HEALTH EFFECTS

INTRODUCTION

The primary purpose of this chapter is to provide public health officials, physicians, toxicologists, and other interested individuals and groups with an overall perspective on the toxicology of ethylene oxide. It contains descriptions and evaluations of toxicological studies and epidemiological investigations and provides conclusions, where possible, on the relevance of toxicity and toxicokinetic data to public health. When available, mechanisms of action are discussed along with the health effects data; toxicokinetic mechanistic data are discussed in Section 3.1.

A glossary and list of acronyms, abbreviations, and symbols can be found at the end of this profile.

To help public health professionals and others address the needs of persons living or working near hazardous waste sites, the information in this section is organized by health effect. These data are discussed in terms of route of exposure (inhalation, oral, and dermal) and three exposure periods: acute (≤14 days), intermediate (15–364 days), and chronic (≥365 days).

As discussed in Appendix B, a literature search was conducted to identify relevant studies examining health effect endpoints. Figure 2-1 provides an overview of the database of studies in humans or experimental animals included in this chapter of the profile. These studies evaluate the potential health effects associated with inhalation, oral, or dermal exposure to ethylene oxide, but may not be inclusive of the entire body of literature. A systematic review of the scientific evidence of the health effects associated with exposure to ethylene oxide was also conducted; the results of this review are presented in Appendix C.

Animal inhalation studies are presented in Table 2-1 and Figure 2-2, and animal oral studies are presented in Table 2-2 and Figure 2-3; limited dermal data were identified for ethylene oxide.

Levels of significant exposure (LSEs) for each route and duration are presented in tables and illustrated in figures. The points in the figures showing no-observed-adverse-effect levels (NOAELs) or lowest-observed-adverse-effect levels (LOAELs) reflect the actual doses (levels of exposure) used in the studies. LOAELs have been classified into “less serious” or “serious” effects. “Serious” effects (SLOAELs) are those that evoke failure in a biological system and can lead to morbidity or mortality (e.g., acute respiratory distress or death). “Less serious” effects are those that are not expected to cause significant dysfunction or death, or those whose significance to the organism is not entirely clear. ATSDR acknowledges that a considerable amount of judgment may be required in establishing whether an endpoint should be classified as a NOAEL, “less serious” LOAEL, or “serious” LOAEL, and that in some cases, there will be insufficient data to decide whether the effect is indicative of significant dysfunction. However, the Agency has established guidelines and policies that are used to classify these endpoints. ATSDR believes that there is sufficient merit in this approach to warrant an attempt at distinguishing between “less serious” and “serious” effects. The distinction between “less serious” effects and “serious” effects is considered to be important because it helps the users of the profiles to identify levels of exposure at which major health effects start to appear. LOAELs or NOAELs should also help in determining whether or not the effects vary with dose and/or duration, and place into perspective the possible significance of these effects to human health. Levels of exposure associated with cancer (Cancer Effect Levels, CELs) of ethylene oxide are indicated in Tables 2-1 and 2-2 and Figures 2-2 and 2-3.

A User’s Guide has been provided at the end of this profile (see Appendix D). This guide should aid in the interpretation of the tables and figures for LSEs and MRLs.

The health effects of ethylene oxide have been evaluated in a number of occupational cohorts and a variety of animal studies. As illustrated in Figure 2-1, the inhalation exposure route was employed in the majority of animal studies; inhalation was assumed to be the predominant exposure route in the occupational cohort studies. The most examined endpoints in animal studies were body weight and neurotoxicity. Cancer was the most examined endpoint in epidemiological studies.

Human and/or animal studies suggest that relatively sensitive noncancer targets of ethylene oxide include respiratory, hematological, endocrine, neurological, reproductive, and developmental endpoints (see Appendix C for more information on the systematic review and for definitions of evidence levels). Human and animal studies have also reported carcinogenic effects.

• Respiratory Endpoints: Respiratory effects are a presumed health effect for humans based on a moderate level of evidence in workers and a high level of evidence in experimental animal studies. Compromised respiratory function has been reported in workers exposed to high levels of ethylene oxide. Inhalation studies in experimental animals have reported several respiratory effects including labored breathing, nasal discharge, pulmonary lesions, rhinitis, and pulmonary edema.
• Hematological Endpoints: Hematological effects are a suspected health effect for humans based on a moderate level of evidence in animal studies. Repeated exposure of experimental animals to ethylene oxide vapor has resulted in hematological effects such as decreases in hemoglobin, hematocrit, erythrocyte count, packed cell volume, and/or increased reticulocytes.
• Endocrine Endpoints: Adrenal gland effects are a suspected health effect in humans based on a moderate level of evidence in animal studies. Adverse effects (gross and/or histopathologic changes in the adrenal gland) have been observed in guinea pigs following acute- and/or chronic-duration exposure to ethylene oxide vapor.
• Neurological Endpoints: Neurological effects are a presumed health effect in humans based on a low level of evidence in occupational exposure studies and a high level of evidence in animal studies. Clinical signs of neurotoxicity (e.g., neuropathy, weakness in extremities, impaired hand-eye coordination, cognitive dysfunction, memory loss, headache, lethargy) were reported among workers exposed to ethylene oxide for various durations. Sural nerve biopsies revealed axonal degeneration and regeneration in two studies. Neurological effects such as ataxia, impaired sensory reflexes, hindlimb paralysis, and/or degenerative histopathologic lesions have been observed among laboratory animals exposed to ethylene oxide by inhalation.
• Reproductive Endpoints: Male reproductive effects are a presumed health effect in humans based on a high level of evidence in animal studies. Animal studies provide convincing evidence of ethylene oxide-induced effects on the male reproductive system (e.g., decreases in male reproductive organ weights, germ cell survival, and sperm count; histopathologic lesions).
• Developmental Endpoints: Ethylene oxide is a presumed developmental toxicant in humans based on animal studies that demonstrated ethylene oxide-induced developmental effects such as depressed fetal weight, delayed ossification, fetal fluid retention, and ocular defects.
• Cancer: The carcinogenicity of ethylene oxide has been evaluated in a number of cohorts involved in ethylene oxide production and/or uses in sterilization. Results from some cohort studies suggest that exposure to ethylene oxide may increase the risk of selected cancer types (e.g., lymphohematopoietic cancer, leukemia, breast cancer). In laboratory animals exposed by inhalation, ethylene oxide was associated with a variety of cancer types (leukemia, mesotheliomas, lymphomas, tumors of lungs, Harderian gland, and female mammary gland and reproductive organs). Forestomach cancer (at the application site) was reported in rats treated by the oral exposure route.

Figure 2-1 Overview of the Number of Studies Examining Ethylene Oxide Health Effects**Includes studies discussed in Chapter 2. A total of 124 studies (including those finding no effect) have examined toxicity; most studies examined multiple endpoints. Chronic exposure was considered the most prevalent exposure duration for human occupational studies. Most studies examined the potential body weight, neurological, and cancer effects of ethylene oxide Fewer studies evaluated health effects in humans than animals (counts represent studies examining endpoint) (Figure not reproduced here; see the original article.)

DEATH

Results from a number of epidemiological studies involving occupational exposure to ethylene oxide and risk of death from various cancer types (particularly all cancers, lymphohematopoietic cancer, leukemia, stomach cancer, pancreatic cancer, and brain and nervous system cancer) are summarized in Section 2.19.

Information regarding death in experimental animals following inhalation exposure to ethylene oxide is available. Limited information is available for the oral exposure route. No information was located for the dermal exposure route.

For rats exposed once to ethylene oxide vapor, 1-hour LC₅₀ values were 5,748 and 4,439 ppm for males and females, respectively; 4-hour LC₅₀ values were 1,972 and 1,537, respectively (Snellings et al. 2011). A 4-hour LC₅₀ of 1,460 ppm was reported for male rats (Jacobson et al. 1956). Mice and dogs were more sensitive to acute lethality than rats; reported 4-hour LC₅₀ values were 835 and 960 ppm, respectively (Jacobson et al. 1956). Lethality occurred in rats, mice, rabbits, guinea pigs, and monkeys following single or repeated inhalation exposure at 204–841 ppm for periods up to 6 weeks (Hollingsworth et al. 1956; Jacobson et al. 1956; NTP 1987). Studies that employed repeated exposure of rats for up to 2 years reported increased mortality at exposures as low as 100 ppm (Garman et al. 1986; Lynch et al. 1984a, 1984b; Snellings et al. 1984b).

Hollingsworth et al. (1956) reported 100% mortality among rats (5/sex) gavaged once at 200 mg/kg. Dunkelberg (1982) reported decreased length of survival among 50 female rats gavaged at 30 mg/kg/day, 2 times/week, for up to 150 weeks.

BODY WEIGHT

No studies were located regarding body weight effects in humans following inhalation, oral, or dermal exposure to ethylene oxide.

Information for the inhalation exposure route is available for body weight effects in experimental animals. Limited information is available for the oral exposure route. No information was located for the dermal exposure route.

Decreased weight gain was observed in rats 14 days after a single exposure to 500 ppm ethylene oxide; weight gain was decreased by 16 and 12% in males and females, respectively, compared to controls (EPA 2005a). Depressed body weight was reported in a few acute-duration studies of rats or mice repeatedly exposed to ethylene oxide vapor at levels in the range of 357–1,200 ppm (Hollingsworth et al. 1956; Saillenfait et al. 1996). Following a 14-week repeated exposure to 200 ppm ethylene oxide, weight gain was decreased by 16 and 12% in male and female rats, respectively (EPA 2005b). Decreases in final body weights (10–20%) were observed in rats and guinea pigs exposed to 113–204 ppm ethylene oxide for up to 226 days (Hollingsworth et al. 1956). Hollingsworth et al. (1956) also reported “markedly subnormal growth” (magnitude not reported) among rats, mice, rabbits, guinea pigs, and/or monkeys intermittently exposed to ethylene oxide vapor at 357 ppm for up to 85 days. Other acute- and intermediate-duration inhalation studies of rats, mice, and/or rabbits found no body weight effects at exposure levels in the range of 100–500 ppm (EPA 1994; Jacobson et al. 1956; Matsuoka et al. 1990; Mori et al. 1991a, 1991b; Neeper-Bradley and Kubena 1993; NTP 1987; Ohnishi et al. 1985; Saillenfait et al. 1996; Snellings et al. 1982a, 1982b). In 2-year rat studies that employed repeated exposure to ethylene oxide vapor, 12–18% depressed body weight gain was reported at an exposure level of 100 ppm (Lynch et al. 1984a, 1984b; Snellings et al. 1984b). No body weight effects were observed for mice repeatedly exposed at up to 100 ppm for 102 weeks (NTP 1987).

Hollingsworth et al. (1956) found no body weight effects among rats gavaged once at 100 mg/kg/day. Repeated dosing of female rats at 100 mg/kg/day resulted in an unspecified magnitude of weight loss; the NOAEL was 30 mg/kg/day.

RESPIRATORY

Limited data are available regarding ethylene oxide-related respiratory effects in humans. Inhalation of ethylene oxide is irritating to mucous membranes including those associated with the respiratory system. Inhalation exposure of workers to high concentrations of ethylene oxide for brief periods has resulted in bronchitis, pulmonary edema, and emphysema (Thiess 1963). Deschamps et al. (1992) reported a case of persistent asthma in a 35-year-old male who was highly exposed to ethylene oxide from a leaky railroad tank. Pulmonary function remained compromised when tested 1 and 3 years after the accidental exposure. There was no evidence of increased risk of death from non-malignant respiratory disease within various cohorts of workers involved in production or use of ethylene oxide (Bisanti et al. 1993; Coggon et al. 2004; Hogstedt 1988; Morgan et al. 1981; Steenland et al. 1991; Swaen et al. 2009; Wong and Trent 1993).

Information is available regarding respiratory effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Dyspnea was observed after 4 hours of exposure of mice to ethylene oxide vapor at a lethal exposure level of 800 ppm (NTP 1987). Adverse respiratory effects (e.g., dyspnea, pulmonary edema, pulmonary hemorrhage and congestion, “severe lung injury”) were reported for experimental animals (rats, mice, and/or guinea pigs) exposed to 357–841 ppm ethylene oxide vapor for acute durations (Hollingsworth et al. 1956; NTP 1987). Rhinitis was also observed in mice exposed to 400 ppm ethylene oxide for up to 2 weeks (NTP 1987). Intermediate-duration studies reported labored breathing and nasal discharge in rats exposed to 406 ppm for 6 weeks (Jacobson et al. 1956), an increase in relative lung weight in rats and guinea pigs exposed to 113–204 ppm for up to 226 days (Hollingsworth et al. 1956), rhinitis in mice exposed to 200 ppm for up to 14 weeks (NTP 1987), and pulmonary congestion and alveolar collapse in dogs exposed to 292 ppm for 6 weeks (Jacobson et al. 1956). Acute bronchopneumonia, chronic pneumonia, pulmonary edema, and suppurative rhinitis were observed in rats exposed at 50 ppm for 104 weeks (Lynch et al. 1984a, 1984b). However, all groups of rats in this study (including controls) experienced a pulmonary bacterial infection as early as 8 months into the treatment period and were treated at months 8, 16, and 20. The infection likely played a significant role in the reported respiratory effects. There were no indications of exposure-related respiratory effects in rats exposed to 100 ppm as part of a 2-generation reproduction study (EPA 1994) or in mice repeatedly exposed at up to 100 ppm for 102 weeks (NTP 1987).

CARDIOVASCULAR

Limited data are available regarding ethylene oxide-related cardiovascular effects in humans. There was no evidence of increased risk of death from cardiovascular or cerebrovascular diseases within various cohorts of workers involved in production or use of ethylene oxide (Bisanti et al. 1993; Coggon et al. 2004; Hagmar et al. 1991; Hogstedt 1988; Kiesselbach et al. 1990; Morgan et al. 1981; Olsen et al. 1997; Steenland et al. 2004; Swaen et al. 2009; Wong and Trent 1993).

Limited information is available regarding cardiovascular endpoints in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Lynch et al. (1984b) found no effects on electrocardiograms of male monkeys during repeated exposure to ethylene oxide vapor at 50 or 100 ppm for up to 2 years. There was no histopathological evidence of ethylene oxide-induced cardiovascular effects in mice repeatedly exposed at 100 ppm for 102–104 weeks (Lynch et al. 1984a, 1984b; NTP 1987).

GASTROINTESTINAL

Limited information is available regarding the potential for ethylene oxide-induced gastrointestinal effects in humans. Nausea and vomiting have been reported, but these are considered to be secondary effects due to neurotoxicity rather than a primary effect of inhaled ethylene oxide on the gastrointestinal tract. There was no evidence of increased risk of death from gastrointestinal effects within cohorts of workers with potential for ethylene oxide exposure (Bisanti et al. 1993; Morgan et al. 1981; Steenland et al. 1991; Wong and Trent 1993).

Limited information is available regarding gastrointestinal effects in experimental animals following inhalation or oral exposure to ethylene oxide. No information was located for the dermal exposure route.

There was no histopathological evidence of gastrointestinal effects in rats or mice repeatedly exposed to ethylene oxide vapor at 100 ppm for 102–104 weeks (Lynch et al. 1984a, 1984b; NTP 1987). Hollingsworth et al. (1956) reported gastric irritation in female rats following repeated gavage exposure at 100 mg/kg/day; the NOAEL was 30 mg/kg/day.

HEMATOLOGICAL

Limited human data are available. Joyner (1964) reported no effects on hemoglobin levels or red blood cell (RBC) or white blood cell (WBC) counts in workers exposed to ethylene oxide at about 5–10 ppm for approximately 10 years.

Information is available regarding hematological effects in experimental animals following inhalation or oral exposure to ethylene oxide. No information was located for the dermal exposure route.

Decreases in hemoglobin, hematocrit, erythrocyte count, packed cell volume, and/or increased reticulocytes were reported in experimental animals (rats, mice, dogs) repeatedly exposed to ethylene oxide vapor at 250–500 ppm for 6–13 weeks (Fujishiro et al. 1990; Jacobson et al. 1956; Snellings et al. 1984a). Hematological alterations indicative of normochromatic anemia were also observed in dogs exposed to 102 ppm for 6 months (Jacobson et al. 1956). One study reported splenic extramedullary hematopoiesis in rats repeatedly exposed at 50 or 100 ppm for 104 weeks and noted splenic focal fibrosis at 100 ppm; however, findings are confounded by a concurrent infection in the rat colony (Lynch et al. 1984a, 1984b). There was no histopathological evidence of ethylene oxide-induced hematological effects in mice intermittently exposed at 100 ppm for up to 102 weeks (NTP 1987).

Hollingsworth et al. (1956) found no evidence of exposure-related hematological effects in female rats repeatedly gavaged with ethylene oxide at up to 30 mg/kg/day.

MUSCULOSKELETAL

No studies were located regarding musculoskeletal effects in humans exposed to ethylene oxide.

Information is available regarding musculoskeletal effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Histopathologic indicators of muscular atrophy were reported for experimental animals (rats, rabbits, monkeys, dogs) repeatedly exposed to ethylene oxide vapor for 6 weeks to 226 days at exposure levels in the range of 204–357 ppm (Hollingsworth et al. 1956; Jacobson et al. 1956). There was no indication of musculoskeletal effects in mice repeatedly exposed to ethylene oxide vapor for up to 14 weeks at 250 ppm (Snellings et al. 1984a). However, multifocal myopathy in skeletal muscle was reported in rats following inhalation exposure at 100 ppm for 104 weeks; the NOAEL was 50 ppm (Lynch et al. 1984a, 1984b). See Section 2.15 for a discussion of neuromuscular alterations.

HEPATIC

Information regarding hepatic effects in humans after inhalation exposure to ethylene oxide is limited to a report by Joyner (1964). The results suggested that workers exposed to about 5–10 ppm for 10.7 years did not have major signs of hepatic toxicity such as jaundice or palpable liver.

Limited information is available regarding hepatic effects in experimental animals following inhalation or oral exposure to ethylene oxide. No information was located for the dermal exposure route.

Hollingsworth et al. (1956) reported slight discoloration and fatty degeneration in livers from rats and guinea pigs exposed to ethylene oxide vapor at 841 ppm for two or three 7-hour exposures. Snellings et al. (1984a) reported an elevation in the liver to body weight ratio in female mice exposed to ethylene oxide at 250 ppm for 11 weeks; however, histological examination showed that the livers were normal at this and all other lower exposure levels for both sexes in this study. No evidence for hepatic effects were seen among rats or mice repeatedly exposed at 100 ppm during a 10-week premating period (EPA 1994) or for 50 or 100 ppm for up to 102–104 weeks (Lynch et al. 1984a, 1984b; NTP 1987).

Slight liver damage (no further details) was reported in rats following repeated gavage exposure to ethylene oxide at 100 mg/kg/day, but not in animals dosed at ≤30 mg/kg/day (Hollingsworth et al. 1956).

RENAL

Information regarding renal effects in humans after inhalation exposure to ethylene oxide is limited to a report by Joyner (1964) which indicates that there was no evidence of nephritis or other parenchymal disease among workers exposed to ethylene oxide at 5–10 ppm for a mean exposure time of 10.7 years.

Information is available regarding renal effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Hollingsworth et al. (1956) reported enlargement, slight congestion, and cloudy swelling of convoluted tubules in kidneys from rats and guinea pigs exposed to ethylene oxide vapor at 841 ppm for two or three 7-hour exposures. In a 14-week study of mice repeatedly exposed to ethylene oxide vapor, renal tubular degeneration was observed in 5/10 and 6/10 males exposed at 100 and 200 ppm, respectively, and in 8/10 females exposed at 200 ppm (NTP 1987). Renal tubular necrosis was observed in 1/10 males of the 100 and 200 ppm exposure groups. Lethal exposure levels (400 and 600 ppm) resulted in increased incidences of necrosis in both sexes exposed for up to 2 weeks. There were no indications of renal effects in rats or mice exposed at 50 or 100 ppm for up to 102–104 weeks (Lynch et al. 1984a, 1984b; NTP 1987).

DERMAL

Data related to human dermal exposure to ethylene oxide are generally associated with case reports of industrial accidents, some of which occurred in the 1930s and 1940s. Concentrated ethylene oxide evaporates rapidly from the skin and produces a freezing effect, often compared to frostbite, leaving burns ranging from first- to third-degree severity (Taylor 1977). Workers drenched with a 1% solution developed large vesiculated blisters (Sexton and Henson 1949).

A study using volunteers by Sexton and Henson (1950) showed that the magnitude of skin injury was related to the concentration of ethylene oxide in solution but peaked at about 50%. This was attributed to the rapid evaporation of the more concentrated solutions, which prevented more prolonged skin contact.

Case reports of patients whose intact skin or wounds had contact with gauze or other hospital supplies that had been sterilized with ethylene oxide indicated that the observed skin reactions included erythema, blister formation, scaling, crusted ulcerations, and second-degree burns (Alomar et al. 1981; Hanifin 1971; Karacalar and Karacalar 2000).

Shupack et al. (1981) demonstrated that human skin reactions to ethylene oxide in patch materials were directly related to the total dose.

Limited information is available regarding dermal effects in experimental animals following inhalation or dermal exposure to ethylene oxide. No information was located for the oral exposure route.

NTP (1987) found no evidence of dermal effects in mice repeatedly exposed to ethylene oxide vapor for up to 102 weeks at up to 100 ppm. Dermal application of ethylene oxide (10 and 50% aqueous solutions) to rabbits for ≥6 minutes resulted in hyperemia and edema (Hollingsworth et al. 1956). Dermal application of undiluted ethylene oxide to the back of rabbits (0.5 mL for 4 hours under occluded conditions) resulted in severe erythema and edema, subdermal hemorrhages, and chemical burns during 72 hours post-application; the chemical was considered corrosive (Celanese Chem Co. 1972).

OCULAR

There is some evidence that occupational exposure to high levels of ethylene oxide can result in cataracts and corneal burns (McLaughlin 1946; Thiess 1963). For example, cataracts developed in four sterilizer operators who were exposed to ethylene oxide from a leaking sterilizer for up to 2 months (Gross et al. 1979; Jay et al. 1982). Because these persons could intermittently smell the fumes, a level of ≥700 ppm was estimated by the authors in retrospect. Although none of the patients were examined before this accidental exposure, the occurrence of cataracts was viewed as unlikely to be a chance occurrence in all four persons in this age range (31–35 years old) who had no systemic or ocular disease that might be associated with cataract formation.

Ocular instillation of a 1% solution of ethylene oxide-to eyes of rabbits resulted in slight corneal cloudiness (severity score 0.8–1.9; maximum possible score of 4.0) during 48 hours post-instillation (McDonald et al. 1977). Lynch et al. (1984a, 1984b) reported cataracts in monkeys and rats intermittently exposed to ethylene oxide vapor for 2 years. Incidences in 50 and 100 ppm groups of monkeys were 2/11 and 3/12, respectively, compared to 0/11 among controls. Incidences in 50 and 100 ppm groups of rats were 3/79 and 9/78, respectively, compared to 2/77 among controls.

ENDOCRINE

No information was located regarding endocrine effects in humans associated with ethylene oxide exposure.

Information is available regarding endocrine effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Hollingsworth et al. (1956) reported pale coloration and enlargement of adrenals, and numerous fat vacuoles in the adrenal cortex from rats and guinea pigs exposed two or three times to ethylene oxide vapor at 841 ppm for 7 hours per exposure. Lynch et al. (1984a, 1984b) reported multifocal cortical vacuolation and hyperplasia in adrenal glands from rats intermittently exposed to ethylene oxide vapor at 50 ppm for up to 104 weeks; however, findings are confounded by a concurrent infection in the rat colony. There was no histopathological evidence of ethylene oxide-induced effects on the thyroid, parathyroid, adrenals, or pituitary gland of mice intermittently exposed at 100 ppm for up to 102 weeks (NTP 1987).

IMMUNOLOGICAL

The immunological effects of human inhalation exposure to ethylene oxide were studied in workers employed for up to 14 years in an ethylene oxide manufacturing plant. Workplace concentrations were generally <0.05 ppm (the detection limit of the analytical method) with occasional peaks of 8 ppm during the 4 years that the air was monitored. There was no effect on any of the blood parameters relating to immune function that were investigated, including T and B lymphocyte counts, lymphocyte activation, and serum IgG, IgM, and IgA levels (Van Sittert et al. 1985).

Thiess (1963) did not observe skin sensitization in ethylene oxide plant workers (average exposure: 10.4 years) who were challenged with a single dermal application of 1% ethylene oxide. However, ethylene oxide was implicated as a skin sensitizer in studies of volunteers following dermal exposure (Sexton and Henson 1950; Shupack et al. 1981). Contact dermatitis and delayed-type hypersensitivity dermatitis have been observed in case reports of ethylene oxide-exposed health care workers and patients (Alomar et al. 1981; Belen and Polat 2015; Brashear et al. 1996; Caroli et al. 2005; Dagregorio and Guillet 2004; Kerre and Goossens 2009; Lerman et al. 1995; Romaguera and Vilaplana 1998).

Limited information is available regarding immunological effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Thymic lymphocytic hypoplasia was reported for male and female mice intermittently exposed to ethylene oxide vapor at 400 ppm for up to 2 weeks; at 600 ppm, necrosis was observed in the thymus (males and females) and spleen (males) (NTP 1987).

NEUROLOGICAL

Neurological effects have frequently been reported in association with human and animal exposure to ethylene oxide via inhalation at a wide range of concentrations and exposure durations.

In humans exposed to ethylene oxide in occupational settings, headache, nausea, and vomiting were reported (Blackwood and Erskine 1938; Sexton and Henson 1949; von Oettingen 1939). Reliable measured or estimated exposure levels were not available in these situations.

Case studies of neurological effects in workers exposed to ethylene oxide have been reported. These studies are insufficient to establish a causal relationship between exposure to ethylene oxide and neurological effects in humans. Neuropathy, impaired hand-eye coordination, cognitive dysfunction, memory loss, headache, and hand numbness were reported in case studies of workers exposed to ethylene oxide for various durations (Brashear et al. 1996; Crystal et al. 1988; Dretchen et al. 1992; Estrin et al. 1987; Finelli et al. 1983; Kuzuhara et al. 1983; Salinas et al. 1981; Schröder et al. 1985; Zampollo et al. 1984). These effects were seen at estimated average exposure levels as low as 3 ppm; however, short-term exposures may have been as high as 700 ppm for some of these workers. Sural nerve biopsies revealed axonal degeneration and regeneration in two studies (Kuzuhara et al. 1983; Schröder et al. 1985).

Information on the neurological effects of inhalation exposure to ethylene oxide has also been derived from case studies of longer-term occupational exposure. Headaches, nausea, vomiting, clumsiness, blunting of the senses, lethargy, numbness, and weakness in the extremities were reported among four sterilizer operators exposed to ethylene oxide for up to 2 months on an intermittent basis at levels of approximately 700 ppm (estimated by the authors based on the fact that the exposed workers could smell the vapors emitted from a leaking apparatus) (Gross et al. 1979). One of the operators experienced recurrent major motor seizures at 20–30-minute intervals near the end of the work shift; nerve conduction testing indicated sensorimotor neuropathy.

Information is available regarding neurological effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Incoordination and semiconsciousness were reported during a 4-hour exposure of mice to ethylene oxide vapor at 1,600 ppm (NTP 1987). Decreased alertness and motor activity was decreased in male rats following a single exposure to 300 ppm and in female rats exposed to 500 ppm (EPA 2005a). Repeated inhalation exposures of experimental animals resulted in neurological effects at similar or lower exposure levels. Effects including impaired sensory and motor function (particularly in hindlimbs), decreased grip strength, altered gait, slight tremors, various degrees of hindlimb paralysis, and peripheral neuropathy have been reported in experimental animals intermittently exposed to ethylene oxide vapor at 100–500 ppm for periods in the range of 48–226 days (EPA 2005b; Hollingsworth et al. 1956; Jacobson et al. 1956; Kaido et al. 1992; Matsuoka et al. 1990; Ohnishi et al. 1985). Snellings et al. (1984a) also reported neurological effects (hunched posture and decreased motor activity) in mice at exposures ≥50 ppm; however, these findings could not be specifically attributed to ethylene oxide exposure due to inadequate descriptions of methods and the limited evaluations of small numbers of animals.

In a 9-month study of rats exposed to ethylene oxide at 250 ppm, retarded growth and maturation of myelinated fibers and mild axonal degeneration in hindleg nerves were observed in the absence of clinical signs of neuropathy (Ohnishi et al. 1986). There were no indications of ethylene oxide exposure-related neurological effects in a 2-generation study of rats exposed to concentrations as high as 100 ppm (EPA 1994) or in a 2-year study of mice intermittently exposed to airborne concentrations as high as 100 ppm (NTP 1987). Lynch et al. (1984a) reported axonal dystrophy in the brain of 1/2 control monkeys and 2/2 and 1/2 monkeys exposed at 50 and 100 ppm, respectively. Demyelination was reported in 1/2 monkeys at the 50 and 100 ppm exposure levels. However, the results could not be specifically attributed to ethylene oxide exposure due to the small number of animals evaluated and the reported histopathologic brain lesion in a control animal.

Nagata et al. (1992) designed a study to investigate potential mechanisms of ethylene oxide neurotoxicity in the rat. Groups of male Wistar rats (5/group) were exposed to ethylene oxide vapor at 0 or 500 ppm for 6 hours/exposure, 3 days/week, for 15 weeks. Following the final exposure period, ³⁵S-methionine was injected into the right dorsal root ganglion to evaluate rapid anterograde axonal transport. The velocity in the ethylene oxide-exposed rats was 33% slower than that of controls. Morphometric analysis of selected portions of sural and peroneal nerve preparations revealed significantly greater incidental degeneration of myelinated fibers from the ethylene oxide-exposed rats than from controls. The study authors suggested that the morphological changes and decreased axonal transport velocity might play a causative role in the development of peripheral neuropathy from chronic ethylene oxide exposure.

REPRODUCTIVE

Possible associations between exposure to ethylene oxide and spontaneous abortion have been explored in epidemiological studies of sterilizer workers (Gresie-Brusin et al. 2007; Hemminki et al. 1982; Rowland et al. 1996). Limitations in these studies preclude drawing conclusions regarding the associations between ethylene oxide exposure and pregnancy outcomes.

Hemminki et al. (1982) evaluated possible associations between exposure to ethylene oxide and spontaneous abortion in a retrospective study of 1,443 sterilizer workers in hospitals in Finland. Information on exposures was obtained from questionnaires sent to supervising nurses and information on pregnancy outcomes and other potential confounding factors was obtained from worker self-surveys. Rates of spontaneous abortion were adjusted for age, parity, decade of pregnancy, smoking, and consumption of coffee and alcohol. Rates of spontaneous abortion were 15.1% in workers who were reported to have ethylene oxide exposure during pregnancy (n=545), 11.3% in workers whose exposure to ethylene oxide during pregnancy was recorded as uncertain (n=293), and 4.6% in workers without reported exposure to ethylene oxide during pregnancy (n=605). Estimates of variance on these rates were not reported; however, rates in the exposed and uncertain exposure groups were reported as significantly different from the group not exposed (p<0.001). Rates were significantly higher (p<0.01) in workers who reported that they were exposed to ethylene oxide but not to glutaraldehyde or formaldehyde (16.1%, n=1,068). Exposure levels were not measured in this study; however, surveys of Finnish hospital sterilizing units found 8-hour weighted mean concentrations that ranged from 0.1 to 0.5 ppm with a highest measured concentration of 250 ppm. Major limitations in this study include absence of measured exposure levels and reliance on self-administered questionnaires for data on exposure, outcomes, and potential confounders.

Rowland et al. (1996) evaluated possible associations between ethylene oxide exposure and the risk of spontaneous abortion, preterm birth, and post-term birth in a retrospective study of 1,320 female dental assistants. The study included 32 sterilizer operators and 1,288 referents with no reported exposure to ethylene oxide. Information on exposures, pregnancy outcomes, and other potential confounding factors was obtained from self-surveys. After adjusting for age, relative risks (RRs) were 2.5 (95% confidence interval [CI] 1.0–6.3) for spontaneous abortion, 2.7 (95% CI 0.8–8.8) for preterm birth, and 2.1 (95% CI 0.7–5.9) for post-term birth. The RR for any of these outcomes was 2.5 (95% CI 1.0–6.1) after adjustment for age and exposure to nitrous oxide and elemental mercury during preparation of mercury amalgam restorations. The major limitations of this study were absence of measurements of exposure levels and reliance on self-reporting for data on potential exposures to ethylene oxide, outcomes, and potential confounding factors.

Gresie-Brusin et al. (2007) evaluated risks of spontaneous abortion and pregnancy loss in a retrospective study of 98 singleton pregnancies among women with ethylene oxide exposure in sterilizing units of 22 hospitals in South Africa. The study subjects were grouped according to “high” exposure (sterilizer operators, n=19) and “low” exposure (not directly involved in ethylene oxide sterilization, n=79). The median level of ethylene oxide measured with personal monitors of sterilizer operators was below the limit of detection (0.01 ppm) and the mean was 1.03 ppm (standard deviation [SD] 4.2). Information on pregnancy outcomes and other potential confounding factors was obtained from surveys conducted by trained interviewers. Prevalence odds ratios (PORs) were 20.8 (95% CI 2.1–199.3; 4 of 19 in the high-exposure group) for risk of spontaneous abortion and 8.6 (95% CI 1.8–43.7; 6 of 19 in the high-exposure group) for pregnancy loss. The study explored various potential confounders, including age and height, pregnancy rank and gestation length, education, antenatal care, emotional stress, smoking (active and passive), exposure to carbon monoxide and anesthetic gases, alcohol consumption, high blood pressure, diabetes and other medical conditions, and physical activity. Of these, maternal height, antenatal care, and emotional stress were associated with exposure to ethylene oxide and none were associated with pregnancy outcomes. The main limitation of this study was its relatively small size (19 in the high-exposure group), which may have contributed to the relatively wide CIs on the PORs.

Information is available regarding reproductive effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Increased incidence of resorptions, decreased numbers of pups per litter, and decreased numbers of fetuses born relative to numbers of implantation sites were reported in a study of female rats intermittently exposed to ethylene oxide vapor at 150 ppm for 3 weeks premating and during gestation days 1–16 (NIOSH 1982). In a 2-generation study, increased post-implantation loss was observed in F0 rats exposed to 33 ppm (EPA 1994) and decreased numbers of live pups per litter were observed at 100 ppm in the F1 and F2 generations (EPA 1994). Exposure-related effects on male reproductive organs (decreases in testicular and epididymal weights, germ cell survival, sperm count; histopathologic lesions in seminiferous tubules) have been reported for rats following intermittent inhalation exposure to ethylene oxide vapor for 6–13 weeks at exposure levels in the range of 250–500 ppm (Kaido et al. 1992; Mori et al. 1991a, 1991b). Intermittent inhalation exposure of monkeys to ethylene oxide vapor at 50 ppm (the lowest exposure level tested) for 24 months resulted in decreases in sperm count (28% less than controls) and motility (32% less than controls); however, reproductive function was not tested (Lynch et al. 1984a). There was no evidence of histopathological lesions in reproductive organs of rats or mice intermittently exposed for 102–104 weeks at 100 ppm (highest exposure level tested) (Lynch et al. 1984a, 1984b; NTP 1987).

DEVELOPMENTAL

No studies were located regarding developmental effects in humans after inhalation exposure to ethylene oxide.

Information is available regarding developmental effects in experimental animals following inhalation exposure to ethylene oxide. No information was located for oral or dermal exposure routes.

Decreases in fetal body weight and crown-rump length and increased incidence of reduced ossification were reported following intermittent exposure of maternal rats to ethylene oxide vapor at 150 ppm (the only exposure level tested) for 3 weeks premating and during gestation days 1–16 (NIOSH 1982). Depressed fetal weight (3–9% less than controls) was noted in other studies of maternal rats intermittently exposed to ethylene oxide vapor in the range of 100–800 ppm during gestation (Neeper-Bradley and Kubena 1993; NIOSH 1982; Saillenfait et al. 1996; Snellings et al. 1982a). Decreases in the number of pups/litter and the ratio of fetuses born to implantation sites were reported in a study of rats intermittently exposed to ethylene oxide vapor at 100 ppm for 12 weeks prior to mating and throughout gestation and lactation periods (Snellings et al. 1982b). In a 2-generation study, decreased pup body weight was observed in the F1 and F2 generations at 33 and 100 ppm (EPA 1994). Saillenfait et al. (1996) reported increased incidence of dilation in the renal pelvis and ureter of rat fetuses following intermittent maternal exposure at 1,200 ppm during gestation days 6–15. Fetal defects (predominantly hydrops and ocular defects) were reported following maternal exposure of mice to ethylene oxide vapor at 1,200 ppm for a single 1.5-hour exposure at timepoints between 1 and 25 hours postmating (Rutledge and Generoso 1989). There were no indications of ethylene oxide exposure-related developmental effects in rabbit fetuses following intermittent inhalation exposure of their mothers at 150 ppm during gestation days 7–19 or 1–19 (NIOSH 1982).

OTHER NONCANCER

No information was located regarding other noncancer effects in humans or animals following exposure to ethylene oxide.

CANCER

Carcinogenicity Assessments. The HHS has classified ethylene oxide as known to be a human carcinogen (NTP 2016) based on sufficient evidence of carcinogenicity from studies in humans (increased risk of cancer in workers exposed to ethylene oxide during its synthesis, production, and use), and evidence for a common mechanism of carcinogenesis in humans and experimental animals (similar genetic damage in cells of animals and workers exposed to ethylene oxide).

EPA (2016) characterized ethylene oxide as “carcinogenic to humans” by the inhalation exposure route, based on the total weight of evidence. The lines of evidence included “(1) strong, but less than conclusive on its own, including epidemiological evidence of lymphohematopoietic cancers and breast cancer in EtO- [ethylene oxide] exposed workers, (2) extensive evidence of carcinogenicity in laboratory animals, including lymphohematopoietic cancers in rats and mice and mammary carcinomas in mice following inhalation exposure, (3) clear evidence that EtO is genotoxic and sufficient weight of evidence to support a mutagenic mode of action for EtO carcinogenicity, and (4) strong evidence that the occurrence of key precursor events are anticipated to occur and progression to tumors in humans and progress to tumors, including evidence of chromosome damage in humans exposed to EtO.” EPA derived unit risk estimates were based on results from the National Institute for Occupational Safety and Health (NIOSH) study (Steenland et al. 2003, 2004) that evaluated cancer risk in a cohort of workers exposed to ethylene oxide. The authors found positive exposure responses for breast cancer mortality in females for log cumulative exposure and a 20-year lag. Odds ratios (ORs) increased with increasing cumulative exposure level, with elevated ORs in the highest quartile (>123,22 ppm-days), 3.13 (1.42, 6.92). There was also a positive exposure-response for lymphoid tumors (non-Hodgkin’s lymphoma, myeloma, and lymphocytic leukemia) in both sexes for cumulative exposure, with elevated ORs in males in the highest quartile (>123,22 ppm-days), 3.76 (1.03, 13.64). The adult-based unit risk estimates were 2.6×10^-3 per µg/m³ 4.8×10^-3 per ppb) for lymphoid cancer, 7.0×10^-4 per µg/m³ (1.3×10^-3 per ppb) for female breast cancer (15-year lag), and 3.0×10^-3 per µg/m³ (5.5×10^-3 per ppb) for both cancer types combined. Application of standard age-dependent adjustment factors yields a full lifetime total cancer unit risk estimate of 5.0×10^-3 per µg/m³ (9.1×10^-3 per ppb) (EPA 2016). The commensurate lifetime chronic (lower bound) exposure level of 2×10^-4 µg/m³ (1×10^-4 ppb) corresponds to an increased cancer risk of 10^-6 (1 in 1,000,000). The unit risk estimate was developed for environmental ethylene oxide exposures up to about 40 µg/m³ (20 ppb) and is not applicable to higher exposure levels that may occur in occupational exposure scenarios. Maximum likelihood estimates of extra risk of lymphoid cancer and breast cancer (combined) occupational exposure scenarios in the range of 0.1–1 ppm for an 8-hour time-weighted average (TWA) for 35 years range from 0.037 to 0.11 (upper bound estimates 0.081–0.22).

IARC has produced several reports on the carcinogenicity of ethylene oxide (IARC 1976, 1985, 1987, 1994, 2008, 2012). IARC has designated ethylene oxide as carcinogenic to humans (Group 1) (IARC 1987, 2012) based on limited evidence for a causal association between exposure to ethylene oxide and lymphatic and hematopoietic cancers and breast cancer in humans, sufficient evidence for carcinogenicity in animals, and strong evidence for a genotoxic mechanism of action for ethylene oxide carcinogenicity.

The Texas Commission on Environmental Quality (TCEQ 2020) evaluated the carcinogenicity of ethylene oxide and concluded that ethylene oxide is likely to be carcinogenic to humans. The TCEQ also determined that the epidemiological data support an association between ethylene oxide exposure and lymphohematopoietic tumors but do not support an association with breast cancer.

Occupational Studies. A number of epidemiological studies have examined possible associations between occupational exposure to ethylene oxide and risk of cancer. Ethylene oxide was first produced using a chlorohydrin process (reaction of ethylene gas with hypochlorous acid to produce ethylene chlorohydrin, which was reacted with calcium oxide to produce ethylene oxide). Workers involved with this process were exposed to a variety of chemicals, including organochlorine byproducts. The chlorohydrin process was inefficient and was gradually replaced with a direct vapor phase oxidation process (oxidation of ethylene gas in the presence of oxygen and silver catalyst at 10–30 atmospheres of pressure). Workers involved in the oxidation process did not experience exposures to the various chemicals used or produced during the chlorohydrin process. Available epidemiological studies include evaluation of cancer risk among cohorts of production workers involved in the chlorohydrin process and/or the ethylene oxidation process. Sterilization processes using ethylene oxide do not include exposures to the variety of chemicals encountered in the chlorohydrin process of ethylene oxide production.

As noted above, IARC and EPA have evaluated ethylene oxide for carcinogenicity (EPA 2016; IARC 2008, 2012). Both agencies concluded that the most convincing evidence for increased risk among workers exposed to ethylene oxide is for lymphohematopoietic cancers and female breast cancer. The discussion of the epidemiological database in this toxicological profile focuses on evaluations of ethylene oxide and risk of lymphohematopoietic cancers, leukemia, myeloma/multiple myeloma, non-Hodgkin’s lymphoma, lymphosarcoma/reticulosarcoma, and breast cancer. Results from a few epidemiological studies and animal studies suggested associations between ethylene oxide exposure and cancer at other sites (e.g., stomach, pancreas, nervous system).

Selected study details for cohorts evaluated for possible associations between exposure to ethylene oxide and selected cancer endpoints are presented in Table 2-3. Study results for leukemia, non-Hodgkin’s lymphoma, breast cancer, lymphohematopoietic cancer, myeloma/multiple myeloma, and lymphosarcoma/reticulosarcoma are summarized in Figures 2-4, 2-5, 2-6, 2-7, 2-8, and 2-9, respectively. Note that lymphosarcomas and reticulosarcomas are old cancer classifications; these cancer types are now classified as non-Hodgkin’s lymphoma. However, the cancer types discussed below are classified as reported by study authors.

Figures 2-4, 2-5, 2-6, 2-7, 2-8, and 2-9 include information on cohort size, exposure via production or use in sterilization processes, number of observed cancers/expected number, and the plotted risk estimates with 95% CIs. Standard mortality ratios and standard incidence ratios compare rates of cancer deaths or incidence rates in the study population with national rates of cancer death or incidence. Analyses comparing different groups within the same study population to national rates of cancer death or incidence may result in different findings. The footnote (¹) is used to identify cohorts exposed during ethylene oxide production via the chlorohydrin process. The footnote (²) is used to identify cohorts exposed during ethylene oxide production via the nonchlorohydrin process. Some studies included cohorts that may have been exposed during production of ethylene oxide by the chlorohydrin process in earlier years and by direct oxidation during later years; therefore, these cohorts were not assigned a footnote. The study of Kiesselbach et al. (1990) did not specify a method of ethylene oxide production. A bias in occupational studies is the “healthy worker effect” which can result in a worker population having a lower mortality or morbidity rate compared to the general population resulting in bias towards the null.

Meta-Analyses of Epidemiological Studies. Shore et al. (1993) conducted a review and meta-analysis on cohorts from previously-published studies. The meta-estimate of effect size was the sum of observed cancer deaths for all studies divided by the sum of expected cancer deaths for all studies (E/O ratio, synonymous with standardized mortality ratio [SMR]). This metric gives more weight to studies that have larger numbers of expected cancers. CIs on E/O were calculated using an approximate variance (V) estimate, 1/E; e^{ln(E/O)-1.96·sqrt(V)}. Variance was adjusted for χ² heterogeneity. The meta-analysis for leukemia was conducted using data from Bisanti et al. (1993), Gardner et al. (1989), Hagmar et al. (1991), Hogstedt (1988), Kiesselbach et al. (1990), Teta et al. (1993), Wong and Trent (1993), and an unpublished study (update to Greenberg et al. 1990). The reported meta-O/E ratio was 1.06 (95% CI 0.73–1.48) based on 31 incident cases or deaths versus 29.31 expected. No trend in O/E was evident when the meta-data were analyzed by exposure frequency, duration, or cumulative exposure. The meta-analysis for non-Hodgkin’s lymphoma was conducted using data from Bisanti et al. (1993), Gardner et al. (1989), Hagmar et al. (1991), Hogstedt (1988), Teta et al. (1993), Thiess et al. (1981), Wong and Trent (1993), and an unpublished study (update to Greenberg et al. 1990). The reported meta-O/E ratio was 1.35 (95% CI 0.93–1.90) based on 31 incident cases or deaths versus 22.93 expected. No trend in O/E was evident when the meta-data were analyzed by exposure frequency, duration, or cumulative exposure.

Teta et al. (1999) updated the meta-analysis of Shore et al. (1993) by including results from Hagmar et al. (1995) and Olsen et al. (1997) and excluding results from Greenberg et al. (1990), Morgan et al. (1981), and Thiess et al. (1981). The meta-estimate of the effect size was the same as that used in Shore et al. (1993), reported as SMR, rather than E/O. The updated meta-analysis included nearly 33,000 workers and >800 cancers. For leukemia and non-Hodgkin’s lymphoma risks, reported SMRs were 1.08 (95% CI 0.61–1.93) based on 35 deaths versus 32 expected, and 1.34 (95% CI 0.96–1.89) based on 33 deaths versus 25 expected, respectively; neither leukemia nor non-Hodgkin’s lymphoma results exhibited positive trends with duration, intensity, or latency. No trend in SMR was evident when the meta-data were analyzed by exposure frequency, duration, or cumulative exposure. The study authors considered the results inconclusive for leukemia and for non-Hodgkin’s lymphoma.

Marsh et al. (2019) conducted a systematic literature review of occupational exposure to ethylene oxide and risk of lymphohematopoietic cancer and breast cancer and identified 13 studies that were included in a meta-analysis. Marsh et al. (2019) limited their quantitative analysis to comparisons of national or broad geographic rates of cancer death or incidence (e.g., SMR/SIR) rather than estimating exposure-risk relationships within each cohort. Overall meta-relative risks (meta-RRs) were 1.48 (95% CI 1.07–2.05) for lymphohematopoietic cancer and 0.97 (95% CI 0.80–1.18) for breast cancer. For lymphohematopoietic cancer, the study authors reported meta-RRs of 1.46 (95% CI 0.85–2.50) among ethylene oxide production workers and 1.07 (95% CI 0.87–1.30) among ethylene oxide sterilization workers. Higher risks of lymphohematopoietic cancer were noted for earlier published studies, compared to more recent studies. Marsh et al. (2019) considered studies published in the 2000s to have used more sound experimental methods and are, therefore, more informative than earlier studies.

Case Reports. Hogstedt et al. (1979a) reported 3 cases of leukemia within a group of 230 workers at a Swedish facility where hospital equipment was sterilized using ethylene oxide. The study authors estimated an 8-hour time-weighted average (TWA) ethylene oxide level of 20±10 ppm in a storage hallway, which was considered higher than levels in the sterilization area. According to national statistics, 0.2 leukemia cases would have been expected. In a 5-year update of the study, 4 cases of lymphohematopoietic cancer (0.3 expected) were observed among 203 of the workers employed ≥1 year (Hogstedt et al. 1986).

Tompa et al. (1999) reported 8 cases of breast cancer within a group of 98 nurses exposed to ethylene oxide for 5–15 years in a hospital in Hungary. Reported ethylene oxide concentrations in the working area ranged from 5 to 150 mg/m³ (2.75–82.5 ppm). However, the study did not account for natural low dose radioactivity from radon in local drinking water, genetic predisposition, or effects of potential environmental, occupational, and/or lifestyle confounders.

Swaen et al. (1996) investigated a cluster of 10 cases of Hodgkin’s lymphoma within a large chemical manufacturing complex in Belgium. Among 214 different chemical substances evaluated, 5 chemicals exhibited elevated ORs (ammonia, benzene, ethylene oxide, sodium hydroxide, and oleum). The cluster of cases could not be specifically associated with ethylene oxide exposure.

Figure 2-4 Summary of Studies Evaluating Leukemia in Workers Exposed to Inhaled Ethylene Oxide 1 = production via chlorohydrin process; CI = confidence interval; EO = ethylene chloride; N = cohort size; NS = not specified; O/E = observed deaths (incidences)/expected; SIR = standardized incidence ratio; SMR = standardized mortality ratio; Production = workers involved in ethylene oxide production; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Figure 2-5 Summary of Epidemiological Studies Evaluating Non-Hodgkin’s Lymphoma in Workers Exposed to Inhaled Ethylene Oxide CI = confidence interval; N = cohort size; NS = not specified; O/E = observed deaths (incidences)/expected; SIR = standardized incidence ratio; SMR = standardized mortality ratio; Production = workers involved in ethylene oxide production; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Figure 2-6 Summary of Epidemiological Studies Evaluating Breast Cancer in Workers Exposed to Ethylene Oxide* *SMRs/SIRs are presented in the figure for ease of comparison across multiple studies and datasets. Additional analyses below use internal comparisons, which may reduce potential confounding from the healthy worker effect. Mikoczy et al. (2011)a found a positive association between inhaled ethylene oxide and breast cancer incidence: Incidence rate ratio (IRR) for upper exposure quartiles (Q) versus Q1-2: Q3: 2.76 (95% CI: 1.20, 6.33); Q4: 3.55 (95% CI: 1.58, 7.93).Steenland et al. (2004) found a positive association between inhaled ethylene oxide and breast cancer mortality: Cox regression coefficient for cumulative exposure with 20-year lag, (trend: p=0.01); odds ratio for upper quartile (Q) versus unexposed: Q4 3.13 (95% CI: 1.42, 6.92).Steenland et al. (2003) found positive associations between inhaled ethylene oxide and breast cancer incidence: Cox regression coefficient for log cumulative exposure with 15-year lag, (trend: p=0.05); odd ratio for upper quartile (Q) versus unexposed: Q5 1.87 (95% CI: 1.12–3.10). CI = confidence interval; N = cohort size; NS = not specified; O/E = observed deaths (incidences)/expected; RR = rate ratio; SIR = standardized incidence ratio; SMR = standardized mortality ratio; Production = workers involved in ethylene oxide production; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Figure 2-7 Summary of Epidemiological Studies Evaluating Lympho-Hematopoietic Cancer in Workers Exposed to Inhaled Ethylene Oxide* *SMRs/SIRs are presented in the figure for ease of comparison across multiple studies and datasets. Additional analyses noted below use internal comparisons, which may reduce potential confounding from the healthy worker effect. Steenland et al. (2004) found a positive association between inhaled ethylene oxide and lympho-hematopoietic cancer: Cox regression coefficient for continuous log cumulative exposure (trend: p = 0.02); odds ratio for upper quartiles (Q) versus unexposed: Q4(males): 3.76 (95% CI: 1.03, 13.64). 1 = production via chlorohydrin process; CI = confidence interval; EO = ethylene oxide; N = cohort size; NS = not specified; O/E = observed deaths (incidences)/expected; PO = propylene oxide; SIR = standardized incidence ratio; SMR = standardized mortality ratio; Production = workers involved in ethylene oxide production; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Figure 2-8 Summary of Epidemiological Studies Evaluating Myeloma-Multiple Myeloma in Workers Exposed to Inhaled Ethylene Oxide CI = confidence interval; N = cohort size; O/E = observed deaths (incidences)/expected; SIR = standardized incidence ratio; SMR = standardized mortality ratio; Production = workers involved in ethylene oxide production; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Figure 2-9 Summary of Epidemiological Studies Evaluating Lymphosarcoma/Reticulosarcoma in Workers Exposed to Inhaled Ethylene Oxide CI = confidence interval; N = cohort size; NS = not specified; O/E = observed deaths (incidences)/expected; SMR = standardized mortality ratio; Use = workers exposed via ethylene oxide sterilization process (Figure not reproduced here; see the original article.)

Animal Studies. Animal studies examining the carcinogenicity of ethylene oxide have reported increased incidences of several tumor types. Significantly increased incidence of brain gliomas (5/79 versus 0/76 controls; p<0.05) and peritoneal mesotheliomas (21/79 versus 3/78 controls; p<0.01) were reported among male Fischer 344 rats intermittently exposed by inhalation for up to 2 years at 100 ppm (Lynch et al. 1984a, 1984b). An increase in the incidence of mononuclear cell leukemia was also observed at 50 ppm (38/79 versus 24/77 controls, p<0.01).

Neoplastic changes reported in rats following exposure to ethylene oxide vapor for up to 2 years included splenic mononuclear cell leukemia, peritoneal mesothelioma, subcutis fibroma, and primary brain neoplasms (Garman et al. 1985, 1986; Snellings et al. 1984b). Significantly increased incidence of spleen mononuclear cell leukemia was observed in female rats exposed to ethylene oxide vapor for up to 2 years at 100 ppm (58% versus 8 or 11% controls; p<0.01). Although the incidence was only statistically significantly elevated at 100 ppm, increased incidences were also observed at 10 ppm (11/54 versus 11/115 in controls) and 33 ppm (14/48). When incidences among rats dying early or killed in moribund condition were included, there were significant mortality-adjusted trends for mononuclear cell leukemia in males and females at ≥33 ppm (incidence data not reported). Male rats of the 100 ppm exposure group exhibited significant increases in peritoneal mesothelioma when rats that died or were killed in moribund condition were included (no incidence reported) and subcutis fibroma (36% versus 2 and 4% control; 0.01>p>0.001). A significant trend for primary brain neoplasms was observed in males (p<0.01) and females (p<0.05).

In a 2-year inhalation study of male and female B6C3 F1 mice, ethylene oxide exposure at 100 ppm resulted in significantly increased incidence of alveolar/bronchiolar carcinoma (16/50 versus 6/50 controls, p=0.048 according to life table analysis) in males, and significantly increased incidences of alveolar/bronchiolar adenoma (17/49 versus 2/49 controls; p=0.001) and alveolar/bronchiolar carcinoma (7/49 versus 0/49 controls, p=0.019) in female mice (NTP 1987). Incidences of Harderian gland papillary cystadenoma were increased at 50 ppm in males (9/44, p=0.014) and females (6/46, p=0.052) and at 100 ppm in males (8/42, p=0.021) and females (8/47, p=0.039). The 100 ppm group of female mice also exhibited marginally significantly increased incidence of malignant lymphoma in the hematopoietic system (22/49 versus 9/49 controls; p=0.049) and uterine adenocarcinoma (5/49 versus 0/49 controls; p=0.058). The 50 ppm (but not 100 ppm) group of female mice also exhibited increased incidences of hepatocellular adenoma (8/48 versus 1/49; p=0.021) and mammary gland adenocarcinoma or adenosquamous carcinoma combined (8/48 versus 1/49 controls; p=0.020). Picut et al. (2003) re-evaluated the pathology and tumor incidence data from NTP (1987) and confirmed the results.

Significantly increased incidences of lung adenomas were reported among female A/J mice intermittently exposed to ethylene oxide vapor for 6 months at 70 or 200 ppm; lung adenoma incidences were 16/28 and 25/29, respectively, compared to 8/30 controls (Adkins et al. 1986). Incidences were statistically significantly increased in both ethylene oxide-exposed groups. In a replicate study that included controls and 200 ppm groups, incidences of lung adenomas in surviving mice were 9/29 (29%) and 12/28 (42%), respectively.

Dose-related increased incidence of malignant tumors of the forestomach (the application site for oral gavage) was reported among female rats administered ethylene oxide by gavage at 7.5 or 30 mg/kg/day for 2 days/week for 150 weeks; incidences were 8/50 and 31/50, respectively, compared to no stomach tumors among 50 vehicle and 50 untreated controls (Dunkelberg 1982). A total of 37 of the 39 tumors were squamous cell carcinomas.

In a lifetime skin painting study, application of a 10% solution of ethylene oxide to the backs of mice did not result in skin tumors (Van Duuren et al. 1965).

GENOTOXICITY

Ethylene oxide has been demonstrated to be genotoxic in human and animal studies in vivo and in a wide variety of test systems in vitro. Extensive reviews are available regarding the genotoxicity of ethylene oxide (see EPA 2016; IARC 1994, 2008, 2012). Studies evaluating the genotoxicity of ethylene oxide in humans and in vivo studies using experimental test species are summarized in Tables 2-4 and 2-5, respectively. Results from selected in vitro assays are summarized in Table 2-6. Available data collectively demonstrate the mutagenicity and clastogenicity of ethylene oxide both in vitro and in vivo. Ethylene oxide induced gene mutation, chromosomal aberrations, sister chromatid exchange, micronucleus formation, deoxyribonucleic acid (DNA) strand breaks, unscheduled DNA synthesis, and cell transformation in vitro. Ethylene oxide induced gene mutation, specific locus mutation, chromosomal aberrations, sister chromatid exchange, micronucleus formation, dominant lethal mutation, and heritable translocation in test species and/or occupationally-exposed humans. Although some conflicting results were observed in occupational studies, results of human studies support that ethylene oxide is genotoxic in humans. Preston (1999) noted that results may vary due to numerous confounding factors, including the time of testing relative to exposure. However, despite some negative results in human studies, IARC (2012) concluded the following regarding the genotoxicity of ethylene oxide: “Ethylene oxide consistently acts as a mutagen and clastogen at all phylogenetic levels, it induces heritable translocations in the germ cells of exposed rodents, and a dose-related increase in the frequency of sister chromatid exchange, chromosomal aberrations and micronucleus formation in the lymphocytes of exposed workers.”

In addition to these genotoxic effects, in vitro studies in mammal tissues, in vivo studies in rats and mice, and studies in humans have demonstrated the formation of DNA adducts. Ethylene oxide is an alkylating agent that forms adducts with DNA, ribonucleic acid (RNA), and proteins. The primary DNA adduct formed is N7-(2-hydroxyethyl)guanine (7-HEG). Other DNA adducts have been found in lesser amounts, these include N3-(2-hydroxyethyl)adenine and O⁶-(2-hydroxyethyl)guanine (EPA 2016; IARC 2012). 7-HEG has been detected in various tissues of rats exposed via inhalation to ethylene oxide for up to 4 weeks (Walker et al. 1990). Duration-related increases were observed in the brain, lungs, spleen, kidneys, leukocytes, liver, and testes. In addition to DNA adducts produced from exposure to exogenous ethylene oxide, DNA adducts can also form from endogenously produced ethylene oxide.