amount that actually is absorbed into the body and has the potential for interaction with biological systems. Dose is important, not only because of the other various sources of ethylene oxide exposure in the environment, but also because humans produce ethylene oxide within the body during natural metabolic processes. Are the exposure and any subsequent dose lower than what is naturally in our bodies? A comprehensive dose-response assessment would provide clarity on the effects of exposure. To further understand body burden, there are established methods for identifying ethylene oxide in the body using biomonitoring. For example, ACGIH has an established biological exposure index (BEI) for ethylene oxide exposure from occupational use. Both exposure and dose assessment provide a well-rounded ethylene oxide risk perspective.
The future for individual health claims as they relate to ethylene oxide exposure remains uncertain. The risk assessment paradigm continues to evolve for overall risk as well as site-specific understandings of risk.
Concurrent with the health risk assessment process, the federal process on reducing ambient emissions moves forward. On March 14, 2024, the EPA announced the ethylene oxide National Emission Standards for Hazardous Air Pollutants (NESHAP) final amendments for commercial sterilizers. The goal of this change is to reduce ethylene oxide emissions and lifetime cancer risk surrounding a number of commercial sterilization facilities. The sterilization amendments were closely followed by an April 9, 2024, EPA final rule aimed at reducing ethylene oxide emissions from chemical manufacturing facilities. Both of these rulemaking efforts demonstrate a targeted approach for reducing overall ethylene oxide community exposures.
However, as with any change, there can be downstream effects. Are any health claims made at the individual level supported by science? And how are changes in the sterilization processes affecting the medical equipment and devices that are implanted in humans? Are there risks for incomplete sterilization because of process restrictions or changes? It is important to continue to strive toward using the best scientific methods for evaluating exposure and risk to answer these causal questions.
