United States
Environmental Protection
Agency
Office of Health and
Environmental Assessment
Washington DC 20460
EPA/600/8-84/009F
June 1985
Final Report
Research and Development
&EPA
Health Assessment
Document for
Ethylene Oxide
Final
Report
-------
EPA-600/8-84-009F
Health Assessment Document
for Ethylene Oxide
U S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Research Triangle Park, NC 2771 1
-------
DISCLAIMER
This document has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
11
-------
PREFACE
The Office of Health and Environmental Assessment has prepared this
health assessment to serve as a "source document" for EPA use. The health
assessment document was originally developed for use by the Office of Air
Quality Planning and Standards to support decision-making regarding possible
regulation of ethylene oxide as a hazardous air pollutant. However, the scope
of this document has since been expanded to address multimedia aspects.
In the development of the assessment document, the scientific literature
has been inventoried, key studies have been evaluated and summary/conclusions
have been prepared so that the chemical's toxicity and related characteristics
are qualitatively identified. Observed effect levels and other measures of
dose-response relationships are discussed, where appropriate, so that the
nature of the adverse health responses are placed in perspective with observed
environmental levels.
iii
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS
AUTHORS;
The EPA's Office of Health and Environmental Assessment (OHEA) is
responsible for the preparation of the health assessment document. The OHEA
Environmental Criteria and Assessment Office (ECAO-RTP) had overall
responsibility for coordination and direction of the document (Dr. Robert M.
Bruce, Project Manager). The chapters addressing physical and chemical
properties, sampling and analysis, air quality, and biological effects in
animals and man were originally written and revised by Syracuse Research
Corporation with the exception of chapters or sections addressing
mutagenicity, teratogenicity and reproductive effects and carcinogenicity.
The air quality chapters (5, 6, 7) were reviewed by Radian Corporation under
contract to the Office of Air Quality Planning and Standards and recommen-
dations proposed.
The principal authors of the chapters or sections prepared by Syracuse
Research Corporation are:
D. Anthony Gray, Ph.D.
Life and Environmental Sciences Division
Syracuse Research Corporation
Syracuse, NY
Bruce Harris, Ph.D
Life and Environmental Sciences Division
Syracuse Research Corporation
Syracuse, NY
Stephen Bosch
Life and Environmental Sciences Division
Syracuse Research Corporation
Syracuse, NY
IV
-------
Joseph Santodonato, Ph.D, CIH
Life and Environmental Sciences Division
Syracuse Research Corporation
Syracuse, NY
The OHEA Carcinogen Assessment Group (CAG) was responsible for
preparation of the sections on carcinogenicity. Participating members of the
CAG are listed below (principal authors of present carcinogenicity materials
are designated by an asterisk(*).
Roy E. Albert, M.D. (Chairman)
Elizabeth L. Anderson, Ph.D.
Larry D. Anderson, Ph.D.
Steven Bayard, Ph.D.*
David L. Bayliss, M.S.
Robert P. Beliles, Ph.D.
Chao W. Chen, Ph.D.
Margaret M.L. Chu, Ph.D.
Herman J. Gibb, B.S., M.P.H.*
Bernard H. Haberman, D.V.M., M.S.
Charalingayya B. Hiremath, Ph.D.
Robert E. McGaughy, Ph.D.
Jean C. Parker, Ph.D
Dharrn V. Singh, D.V.M. , Ph.D.* -
Todd W. Thorslund, Sc.D.
The OHEA Reproductive Effects Assessment Group (REAG) was responsible for
the preparation of sections on rnutagenicity, teratogenicity and reproductive
effects. Participating members of REAG are listed below (principal authors of
present sections are indicated by an asterisk). The Environmental Mutagen
Information Center (EMIC) , in Oak Ridge, TN, identified literature bearing on
the nutagenicity of ethylene oxide.
Eric Clegg, Ph.D.*
John R. Fowle, III, Ph.D.*
Ernest R. Jackson, M.S.
Casey Jason, M.D.
David Jacobson-Kram, Ph.D.
K.S. Lavappa, Ph.D.
Sheila L. Rosenthal, Ph.D.
-------
Carol N. Sakai, Ph.D.*
Carmella Tellone, B.S.*
Vicki-Vaughan Dellarco, Ph.D.
Peter E. Voytek, Ph.D (Director)
REVIEWERS
The following individuals provided peer review of this draft or earlier
drafts of this document:
U.S. Environmental Protection Agency
Karen Blanchard
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC
Robert M. Bruce, Ph.D.
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Research Triangle Park, NC
James W. Falco, Ph.D.
Office of Health and Environmental Assessment
Exposure Assessment Group
Washington, D.C.
Lester D. Grant, Ph.D.
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Research Triangle Park, NC
Joseph Padgett
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC
William E. Pepelko
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Cincinnati, OH
Jerry F. Stara, D.V.M.
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Cincinnati, OH
VI
-------
Consultants and Reviewers
I.W.F. Davidson, Ph.D.
Bowman Gray School of Medicine
Wake Forest University
Winston Salem, NC
Larry Fishbein, Ph.D.
National Center for Toxicological Research
Jefferson, AR
Richard N. Hill, M.D., Ph.D.
Office of Toxic Substances
U.S. Environmental Protection Agency
Washington, DC
Derek Hodgson, Ph.D.
University of North Carolina
Chapel Hill, NC
George R. Hoffman, Ph.D.
Holy Cross College
Worcester, MA
Rudolph J. Jaeger, Ph.D.
Consultant Toxicologist
7 Bogert Place
Westwood, NJ
Marshall Johnson, Ph.D.
Thomas Jefferson Medical College
Philadelphia, PA
Edmond J. LaVoie, Ph.D.
American Health Foundation
Valhalla, NY
P.O. Lotilaker, Ph.D.
Fels Research Institute
Temple University Medical Center
Philadelphia, PA
Sam Shibko, Ph.D.
Health and Human Services
Division of Toxicology
Washington, DC
Charles M. Sparacino, Ph.D.
Research Triangle Institute
Research Triangle Park, NC
Vll
-------
Danial S. Straus, Ph.D.
University of California
Riverside, CA
Robert Tardiff, Ph.D.
1423 Trapline Court
Vienna, VA
Norman M. Trieff, Ph.D.
University of Texas Medical Branch
Department of Pathology, UTMB
Galveston, TX
Benjamin Van Duuren, Ph.D.
New York University Medical Center
550 First Avenue
New York, NY
Janes R. Withey, Ph.D.
Department of National Health and Welfare
Tunney's Pasture
Ottawa, Ontario
Canada, KIA 01Z
SCIENCE ADVISORY BOARD
ENVIRONMENTAL HEALTH COMMITTEE
The content of this health assessment document on ethylene oxide was
independently peer-reviewed in public session by the Environmental Health
Committee of the Environmental Protection Agency's Science Advisory Board.
CHAIRMAN, ENVIRONMENTAL HEALTH COMMITTEE
Dr. Herschel E. Griffin, Professor of Epidemiology, Graduate School of
Public Health, 6505 Alvarado Road, San Diego State University, San Diego,
California 92182-0405
EXECUTIVE SECRETARY. SCIENCE ADVISORY BOARD
Dr. Daniel Byrd III, Executive Secretary, Science Advisory Board,
A-101 F, U.S. Environmental Protection Agency, Washington, D. C. 20460
Vlll
-------
MEMBERS
Dr. Seymour Abrahamson, Professor of Zoology and Genetics, Department of
Zoology, University of Wisconsin, Madison, Wisconsin 53706
Dr. Morton Corn, Professor and Director, Division of Environmental Health
Engineering, School of Hygiene and Public Health, The Johns Hopkins
University, 615 N. Wolfe Street, Baltimore, Maryland 21205
Dr. John Doull, Professor of Pharmacology and Toxicology, University of
Kansas Medical Center, Kansas City, Kansas 66103
Dr. Jack D. Hackney, Chief, Environmental Health Laboratories, Professor
of Medicine, Rancho Los Amigos Hospital Campus of the University of
Southern California, 7601 Imperial Highway, Downey, California 90242
Dr. Marvin Kuschner, Dean, School of Medicine, Health Science Center,
Level 4, State University of New York, Stony Brook, New York 11794
Dr. Daniel Menzel, Director and Professor, Pharmacology and Medicine,
Director, Cancer Toxicology and Chemical Carcinogenesis Program, Duke
University Medical Center, Durham, North Carolina 27710.
Dr. Steven M. Rappaport, Associate Professor of Industrial Hygiene,
School of Public Health, Department of Biomedical and Environmental
Health Sciences, University of California, Berkeley
Dr. Michael J. Symons, Professor, Department of Biostatistics, School of
Public Health, University of North Carolina, Chapel Hill, North Carolina
27711
Dr. D. Warner North, Principal, Decision Focus Inc., Los Altos Office
Center, Suite 200, 4984 El Camino Real, Los Altos, California 94022
Dr. Bernard Weiss, Professor, Division of Toxicology, P.O. Box RBB,
University of Rochester, School of Medicine, Rochester, New York 14642
Dr. Ronald Wyzga, Electric Power Research Institute, 3412 Hillview
Avenue, P.O. Box 1041, Palo Alto, California 94303
Dr. Edward F. Ferrand, Assistant Commissioner for Science and Technology,
New York City Department of Environmental Protection, 51 Astor Place, New
York, New York 10003
Dr. Ronald D. Hood, Professor, Developmental Biology Section, Department
of Biology, The University of Alabama, and Principal Associate, R.D. Hood
and Associates, Consulting Toxicologists, P.O. Box 1927, University,
Alabama 35486
IX
-------
SPECIAL CONSULTANTS
Dr. Robert R. Maronpot, Head, Experimental Pathology, National Toxicology
Program, NIEHS, P.O. Box 12233, Research Triangle Park, North Carolina
27709
Dr. Walderico M. Generoso, Senior Scientist, Biology Division, Oak Ridge
National Laboratory, Oak Ridge, Tennessee 37830
-------
TABLE OF CONTENTS
LIST OF TABLES xv
LIST OF FIGURES xix
LIST OF ABBREVIATIONS xx
1 . SUMMARY AND CONCLUSIONS 1-1
2. INTRODUCTION 2-1
3. PHYSICAL AND CHEMICAL PROPERTIES 3-1
3.1 SYNONYMS AND CAS NUMBER 3-1
3.2 STRUCTURAL AND MOLECULAR FORMULAS 3-1
3.3 TORSIONAL ANGLES AND BOND DISTANCES 3-1
3.4 PHYSICAL PROPERTIES OF PURE ETHYLENE OXIDE 3-2
3.4.1 Description 3-2
3.4.2 Molecular Weight 3-2
3.4.3 Melting Point 3-2
3 .4.4 Boiling Point 3-2
3.4.5 Boiling Point Change with Pressure Change 3-2
3.4.6 Density 3-2
3.4.7 Coefficient of Cubical Expansion 3-2
3.4.8 Refractive Index 3-3
3.4.9 Vapor Pressure 3-3
3.4.10 Aqueous Solubility 3-3
3-4.11 Freezing Point of Aqueous Solutions 3-4
3.4.12 Boiling Point of Aqueous Solutions 3-4
3.4.13 Flash Point 3-4
3.4.14 Flash Point of Aqueous Solutions 3-5
3.4.15 Explosive Limits in Air, Voluras % 3-5
3.4.16 Heat of Combustion at 25°C 3-5
3.4.17 Log Octanol/Water Partition Coefficient 3-5
3.4.18 Ultraviolet Spectroscopic Data 3-5
3.5 PHYSICAL PROPERTIES AND DESCRIPTION OF COMMERCIAL
ETHYLENE OXIDE 3-5
3.6 CHEMICAL PROPERTIES 3-7
3.6.1 Reduction 3-7
3.6.2 Clathrate Formation 3-7
XI
-------
TABLE OF CONTENTS (cont. )
Page
3.6.3 Polymerization 3-7
3.6.4 Other Reactions 3-8
3.6.5 Hydrolysis and Related Reactions 3-8
3.6.6 Free Radical Reactions 3-17
4. SAMPLING AND ANALYTICAL METHODS 4-1
4 .1 SAMPLING 4-1
4 .2 ANALYSIS 4-6
5. SOURCES IN THE ENVIRONMENT 5-1
5 .1 PRODUCTION 5-1
5.1.1 Quantities Produced 5-1
5.1.2 Producers, Production Sites, and Distribution 5-1
5.1.3 Production Methods and Processes 5-1
5.2 USES OF ETHYLENE OXIDE 5-13
5.2.1 Ethylene Glycol. 5-13
5.2.2 Nonionic Surface-Active Agents 5-15
5.2.3 Di-, Tri-, and Polyethylene Glycols 5-15
5.2.4 Glycol Ethers 5-16
5.2.5 Ethano la mines 5-16
5.2.6 Miscellaneous Applications.., 5-16
5.2.7 Discontinued Uses of Epoxides 5-18
5.2.8 Projected or Proposed Uses 5-18
5.2.9 Alternatives to Uses for Ethylene Oxide. 5-18
5.3 POTENTIAL FOR ENVIRONMENTAL CONTAMINATION 5-19
5.3.1 Air Emissions from Production. <. 5-19
5.3.2 Handling, Transport, and Storage 5-22
5.3.3 Potential Environmental Formation 5-23
5 .4 SUMMARY 5-25
6. ENVIRONMENTAL FATE, TRANSPORT, AND DISTRIBUTION 6-1
6 .1 INTRODUCTION 6-1
6.2 ETHYLENE OXIDE FATE IN WATER 6-1
6.3 ETHYLENE OXIDE FATE IN SOIL 6-3
xn
-------
TABLE OF CONTENTS (cont. )
6.4 ETHYLENE OXIDE FATE IN THE ATMOSPHERE 6-4
6.5 DEGRADATION IN COMMODITIES AND MANUFACTURED PRODUCTS 6-6
6.6 BIO ACCUMULATION IN AQUATIC ORGANISMS 6-9
6.7 SUMMARY 6-9
7. ENVIRONMENTAL LEVELS AND EXPOSURE 7-1
7 .1 INTRODUCTION 7-1
7.2 ENVIRONMENTAL LEVELS 7-1
7.3 EXPOSURE 7-4
7 .4 SUMMARY 7-4
8. ECOLOGICAL EFFECTS 8-1
8.1 MICROORGANISMS AND INSECTS 8-1
8 .2 PLANTS 8-2
8 .3 AQUATIC ORGANISMS 8-2
9. BIOLOGICAL EFFECTS IN ANIMALS AND MAN 9-1
9 .1 PHARMACOKINETICS 9-1
9.1.1 Absorption 9-1
9.1.2 Distribution 9-1
9.1.3 Metabolism 9-3
9.1.4 Elimination 9-4
9.2 ACUTE, SUBCHRONIC, AND CHRONIC TOXICITY 9-6
9.2.1 Effects in Humns 9-6
9.2.2 Effects in Animals 9-14
9.2.3 Summary of Toxicity 9-24
9.3 TERATOGENICITY AND REPRODUCTIVE TOXICITY 9-26
9.3.1. Teratogenic Effects 9-26
9.3.2. Reproductive Effects 9-31
9.3.3. Testicular Effects 9-33
9.3.4. Adverse Reproductive Outcome in Humans 9-35
9.3.5. Summary of Teratogenic ity and Reproductive
Tox ic ity 9-38
Xlll
-------
TABLE OF CONTENTS (cont. )
9.4 MUTAGENICITY 9-H6
9.4.1. Gene Mutation Studies 9-i»9
9.4.2. ChromosonB Aberration Studies 9-64
9.4.3. Chromosome Mutations in Human Populations 9-77
9.4.4. Other Studies Indicative of Genetic Damage 9-81
9.4.5. Summary and Conclusions of the Mutagenicity of
Ethylene Oxide 9-92
9 .5 CARCINOGENICITY 9-94
9.5.1 Animal Studies 9-94
9.5.2 Epidemic logic Studies 9-118
9.5.3 Quantitative Estimation 9-136
9.5.4 Summary 9-162
9.5.5 Conclusions 9-168
Appendix 9A: Comparison of Results by Various Extrapolation
Models A1
Appendix 9B: International Agency for Research on Cancer
Classification System for the Evaluation of the
Carcinogenic Risk of Chemicals to Humans B1
10. REFERENCES 10-1
xiv
-------
LIST OF TABLES
Table
3-1
3-2
3-3
3-4
4-1
5-1
5-2
5-3
5-4
5-5
5-6
8-1
9-1
9-2
9-3
9-4
9-5
9-6
Hydrolysis Kinetics of Ethylene Oxide
Specific Rates of Reaction of Anions and Lewis Bases with
Ethylene Oxide
Breakthrough and Safe Sampling Volumes for Propylene Oxide
Ethylene Oxide Production
Ethylene Oxide Producers, Plant Sites, Capacities, Processes,
and Tec hnology
Ranges of Reaction Systems Variables in the Direct
Air -Oxidation of Ethylene Oxide
Ranges of Reaction Systems Variables in the Direct
Oxygen-Oxidation of Ethylene Oxide
Typical Vent Gas Composition for Both Air- and Oxygen-Based
Ethylene Oxide Plants
Acute Aquatic Toxicity of Ethylene Oxide
Acute Toxicity of Ethylene Oxide
Subchronic Toxicity of Ethylene Oxide
Summary of Studies on Teratogenicity and Reproductive
Tox ic ity
Summary of Mutagenicity Testing of Ethylene Oxide: Gene
Mutations in Bacteria
Summary of Mutagenicity Testing of Ethylene Oxide: Gene
Mutations Tests in Lower Plants. (Yeast)
Summary of Mutagenicity Testing of Ethylene Oxide: Mutation
Te st s in Higher Plants
Page
3-6
3-9
3-12
3-13
4-2
5-2
5-3
5-9
5-11
5-14
5-21
8-3
9-2
9-17
9-39
9-47
9-53
9-55
XV
-------
LIST OF TABLES (cont. )
Table
9-7
9-8
9-9
9-10
9-11
9-12
9-13
9-14
9-15
9-16
9-17
9-18
9-19
Summary of Mutagenicity Testing of Ethylene Oxide: Gene
Summary of Mutagenicity Testing of Ethylene Oxide: Mammalian
Cells in Culture
Summary of Mutagenicity Testing of Ethylene Oxide: Dominant
Lethal Tests
Summary of Mutagenicity Testing of Ethylene Oxide: Heritable
Summary of Mutagenicity Testing of Ethylene Oxide: Chromosome
Aberra tion Te st s
Summary of Mutagenicity Testing of Ethylene Oxide:
Micronucleus Tests
Summary of Mutagenicity Testing of Ethylene Oxide: Chromosome
Summary of Mutagenicity Testing of Ethylene Oxide: SCE
Formation in Human Populations
Summary of Mutagenicity Testing of Ethylene Oxide: SCE
Formation in Experimental Studies
Summary of Mutagenicity Testing of Ethylene Oxide: Unscheduled
DNA Synth esis
Design Summary for Carcinogenicity Testing of Ethylene Oxide
Tumor Induction by Intragastric Administration of Ethylene
Oxide in Female Sprague-Dawley Rats
Cumulative Percentages of Male Fischer 344 Rats that Died
or were Sacrificed in a Moribund Condition After Exposure
to Ethylene Oxide Vapor
Page
9-58
9-62
9-65
9-67
9-68
9-69
9-70
9-82
9-84
9-85
9-97
9-97
9-101
9-20 Cumulative Percentages of Female Fischer 344 Rats that Died
or were Sacrificed in a Moribund Condition After Exposure
to Ethylene Oxide Vapor 9-102
xvi
-------
LIST OF TABLES (cont. )
Table
9-21 Cumulative Percentages of Male Fischer 344 Rats that were
Alive at the Beginning of Month 17, but Died or ware
Sacrificed in a Moribund Condition After Subsequent
Exposure to Ethylene Oxide Vapor 9-103
9-22 Cumulative Percentages of Female Fischer 344 Rats that were
Alive at the Beginning of Month 17, but Died or were
Sacrificed in a Moribund Condition After Subsequent
Exposure to Ethylene Oxide Vapor 9-104
9-23 Summary of Selected Tumor Incidence Comparisons for Male
and Female Fischer 344 Rats Exposed to Ethylene Oxide for
Two Years 9-106
9-24 Ethylene Oxide 2-Year Vapor Inhalation Study: 24-Month Final
Sacrificed Frequency of Expo sure-Re la ted Neoplasms
for 110- to 116-Week-Old Fischer 344 Rats 9-107
9-25 Ethylene Oxide 2-Year Vapor Inhalation Study: Frequency of
Exposure Related Neoplasms at 24-Month Final Sacrificed and
in Fischer 344 Rats Dying Spontaneously or Euthanized
Wh en Mo rib und 9-108
9-26 Ethylene Oxide 2-Year Vapor Inhalation Study: Frequency of
Brain Neoplasms in Fischer 344 Rats 9-112
9-27 Ethylene Oxide 2-Year Vapor Inhalation Study: Frequency of
Primary Brain Neoplasms Types in Fischer 344 Rats 9-113
9-28 Leukemia Incidence in Male Fischer 344 Rats Exposed
to Ethylene Oxide for 2 years 9-115
9-29 Incidence of Neoplastic Lesions in Male Fischer 344
Rats Exposed to Ethylene Oxide for 2 years 9-117
9-30 Observed and Expected Number of Deceased Among 153 Women
and 50 Men with Continuous or Intermittent Exposure to
Ethylene Oxide 9-127
9-31 Comparison of Observed Numbers of Cancer Deaths in
BASF-Aktiengesellschaft, Ludwigshafen Plants 1928-80
for Persons Having 10 Years of Observation Following
Exposure to Alkylene Oxide With That Expected Based
on Mortality Statistics for Rhinehessia-Palatinate
1970-75, Ludwigshafen 1970-75, and the Federal
Republic of Germany 1971-74, By ICD Code and Cause
of Death 9-130
xvii
-------
LIST OF TABLES (cont.)
Table
9-32 Related Risks of Death From Cancer in the Alkylene Oxide
Cohort as Compared With the Styrene Cohort, By Age 9-132
9-33 Bushy Run Ethylene Oxide Inhalation Study in Fischer 344 Rats.
Incidence of Peritoneal Mssothelioma and Brain Glioma in Males,
and MDnonuclear Cell Leukemia and Brain Glioma in Females by
Dose Among Survivors to First Tumor. Maximum Liklihood
Estimates of Linear Term and 95$ Upper-Limit q-|* 9-150
9-34 NIOSH Ethylene Oxide Inhalation Study In Nfele Fischer 344 Rats.
Incidence of Peritoneal Mesothelioma and Brain Glioma by Dose
Among Total Examined. Estimates of 95% Upper-Limit Risk Based
on Human Equivalent Dose 9-154
9-35 Leukemia (ICD 204-207) Incidence and Nbrtality: Ethylene Oxide
Epidemiology. Included are Relative Risks, 95% Confidence
Limits, Nominal Exposure Estimates, and 95/8 Confidence Limits
on Unit Risk 9-156
9-36 Relative Carcinogenic Potencies Among 54 Chemicals Evaluated
by the Carcinogen Assessment Group as Suspect Human
Carcinogens 9-163
xviii
-------
LIST OF FIGURES
Figure Page
5-1 Schematic for Air-Based Ethylene Oxidation 5-6
9-1 Mutagenic Response of Salmonella typhimurium Strain
TA1535 Exposed to Ethylene Oxide 9-50
9-2 Mutagenic Response of CHO Cells to EtO 9-63
9-3 Percentages of Male and Female Fischer 344 Rats with
Histologically Confirmed Mononuclear Cell Leukemia
at 24-Month Sacrifice 9-111
9-4 Histogram Representing the Frequency Distribution of the
Potency Indices of 54 Suspect Carcinogens Evaluated
by the Carcinogen Assessment Group 9-161
xix
-------
LIST OF ABBREVIATIONS
BCF Bio concentration factor
BOD Biochemical oxygen demand
CNS Central nervous system
DMSO Dimethyl su If oxide
DNA Deoxyribonucleic acid
GC Gas chromatography
GSH Gluthathione
LC,.,. Concentration lethal to 50% of recipients
50
LDp.,. Dose lethal to 50% of recipients
MS Mass spec tome try
ppb Parts per billion
ppm Parts per million
PVC Poly vinyl chloride
SCE Sister chromatid exchange
TWA Time-weighted average
v/v VoluriE per volume
xx
-------
1. SUMMARY AND CONCLUSIONS
Ethylene oxide is a colorless, flammable gas at ambient temperature. It
is soluble in water (195 m£ vapor in 1 mSL water at 20°C) and is a highly
reactive chemical. The chemical reaction of ethylene oxide that is environ-
mentally significant is the hydrolytic reaction in aqueous media.
The most suitable chemical method available for the analysis of ethylene
oxide in the atmosphere is its collection on a sorbent cartridge and subsequent
analysis by gas chroraatography with flame ionization or mass spectrometric
detector. Biological samples are analyzed by the purge and trap method,
followed by gas chromatography with flame ionization or mass spectrometric
detection.
Ethylene oxide is produced almost exclusively by direct oxidation of
ethylene. Its 1981 production volume was 4937 million pounds. The largest
single use of ethylene oxide is in the synthesis of ethylene glycol. Small
amounts of ethylene oxide are used as a sterilant and in the manufacture of
pesticides, Pharmaceuticals, and medicinal devices.
The major emission sources from production facilities of ethylene oxide
are the main process vents and purge gas vents. Total air emission from
production in 1978 has been estimated to be 2 million pounds. Although only
small amounts of ethylene oxide are used for consumer products, this represents
a considerable potential for human exposure.
In aquatic media, ethylene oxide will degrade by hydrolysis with a half-
life of = 12-14 days. Evaporation from aquatic media will also be a significant
loss process. There is no conclusive evidence that microbial degradation is
significant in aquatic media. The fate of ethylene oxide in soil will probably
1-1
-------
be similar to that in water (its fate in the atmosphere is not obvious from
the available data). Available rate constants for its reaction with hydroxyl
radicals and oxygen atom (^P) and its reaction in smog chambers predict that
ethylene oxide will persist in the atmosphere. No reports measuring ambient
levels of ethylene oxide have been found. Only one report exists for its
detection in ambient aquatic media. Several studies have detected the presence
of ethylene oxide in commodities and commercial goods including food, medical
supplies and drugs.
The pharmacokinetics of ethylene oxide have not been studied extensively.
Only one study was found about the absorption of this chemical; it concerned
the inhalation exposure of rats. The toxicity data suggest that absorption
occurs via the respiratory and gastrointestinal tracts. During inhalation
exposure, the highest concentration of ethylene oxide was associated with the
protein fraction of the lungs, while ethylene oxide that reaches the systemic
circulation is distributed widely to various tissues (liver, kidney, lung,
testes, brain, spleen and intestinal mucosa). Ethylene oxide is eliminated
primarily by the kidneys with the metabolite, ethylene glycol, as well as
glutathione conjugates identified in the urine. Ethylene oxide also reacts
with cellular macromolecules, and reaction with DNA results in small quantities
of 7-hydroxyethylguanine in the urine. The half-life of ethylene oxide has
been estimated to be between = 10 and 30 minutes, indicating rapid removal of
absorbed compound. Macromolecular-bound products and metabolites such as
ethylene glycol are removed more slowly.
The effects of acute exposure of humans to ethylene oxide have been
described in case reports and, to a more limited extent, in control studies.
Case reports have indicated that headaches, nausea, vomiting, dyspnea and
1-2
-------
respiratory irritation occur from exposure to the vapors of ethylene oxide.
Dermal contact with the liquid, aqueous solutions or clothing containing
absorbed ethylene oxide results in skin burns and possibly sensitization.
Control studies in humans indicate that exposure to 2200 pptn of ethylene oxide
was slightly irritating, while exposure to 22,000 ppm adversely affected
membranes in the nose. Studies of dermal contact indicate that 50% solutions
are optimum for producing chemical burns. Some effects may be delayed several
hours subsequent to exposure.
In acute studies in laboratory animals, the LC5Q values for inhalation
exposure (rats, mice and dogs) range from 835-5000 ppm for a 4-hour exposure,
and the oral LD5Q value (rabbits, guinea pigs and rats) range from 100-631
mg/kg body weight. Gross symptoms of toxicity were respiratory irritation,
salivation, nausea, vomiting, diarrhea, convulsions, and death. Pathologic
findings included lung, liver, and kidney damage. Ethylene oxide was an
irritant in dermal studies, but failed to result in sensitization in guinea
pigs. Acute effects in experimental animals appear to be similar to those
reported in case reports of human exposure to ethylene oxide.
The subacute and chronic effects of ethylene oxide in man are not well-
documented nor readily available from clinical case reports. Case reports of
workers exposed repeatedly to ethylene oxide indicate that neurotoxicity
occurred consistent with sensorimotor neuropathy. Some signs of neuropathy
appeared to persist after cessation of exposure. In occupational epidemiology
studies, the non-neoplastic results were reported to be an increase in
lymphocyte count, a decrease in hemoglobin value (levels of exposure were not
reported) and an increased incidence of deaths resulting from diseases of the
circulatory system (exposure level <50 mg/m-*); however, differences in study
1-3
-------
design may account for these apparent inconsistencies. From the available
data, few conclusions can be drawn regarding the systemic toxic effects of
chronic exposure to ethylene oxide.
In animal studies, repeated exposures to high concentrations of ethylene
oxide, 1533-643 mg/m3, resulted in the death of rats and mice. At the lower
exposure, guinea pigs and monkeys showed signs of neurotoxicity and growth
depression. Decreased body weight gain was observed in rats at exposures as
low as 60 mg/m3, 6 hours/day, 5 days/week for 2 years, while signs of neuro-
toxicity, hunched posture and reduced locomotion were observed at doses as low
as 90 mg/m3, 5 hours/day for 10 weeks. In rats and mice, the no effect level
appeared to be 18 mg/m3. Similar signs of toxicity were observed after oral
exposure to ethylene oxide, with weight loss observed at 100 mg/kg body weight
and no effects observed at <10 mg/kg body weight (15-20 doses). Studies in
animals support the observation of neurotoxicity described in human case
reports.
The ability of ethylene oxide to cause teratogenic or adverse reproductive
effects has been examined in a number of species (mouse, rat, rabbit, monkey
and human) by two routes of administration (inhalation and intravenous). Rats,
but not rabbits, exposed to 150 ppm ethylene oxide administered by inhalation
displayed signs of maternal toxicity and toxicity to the developing conceptus.
Ethylene oxide (150 mg/kg) administered intravenously to mice caused maternal
toxicity and developmental toxicity. Intravenous administration of ethylene
oxide in rabbits (9, 18 and 36 mg/kg) produced embryotoxicity associated with
maternal toxicity. Ethylene chlorohydrin (ECH), a reaction product of ethylene
oxide, produced adverse effects on maternal and fetal well-being in mice but
not in rabbits at 120 mg/kg administered intravenously and produced adverse
1-4
-------
developmental effects without significant toxicity when administered ECH at 60
mg/kg intravenously. Ethylene oxide (100 ppm) administered by inhalation in a
one-generation study caused severe adverse effects including a higher incidence
of infertility, longer gestational periods, a decrease in the number of pups
born, and a decrease in the number of implantation sites. The same laboratory
observed lowered fetal weights, but not a substantial level of malformations in
response to 100 ppm ethylene oxide administered to rats by inhalation on
gestation days 6-15. Testicular degeneration was observed in hamsters and rats
inhaling 204 to 357 ppm ethylene oxide. More recently, adverse effects on
sperm concentration and motility but not morphology in Cynomologus monkeys
exposed to 50 and 100 ppm ethylene oxide by inhalation were reported. An
epidemiologic study of nursing personnel exposed to ethylene oxide found an
association between ethylene oxide exposure and spontaneous abortion.
In conclusion, ethylene oxide produces adverse reproductive and teratogenic
effects in both females (maternal toxicity, depression of fetal weight gain,
fetal death, fetal malformation) and males (reduced sperm numbers 'and sperm
motility) if the concentration of the chemical reaching the target organ is
sufficiently high or if exposure at lower levels is sufficiently long. Thus,
the experiments in which ethylene oxide was injected intravenously have produced
more detrimental effects than the short-term inhalation experiments. However,
even short-term inhalation experiments have resulted in suggestive evidence of
detrimental effects.
Ethylene oxide has been shown to induce gene mutations in bacteria, fungi,
higher plants, Drosophila, and cultured mammalian cells in tests conducted
without the use of exogenous hepatic metabolic activation systems. It is
therefore a direct-acting rautagen. Ethylene oxide has also been shown to
1-5
-------
induce dominant lethal effects in mice and rats; chromosomal aberrations in
higher plants, Drosophila, mice, and rats; and micronuclei in mice and rats.
Based on these positive findings in different test systems, ethylene oxide is
judged to be capable of causing chromosomal aberrations. It has also been
shown to induce sister chromatid exchange (SCE) in rabbits, rats, and humans.
Tissue distribution studies have shown that ethylene oxide reaches the
gonads. This result is consistent with evidence that ethylene oxide causes
unscheduled DNA synthesis (UDS) in germ cells of male mice and heritable
mutations in insects and rodents (i.e., sex-linked recessive lethals and
heritable translocations in Drosophila, dominant lethals in rats and mice,
and heritable translocations in mice). Ethylene oxide can therefore be
regarded as mutagenic both in somatic cells and in germ cells.
Based on the available data, there is overwhelming evidence that ethylene
oxide is a direct-acting mutagen that has the potential to cause mutations in
the cells of exposed human tissue. The observations that ethylene oxide
reaches and reacts with mammalian gonadal DNA, and causes heritable mutations
in intact mammals, indicates that it may be capable of causing heritable
mutations in man provided that the pharmacokinetics of ethylene oxide in humans
also results in its distribution to the DNA of germ cells.
Both human and experimental animal data are available to assess the
carcinogenicity of ethylene oxide. The human evidence suggests an association
between exposure and cancer incidence, while the animal evidence is more sub-
stantial.
Three epidemiologic studies showed a significant association between
ethylene oxide exposure and the occurrence of cancer. Two of the studies
found an excess risk of leukemia associated with ethylene oxide exposure.
1-6
-------
While these studies have shortcomings and are not definitive, they do, never-
theless, constitute "limited," bordering on inadequate, epideraiologic evidence
for human carcinogenicity using the U.S. Environmental Protection Agency (EPA)
Proposed Guidelines for Carcinogen Risk Assessment. Two long-term inhalation
studies in rats show statistically significant responses for leukemia, brain
tumors and peritoneal mesothelioma. In addition, positive results for the
carcinogenicity of ethylene oxide have been obtained by subcutaneous injection
in mice and by intragastric administration in rats. The animal evidence for
the carcinogenicity of ethylene oxide is judged to be "sufficient" using the
EPA weight-of-evidence classification system.
On the basis of the human, animal, and mutagenic data cited herein,
ethylene oxide is classified as being "probably carcinogenic to humans" and
belonging in EPA Group Bl. This classification is qualified as bordering on
Group B2, however, because of limitations in the human evidence. According to
the International Agency for Research on Cancer (IARC) guidelines for evaluating
carcinogen evidence, ethylene oxide would be classified as Group 2A, meaning
that ethylene oxide is a "probable human carcinogen," but bordering on Group 2B
because of limitations in the human evidence. (See Appendix 9B for a descrip-
tion of the IARC classification system.)
Presuming that ethylene oxide is carcinogenic in humans, upper-limit in-
cremental unit risk and potency estimates have been extrapolated from the 2-
year rat inhalation studies. These estimates are upper limit in the sense that
a true risk level cannot be pinpointed because of uncertainties in low-dose
extrapolation, and, therefore, a modelling technique is employed which produces
a statistical upper-bound estimate of risk while retaining biological plausi-
bility. Upper limit means that the true risk is not likely to exceed the
1-7
-------
calculated value and may be lower. These estimates are within the range of
uncertainty of those derived from the human studies. Based on leukemias and
brain gliomas in rats, extrapolation to humans yields an upper-limit incremental
unit risk estimate of 1.0 x 10~^, for lifetime cancer risk resulting from
continuous exposure to air that contains an ethylene oxide concentration of
1 y g/rn-^. The relative potency index for ethylene oxide, which is based on
both the upper-limit unit risk value and the molecular weight, is in the lower
part of the third quartile of 54 chemicals that the CAG has evaluated as
potential or known human carcinogens.
1-8
-------
2. INTRODUCTION
EPA's Office of Research and Development has prepared this health assess-
ment to serve as a "source document" for Agency use. This health assessment
was originally developed for use by the Office of Air Quality Planning and
Standards to support decision-making regarding possible regulation of ethylene
oxide under Section 112 of the Clean Air Act. However, based on the expressed
interest of other Agency offices, the scope of this document was expanded to
address ethylene oxide in relation to sectors of the environment other than
air. It is fully expected that this document will serve the information needs
of many government agencies and private groups with health-related interests
in ethylene oxide.
In the development of the assessment document, existing scientific liter-
ature has been surveyed, key studies have been evaluated and conclusions have
been prepared so that the chemical's toxicity and related characteristics are
identified.
The document considers all sources of ethylene oxide in the environment,
the likelihood of human exposure, and the possible effect on man and lower
organisms from absorption. The information found in the document is integrated
into a format designed for risk assessment use. However, the information in
this document on environmental levels and exposure is not intended, nor should
it be used, to support any conclusions regarding risks to public health. When
appropriate, the authors of the document have attempted to identify gaps in
current knowledge that limit risk evaluation capabilities.
The literature searches that support this document vary somewhat. The
document is current to February 1984 with the following exceptions: the
2-1
-------
mutagenicity section is current to January 1984, and the carcinogenicity
section is current to January 1985.
2-2
-------
3. PHYSICAL AND CHEMICAL PROPERTIES
3.1. SYNONYMS AND CAS NUMBER
Synonyms: 1,2-epoxyethane
ethylene oxide
oxirane
CAS Number: 75-21-8
3.2. STRUCTURAL AND MOLECULAR FORMULAS
Structural formula:
Molecular formula:
3.3. TORSIONAL ANGLES AND BOND DISTANCES (Hirose, 197*0a
Torsional
Angles
-------
3.4. PHYSICAL PROPERTIES OF PURE ETHYLENE OXIDE
3.^.1. Description. Ethylene oxide is a colorless, flammable gas which
condenses at low temperatures to a colorless, clear, mobile liquid (Cawse
et al., 1980; Hawley, 1981).
3.4.2. Molecular Weight.
44.05 (Weast, 1972)
3.4.3. Melting Point.
-111°C (Weast, 1972)
3.4.4. Boiling Point.
10.4°C (at 101.3 kPa = 1 atm) (Cawse et al., 1980)
3.4.5. Boiling Point Change with Pressure Change (Abp/pressure at 100 kPa).
0.25 K/kPa (Cawse et al., 1980)
0.033 K/torr (Cawse et al., 1980)
3.4.6. Density.
dj°: 0.8824 (Weast, 1972)
3.4.7. Coefficient of Cubical Expansion (at 20°C, per °C).
0.00161 (Cawse et al., 1980)
3-2
-------
3.4.8. Refractive Index (at 7°C).
1 .3597 (Weast, 1972)
3.4.9. Vapor Pressure (Cawse et al. , 1980).
Temperature Vapor Pressure
°C kPa Torr
-40
-30
-20
-10
0
10
20
30
40
50
60
70
80
90
100
8.35
15.05
25.73
42.00
65.82
99.54
145.8
207.7
288.4
391.7
521.2
681.0
875.4
1108.7
1385.4
62.6
112.9
193.0
315.0
493.7
746.6
1093
1558
2163
2938
3909
5108
6566
8315
10390
3.4.10. Aqueous Solubility3 (Cawse et al., 1980).
Pressure Temperature
kPa torr 5°C 10°C 20°C
20
27
40
53
67
80
93
101
150
202.5
300.0
397.5
502.5
600.0
697.5
757.5
45
60
105
162
240
NT '
NT
NT
33
46
76
120
178
294
NT
NT
20
29
49
74
101
134
170
195
Solubility in m£ vapor/mJ, water, vapor volume
at 0°C and 1 atm
NT = Not tested
3-3
-------
3.4.11. Freezing Point of Aqueous Solutions (Cawse et al., 1980)
Ethylene Oxide
Freezing Point
Weight %
0
2.5
5
10
15
20
30
40
50
60
70
80
90
100
Mole
0
1.
2.
4.
6.
9.
14.
21.
29.
38.
48.
62.
78.
100
*
0
1
4
7
3
9
4
0
0
8
1
6
o
0.0
-0.9
-1.6
5.6
8.9
10.4
11.1
10.4
9.3
7.8
6.0
3.7
0.0
-112.5
C
(eutectic)
(max)
3.4.12. Boiling Point of Aqueous Solutions (Cawse et al., 1980)
Ethylene Oxide
Boiling Point
Weight %
0
2.5
5
10
15
20
30
40
50
60
70
80
90
100
Mole %
0
1.0
2.1
4.4
6.7
9.3
14.9
21.4
29.0
38.0
48.8
62.1
78.6
100
°C
100
70
58
42.5
38
32
27
21
19
16
15
13
12
10.4
3.4.13. Flash Point (tag open cup).
<-l8°C (Cawse et al., 1980)
3-4
-------
. Flash Point of Aqueous Solutions (Cawse et al., 1980),
Ethylene Oxide Flash Point
Weight % Closed Cup (°C)
1 31
3 3
5 -2
3.4.15. Explosive Limits in Air, Volume % (Cawse et al., 1980).
Upper Limit 100?
Lower Limit 3%
3.4.16. Heat of Combustion at 25°C (Cawse et al., 1980).
5.17 kJ/mol
1.24 IcCal/mol
3.4.17. Log Octanol/Water Partition Coefficient.
-0.30 (Hansch and Leo, 1979)
3.4.18. Ultraviolet Spectroscopic Data (Weast, 1972).
Xgas = 169 nm
max
log e = 3.58
Xgas = 171 nm
max
log e = 3.57
3.5. PHYSICAL PROPERTIES AND DESCRIPTION OF COMMERCIAL ETHYLENE OXIDE
The physical properties and specifications for commercial ethylene oxide
are described in Table 3-1.
3-5
-------
TABLE 3-1
Manufacturers' Specifications for Ethylene Oxide
a,b
Purity, wt % rain
Water, wt % max
Aldehydes, as acetaldehyde, wt % max
Acidity, as acetic acid, wt % max
C0_, wt % max
uo Total Cl as Cl~, wt % max
CTi
Nonvolatile residue, g/100 m£, max
Color, APHA, max
Residual Odor
Appearance
Acetylene, max
BASF
99.95
0.005
0.005
0.002
0.005
0.005
0.010
10
NA
NA
NA
Celanese
99.95
0.02
0.01
0.002
NA
NA
0.01
10
none
clear
NA
Dow
NA
0.03
0.005
0.002
0.002
0.005
0.01°
5
NA
NA
0.0005
Jefferson
NA
0.03
0.025
0.005
NA
nil
0.01
NA
none
clear
nil
Shell
NA
0.03
0.010
0.0020
NA
NA
0.010
10
none
clear
NA
Wyandotte
NA
NA
0.003
0.002
0.005
0.0005
0.01
10
mild
NA
NA
Source: U.S. EPA, 1980
This information was obtained from the respective manufacturer's product data sheets, available from each
manufacturer on request.
Presently, 0.005 g/100 m£ in Dow ethylene oxide (Kurginski, 1979)
NA = Not available; wt = weight; max = maximum; min = minimum
-------
Commercial grade ethylene oxide has a purity >99.9/f. Specific impurities
include trace quantities of water, aldehydes (specified as acetaldehyde), acid
(specified as acetic acid), chloride, and an unspecified residue. Since
commercial grade ethylene oxide is virtually pure, its physical properties are
the same as those described above for pure ethylene oxide.
3.6. CHEMICAL PROPERTIES
The majority of information contained in this section was taken from
Cawse et al. (1980).
Ethylene oxide is a highly reactive epoxide. Industrially, it is used
principally as an intermediate for a wide variety of compounds. Most of its
reactions involve opening the epoxide ring. An exception is the formation of
oxonium salts with strong anhydrous mineral acids.
3.6.1. Reduction. Catalytic hydrogenation or chemical reduction of ethylene
oxide results in the formation of ethanol.
3.6.2. Clathrate Formation. Ethylene oxide and water form a stable clathrate
containing 6.38-6.80 molecules of ethylene oxide to 46 units of water in the
unit cell. The maximum observed melting point for these compounds is 11.1°C
(Section 3.4.11).
3-6.3. Polymerization. Low molecular weight polymers can be formed by the
reaction of ethylene oxide and water or alcohols. The average molecular
weight of these polymers (polyethylene glycols) ranges from 200-14,000,
depending upon the reaction conditions. High polymers, with molecular weights
3-7
-------
ranging from 90,000 to U x 10 , are formed by coordinate anionic
polymerization. This reaction involves the coordination of a metallic
compound with ethylene oxide to initiate the reaction. Numerous
organometallic and alkaline earth compounds and mixtures are used as
catalysts. This process is important in the formation of non-volatile
residues during ethylene oxide storage (Section 3.5). The main catalyst for
this process is rust, and no inhibitor has been found.
3.6.4. Other Reactions. Table 3-2 lists some other representative reactions
of ethylene oxide.
3.6.5. Hydrolysis and Related Reactions. Epoxides degrade in water by
hydrolysis and related ionic reactions and, possibly, by radical oxidations.
The hydrolysis chemistry involves cleaving a carbon-oxygen bond of the cyclic
ether to form ethylene glycol. Bronsted et al. (1929) noted the pathways for
ethylene oxide hydrolysis in aqueous hydrochloric acid, describing hydrolysis
as a combination of a noncatalytic reaction (herein referred to as the
spontaneous hydrolysis) and an acid-catalyzed hydrolysis. Reaction with
chloride ion was similar to hydrolysis in that chloride and epoxide reacted
without catalysis and with acid catalysis.
Long and Pritchard (1956) demonstrated that epoxide hydrolysis was also
base catalyzed. For any epoxide, the degradation pathways are as follows for
the neutral (I), acid-catalyzed (II), and alkali-catalyzed hydrolyses (III):
k
(I) C HaO + HpO ——>HOCH CH OH (II) C^H.O + HO k.
>• HOCH0CH.OH
H 0+ ^ *
3-8
-------
TABLE 3-2
Typical Reactions of Ethylene Oxide
1. Crown Ethers
n H-C - CH0 Catalyst> cyclic {CH-CH n^
2\/2 2 2 "
0
2. Hydrolysis
H C - CH0 + H00 > HOCH0CH0OH
2\/ 2 2 2 2
0
3. Reaction with Alcohols
H0C - CH~ + ROH > R-0-CH.CH_OH -> R_04CH.CH00^ H
£ \ / d. 22 oxide 2 2 n
0
4. Reaction with Organic Acids and Acid Anhydrides
RCOOCOR + H_C - CH0 > RCOOCH.CH0OH ethyj-ene> RC004CH0CH00^ H
d \ t d 22 oxide 2 2 n
0
5. Reaction with Ammonia and Primary and Secondary Amines
R-NH -.- H_C - CH_ - > R-NH-CH0CH0OH -> R_NH4CH0CH00^ H
d d. \. / d 22 oxide 2 2 n
0
6. With Hydrogen Sulfide and Mercaptans (e.g., glutathione, cystine)
H2C - CH2 + RSH >
0
3-9
-------
TABLE 3-2 (cont.)
7. Reaction with Pyridine (and possibly other nitrogen heterocycles)
+ H0C - CH0 — > /(5V-CH0-CH9OH + OH~ > /C^N + HOCH?CH?OH
\ / ^—/ ^ f
0
8. With Phenols
- CH2
0
9. With Hydrogen Cyanide
H C - CH + HCN > HOCH CH CN > CH =CH-CN
\ / acrylonitrile
0
3-10
-------
(III) C2H40 + H20 —*-+ HOCH2CH2OH
OH~
Table 3-3 summarizes hydrolysis data for ethylene oxide. The temperature
coefficients for the rate constants are the following:
log kA = 10.753 + log T -0.0255/R - 79.5/RT (Long et al., 1957)
log kN = 7.726 - 79.5/RT (Lichtenstein and Twigg, 1948)
log kg = 9.312 - 75.3/RT (Lichtenstein and Twigg, 1948)
Epoxides can also react with nucleophiles (anions or Lewis bases). The
chemistry, although similar to hydrolysis, is more complex. The epoxide ring
can be cleaved by spontaneous reaction or by acid-catalyzed reaction:
I
-C k -COH
'^0 + X~ + HO —2-+ I + OH~
— U d. — L/A
I
I
~C \ kx ~COH
"l
Table 3-4 summarizes specific rate constants for reactions of ethylene oxide
with various anions. The consensus is that the spontaneous reaction is S.,2,
but disagreement exists over whether the acid-catalyzed epoxide ring opening
is A1-like or A2-like (Long et al., 1957; Lamaty et al., 1975; Pritchard and
Long, 1956; Pritchard and Siddiqui, 1973; Virtanen and Kuokkanen, 1973). A
discussion of the mechanism is beyond the scope of this review.
3-11
-------
TABLE 3-3
Hydrolysis Kinetics of Ethylene Oxide
Temperature
(K)
293
293.2
298
298
298
298
298
298
NR
303.2
kA x 103
(M-IS-I)
5.34a
NR
9.3°
NR
NR
NR
9
NR
10. Od
16. 9e
Specific Rate Constants
kN x 107
(S-1)
3.6ia
4.2b
6.75°
5.62f.S
6.17f'h
6.6lf,i
5.56J
5.8^
NR
NR
kB x 104
(M-1S-1)
NR
0.65b
1.0d
NR
NR
NR
1.1
NR
NR
NR
aBronsted et al., 1929
Lichtenstein and Twigg, 1948
GEastham and Latreraouille, 1952
Pritchard and Long, 1956
eLong et al., 1957
fConway et al., 1983
sRiver water pH 7.4
h
i
Sterile river water pH 7.4
Sterile distilled water
JLong and Pritchard, 1956
kKoskikallio and Whalley, 1959
NR = Not reported 3_12
-------
TABLE 3-4
Specific Rates of Reaction of Anions and Lewis Bases with Ethylene Oxide
Lewis Base
or Anion
C1~
Br~
Pyridine
Temperature
K
293
298
298
300
293
298
291
I06ky
U/mol-sec,)a
NR
NR
0.3056
NR
NR
NR
200 (water)d
102kx
U2/mol2-sec)a
2.17 (water)5
3.67 (water)b
NR
8.23 (50% aqueous
ethanol)0
8.67 (water)5
14.5 (water)5
NR
k = neutral reaction; k = acid catalyzed
y x
3Bronsted et al., 1929
Lamaty et al., 1975
Pritchard and Siddiqui, 1973
Conway et al., 1983
NR = Not reported
3-13
-------
Some products of epoxide reaction with Lewis bases or with anions are not
stable. For example, tertiary amines, such as pyridine, are capable of
catalyzing epoxide hydrolysis to glycol:
, HOCH2CH2 NC5 H5 -^ HOCH2CH2OH
CH
Aqueous chemical degradation in the environment can be estimated from the
contributions of hydrolysis (Equation 1) and anion reactions (Equation 2):
-dC
dt - V*N - *A"H3Cf " *B°OH~'"epox (1)
-dC
= (k .C,, + k .C-.C,, 0+)C (2)
epox
dt
where C.., k ., and k . refer to the concentration and specific rate constants
AI y^- xi
for each anion or Lewis base. The overall degradation rate is the sum of all
contributions, as given in Equation 3:
dC
Z(ky. + KxiCH0+)CA.]C (3)
The relative importance of chemical hydrolysis vs. reaction with chloride
ion was assessed for ethylene oxide. Degradation half-lives and product
distributions (chlorohydrin-to-glycol ratios) were estimated for freshwater
3-14
-------
and marine water (NaCl concentration of 3% or 0.513 M). The following
specific rate constants from Tables 3-3 and 3-4 were used:
kN 0.661 x 10~6 s~1
kA 9 x 10~3 M"1 s"1
kg 1 x 10~4 M"1 s"1
k P1 0.305 x 10~6 M~1 s~1
y ? ? ?
3.67 x 10~^ M s"
Estimates were calculated for pH 5.0, 7.0, and 9.0, which is approximately the
pH range of natural waters. Half-lives for chemical degradation and the
chlorohydrin/glycol ratios (for sea water reactions) are summarized below:
Calculated Ethylene Oxide
Half-Life at 298 K
(hours)
pH 5 7 9
Freshwater 256 291 291
Saline Solution
0.85? (physiological) 273
1/K 240 270 270
3% (marine) 212 236 236
Conway et al. (1983) used buffered (pH 7) sterile solutions of 0, 1, and 3%
NaCl to hydrolyze ethylene oxide and reported half-lives of 314, 265, and 224
hours, respectively. The half-lives for river water (pH 7.4), sterile river
water (pH 7.4), and sterile distilled water were 341, 310, and 293 hours. The
chlorohydrin/glycol ratios experimentally determined by Conway et al. (1983)
were 0.11 and 0.23 for 1 and 3% saline solutions.
These data provide some understanding of the fate of ethylene oxide in
biological fluids. The hydrolysis half-life in physiological saline (0.85?
NaCl) is 273 hours or 11.4 days. This long a half-life would clearly allow
3-15
-------
for other reactions to take place. As an example, the half-life for the
ethylene oxide reaction with pyridine in water is 58 minutes. Other
nucleophiles present in biological systems (e.g., RS~, PhNH_) are known to be
more nucleophilic than pyridine and may react with ethylene oxide much more
rapidly than water or chloride.
Hydrolysis or hydrolysis-type reactions are the most significant
industrial reactions of ethylene oxide. Ethylene glycol is the hydrolysis
product; higher glycols (diethylene, triethylene, and polyethylene glycols)
and glycol ethers result from the reaction of ethylene oxide with glycols and
alcohols, respectively. Glycol esters of carboxylic acids and phenols, and
ethers of cellulose, starch, and other polyols are also prepared through
hydrolysis-like reactions. For example, reaction of ethylene oxide and
nonylphenol yields nonylphenoxypolyethoxyethanol, a non-ionic, surface-active
agent (Blackford, 1976a).
n CH2 CH2 + C9HigOH > CgH^OCCH^O^H
Ethylene oxide reacts with amines by pathways similar to reactions with
hydroxyl compounds. Reaction of ethylene oxide and ammonia yields the commer-
cially important ethanolamines:
0
NH3 + nH2C - CH2 > H£N 4 CH2CH20 ^ H
where n is typically 1 to 4. Choline is prepared by reacting trimethylamine
with ethylene oxide (Jukes, 1964):
3-16
-------
(CH ) N + H2C - C
Some ionic reactions of ethylene oxide are listed in Table 3-2.
3.6.6. Free Radical Reactions. The free-radical chemistry of ethylene
oxide is of particular importance in determining its fate in the atmosphere.
The most important free-radical reaction is the reaction with hydroxyl
radical.
Only one study was found in the available literature. Fritz et al.
(1982) reported the results of a study utilizing a laser photolysis/resonance
fluorescence unit designed to study the reactions of OH radicals with
anthropogenic pollutants. They generated hydroxyl radical by HNO_ photolysis
and studied the reaction at 297, 377, and 435 K, at 10 torr (Ar). The
following relations were reported (mean + 36 ):
and
v\ = (8.0 + 1.6) x 10 cm /molec-sec
K )
k(T) = (1-1 ± O-1*) x 10~11 exp (-1460/T) cm3/molec-sec
The mechanism involved hydrogen abstraction followed by ring opening, reaction
with oxygen, nitrogen oxide and finally decomposition to carbon monoxide and
formaldehyde. Ring opening may take place either before oxygen addition or
after NO reaction.
3-17
-------
4. SAMPLING AND ANALYTICAL METHODS
4.1. SAMPLING
The state-of-the-art in air sampling utilizes solid sorbents. Samples
can subsequently desorb by solvent or thermal means. Critical factors in the
method are the capacity of the sorbent to retain the epoxide during the
collection and the complete desorption of the epoxide.
Brown and Purnell (1979) evaluated Tenax GC sampling tubes for use in
ambient air monitoring studies and found them to be inappropriate for ethylene
oxide. Although most of the 71 compounds tested were adequately retained,
ethylene oxide was not, having the third poorest retention.
Pellizzari et al. (1976) evaluated Tenax GC and other sorbents for
sampling atmospheric propylene oxide, a compound chemically very similar to
ethylene oxide. Table 4-1 compares the breakthrough volumes for several sor-
bents. The effect of humidity on the breakthrough volume was tested for Tenax
GC: breakthrough volume remained unchanged in the range of 4.0-4.5 &/g when
humidity was increased from 41-92$. Storage time affected the recovery of
diepoxybutane (300 ng) from Tenax GC cartridges (desorbed thermally and
analyzed by GC. When analysis was immediate, recovery was 100/&. After the
loaded cartridge was stored for 1 week, recovery dropped to 76/6. Combined
transport (6 days) and storage yielded recoveries of 75 and 6^% after 1 and 2
weeks, respectively. Since Brown and Purnell (1979) and Pellizzari et al.
(1976) used comparable methods for determining the breakthrough volume, it
appears that propylene oxide and ethylene oxide behave similarly. Brown and
Purnell (1979) have noted that, under the conditions of the test (5-600
4-1
-------
TABLE 4-1
Breakthrough and Safe Sampling Volumes for Propylene Oxide
with Several Sorbents
Sorbent Breakthrough Volume
H/g (sorbent)a
PEL Carbon
PCB Carbon
SAL9190
MI808
Tenax GC (35/60)°
Porapak Q (100/120)
Chromosorb 101 (60/80)
Chromosorb 102 (60/80)
Chromosorb 101 (60/80)
36
40
40
24
4
4
4
8
>36
Safe Sampling Volume
a/g)b
9
10
10
6
1
1
1
2
9
Pellizzari et al., 1976
DBrown and Purnell, 1979
'Mesh size
4-2
-------
m£/minute flow rate, <100 ppra vapor concentration, <20°C, and <95% relative
humidity) the breakthrough volume is not less than 50% of the retention
volume, and a safe sampling volume is 50% of the retention volume. If
propylene oxide behavior is analogous to ethylene oxide, the reported levels
of ethylene oxide present in air could not be determined accurately, since the
great majority of monitoring studies use air samples larger than the break-
through volume for ethylene oxide.
The National Institute for Occupational Safety and Health (NIOSH) has
published standard procedures for ethylene oxide collection in air (NIOSH,
1977). Their procedure calls for the sampling of 5 £ of air through glass
tubes packed with activated coconut-shell charcoal. For ethylene oxide, two
tubes mounted in series are used, containing 400 and 200 mg of charcoal. If
the back-up tube (200 mg) contains >25% of the epoxide, the analysis is not
considered valid. Ethylene oxide should be desorbed from the charcoal with
2.0 m£ of carbon disulfide; aliquots are then analyzed by GC with flame-ioni-
zation detection. In NIOSH (1977) tests on the analytical parameters,
ethylene oxide was sampled at concentrations from 41-176 mg/m (23-98 ppm);
precision (CVT) was 0.103 (or standard deviation of 9.3 mg/m ), and accuracy
was 0.9$ lower than the "true" value. NIOSH (1977) recommended this method
for industrial hygiene monitoring at sample concentrations of 20-270 mg/m .
The monitoring of occupational exposure to ethylene oxide by adsorption
through activated carbon and subsequent desorption with carbon disulfide was
reported by Qazi and Ketcham (1977). These investigators evaluated several
carbon and noncarbon adsorbants and concluded that Columbia JXC activated
carbon was most suitable for the collection of ethylene oxide in air. The
breakthrough volume for ethylene oxide with this adsorbant was dependent on
4-3
-------
both the flow rate and the moisture content of the air. At relative
humidities >60% and sampling rates of 20-25 nd/minute, the breakthrough volume
was <10 H. The quantification of ethylene oxide was done by GC with a
TERGITOL TMN or UCON LB550X column interfaced with flame-ionization detector.
At concentration levels of 0.5-5.0 ppm, the average recovery of ethylene oxide
by this method was 97%, with a relative standard deviation and error of 3.8
and 2.956, respectively. The lower detection limit of the method was 0.15 ppm
with a sample volume of 10 9,.
The quantification of ethylene oxide and its two volatile metabolites, 2-
chloroethanol and ethylene glycol, at a concentration level of 1-10 ppb in
biological samples was attempted by EOIC (1984). A purge and trap method
consisting of nitrogen gas bubbling through purge cells, and charcoal traps
for collecting the transferred volatiles was found most suitable for
biological samples. The subsequent quantification of ethylene oxide and its
metabolites was done by thermal desorption of the charcoal trap and analysis
on interfaced GC/MS operated on a selected ion-monitoring mode. Both packed
carbowax 20 M and Earbopack/THEED GC columns were used; however, this
technique provided non-reproducible data due to deterioration (peak tailing)
of the GC columns with time.
Romano and Renner (1975) described the results of a six-laboratory inter-
comparison of three methods for sampling ethylene oxide concentrations in
surgical equipment. The study was administered through a Subcommittee on
Ethylene Oxide Sterilization of the Association for Advancement of Medical
Instrumentation. The three sampling methods were vacuum extraction with
sample freezeout, headspace analysis, and acetone extraction. The vacuum-
freezeout technique requires distillation of volatiles from the sample, and
4-4
-------
freezing them in a cold trap. The sample is then vaporized and an aliquot is
removed with a vacuum syringe for GC analysis. This method requires greater
time and equipment than the other techniques and is subject to errors from
equipment leaks; however, it is the most sensitive method, and, since the
sample injected into the GC is a vapor, column life is prolonged. Acetone
extraction consists of partitioning the epoxide between the sample and the
acetone solvent. Its advantage is its simplicity. Its disadvantages include
its inability to quantitatively extract epoxide, impurities from the solvent
and the plastics, the reduced lifetime of GC columns, and low sensitivity. In
headspace analysis, the sample is placed into a vial equipped with a septum
for gas withdrawal by syringe. The epoxide partitions between the sample and
headspace gases. Romano et al. (1973) reported that the headspace technique
has a lower limit of 0.1 ppm and that the technique can be automated. The
advantages of this technique include its ease of performance, speed,
sensitivity, and relatively long column life. Its disadvantage is that leaks
in septa, vial caps, etc., can yield low measurements. Among the three
overall methods, Romano and Renner (1975) found no significant differences,
though they did find slight differences between laboratories.
Ben-Yehoshua et al. (1971) analyzed fruit pulp by blending it with 50 m£
of analytical grade acetone for 30 seconds and filtering the homogenate to
clarity. The samples were then stored at -10°C in bottles with self-sealing
stoppers. Measurements (by GC) of added ethylene oxide and its residues were
accurate to +5%. Scudamore and Heuser (1971) extracted wheat flour and other
commodities, including coconut, sultanas, lentils, and ground nuts with 5:1
(v/v) analytical grade acetone:water. The extraction used as little as 3 m£
solvent/g sample. A contact time of 24 hours was sufficient to yield ethylene
-------
oxide recoveries (by GC) of >95%. Pfeilsticker et al. (1975) extracted 10 g
of grain (not crushed) with 5 mJl of methanol using continuous agitation for 24
hours. Recovery of ethylene oxide was 73% (25+1.7).
Brown (1970) sampled and analyzed surgical materials (plastic and rubber)
for ethylene oxide residues by means of a three column chromatography system.
This system could separate ethylene oxide and its degradation product,
ethylene chlorohydrin. Epoxides were extracted with £-xylene (3 days of
contact) or _co-sweep distillation. The £-xylene solution was passed through
one column of Florisil; ethylene chlorohydrin remained fixed in the column and
ethylene oxide passed through. The solution was then passed through the
second acid-celite column, which converted any ethylene oxide to ethylene
chlorohydrin. A third Florisil column retained the ethylene chlorohydrin,
which was subsequently eluted with petroleum ether. The sample was
concentrated and analyzed by GC. Brown (1970) reported values as low as 1.8
ppm, but the accuracy, precision, and minimum detection limit were not
described.
4.2. ANALYSIS
To date, GC analysis for ethylene oxide has used only flame-ionization or
thermal-conductivity detection. Neither detection system is selective, so the
epoxides must be separated from all interfering components, and the analytical
column must be chosen around potential interferences. Columns used for
epoxide analysis have included uncoated Poropak Q, QS, and R, and Chromosorb
102 (Taylor, 1977a,b; Ben-Yehoshua and Krinsky, 1968; Steinberg, 1977), and a
variety of coated columns. The most common liquid phases appear to be SE-30,
Carbowax 20M, and polypropylene glycol (Ben-Yehoshua and Krinsky, 1968;
4-6
-------
Casteignau and Halary, 1972; Steinberg, 1977; Hughes et al., 1959). Bertsch
et al. (1974) used a 100m x 0.5mm capillary column coated with Emulphor ON
870. The GC methods in current use appear capable of epoxide analysis at the
ppm level.
The EOIC (1984) used a GC/MS/SIM system for the quantification of
ethylene oxide and its volatile metabolites in biological samples. More
recent techniques involving a negative-ion atmospheric pressure chemical
ionization MS/MS system developed by Sciex of Canada were used by U.S. EPA
(1984) for the on-site monitoring of pollutants in the atmosphere. This
system was used for the identification, but not for the quantification, of
ethylene oxide in a synthetic gas mixture.
Other analytical methods include various wet chemical techniques.
Epoxides can be analyzed by ring opening with specific reagents and subsequent
analysis for that reagent or one of its products (Dobinson et al., 1969). For
example, Mishmash and Meloan (1972) reported what may be the most recent use
of this approach. Butylene oxide was hydrolyzed to its glycol, and the glycol
was oxidized with periodic acid. Residual oxidant was analyzed by adding Cdl?
-starch and measuring the starch-I., complex concentration at 590 nm. They
claimed a detection limit in the nmole range.
4-7
-------
5. SOURCES IN THE ENVIRONMENT
5.1. PRODUCTION
5.1.1. Quantities Produced. Production volumes and sales quantities for
ethylene oxide are listed in Table 5-1 for the years 1972 to 1982.
5.1.2. Producers, Production Sites, and Distribution. The producers, produc-
tion sites, and annual capacities of ethylene oxide are listed in Table 5-2.
ICI Americas is building a new ethylene oxide plant in Bayport, Texas; the
naraeplate capacity is 520 million pounds/year (Anonymous, 198la). Dow will
add 400 million pounds/year capacity onto its Plaquemine, Louisiana, facility
during the fourth quarter of 1983. Union Carbide is building a 400 million
pounds/year unit in Alberta, Canada, slated to be on stream in 1985. PPG
Industries and DuPont are conducting a feasibility study to determine whether
to move the former's idle Guayanilla, Puerto Rico, facility (rated at 300
million pounds/year) to Beaumont, Texas, to be operated by both companies.
5.1.3. Production Methods and Processes.
5.1.3-1. INTRODUCTION ~ The majority of the information in this section
was obtained from Cawse et al. (1980).
Ethylene oxide is produced almost exclusively by direct oxidation, using
either air or oxygen. Other processes cannot compete with the lower operating
costs of direct oxidation. Only one plant in the United States currently has
5-1
-------
TABLE 5-1
Ethylene Oxide Production3)b
Year
1982d
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
Production
5200 (2359)
4937 (2240)
5220 (2368)
5665 (2570)
5012 (2273)
4364 (1980)
4184 (1898)
4467 (2026)
3893 (1766)
4167 (1890)
3962 (1797)
Sales0
NA
NA
531 (241)
560 (254)
525 (238)
549 (249)
439 (199)
409 (186)
457 (207)
501 (227)
454 (206)
aSource: USTC, 1974, 1975; USITC, 1976, 1977a, 19775, 1978, 1979, 1980, 1981
All quantities are expressed in millions of pounds; SI units in millions of
kilograms are given in parentheses.
Q
The difference between production and sales does not enter the merchant
marketplace.
Projected (Source: Anonymous, 1982)
NA = Not available
5-2
-------
TABLE 5-2
Ethylene Oxide Producers, Plant Sites, Capacities, Processes, and Technology^
I
uo
Company
BASF Wyandotte, Indust. Chem. Group
Basic Chems. Div.
Calcasieu Chem. Corp.0
Celanese Corp.
Celanese Chem. Co., Inc.
Dow Chemical U.S.A.
Eastman Kodak Co.
Eastman Chemical Prod., Inc.
Subsid. Texas Eastman Co.
ICI Americas, Inc., Petrochems. Div.
Location
Geismar, LA
Lake Charles, LA
Clear Lake, TX
Freeport, TX
Plaquemine, LA
Longview, TX
Bayport, TX
Annual
Capacity^
481 (216)
225 (101)
425 (191)
26Qd (117)
450e (203)
195 (88)
520 (234)f
Process
Oxidant
oxygen
oxygen
oxygen
air
air
oxygen
NA
Technology
Shell
Shell
Shell
Dow
Dow
Shell
NA
Inter-North, Inc.
Northern Petrochem. Co.,
Subsid. Petrochems. Div.
Olin Corp., Olin Chems. Group
PPG Industries, Inc.
Chems. Group, Chem. Div.-U.S.
Joliet, IL 230 (104) oxygen Scientific Design
Brandenburg, KY 110 (50) oxygen Shell
Beaumont, TX
155 (70)
air
Scientific Design
-------
TABLE 5-2 (cont.)
Company
Shell Chemical Co.
Sun Olin Chemical Co.
Texaco, Inc.
Texaco Chemical Co., Div.
Union Carbide Corp.
Chems. and Plastics Div.
Union Carbide Carbie, Inc., Subsid.
Location
Geismar, LA
Claymont, DE
Port Neches, TX
Seadrift, TX
Taft, LA
Ponce, PR
Annual
Capacity13
700
100
700
1000
1250
640
(315)
(45)
(315)
(450)
(563)
(288)
Process
Oxidant
oxygen
oxygen
air
air
air
air
Technology
Shell
Shell
Scientific Design
Union
Union
Union
Carbide
Carbide
Carbide
aSources: Anonymous, 198la; SRI International, 198la,b; Cawse, 1980
Capacities are expressed in millions of pounds; capacities in millions of kilograms are in parentheses.
GPlant is on indefinite standby as of January 31, 1981 (Anonymous, 198la).
Approximately 200 million pounds/year (90 million kg/year) additional capacity can be obtained from a
chlorohydrin unit used for propylene oxide production.
Expansion of 400 million pounds/year (180 million kg/year) is due in the fourth quarter of 1983.
f
Under construction
NA = Not available
-------
chlorohydrin capacity (Dow at Freeport, Texas; see Table 5-2). The major
drawback of the direct oxidation process is the loss of =25-30? of the
ethyl ene to carbon dioxide and water.
5.1.3.2. DIRECT OXIDATION — The overall reaction for direct oxidation
can be represented as follows:
CH2
0
5.1.3.2.1. Air-Based Oxidation — The schematic for air-based ethylene
oxidation is presented in Figure 5-1 . Little detailed information is avail-
able concerning process technology; however, the salient features are
presented below.
In the first section, air and ethylene are fed into the recycle gas
stream (the recycle gas contains unreacted starting material from the main
absorber). The recycle stream is fed into a bank of tubular main reactors,
the number of reactors depending chiefly on the capacity of the plant,
activity of the catalyst, and size of the reactors. In the main reactor, the
ethylene is oxidized to ethylene oxide, carbon dioxide, and water, as well as
to minor components such as formaldehyde and acetaldehyde.
Ethylene conversion to ethylene oxide per pass in the main reactors is
20-50?. Oxidation inhibitors (e.g., vinyl chloride, ethylene dichloride) are
added to retard carbon dioxide formation. The process stream leaving the
reactor may contain 1-2 mole % ethylene oxide. This hot effluent gas is
cooled to around 35-40° C and fed to the main absorber.
5-5
-------
tor
•btorbar raactor
Puff*
absorber
Dvtorbor
Stripper Refiner
Steam
Coolant
FIGURE 5-1
Schematic for Air-Based Ethylene Oxidation
Source: Schultze, 1965
-------
The main absorber uses cold water to dissolve the ethylene oxide, some
carbon dioxide, and traces of hydrocarbons and aldehydes. The unabsorbed gas
is split overhead. The largest portion is used as recycle gas, and to cool
the effluent stream from the main reactor; the gas then enters the main
reactor. A much smaller portion of the absorber effluent gas is fed as the
main stream to the secondary or purge reactor. The effluent from the purge
reactor is heat exchanged with the main stream and sent to the purge absorber
which operates in the same manner as the main absorber.
The purge reactor system reacts a large portion of the ethylene present
in the purge gas from the main reactor which must be vented from the main
reactor so that inert gases (principally nitrogen and carbon dioxide) do not
accumulate. Although Figure 5-1 shows a two stage air-based plant with a
single purge reactor, some large plants have three or more stages to improve
the overall yield. These plants merely place another purge reactor and
absorber in series.
In some plants, the ethylene content of the vent gas is sufficiently high
to make energy recovery economical. This not only produces valuable power
from the vent gas, but also reduces the hydrocarbon emissions from the
process.
The remainder of the process involves purification. The ethylene oxide-
water solution from the absorbers is heat-exchanged and sent to the desorber,
where the ethylene oxide is steam stripped under reduced pressure. The
ethylene oxide is collected at the top and compressed for further purifica-
tion, while the stripped water is recirculated to the main and purge
absorbers.
5-7
-------
The ethylene oxide from the desorber still contains some carbon dioxide,
nitrogen, aldehydes, and traces of ethylene and ethane, and must be sent to
the stripper. Here, the light gases are separated overhead and vented, while
the partially purified ethylene oxide is taken from the bottom of the stripper
and sent to the mid-section of a final refining column. The ethylene oxide
from the refining section should have a >99.5 moIe-% purity.
The specific conditions used to operate ethylene oxide plants are
proprietary information; however, the general ranges suggested by the
literature and patent reviews have been summarized by Cawse et al. (1980) and
are presented in Table 5-3.
5.1.3.2.2. Oxygen-Based Oxidation — The differences in oxygen-based and
air-based oxidation processes are almost entirely the result of the change in
oxidants. The main difference is that the purge reactor is absent in the
oxygen-based process and a carbon dioxide removal unit and an argon vent are
added. In the air-based cycle, the low per-pass conversion, the necessity of
complete ethylene oxide removal in the absorber, and the accumulation of
nitrogen necessitates a substantial purge system. Because of this, a staged
reaction-absorption system is required. Since the oxygen-based process uses
essentially pure oxygen, the recycle gas is almost entirely unconverted
ethylene; hence, there is no need for a purge system. Carbon dioxide,
however, is still produced in the oxygen system, and because it has a negative
effect on catalyst selectivity, the carbon dioxide must be removed. In
addition to the carbon dioxide removal unit, an argon vent is required. Argon
is a major impurity in oxygen and can build up to the extent of 30-40 rnole-^.
5-E
-------
TABLE 5-3
Ranges of Reaction System Variables in the Direct
Air-Oxidation of Ethylene Oxidea
Variable Range
ethylene, mole % 2-10
oxygen, mole % 4-8
carbon dioxide, mole % 5-10
ethane, mole % 0-1.0
temperature, °C 220-277
pressure, MPa (psi) 1-3 (145-435)
space velocity , h~ 2000-4500
pressure drop, kPa (torr) 41-152 (308-1140)
conversion, % 20-65
selectivity or yield (mole basis, %) 63-75
aSource: Cawse et al., 1980
The space velocity is the standard volume of the reactant stream fed per unit
time divided by the volume of reactor space filled with catalyst.
h = hour
5-9
-------
In spite of this additional purge, the total vent stream from an oxygen-based
plant is much smaller than that of an air-based plant.
As is the case with an air-based unit, the main process vent stream
usually contains a high concentration of hydrocarbons. In such a case, the
purge stream can be used for energy recovery. The operating ranges for an
oxygen-based process are summarized in Table 5-4.
The choice of oxygen versus air as the oxidant is based strictly on
economics; in general, for small-to-raedium capacity units (<50,000 tons/year),
oxygen-based plants have lower capital cost even with the necessary air separ-
ation facility. For medium-to-large plants (75,000-150,000 tons/year), the
air process investment is smaller unless oxygen can be purchased from a very
large air separation facility. Operating costs of the facilities can differ
significantly and are based on the cost of ethylene, oxygen, catalyst, and
energy.
5.1.3.2.3. Chlorohydrin Processes — The chlorohydrin process was the
main method of ethylene oxide manufacture until 1957. In 1972, the Dow
Chemical Company converted the remaining chlorohydrin capacity plants to the
production of propylene oxide, and the process was not used again for ethylene
oxide production until 1975. The Dow Chemical Company has built-in
flexibility for using the chlorohydrin process to produce either propylene
oxide or ethylene oxide. Since 1975, part of this capacity has been used for
ethylene oxide. During 1975, the Dow Chemical Company made between 25 and 50
million pounds of ethylene oxide via the chlorohydrin process (Blackford,
1976b). The chlorohydrin process is attractive commercially only when a good
supply of captive low-cost chlorine and lime or caustic soda is available.
5-10
-------
TABLE 5-4
Ranges of Reaction System Variables in the
Direct Oxygen-Oxidation of Ethylene Oxide3
Variable Range
ethylene, mole % 15-40
oxygen, mole % 5-8.5
carbon dioxide, mole % 5-15
ethane, mole % 0-2
argon, mole % 5-15
nitrogen, mole % 2-60
methane, mole % 1-60
temperature, °C 220-275
pressure, MPa (psi) 1-2.2 (145-319)
space velocity , h~ 2000-4000
conversion, % 7-15°
selectivity or yield (mole basis, $) 70-77
aSource: Cawse et al., 1980
The space velocity is the standard volume of the reactant stream feed per
unit time divided by the volume of reactor space filled with catalyst.
°At 30 mole % ethene
h = hour
5-11
-------
Also, satisfactory markets or disposal facilities are needed for the major by-
products (Schultze, 1965).
The chlorohydrin process starts with conversion of ethylene to ethylene
chlorohydrin with hypochlorous acid. The chlorohydrin is converted to
ethylene oxide by dehydrochlorination with slaked lime. Two major by-
products, 1 ,2-dichloroethane ( = 100-150 pounds/1000 pounds ethylene oxide) and
b is (2-chlo roe thyl) ether (=70-90 pounds/1000 pounds ethylene oxide), are formed
during the chlorohydrin formation; acetaldehyde (5-10 pounds/1000 pounds
ethylene oxide) is produced during the dehydrochlorination.
The formation of ethylene oxide from ethylene chlorohydrin can be repre-
sented by the following equation:
2 HOCH0CH0C1 + Ca(OH)_ >• 2 0HC-CH0 + CaCl_ + 2H00
d d
-------
is refined in two columns; the first column removes chlorinated hydrocarbons
and the second column removes acetaldehyde.
5.2. USES OF ETHYLENE OXIDE
A description of the various uses of ethylene oxide is given below:
Billion Pounds3 Percent of Total
Ethylene glycol 3.2 62
Nonionic surface-active agents 0.62 12
Glycol ethers 0.31 6
Ethanolamines 0.26 5
Miscellaneous applications 0.78 15
(higher glycols, urethane
polyols, sterilant, fumigant,
export)
Source: Anonymous, 1981 a
Based on 1982 production estimates of 5.2 billion pounds.
The major users and use sites for ethylene oxide are listed in Table 5-5. As
can be seen from this table, a very large percentage of production is captive-
ly consumed by the primary manufacturers. A general description of the
various uses of ethylene oxide is presented below.
5.2.1. Ethylene Glycol. By far, the largest single use of ethylene oxide is
its use captively as an intermediate in the synthesis of ethylene glycol,
which is currently produced by hydration of ethylene oxide. Current industry
capacity to produce ethylene glycol is 5815 million pounds annually
(Anonymous, 198lb). The growth in consumption of ethylene oxide has largely
depended on its use as an intermediate for ethylene glycol production.
5-13
-------
TABLE 5-5
Users and Use Sites of Ethylene Oxide*
Company
Ethylene Glycol Diethylene Ethanol-
Location Glycol Ethers Glycol amine
Triethylene Polyethylene
Glycol Glycol
BASF Wyandotte Corp.
Calcasieu Chem.
Celanese Chem.
Dow Chem.
Eastman Kodak
Northern Petrochem.
Olin Corp.
PPG Ind.
Shell Chem.
Texaco Jefferson Chem.
Union Carbide
Ashland Chem.
Hoadag Chem.
Geismar, LA
Wyandotte, MI
Lake Charles, LA
Clear Lake, TX
Freeport, TX
Plaquemine, LA
Midland, MI
Lcngview, TX
Morris, IL
Brandenburg, KY
Beaumont, TX
Guayanilla, PR
Geismar, LA
Port Neches, TX
Seadrift, TX
Taft, LA
Penuelas, PR
Texas City, TX
Institute and
S. Charleston, WV
Janesville, WI
Skokie, IL
•Source: SRI International, 1977
+ indicates user of ethylene oxide; - indicates non-users of ethylene oxide
-------
(Blackford, 19?6b). Ethylene glycol is used mainly in polyester production
and antifreeze formulations (Anonymous, 198lc).
5.2.2. Nonionic Surface-Active Agents. Of the nonionic surface-active agents
synthesized from ethylene oxide, -25% are of the cyclic variety, while -75%
are of the acyclic variety. In the cyclic group, ethylene oxide is used to
make ethoxylate alkyl phenols and alkylphenol-formaldehyde condensates.
Production of ethoxylated nonylphenol is probably the largest volume product
of the cyclic group; another large-volume product is ethoxylated dodecyl-
phenol. These surface-active agents are primarily used in detergents. The
acyclic surface-active category includes ethylene oxide used in the synthesis
of surface-active polyethylene glycol esters, ethoxylated alcohols, polyether
polyols, ethoxylated fats and oils, and miscellaneous ethoxylated products,
such as raercaptans, glycols, and polyols (Cogswell, 1980). Industry estimates
that ethylene oxide consumption for acyclic surface-active agents is expected
to increase. The manufacture of ethoxylated linear alcohols, used in heavy-
duty liquid detergents, will account for most of this growth (Cogswell, 1980).
5.2.3- Di-, Tri-, and Polyethylene Glycols. Ethylene oxide and ethylene
glycol react to form diethylene glycol, triethylene glycol, and polyethylene
glycol. Diethylene and triethylene glycols are obtained mainly as by-products
of ethylene glycol manufacture. Diethylene glycol is used to produce poly-
ester resins, as a textile lubricant, and in solvent extraction. Triethylene
glycol is used as a humectant and In natural gas dehydration, vinyl plasti-
cizers, and polyesters. Industry capacity to make diethylene glycol is 472
5-15
-------
million pounds/year; capacity to make triethylene glycol is =145 million
pounds/year (SRI International, 1977).
5.2.4. Glycol Ethers. Ethylene oxide is combined with alcohols to manufac-
ture glycol monoethers. These include ethylene glycol monomethyl, monoethyl,
and rnonobutyl ethers and diethylene and triethylene monoethyl, monornethyl, and
monobutyl ethers (Cogswell, 1980). Glycol ethers are used mainly as solvents.
Industry capacity to make glycol ethers is 865 million pounds annually (SRI
International, 1977).
5.2.5. Ethanolamines. Ethylene oxide reacts with ammonia to form a mixture
of mono-, di-, and triethanolamines. The proportion of each of three ethanol-
amines is dependent upon the ratio of reactants used. About 25-30/5 of all
ethanolamines are used in soaps and detergents, 5-20% in scrubbing acid gases
(especially in the synthesis of ammonia), 10? by the metal industry, 8% by the
textile industry and 5-15$ in toilet goods (Blackford, 1976b). The remainder
is used in various other applications.
5.2.6. Miscellaneous Applications. Ethylene oxide is consumed in the synthe-
sis of numerous commercial chemicals. The largest amount in the miscellaneous
group described above goes into the production of polyether polyols for
flexible polyurethane foams. In 1978, =100 million pounds (45 million kg) of
ethylene oxide were consumed in making these polyols (Cogswell, 1980).
Approximately 17 million pounds of ethylene oxide is used annually to
make the medicinals, choline and choline chloride; another 10 million pounds
of ethylene oxide is used annually in the manufacture of hydroxyethyl starch,
5-16
-------
a semi-synthetic gum used in textile sizing and in adhesives. The production
of hydroxyethyl cellulose, another adhesive additive, produced by the reaction
of cellulose with ethylene oxide uses -25 million pounds (11 million kg) of
ethylene oxide annually (Cogswell, 1980). Arylethanolamines are made by
reacting ethylene oxide with either aniline or aniline derivatives. It is
estimated that -3 million pounds (1.1 million kg) of ethylene oxide are used
annually to make arylethanolamines (Cogswell, 1980). They are used as
intermediates in the production of monoazo dyestuffs.
Acetal copolymer resins are produced by the catalytic copolymerization of
1,3,5-trioxane with a cyclic ether such as ethylene oxide. Ethylene oxide
consumed in this use is believed to have been -2 to 3 million pounds/year
(0.9-1.1 million kg) from 1977-1978 (Cogswell, 1980).
Ethylene oxide is used to produce ethoxylated cationic surface-active
agents (non-ionic surface-active agents are discussed in Section 5.2.2).
Several million pounds of ethylene oxide are used annually to produce cationic
agents such as ethoxylated (coconut oil alkyl) amine, ethoxylated (tallow
alkyl) amine, and various ethoxylated fatty acid amino amides (Blackford,
1976b).
Small amounts of ethylene oxide are also used as a fumigant, as a
sterilant for food and cosmetics, and in hospital sterilization (Gilmour,
1978). In 1975, an estimated 0.1 million pounds of ethylene oxide were used
for fumigant purposes (Landels, 1976). Dow Chemical (Kurginski, 1979) has
estimated that
-------
5.2.7. Discontinued Uses of Epoxides. The only significant discontinued use
of ethylene oxide known is the production of acrylonitrile. Until 1953 (when
acetylene was first used), all acrylonitrile was produced by the catalytic
dehydration of ethylene cyanohydrin that was prepared from ethylene oxide and
hydrogen cyanide. The reaction may be represented as follows:
HCN - »- HOCH2CH2CN - >• CH2 = CHCN
0
In 1956, American Cyanamid Company closed its 35 million pounds/year plant at
Warners, New Jersey, which was based on this process. From then until 1966,
when it was discontinued, this process was used only by Union Carbide at
Institute, West Virginia (Blackford, 1974). In 1965, Union Carbide consumed
90 million pounds of ethylene oxide to make acrylonitrile.
5.2.8. Projected or Proposed Uses. Wood treatment is a potentially important
market for epoxides (Anonymous, 1977). The USDA Forest Product Laboratory has
reported that treating southern yellow pine with epoxides (including ethylene
oxide, propylene oxide, and butylene oxide) improves its durability. The
treatment adds 20-30$ (by weight) of the epoxide to the wood.
5.2.9. Alternatives to Uses for Ethylene Oxide. More than 99$ of ethylene
oxide produced in the United States is used as a chemical intermediate in
chemical syntheses of glycols and other compounds. Alternatives would require
production routes from raw materials other than ethylene oxide.
Roughly 62% of the ethylene oxide produced is hydrolyzed to ethylene
glycol. A new process for making ethylene glycol directly from ethylene has
5-18
-------
been developed by Halcon, Inc. (Klapproth, 1976). Ethylene is reacted with
acetic acid in the presence of a catalyst to form mono- and diacetates, which
are then hydrolyzed to ethylene glycol. Oxirane Corporation has constructed
an 800 million pounds/year plant based upon this technology in Channelview,
Texas. This represents -25% of the present total industry ethylene glycol
capacity.
For other compounds synthesized from ethylene oxide, no information was
available on synthesis from other raw materials.
About 0.1 million pounds of ethylene oxide are used as a fumigant
annually (Dow Chemical estimates that the volume of ethylene oxide used as a
fumigant is <0.2% of total production, which in 1978 would be equal to <10
million pounds; Kurginski, 1979). It seems possible that alternative
commercial fumigants could replace ethylene oxide in many of its fumigant
uses.
5-3. POTENTIAL FOR ENVIRONMENTAL CONTAMINATION
5.3.1. Air Emissions from Production. Air emissions from direct-oxidation
ethylene oxide plants of all types consist mainly of ethylene, ethylene oxide,
and traces of ethane. The main process vent stream is responsible for most of
the air emissions in both air- and oxygen-based units. In air units, this
vent is located on the last purge reactor absorber and is principally spent
air (Np, 02, and some inert gases), carbon dioxide, traces of ethylene oxide,
and generally <2 mole % hydrocarbons. A catalytic converter is sometimes
added to the main process vent in an air system.
5-19
-------
The analogous vent stream from an oxygen-based system is ^lOO times
smaller and contains a much higher hydrocarbon concentration, and is conse-
quently used as a fuel. Table 5-6 presents approximate concentrations of
typical vent stream contaminants for the main process vent and the purge gas
vent.
For unburned vent gas from an oxygen-based unit, the total hydrocarbon
emissions have been estimated to be = 12 g/kg product. If methane is used as a
diluent and the purge gas incinerated, the emissions can be reduced to -^ g/kg
product. In an air-based unit without catalytic combustion of the purge gas,
hydrocarbon emissions are estimated to be >30 g/kg product. The use of a
catalytic converter can reduce emissions to -15 g/kg product. In a study
conducted for the U.S. EPA, the total ethylene oxide emissions in 1978 were
estimated to be -2 x 10 pounds (9.09 x 10 kg) (Systems Application, Inc.,
1982).
The major aqueous waste is draw-off from separator bottoms (Liepins et
al., 1977). The process water is recycled in ethylene oxide manufacture and
in the primary use of ethylene oxide as an intermediate in ethylene glycol
manufacture (Sittig, 1962, 1965). The aqueous waste from direct oxidation
plants will contain small amounts of glycols, aldehydes, and heavy glycols
(Cawse et al., 1980). No information was available on how much of the process
water eventually is treated, and no details were provided on treatment
methods. The wastewater will have a high BOD, but inorganic and refractory
organics appear to be minimal problems (Sittig, 1962, 1965; Spencer, 1971).
Conventional water treatment (including filtration and flocculation) with a
5-20
-------
TABLE 5-6
Typical Vent Gas Composition for Both Air- and Oxygen-Based
Ethylene Oxide Plants*
Stream
Air-Based
Range, mole %
Oxygen-Based
Main Process Vent
nitrogen
oxygen
methane
ethane
ethylene
ethylene oxide
carbon dioxide
argon
water
85-93
1.0-5
0-0.9
trace-0.2
trace-2.5
0-0.01
5-15
NP
0.1-1.5
2-35
5-7
1-35
trace-0.2
13-35
0-0.01
5-15
5-15
0.1-0.5
C02 Rich Purge Gas (water-free)
nitrogen
oxygen
ethylene and hydrocarbons
ethylene oxide
carbon dioxide
inert compounds
13-25
1-26
2.5-8.0
0-1.0
62-80
NP
NP
0.02
0.3-0.9
NP
99-99.7
0.005-0.015
*Source: Cawse et al., 1980
NP = Not present
5-21
-------
biological treatment appears sufficient to remove the major contaminants of
the process water (Spencer, 1971; Shenderova et al., 1972). There is no solid
waste produced during ethylene oxide manufacture.
5.3-2. Handling, Transport, and Storage. Ethylene oxide could be released as
a result of fugitive emissions or venting during its handling, transport, or
storage. No specific information was available to describe these losses;
information on current practices, procedures, or environmental controls was
sparse and no monitoring information was available. The potential situations
of release of epoxides without attempting to establish their relative
importance have been discussed in the following paragraphs.
Bulk shipments of ethylene oxide are commonly made by railroad 10,000 and
20,000 gallon freight tankers. Shipments are also made in special 55-gallon
drums and by highway truck tankers. Ethylene oxide is stored in bulk
containers, as well as in smaller quantities in 55-gallon drums. No
information was available on the usual emission controls used on storage and
transport containers. "Padded" containers, if used, would conserve vapors
which would otherwise be vented to the atmosphere. Emissions could also occur
during equipment purging in routine maintenance, gauge glass blowdown, or
leaks.
Release is also possible during transfer. In normal practice, railway
tankers are loaded and unloaded directly from or into storage tanks. The
chemical is transferred under nitrogen pressurization (-50 psi) for pumping.
Faulty equipment or over-pressurization could cause epoxide emissions. Small
amounts could be spilled during handling as well.
5-22
-------
One concern in addition to normal working and handling losses is release
from a storage container or transport-related accident. This could range from
a relatively minor incident, such as release through a pressure safety valve
or a rupture disc, to a major accident in which an entire storage container or
tanker ruptures. No information was available to predict how often the minor
release accidents occur or the amount of ethylene oxide they annually release.
Storage, transport, and handling methods have been extensively described
in literature supplied by manufacturers (BASF Wyandotte Corp., 1972; Dow
Chemical Company, 1977; Jefferson Chemical Company, undated a and b; Oxirane
Corporation, undated) and safety information sources (NFPA, 1975; MCA, 1971).
This literature chiefly concerns safety of humans and property. Tank cars for
ethylene oxide and propylene oxide are specified as ICC-105A100W and 105A100.
These are equipped with pressure relief valves which vent excessive pressure
into the atmosphere. The epoxides should be stored in an area detached from
the plant site and storage tanks should be diked. Ethylene oxide tanks should
be equipped with cooling pipes. Tanks must be equipped with pressure relief
valves, but specific instructions on emission control of excess pressure was
not included. Vapor recompression systems could be applied to prevent
emissions (Spencer, 1971).
5.3.3. Potential Environmental Formation. The major source of potential
inadvertent production of ethylene oxide in the environment is probably the
combustion of hydrocarbon fuels. Hughes et al. (1959) used gas-liquid
partition chromatography to separate and identify oxygenated derivatives of
hydrocarbons that were found in the combustion products of hydrocarbon fuels.
Among the oxygenated combustion products identified were ethylene oxide and
5-23
-------
propylene oxide. Barnard and Lee (1972) also identified these compounds in
the oxygenated products of n-pentane combustion. Seizinger and Dimitriades
(1972) suggested that ethylene oxide is a component of automobile exhaust.
They tested the combustion of simple unleaded hydrocarbon components of
gasoline. Stationary sources of hydrocarbon combustion also might emit large
quantities of these compounds into the environment.
Ethylene oxide has been identified in tobacco smoke (Binder and Lindner,
1972; Binder, 1974). It is not uncommon for tobacco to be treated with
ethylene oxide by cigarette manufacturers for its fumigant properties. Binder
and Lindner (1972) determined that the ethylene oxide concentration of
unfumigated tobacco smoke was 0.02 jig/m&, while fumigated tobacco smoke had a
concentration of 0.05 ng/m£ and extensively fumigated tobacco smoke had a
concentration of 0.30 jig/mjl. Binder (1974) determined that the ethylene oxide
content of smoke from unfumigated tobacco was 1 ng/g.
Epoxides are formed in the photochemical smog cycle. Olefins can be
converted to the corresponding epoxides by reaction with an organic peroxide
(Altshuller and Bufalini, 1965). Alkyl peroxides can decompose to yield an
epoxide and oxy radical (NAS, 1976).
Water disinfection procedures might convert olefins to epoxides. Olefin
conversion during chlorination of potable water would proceed by the same
route as for chlorohydrination production of the epoxide (see Section 3-6.5).
However, this process would require the conversion of ethylene to the
chlorohydrin (Morris, 1975; Carlson and Caple, 1977). Since ethylene is very
volatile, it seems unlikely that ethylene remains in water long enough for
this process to occur significantly.
5-24
-------
5.4. SUMMARY
This section discusses production, uses, and emission of ethylene oxide.
Ethylene oxide is produced almost exclusively by direct oxidation of ethylene
using either air or oxygen. Its 1981 production volume was 4937 million
pounds, down from 5220 million pounds in 1980.
The major emission sources from production facilities are the main
process vent for both air and oxygen units and the purge gas vent for air
units; fugitive emissions are also a source of ethylene oxide in the
atmosphere, although no emission estimate is available for this source. Total
air emissions from production have been estimated to be around 2 million
pounds based on 1978 production volume. Ethylene oxide also enters the
atmosphere from handling, storage, and transfer operations, as well as the
disposal of process wastes. There is no solid waste from ethylene oxide
manufacture.
More than 90% of the ethylene oxide produced is used captively as a
chemical intermediate, where there is some potential for environmental
contamination. Up to 10 million pounds are used annually for fumigation and
sterilization; ethylene oxide emissions from such uses might be significant.
Ethylene oxide may also be produced by hydrocarbon combustion (e.g.,
automobile exhaust).
5-25
-------
6. ENVIRONMENTAL FATE, TRANSPORT, AND DISTRIBUTION
6.1. INTRODUCTION
Epoxides are not persistent in the environment. Available information on
their chemical and biological properties characterizes them as highly reac-
tive. The available information was not sufficient to develop a definite
description of their environmental transport characteristics. Interphase
transport from water to air seems to be a slow process, but evaporation of
ethylene oxide applied as a sterilant or a fumigant appears to be a rapid
process. High water solubility and high vapor pressure result in significant
mobility within water or air.
Epoxide degradation has been fairly well characterized, and indicates
that ethylene oxide is reactive in all media. Available information on its
ionic reactions indicates that chemical (see Section 3) and biological
degradation produce the same degradation products. Its degradation in water,
soil, commodities, and manufactured products proceeds through ionic reactions.
Degradation in the atmosphere has not been well-characterized with respect to
processes or products. Available information indicates that it is very reac-
tive in photochemical smog cycle reactions. No information was available on
whether ionic reactions (e.g., with water vapor or water within aerosols)
significantly contributes to its degradation in the atmosphere.
6.2. ETHYLENE OXIDE FATE IN WATER
Ethylene oxide degrades in water by hydrolysis and related nucleophilic
reactions; aqueous radical reactions are not a significant process. The
5-1
-------
hydrolysis chemistry of ethylene oxide has been discussed in Section 3.6.5,
and the information presented there will be used in the present discussion.
Ethylene oxide has a hydrolysis half-life of 12.2 days in pure water,
12.9 days in filtered (0.22 \m filtered) Kanawha River water, and 14.2 days in
unfiltered Kanawha River water (Conway et al. , 1983). The Kanawha River water
had a pH of 7.4 and the initial ethylene oxide concentration was =^70 mg/&.
These variations in hydrolysis rates are well within the error limits of
hydrolysis experiments discussed by Mabey and Mill (1978).
It is interesting to note that the presumed presence of a microbial
population in the unfiltered river water did not decrease the half-life of the
ethylene oxide. Although the microbial concentration was not reported, the
lack of a significant change in degradation rate may indicate that biological
reactions are not significant in river water. Also, it should be noted that a
half-life of 12-14 days allows for exposure of biota and possibly humans to
ethylene oxide, although the addition of hypochlorite in water treatment
plants reduces the likelihood of human exposure, Conway et al. (1983) also
reported that pH variations would have less of an effect on the rate of
hydrolysis than temperature over a pH range of 5-10.
Evaporation from water also appears to be a significant removal process.
Conway et al. (1983) reported the calculated relative desorption coefficient
a, (a, = K. (ethylene oxide)/K,( 00), K, is the desorotion coefficient) to be
d d a a 2 a
0.31, 0.34, and 0.36 for 10, 20, and 30° C water. Experimental values for 22° C
water of 0.36 for no wind and 0.39 for a 5 m/s wind are reasonably consistent
with a calculated value of 0.34 and may be the result of increased turbulence
and wind flow. These values of a, indicate that ethylene oxide will be
desorbed from a body of water with a rate dependent upon the actual oxygen-
6-2
-------
transfer rate in a specific system. The rate of desorption will be less than
that for volatile low solubility organic compounds such as toluene, benzene,
and chloroform, which have an a of =0.65 (Rathburn and Tai, 1981).
Conway et al. (1983) also measured the BOD using 2 mi of domestic
sewage/BOD bottle. They found biooxidation was 5, 22, 40, and 52% (of
theoretical) on days 5, 10, 15, and 20, respectively. They suggested that in
a sewage treatment plant, where the microbial population is much higher,
biodegradation might be very fast; however, from their data it is not possible
to determine whether the chemical degraded is actually ethylene oxide or
whether it is ethylene glycol (from hydrolysis), since the hydrolysis half-
life (=14 days) is similar to the BOD half-life of slightly less than 20 days.
Hendry et al. (1974) reported the rate constant for the reaction of one
epoxide with alkyloxy radical proceeding by a-hydrogen abstraction to be 8.5 x
10 M~ s~ /a-hydrogen, or 3.4 x 10 M~ s~ for ethylene oxide. Given an
-14
alkyloxy radical concentration in ambient water of 10 M, the half-life for
this process is -6 years. Hence, hydrolysis and evaporation appear to be the
dominant fate processes for ethylene oxide, while no definitive statement can
be made regarding its biodegradation.
6.3. ETHYLENE OXIDE FATE IN SOIL
Pertinent data regarding chemical degradation of ethylene oxide in soil
were not located in the available literature. It seems reasonable, given the
composition of soil, that the half-life of ethylene oxide would be shorter in
soil than in water.
6-3
-------
6.4. ETHYLENE OXIDE FATE IN THE ATMOSPHERE
Little direct information on epoxide behavior in the atmosphere was
available; however, some characteristics of ethylene oxide behavior can be
inferred from the data on free-radical chemistry.
The atmospheric reactivity of volatile organic chemicals has been
characterized by their relative reaction rates with hydroxyl radicals in the
gas phase (Cupitt, 1980; Darnall et al., 1976); however, there are a number of
difficulties in determining an atmospheric half-life or lifetime for ethylene
oxide based on hydroxyl radical reactions. One important difficulty is in
choosing the appropriate hydroxyl radical concentration: a number of different
modeling and direct measurement efforts have provided a wide range of values
for both average and altitude specific concentrations of hydroxyl radicals. A
reasonable compromise for an average OH concentration is 1 x 10 molecules
cm based on more recent modeling efforts (Cupitt, 1983). For ground level,
the concentration may be somewhat higher, possibly around 1.3 to 1.4 x 10
molecules cm during the summer (Crutzen and Fishman, 1977; Logan et al.,
1981). Using these two values, a temperature of 300 K, and the Arrhenius
equation of Fritz et al. (1982) (see Section 3.6.5), the lifetime of ethylene
oxide is found to lie somewhere between 100 days (using the upper limit of
the Arrhenius equation) and 215 days (the lower limit) for a hydroxyl radical
r _-j
concentration of 1 x 10 molecules cm, and between 74 days and 159 days
(using the upper limit) for a hydroxyl radical concentration of 1.35 x 10
molecules cm. This lifetime is in sharp contrast with the significantly
shorter values found for other ethers. For example, tetrahydrofuran, a five
membered cyclic ether, has a lifetime of =1 day. Fritz et al. (1982)
suggested that this disparity is due to the distorted sp bonds in ethylene
6-4
-------
oxide that give rise to a hydrogen-abstraction activation energy of 5.8
kcal/raol, higher than the standard 2.8 kcal/mol.
Bogan and Hand (1978) found the absolute rate constant of the reaction of
— 1 A
oxygen atoms [0(-*P)] for ethylene oxide to be (6.3 + 0.18) x 10~
cra3/mcule-sec at 300 K. This rate is several orders of magnitude slower than
that for the hydroxyl radical reaction, and yields a half-life of 1400 years,
-3 h o
given an atmospheric 0(JP) concentration of 2.5 x 10 molecules/cm (Graedel,
1978).
P
Sickles et al. (1980) measured the rate of ozone production in a Teflon
smog chamber to rank 19 compounds relative to propane. The chambers were
irradiated with sunlight outdoors. Purified air, the organic compound to be
tested and N02 were added before sunrise to multiple chambers; ethylene oxide-
to-NC>2 ratio at the onset of the experiment was 4:0.067. Ethylene oxide was
much less reactive than propane. The rank order of reactivities found was:
Acrylonitrile >perchloroethylene >ethanol >ethylacetate >acetone >methanol
>acetic acid >propane >ethylene dichloride >acetylene >chloroforra >dimethyl
formamide >benzaldehyde >methylene chloride >pyridine >*ethylene oxide >methyl
chloroform >phenol >acetonitrile >nitrobenzene. The relative ordering of com-
pounds was similar in an indoor smog chamber study (Dimitriades and Joshi,
1977). The smog chamber study of Joshi et al. (1982) also concluded the low
reactivity of ethylene oxide (half-life >53 hours). All of these results
indicate that ethylene oxide is relatively unreactive in the atmosphere
compared to other ethers.
With the information currently available, no definitive statement can be
made regarding the atmospheric fate or lifetime of ethylene oxide.
6-5
-------
6.5. DEGRADATION IN COMMODITIES AND MANUFACTURED PRODUCTS
Ethylene oxide is registered in the United States for use as a fumigant
or sterilant for several stored-food commodities and manufactured products
(Goncarlovs, 1983). These include its use as a fumigant for bulk food
containers, other food containers, stored grain, stored fruits, stored
processed foods, tobacco products, garments, furs, stored herbs and spices,
furniture, aircraft, buses, railroad cars, and laboratory animal bedding. As
a sterilant, it is used principally on hospital equipment and Pharmaceuticals.
Ethylene oxide is used as a fumigant chiefly to protect stored products from
insect or microbial destruction. The fate of this epoxide and its residue are
especially important in materials, commodities, and products coming into close
contact with humans, such as surgical equipment, Pharmaceuticals, and food
service and packaging materials (Wesley et al., 1965; Alguire, 1973; Holmgren
et al., 1969; Gilmour, 1978).
The study of the fate of ethylene oxide in these materials has
established that it will degrade to glycol and halohydrin or evaporate. The
degradation results from chemical and/or enzymatic activity. The halohydrin
route requires epoxide reaction with inorganic halide. The halide could be
naturally present, added or derived from organic halides. Bromide ion is
often supplied by degraded methyl bromide, another fumigant (Rowlands, 1971;
Lindgren et al., 1968).
Scudamore and Heuser (1971) measured the apparent degradation and
evaporation of ethylene oxide and its residues, ethylene chlorohydrin and
ethylene bromohydrin, over a 1-year period. Apparent first-order specific
rate constants, k, were calculated for epoxide dissipation. The rate
5-6
-------
constant, k, combined losses from the degradation (chemical and metabolic
pathways), k-, and evaporation, k..:
k = kD + kv
The glycols (ethylene and diethylene) were measured once at 6 months or 1 year
after treatment. The parameters considered included the ethylene oxide
treatment (dose and temperature during application), the moisture content of
the commodity, storage temperature and type of storage (closed containers
versus open trays). Ethylene oxide residues dissipated rapidly. While the
estimated half-life was longest at 10°C in sealed containers, it never
exceeded 2 weeks. Increasing the ethylene oxide dose did not have a simple
effect on its loss rate. For the most part, small increases in the dose
slightly decreased the loss rate, while very large increases caused larger
decreases in the rate of loss and, sometimes, non-linear changes. The effect
of moisture content appeared varied and relatively small. Scudamore and
Heuser (1971) also monitored some commercially treated products and found
ethylene halohydrin residues but no ethylene oxide residues. They concluded
that ethylene oxide will usually dissipate from treated commodities, but,
under some circumstances, small quantities could persist for several months.
Stijve et al. (1976) discussed the fate of ethylene oxide applied as a
fumigant to commodities. They suggested that ethylene oxide could be retained
by physical adsorption, but that it would persist not more than a few weeks
before volatilization or reaction with natural constituents of the commodity.
Ben-Yehoshua et al. (1971) examined ethylene oxide residues during the
treatment of dates. They reported a small ethylene oxide loss in the empty
6-7
-------
container and ascribed this to apparent adsorption to the container walls.
The larger losses found with 2.1 kg of dates in the container were due to
ethylene oxide uptake by the fruit. Ethylene oxide loss in treated dates left
in open containers was attributed to degradation (to the chlorohydrin and
glycol) and volatilization.
The available information on the fate of ethylene oxide applied to
manufactured goods was less extensive as that on its fate in commodities. All
available information suggests behavior similar to that discovered in
commodities. Alguire (1973) described losses of ethylene oxide from
polystyrene creamer cups and cream cheese wrappers at ambient temperature and
open to the environment. The ethylene oxide did not degrade on the
polystyrene cups, and was lost solely through out-gassing. More than 90%
evaporated by the first day, and no residual ethylene oxide remained after 5
days. Ethylene oxide loss from cream cheese wrappers consisted primarily of
conversion to ethylene glycol; no ethylene chlorohydrin was detected at any
time. Ethylene oxide was completely gone by the tenth day.
Some studies have identified ethylene chlorohydrin residues in manufac-
tured goods sterilized with ethylene oxide. These studies did not seek any
information on volatilization losses. Brown (1970) identified ethylene oxide
and its derivatives on treated equipment made of rubber, Dacron, and poly-
vinylchloride, but did not detect chlorohydrin on polyethylene equipment.
Holmgren et al. (1969) measured 0-1500 ppm chlorohydrin on 21 ethylene oxide-
treated drugs.
6-8
-------
6.6. BIOACCUMULATION IN AQUATIC ORGANISMS
Specific experimental information regarding the b ioac cumulation of
ethylene oxide in aquatic organisms is not available. Veith et al. (1979)
have suggested the calculation of BCF from the following equation:
log BCF = 0.76 log K -0.23
\J W
where K is the partition coefficient between octanol and water. Using this
o w
equation and the log K of -0.30, reported by Hansch and Leo (1979), the BCF
for whole fish was calculated to be 0.34.
6.7. SUMMARY
This section discusses the results of studies related to the fate of
ethylene oxide in the environment. In water, ethylene oxide will degrade by
hydrolysis and related nucleophilic reactions with a half-life of =12-14 days
at 298 K. Lower temperatures lengthen the half-life; pH changes have minimal
effects. Volatilization will also be a significant process although less so
than for sparingly soluble compounds like toluene, chloroform or benzene.
There is no conclusive evidence that microbial degradation is significant;
however, the biological components of sewage sludge might react rapidly with
ethylene oxide. The fate of ethylene oxide in soil will probably be similar
to that in water; its half-life will probably be shorter.
The fate of ethylene oxide in the atmosphere is not obvious from the
information present in the literature. Rate constants are available for
hydroxyl radical and oxygen atom [0( P)] reactions as well as smog chamber
6-9
-------
studies. All predict that ethylene oxide will persist in the atmosphere, but
the actual lifetime cannot be predicted.
In commodities, food containers, and manufactured goods, ethylene oxide
appears to volatilize or to hydrolyze to glycol or halohydrin with a half-life
of -2 weeks.
6-10
-------
7. ENVIRONMENTAL LEVELS AND EXPOSURE
7.1. INTRODUCTION
The purpose of this document is to present available information relevant
to human health effects that could be caused by this substance.
Any information regarding sources, emissions, ambient air concentrations,
and public exposure has been included only to give the reader a preliminary
indication of the potential presence of this substance in the ambient air.
While the available information is presented as accurately as possible, it is
acknowledged to be limited and dependent in many instances on assumption
rather than specific data. This information is not intended, nor should it be
used to support any conclusions regarding risks to public health.
If a review of the health information indicates that the Agency should
consider regulatory action for this substance, a considerable effort will be
undertaken to obtain appropriate information regarding sources, emissions, and
ambient air concentrations. Such data will provide additional information for
drawing regulatory conclusions regarding the extent and significance of public
exposure to this substance.
7.2. ENVIRONMENTAL LEVELS
Ambient monitoring has portrayed ethylene oxide as an almost non-existent
contaminant of environmental or biological samples. Although ethylene oxide
is rarely identified in monitoring studies, its principal degradation products
(glycols and halohydrins) have been found.
7-1
-------
No monitoring data were available for ethylene oxide in biological
tissues, except for some tissue-distribution studies. Since epoxides are
reactive alkylating agents, it is reasonable to expect that ring-opening
reactions will occur rapidly in biological systems so that the finding of
detectable levels in environmental biota is unlikely (Anderson, 1971).
Only one ambient air monitoring study reporting the presence of ethylene
oxide in air was found in the available literature. Bertsch et al. (1974)
tentatively identified ethylene oxide in the ambient air near the University
of Houston. However, the authors used Tenax as the adsorbant for trapping air
contaminants and its use casts doubt on their tentative identification, since
Tenax does not adequately retain ethylene oxide (see Section 4.1).
U.S. EPA (1976) listed one monitoring observation of ethylene oxide in
water. It was observed in the effluent from a chemical plant in Brandenburg,
Kentucky. No other epoxide observation was reported. U.S. EPA (1976) also
noted observations of ethylene halohydrin, which might have been released in
industrial wastes as such, rather than occurring as residues from epoxide.
No other reports of ethylene oxide in ambient air or water were found,
yet Systems Application, Inc. (1982) reported that the maximum possible
exposure concentration level of ethylene oxide, based on dispersion models,
•3
was 5 |ig/m (2.77 ppb). The justification for this value could be the
reactivity of ethylene oxide or the lack of an adequate sampling method (see
Section 4). Most sampling methods either lose significant amounts of ethylene
oxide on even short-term storage, or use adsorbants with a very poor affinity
for ethylene oxide (e.g., Tenax GC); however, this is certainly not the case
for all studies, especially those using freeze-out techniques. Since
environmental sample concentrations are rarely as high as workplace sample
7-2
-------
concentrations (particularly in the case of a reactive molecule such as
ethylene oxide), the well documented NIOSH (1977) method becomes inadequate.
The problem is compounded by the fact that few monitoring studies are
undertaken to identify only a single compound in the environment. These
studies must assume some compromise between completeness and speed, making it
impossible to optimize conditions for the detection of any one compound.
Several studies have examined the residues of ethylene oxide that has
been applied to commodities and manufactured goods as a fumigant and
disinfectant. Another portion of this report (Section 6.5) describes
investigations on the fate of this epoxide. The information here concerns
residues in actual commercial products.
Scudamore and Heuser (1971) evaluated ethylene oxide and its metabolites
in commercially treated products. While they never detected ethylene oxide in
commercial products, they did find ethylene chlorohydrin residues ranging from
10-70 ppm. Lindgren et al. (1968) reviewed studies on residues from ethylene
oxide treatment, most of which were fate studies rather than ambient
monitoring studies. They suggested that residual epoxide could be present in
commercial products.
Ethylene oxide is a common sterilant for surgical equipment. Its fate in
plastic and rubber surgical equipment parallels its behavior in commodities.
Brown (1970) monitored residues on various hospital equipment sterilized with
ethylene oxide. Ethylene oxide was observed in three samples, one of which
had received treatment =80 days previously. Ethylene chlorohydrin was
detected in 10 samples.
7-3
-------
7.3. EXPOSURE
The available data concerning the environmental levels of ethylene oxide
are insufficient to properly estimate exposure; however, a general overview
can be made. Over 5 billion pounds (>2 billion kg) of ethylene oxide is
produced yearly. The vast majority is used captively as a synthetic
intermediate. Perhaps 10 million pounds (4.5 million kg) is used for
fumigation/sterilization for products that include food commodities, medical
devices, Pharmaceuticals, and cosmetics. This use constitutes the only
documented potential exposure to ethylene oxide, though the extent of this
exposure must be determined. Ethylene oxide also appears to be a product of
incomplete combustion, and has been identified in automobile and diesel
exhaust and in tobacco smoke. It can be formed during the photochemical smog
cycle, but appears to be rapidly destroyed.
7.4. SUMMARY
This section discusses the results of monitoring studies conducted to
measure the levels of pollutants, including ethylene oxide, in the environ-
ment. Very little information is available on ambient monitoring, no
confirmed detection of ethylene oxide in air has been reported, and only one
report exists for water. The lack of more monitoring reports may be because
most, but not all, sampling methods would not detect ethylene oxide even if it
were present. Several studies have examined the persistence and fate of
ethylene oxide in commodities and commercial goods including food, medical
supplies, and drugs.
7-4
-------
8. ECOLOGICAL EFFECTS
8.1. MICROORGANISMS AND INSECTS
Ethylene oxide is used as a fumigant for foods (particularly grains) and
spices, and shows major microbial, insecticidal, and acaricidal activity
(Sykes, 1964; Lindgren and Vincent, 1966).
Fumigation with ethylene oxide has been used to control a wide variety of
bacteria, fungi, rickettsiae and viruses. Sykes (1964), for example, reported
that exposures to gaseous ethylene oxide at concentrations of 1-10$ will kill
Bacillus globigii, Staphylococcus aureus, Escherichia coli, Chromobacterium
prodigiosum, and Mycobacterium phlei within a few hours. Roberts et al.
(19^3) found that 10$ gaseous ethylene oxide will kill Bacillus anthracoides
in 8 hours. Ethylene oxide also has significant sporicidal activity against
dry bacterial spores (Bruch and Koesterer, 1961). Exposure of Bacillus
subtilis spores to 1-2$ vapor concentrations of ethylene oxide killed j>95$ of
the spores within 4 hours. A 5$ gaseous concentration of ethylene oxide
produced 90$ kill of airborne B. globigii spores in <2 hours (Roberts et al.,
19^3). Treatment of agar slants containing yeasts and fungi with 8$ gaseous
ethylene oxide for 3 hours was lethal to these microorganisms (Whelton et al.,
1946). Skeehan (1959) indicated that herpes simplex, vaccinia, and bovine
respiratory viruses are susceptible to saturated ethylene oxide vapor
treatment.
Susceptible insects common to stored products include the flour beetle,
rice weevil, and grain weevil (Lindgren et al., 1954). Ethylene oxide will
kill one-half of the stored-product insect population at a concentration range
8-1
-------
of 6-18 mg/& (Ong, 1948).. Lindgren and Vincent (1966) reported a major
reduction in available tissue glutathione content of Calliphora larvae exposed
to ethylene oxide. Decrease in tissue glutathione via depletion of reduced-SH
groups may be the mechanism of toxicity. The insect toxicity of ethylene
oxide has been ranked by these authors as intermediate between those of
ethylene dibromide and' ethylene dichloride. A bibliography of ethylene oxide
insecticidal properties, citing 185 references, has been published (Young and
Busbey, 1935).
8.2. PLANTS
Pertinent data regarding the effects of exposure to ambient levels of
ethylene oxide on plants were not found in the available literature. As
detailed in Section 9.4, ethylene oxide is capable of inducing mutations and
chromosomal aberrations in plants.
8.3. AQUATIC ORGANISMS
Limited information is available on the toxicity of ethylene oxide to
aquatic organisms. The acute toxicity of ethylene oxide appears to be
moderate, as indicated by LC^s in the range of 84-90 mg/Jl for fish, a mean
48-hour LC5Q of 212 mg/& for Daphnia and 745 mg/Jl for brine shrimp (Table
8-1). kCj-Q values for the hydrolysis product ethylene glycol were >10,000
mg/& for the above species except goldfish (which were not tested) (Conway et
al., 1983). If reacted to form ethylene chlorohydrin, the 96-hour LC,-0 for
fathead minnows was =90 mg/Jl (Conway et al., 1983).
8-2
-------
TABLE 8-1
Acute Aquatic Toxicity of Ethylene Oxidec
LCc;n (95? Confidence limits), mg/£
Test Procedure
range-finding , static, aerated
range-finding , static, sealed
under oxygen
definitive static acute (no
(X> aeration)
static acute
static acute
static acute
Test Organism 24 hr
fathead minnow 274 (150-500)
fathead minnow 86 (50-150)
fathead minnow 90 (63-125)
goldfish 90
Daphnia magna >300
270
260
brine shrimp >500
350
570
48 hr
NA
NA
89 (63-125)
NA
300
137 (83-179)
200 (150-243)
>500
1000
490
96 hr
NA
NA
84 (73-96)
NA
NA
NA
NA
NA
NA
NA
Reference
Conway
et al. ,
Conway
et al. ,
Conway
et al. ,
Bridie
et al. ,
Conway
et al. ,
Conway
et al . ,
1983
1983
1983
1979
1983
1983
Source: Conway et al., 1983
Range-finding tests used 2 fish/test concentration
Definitive tests used 10 fish/test concentration
NA = Not applicable
-------
9. BIOLOGICAL EFFECTS IN ANIMALS AND MAN
9.1. PHARMACOKINETICS
9.1.1. Absorption. Only limited data regarding the absorption of ethylene
oxide were found in the available literature. However, acute toxicity data
suggest that absorption occurs readily via the respiratory and gastrointes-
tinal tracts (Table 9-1).
9.1.2. Distribution. Information concerning the distribution of ethylene
oxide in the body is limited. Two studies have shown that it is found in many
tissues following inhalation exposure or intravenous administration.
Ehrenberg et al. (1974) conducted inhalation studies with radioactively-
labeled [1,2- H] ethylene oxide. Following exposure of mice to 1.15 ppm of
the labeled chemical in air for 75 minutes, the highest levels of radioac-
tivity (in unidentified chemical form) were associated with proteins isolated
from the lungs, kidneys, and liver. Lower levels of radioactivity were
measured in the testes, brain, and spleen, but additional organs were not
analyzed.
Appelgren et al. (1977) carried out whole-body autoradiography on mice
14
injected intravenously with radioactive [ C]-ethylene oxide (label position
unspecified). Preliminary inhalation studies with labeled ethylene oxide
showed a similar tissue distribution of the compound similar to that seen
following intravenous injection, except for a high initial labeling of the
respiratory mucosa (data not shown). Two minutes after the injections, con-
centrations of radioactivity 2-3 times those seen in the blood were observed
9-1
-------
ru
TABLE 9-1
Acute Toxicity of Ethylene Oxide
Route
oral
oral
oral
oral
oral
ihl.
ihl.
ihl.
ihl.
ihl.
i .v.
i .v.
i.p.
i.p.
i.p.
s.c.
Species
rat
rat
rat
guinea pig
rabbit
rat
rat
guinea pig
mouse
dog
rabbit
rat
rat
mouse
rabbit
rabbit
Sex
M
M
M
M,F
M,F
M
M,F
NR
F
M
M,F
M
M,F
M,F
M,F
M,F
Strain
Wistar
NR
NR
NR
NR
white
Sherman
NR
white
beagle
NR
NR
NR
NR
NR
NR
Dose
330 mg/kg
100 mg/kg
200 mg/kg
270 mg/kg
631 mg/kg
1460 ppm/4 hours
4000 ppm/4 hours
7000 ppm/2.5 hours
835 ppm/4 hours
960 ppm/4 hours
178 mg/kg
355 mg/kg
178 mg/kg
178 mg/kg
251 mg/kg
200 mg/kg
Response
LD50
0/5 died
5/5 died
LD50
LD50
LC50
LC50
LClow
LC50
LC50
LD50
LD50
LD50
LD50
LD50
LD50
Reference
Smyth et al., 1941
Hollingsworth et al.,
Hollingsworth et al.,
Smyth et al., 1941
Woodward and Woodward,
Jacobson et al., 1956
Carpenter et al., 1949
Waite et al., 1930
Jacobson et al., 1956
Jacobson et al., 1956
Woodward and Woodward,
Bruch, 1973
Bruch, 1973
Bruch, 1973
Woodward and Woodward,
Woodward and Woodward,
1956
1956
1971
1971
1971
1971
Ihl. = inhalation; i.v. = intravenous; i.p. = intraperitoneal; s.c. = subcutaneous; NR = not reported
-------
in the liver, kidneys, and pancreas. Tissue labeling 20 minutes to 4 hours
after exposure showed high levels of radioactivity in the liver, kidneys,
lungs, intestinal mucosa, epididymis, cerebellum, and testes. Twenty-four
hours after injection, radioactivity was still found in the liver, intestinal
mucosa, epididymis, cerebellum, bronchi, and bone marrow. Since these
observations were made on auto radiographs, quantitative results were not
reported. The extent of bioexchange of the radioactive label into natural
body constituents also could not be determined in this study.
9.1.3. Metabolism. Comprehensive studies designed to fully characterize the
metabolic fate of ethylene oxide have not been conducted.
Significant concentrations of ethylene glycol were detected in the plasma
of four beagle dogs following the intravenous administration of 25 or 75 mg/kg
ethylene oxide on separate occasions (Martis et al. , 1982). Urinary excretion
data indicated that 7-2^% of the administered dose was excreted in the urine
within 24 hours as ethylene glycol; the mean percentages of the low and high
doses that were excreted in the urine were 13-5 +_ 3-5% and 14.2 + 8.1?,
respectively.
14
Two urinary metabolites were detected when [1 ,2- C] ethylene oxide was
administered to Sprague-Dawley rats via a single intraperitoneal injection at
a dosage of 2 mg/kg (Jones and Wells, 1981). The urinary metabolites were S-
(2-hydroxyethyl)-cysteine (9% of the dose) and N-acetyl-S-(2-hydroxyethyl) -
cysteine (33% of the dose), which suggests that the metabolism of ethylene
oxide involved conjugation with glutathione. A small percentage of the dose
14
was exhaled as C0» and as unchanged ethylene oxide (Section 9.1.4).
9-3
-------
In the inhalation study with mice summarized in Section 9.1.2 (Ehrenberg
et al. , 1974), the only urinary metabolite characterized was 7-hydroxyethyl-
guanine, >*iich accounted for a minor amount (0.007$) of the total urinary
radioactivity. Significant alkylation of tissue proteins was found, and
alkylation of DNA was confirmed by the identification of a high specific
activity radiolabeled 7-hydroxye thy Iguan ine. Gumming et al. (1981) reported
large differences in the patterns of initial alkylation as well as removal of
total alkylation products from the DNA of various tissues (i.e., testis,
liver, lung, kidney, spleen) of mice following inhalation exposure to tritium-
labeled ethylene oxide. Thus, ethylene oxide distributes and reacts
extensively throughout the body.
9.1.4. Elimination. In the inhalation study with mice (Ehrenberg et al. ,
1974) using tritium-labeled [1 ,2-%]-ethylene oxide (see Section 9.1.3), it
was found that 78% (mean value) of the absorbed radioactivity was excreted in
the urine within 48 hours. The biological half-life in mice was reported to
be -9 minutes, indicating rapid urinary elimination.
Approximately 43$ of the administered radioactive dose of [1 ,2- C]
ethylene oxide (2 mg/kg, single injection) was excreted in the urine of mice
over 50 hours, most of which (=40/5) appeared within 18 hours of dosing (Jones
and Wells, 1981). Two urinary metabolites, S-(2-hy drox yeth y 1)-cy ste ine and N-
acetyl-^-(2-hydroxyethyl)-cysteine accounted for 9 and 33% of the dose,
respectively. Within 6 hours, 1.5$ of the dose was exhaled as C0_ and ~\% as
unchanged ethylene oxide, but these are not maximum values (exhaled
radioactivity was not sampled at later post-exposure times).
9-4
-------
Martis et al. (1982) investigated the elimination kinetics of intra-
venously administered ethylene oxide in beagle dogs. Four dogs received
single 25 and 75 mg/kg injections of the compound on separate occasions, and
venous blood was sampled for ethylene oxide and ethylene glycol at 0, 0.08,
0.25, 0.5, 1.0, 2.0, 3.0, 4.0, 7.0, and 24 hours after administration. It was
found that the ethylene oxide cleared rapidly from the plasma, and that in all
cases concentrations decreased to <2% of the zero-time value within 5 hours.
The plasma concentration of ethylene oxide declined exponentially, and first-
order rate constants of 0.025 + 0.006 min~1 and 0.023 + 0.010 rain" for the
low and high dosages, respectively, were calculated from the plasma
concentration corresponded to plasma half-lives of 29.3 + 5.7 min and 36.5 +
18.5 min. It was noted that the lack of significant differences in kinetic
parameters (i.e., elimination rate constant, plasma half-life, apparent
distribution volume, total body clearance) at the two dose levels indicates
that the elimination kinetics are not dose-dependent. Ethylene glycol was
formed quite rapidly following the administration of ethylene oxide, and
plasma concentrations reportedly exhibited the characteristics of a metabolite
in a one-compartment model; maximum plasma concentrations of ethylene glycol
were reached by 90 + 24.5 minutes (25 mg/kg) and 120 + 42.4 minutes (75 mg/kg)
post-in ject ion. Plasma concentration-time data for ethylene glycol following
the intravenous injection of 35 and 106 mg/kg of ethylene glycol indicated
half-lives of 177.1 + 29.3 and 264.9 + 90 minutes, respectively.
9-5
-------
9.2. ACUTE, SUBCHRONIC, AND CHRONIC TOXICITY
9.2.1. Effects in Humans.
9.2.1.1. ACUTE EXPOSURE — Case reports indicate that headache, nausea,
vomiting, dyspnea, and/or respiratory irritation are common effects of acute
inhalation exposure to ethylene oxide (Greaves-Walker and Greeson, 1932;
Blackwood and Erskine, 1938; von Oettingen, 1939; Anonymous, 1947; Sexton and
Henson, 19^9; Hollingsworth et al. , 1956; Curme and Johnston, 1952; Salinas et
al. , 1981). Symptoms of poisoning have been reported to be delayed by several
hours following exposure. Similar effects (e.g., marked nausea and profuse
vomiting), as well as mild leukocytes is and blisters (discussed subsequently),
developed in three chemical plant workers who were dermally drenched with 1?
aqueous ethylene oxide solution (Sexton and Henson, 19**9). Inhalation expo-
sure to high concentrations of ethylene oxide for brief periods has been
associated with bronchitis, pulmonary edema, and emphysema (Thiess, 1963), as
well as convulsive movements (Salinas et al. , 1981). In a controlled study of
the effects of ethylene oxide on human volunteers, Greaves-lfelker and Greeson
(1932) observed that ethylene oxide at =2200 ppm was slightly irritating to
four subjects. At a 5-fold higher concentration, the compound had a definite
effect on nasal mucosa within =10 seconds.
Three chemical plant workers drenched with 1/E aqueous ethylene oxide
solution developed marked nausea and profuse vomiting several hours following
exposure (Sexton and Henson, 19^9). Large vesiculated blisters developed in
the areas of exposed skin, and two workers who had complete blood counts taken
showed a mild leukocyte sis.
9-6
-------
Cobis (1977) reported a very low incidence of health-re la ted effects due
to exposure to ethylene oxide in Veteran's Administration medical facilities.
Ethylene oxide was used for sterilization purposes in 162 hospitals and 7
outpatient clinics over an average of 8.2 years. Only 12 employees were
reported to have been involved in exposure incidents, and symptoms included
watering eyes, nausea, and skin irritation. These cases are currently being
followed to determine possible exposure sequelae. The average exposure con-
centration was not given, and it is presumed (although not stated) that the
employees were exposed to ethylene oxide vapor.
The derma to logical effects of ethylene oxide contact have been reviewed
by Taylor (1977). Concentrated ethylene oxide evaporates rapidly from the
skin and produces a freezing effect, resulting in burns ranging from first-
through third-degree severity. Ethylene oxide gas retained in porous
materials that have not been properly aired can produce skin irritation. Foot
burns (Phillips and fey, 19^9) and hand burns (Royce and Vbore, 1955), for
example, have been observed in workers that wore ethylene oxide-sterilized
rubber boots and rubber gloves, respectively. Biro et al. (1974) described a
hospital incident in which 19 women were burned by surgical gowns and drapes
that had been sterilized with ethylene oxide. Joyner (1964) found in a 2-year
retrospective study of medical records that ethylene oxide plant workers had
experienced exposure-re la ted burns.
Sexton and Henson (19^9) described the derma to logical reactions that
occurred in 6 men whose skin was directly exposed to a 1/5 water solution of
ethylene oxide for periods ranging from 15 minutes to 3 hours. The men with
the maximum exposures (2-3 hours) exhibited the most marked cutaneous effects
9-7
-------
(vesicular eruptions), but nausea and vomiting were the only systemic effects
noted.
In a subsequent study, Sexton and Benson (1950) applied 1-100? solutions
of ethylene oxide to the skin of 8 volunteer subjects for tine intervals that
ranged from 20 seconds to 95 minutes. The magnitude of skin injury appeared
to be related to the duration of contact and the concentration. The most
hazardous concentrations of ethylene oxide were in the 50% range, since the
manifestation arbitrarily examined in this study (minimal second-degree burn
demonstrated as an area of erythema with one or more superimposed vesicles)
was produced in 45 seconds with this solution. The degree of skin injury was
proportionately decreased at concentrations both greater and less than 50%.
The lowest ethylene oxide concentration investigated (1?) produced a mild
reaction (erythema) after 50 minutes of exposure. The milder skin reactions
at concentrations >50% were attributed to the fact that the more concentrated
solutions boil vigorously, thus preventing efficient skin penetration; the
more dilute solutions lacked sufficient chemical to cause injury except after
prolonged contact. Delayed skin sensitization developed in 3 of the 8
subjects.
Shupack et al. (1981) demonstrated that human skin reactions were
directly related to total dose when exposures were to ethylene oxide that was
retained in permeable materials. In tests with 12 unsensitized volunteers, it
was found that patch materials that rapidly lose ethylene oxide (i.e., fabric
or rubber) elicited few reactions, even at ethylene oxide levels as high as
3000-5000 ppm after 4-8 hours of contact. Patch materials that lost ethylene
oxide slowly produced mild skin reactions (erythema plus edema) at material
levels as low as 1700 ppm (PVC film) and 1000 ppm (PVC blocks) after similar
9-8
-------
durations of contact. Patches were removed from the subjects after 1, 2, 4,
and 8 hours; it was found that most of the ethylene oxide diffused from the
fabric and rubber patches within 1 hour and from the PVC film patches within 4
hours, but that the PVC block retained a substantial portion of ethylene oxide
residue at 4 hours. In a subsequent experiment the same subjects (i.e., those
previously exposed in the first experiment) were exposed to patch materials
that retained ethylene oxide the longest (thick PVC blocks and petrolatum
applied to Webril pads). It was found that the reactions were most widespread
when the ethylene oxide levels in these materials were =1000 ppm; erythema
appeared in 10 of the 12 PVC block subjects and 10 of the 12 petrolatum
subjects after 4-8 hours of contact, and cleared within 3-4 days. Reactions
were not elicited at nominal levels of 10 or 100 ppm ethylene oxide in PVC or
petrolatum, although one subject who had developed sensitivity to 1000 ppm
ethylene oxide in PVC block in the first experiment showed a mild delayed
reaction to 100 ppm. Little or no reaction developed to patches that
contained ethylene oxide by-products that were present in the original patches
(i.e., ethylene glycol and ethylene chlorohydrin), indicating that ethylene
oxide was the toxic agent.
Although incidental findings in the Sexton and Henson (1950) and Shupack
et al. (1981) experimental studies described above suggest that ethylene oxide
can cause skin sensitization, Thiess (1963) did not observe sensitization in
ethylene oxide plant workers who were challenged with a single dermal applica-
tion of 1% after an average of 10.4 years of occupational exposure. Anaphy-
lactic reactions have been observed in patients using ethylene oxide steril-
ized plastic tubing for hemodialysis (Poothullil et al. , 1975) or cardiac
catheter izat ion (Pessayre and Trevoux, 1978). These symptoms included uti-
9-9
-------
caria, breathlessness, and hypotension. In a follow-up study on a patient
apparently sensitized by contact with hemodialysis tubing, Dolovich and Bell
(1978) illustrated that this patient showed a positive skin test response to
ethylene oxide-serum albumin conjugate, and produced in vitro histamine
release to this antigen. This response indicates that a specific IgE antibody
to ethylene oxide had been induced in this patient.
Clinical reports of hemolysis following use of ethylene oxide sterilized
plastic tubings have also been published (Hirose et al. , 1963; Clarke et al.,
1966). Ethylene oxide, rather than a chemical reaction product, is impli-
cated, since this type of effect can be prevented by extensive aeration of
ethylene oxide sterilized plastic devices.
Ethylene oxide vapors in high concentrations are irritating to the
eyes, but ocular contact with liquid ethylene oxide can cause severe burns. A
workman exposed to ethylene oxide in an unstated manner was reported to have
suffered a corneal burn, but healing was observed within 48 hours following a
cornea 1 denudement procedure (McLaughlin, 1946). Thiess (1963) described two
cases of accidental eye injury with ethylene oxide. A nurse was exposed to a
direct blast of ethylene oxide from a sterilizer cartridge, and developed an
epithelial keratitis of the cornea within 3 hours. Within 24 hours, the eye
was entirely normal. The second case involved a patient who received a squirt
of liquid ethylene oxide (concentration not stated) in the eye and was treated
immediately by extensive washing with water; this resulted in only irritation
of the conjunctivae that persisted for =1 day.
9.2.1.2. SUBCHRONIC AND CHRONIC EXPOSJRE — Limited information is
available on toxic effects of subchronic or chronic ethylene oxide exposure in
9-10
-------
humans. The information is largely derived from clinical case reports from
retrospective mortality studies.
Gross et al. (1979) reported on four cases of apparent ethylene oxide-
induced neurotoxicity. This occurred in a plant in which a sterilizer was
found to have leaked for 2 months of operation. The exact levels of ethylene
oxide were unknown, but the four individuals involved reported that they could
intermittently smell the ethylene oxide gas, indicating roughly a level of
>700 ppm. The length of exposure to ethylene oxide from the leaking
sterilizer was 3 weeks for cases 1 and 2, 2 weeks for case 3, and 2 months for
case 4. Three of the four cases had worked as sterilizer operators for >2
years and were exposed to ethylene oxide from the leaking sterilizer for 2, 3,
or 8 weeks; the fourth had been an operator for only 3 weeks and was exposed
for the duration.
The individual who had been exposed to ethylene oxide for 3 weeks had
noted conjunctival and mucosal irritation and transient blunting of the senses
of smell and taste, and developed headache, nausea, vomiting, and lethargy
that was followed by acute encephalopathy (recurrent major motor seizures at
20-30 minute intervals). Two of the other three operators were symptomatic
(i.e., headaches, numbness and weakness in the extremities, fatiguability, one
case of memory/think ing disturbances) and had abnormal neurological examina-
tion results that were consistent with sensorimotor neuorpathy. Nerve conduc-
tion studies were abnormal in these three operators, including the asympto-
matic patient, and were compatible with the diagnosis of sensorimotor neuro-
pathy. Removal from exposure resulted in relief of symptoms within 2 weeks.
Two of the individuals returned to work under normal conditions of lower ethy-
lene oxide exposure, but improvement in nerve conduction was not observed;
9-11
-------
significant improvement was noted, however, in the third individual who
returned to work in a position without ethylene oxide exposure.
Jensen (1977) reported that three workers using ethylene oxide steril-
izers were hospitalized for neuropathy of the lower linbs. Follow-up
indicated that these effects were reversible.
Jay et al. (1982) found that the four sterilizer operators described
above (Gross et al. , 1979), who were exposed to excessive levels of ethylene
oxide from a leaking sterilizer and developed neurologic abnormalities, subse-
quently developed cataracts. The operator exposed for 2 months developed
bilateral cataracts during the following 2 1/2 years; cataracts were diagnosed
in the other three operators upon examination =3 1/2 years after exposure to
the leaking sterilizer. Eight other men whose work involved exposure to
ethylene oxide sterilizers (6 of the 8 were sterilizer operators), but who
were asymptomatic, were subjected to complete ocular examinations, but
cataracts were not found. Four of the 12 men, two of whom had not worked on
the leaking sterilizer, had increased central corneal thickness with normal
endothelial cell counts when compared with a control group of 12 subjects of
higher average age (41 vs. 33 years). None of the patients were examined
before exposure to ethylene oxide, but the authors believed it unlikely that
cataracts would occur by chance in persons in this age range, particularly
because none of the patients had any systemic or ocular disease that might be
associated with cataract formation.
Hemoglobin values and lymphocyte counts were reported to be significantly
lower and higher, respectively, in a group of Swedish ethylene oxide produc-
tion workers when compared with control subjects (Ehrenberg and Hallstrom,
1967). The design and results of this study are more completely described in
9-12
-------
Section 9.5.2, but it should be noted that the production workers were
reported to have been exposed for 2-20 years (average 15 years) to an unknown
level of the compound.
Joyner (1964) conducted a retrospective morbidity study of 37 male ethy-
lene oxide production plant workers. These workers varied in age from 29-56
years and were exposed to typical concentrations of 5-10 ppm (range 0-55 ppm)
for 5-16 years (mean 10.7 years). Age-matched controls consisted of 41 opera-
tors (mean length of service, 11.7 years) assigned to other production units,
who had past exposure to many different petrochemical industry agents, but had
never exhibited clinical effects attributable to systemic chemical toxicity.
As detailed in Sections 9.3 and 9.5, no significant increase in health
problems relative to controls was found. This evaluation should have been
sufficient to identify major toxic effects of extended low-level ethylene
oxide exposure, although limitations in the design of the study, as well as an
insufficient period of observation, preclude evaluation of more subtle toxic
or carcinogenic responses.
An excess of deaths from specific causes (including all circulatory
causes) other than certain malignancies (Section 9.5) was not observed in a
group of 767 male ethylene oxide workers from the Texaco Chemical Company
Plant in Port Neches, Texas (Morgan et al. , 1981). These cohort members had
been employed for at least 5 years between January 1955 and December 31, 1977,
and an industrial hygiene survey of the plant performed in July, 1977, showed
that the 8-hour TWA exposure to ethylene oxide was well below 50 ppm.
Hogstedt et al. (1979a) conducted a cohort study of mortality among 89
full-time ethylene oxide production workers, 86 intermittently exposed main-
tenance workers, and a group of 66 unexposed control workers during the years
9-13
-------
1961-1977. As described in Section 9.5, exposure patterns were quite complex;
in addition to ethylene oxide (concentrations were generally <50 mg/nr),
workers were exposed at different times to ethylene dichloride, ethylene
chlorohydrin, ethylene, low concentrations of bis(2-chloroethyl)ether, as well
as traces of other chemicals. It was found that the full-time exposed cohort
showed considerable excess mortality when compared with the number expected
based on national statistics. The excess mortality arises mainly from
increased mortality due to stomach cancer and leukemia (Section 9.5), but also
from diseases of the circulatory system. When at least 1 year of exposure and
_>10 years of induct ion-latency time were required for inclusion in the study,
there were 12 observed deaths attributed to the circulatory system (9 due to
coronary heart disease and 3 due to cerebrovascular disease), compared to the
expected incidence of 6.3; this difference was statistically significant
(P<0.05). The excess mortality was of the same magnitude in a restricted
cohort of those with MO years of employment in ethylene oxide production and
20 years of induct ion-latency time (7 observed, 2.2 expected).
9.2.2. Effects in Animals.
9.2.2.1. ACUTE EXPOSURE — The acute toxicity of ethylene oxide is sum-
marized in Table 9-1 . Exposure of mice, rats, guinea pigs, rabbits, and dogs
to lethal levels of ethylene oxide has produced symptoms of mucous membrane
irritation and CNS depression, including lacrimation, nasal discharge, saliva-
tion, nausea, vomiting, diarrhea, respiratory irritation, in coordination, and
convulsions (Sexton and Henson, 19^9; Hollingsworth et al., 1956; Mine et al.,
1981). Animals that survived the initial exposures showed subsequent bron-
9-14
-------
chit is, pneumonia, and loss of appetite, with delayed symptoms of apathy,
dyspnea, vomiting, paralysis (particularly of the hindquarters), periodic con-
vulsions, and death (Waite et al. , 1930; Hollingsworth et al. , 1956). Prompt
deaths are usually due to lung edema; delayed deaths frequently result from
secondary infections in the lungs, although general systemic intoxication may
also be a factor (Hine et al. , 1981 ).
Pathological findings following lethal exposure to ethylene oxide in
mice, rats, and guinea pigs showed congestion of the lungs, hyperemia of the
liver and kidneys, and gray discoloration of the liver (Waite et al. , 1930).
Pathological findings after delayed death caused by ethylene oxide included
emphysema of the lungs, fatty degeneration of the liver, cloudy swelling of
the kidney tubules, and congestion of the spleen and brain (Hollingsworth et
al. , 1956). Intravenously-administered ethylene oxide caused congestion in
all organs of the rabbit (Greaves-Walker and Greeson, 1932). Zamlauski and
Cohen (1976) have reported that infusion of ethylene oxide in the rat at blood
levels of 0.45-4.5 mg/mS, produced a significant decrease (-30$) in glomerular
filtration rate, which indicates effects of ethylene oxide on kidney function.
Ethylene oxide in 10 and 50% aqueous solutions produced hyperemia and
edema in shaved rabbit skin when applied through cotton pads for 1-60 minutes
(Hollingsworth et al. , 1956). Bruch (1973) studied the dermal irritation pro-
perties of 2-10$ aqueous ethylene oxide solutions in guinea pigs and rabbits.
Subcutaneous injection in the guinea pig resulted in ecchymoses and skin
thickening, while intradermal injection and topical application in the rabbit
resulted in mild irritation. Topical or intraderraal administration of 1/J
ethylene oxide (0.5 m£), thrice weekly for 3 weeks, did not result in sensiti-
zation in guinea pigs (Woodward and Woodward, 1971).
9-15
-------
McDonald et al. (1977) studied the ocular effects of varied concentra-
tions of ethylene oxide in saline applied repeatedly over a 6-hour period to
the eyes of rabbits. They observed a dose-dependent increase in congestion,
swelling, discharge, iritis, and corneal cloudiness, indicating the irritating
effect of ethylene oxide on mucous membranes and corneal epithelium. The
maximum nondamaging concentration for this time period was 0.1/6 ethylene
oxide. In another study of ocular irritation in rabbit eyes, Vfoodward and
Woodward (1971) found slight irritation following a single application of 10$
aqueous ethylene oxide (duration of exposure unknown), and a no-effect concen-
tration of 2.1$ ethylene oxide was determined. The higher values determined
in this study are probably the results of a different mode of application and,
therefore, different duration of exposure.
9.2.2.2. SUBCHRONIC AND CHRONIC EXPOSURE -- The subchronic toxicity of
inhaled ethylene oxide has been investigated in a variety of different animal
species by different routes of exposure (Hollingsworth et al., 1956; Jacobson
et al. , 1956). As summarized in Table 9-2, symptoms of poisoning and patho-
logic changes are similar to those observed in acute studies, including lung,
kidney, and liver damage, and neuropathy of the hindquarter and testicular
tubule degeneration in some species.
Hollingsworth et al. (1956) observed neurotoxic effects in animals
following inhalation exposure to 357 ppm ethylene oxide vapor for several
weeks (the exposure for each species is presented in Table 9-2). Rats,
rabbits, and monkeys showed paralysis and atrophy of the muscles of the hind
limbs. These effects were reversible 100-132 days after discontinuation of
exposure. Special studies on monkeys were carried out with repeated (38-94)
9-16
-------
TABLE 9-2
Subchronic Toxioity of Ethylene Oxide
Route
Species
Concentration
Number of
Exposures
Effects
Reference
inhalation
inhalation
UD
--3
inhalation
inhalation
20 rats (10/sex)
16 guinea pigs (8/sex)
5 mice (female)
2 rabbits (1/sex)
1 monkey (female)
30 mice (female, white)
20 rats (male, white)
ppm
100 ppra
20 rats (10/sex)
10 mice (female)
357 ppm
16 guinea pigs (8/sex)
357 ppm
up to 8 in 10 days
(7 h/d; 5 d/wk)
30 (6 h/d; 5 d/wk)
33-38 (7 h/d; 5 d/wk)
123 in 176 days (7 h/d;
5 d/wk)
Death in all animals. Patholo- Hollingsworth
logic changes in lungs, liver and et al., 1956
kidneys similar to those in acute
poisoning.
Weight loss, reddish nasal dis- Jacobson et al.
charge, diarrhea, labored breath- 1956
ing, weakness of the hind legs,
and some deaths (13/20 exposed
and 0/20 control rats, and 2M/30
exposed and 3/30 control mice).
Fifteen additional rats or mice
were examined pathologically;
changes were limited to a few
cases of hemoslderosis in the
spleen that occurred late in the
exposure period.
Death in 10/10 mice (33 exposures) Hollingsworth
and 18/20 rats (38 exposures) et al., 1956
caused by secondary respiratory
infections. Impairment of sensory
and motor function in rats prior to
death, resulting in reversible
hind leg muscle paralysis and
atrophy.
Growth depression, degeneration Hollingsworth
of the testicular tubules with et al., 1956
replacement fibrosis (males),
slight fatty degeneration of the
adrenal cortex (females). No
nervous system effects or
mortality. •
-------
TABLE 9-2 (oont.)
Route Species Concentration
Inhalation 2 monkeys (I/sex) 357 ppm
2 monkeys (males) 357 ppm
Number of
Exposures
38-11 in 60 days
91 in 110 days
(both schedules 7 h/d;
5 d/wk)
Effects
Growth depression and characteris-
tic neurological impairment (e.g.,
hind limb paralysis and muscular
atrophy, poor or nonexistent knee
reflex, extensor reflex and
hindquarter/genitalla pain percep-
tion). No histopathologlc effects
of exposure.
Reference
Holllngsworth
et al.,1956
inhalation
3 dogs (male, Beagle)
290 ppm
MD
I
oo
inhalation
20 rats
20M ppm
inhalation
60 mice (30/sex)
250 ppm
inhalation
8 guinea pig
1 rabbits (2/sex)
2 monkeys (female)
20t ppm
30 (6 h/d; 5 d/wk)
127-133 in 185-193 days
(7 h/d; 5 d/wk)
50-55 (6 h/d; 5 d/wk)
127-157 in 176-226 days
(7 h/d; 5 d/wk)
Two of 3 exposed dogs showed Jacobson et al.,
toxic signs that included vomiting, 1956
slight tremors, transient weakness
of the hind legs and decreases In
red blood cells, hemoglobin, and
hematocrlt. Hematologlc parameters
normal in control dogs. Lungs
showed congestion and alveolar
collapse and fatty changes in the
hindquarters were consistent with
muscular atrophy.
Weight loss, some deaths with Holllngsworth
effects on lungs (congestion, et al., 1956
hemorrhage, emphysema, atelecta-
sis) kidneys and testes (slight
degeneration of some tubules)
(slight cloudy swelling of tubules)
Signs of neuromuscular toxicity, Snellings et al.
decreased red blood cell count, 1981a
packed cell volume and hemo-
globin concentration were ob-
in both sexes. No histopathologic
effects were observed.
No effect on growth or mortality. Hollingsworth
Evidence of paralysis/muscular et al., 1956
atrophy in the rabbits and monkeys.
Slight edema and congestion noted
In rabbits' lungs.
-------
TAELE 9-2 (cent. )
Route Species Concentration
inhalation 20 rats 113 ppm
8 guinea pigs
U rabbits (2/sex)
2 monkeys (females)
inhalation 60 mice (30/sex) 100 ppm
Inhalation 30 mice (females, l*iite) 100 ppm
20 rats (male, Wiite)
inhalation 2tO rats (120/sex) 100 ppm
inhalation 3 dogs (male, Beagle) 100 ppm
inhalation 60 mice (30/sex) 50 ppm
inhalation 20 rats H9 ppm
8 guinea pigs
Number of
Exposures Effects
122-157 in 176-226 days Growth depression and a moderate
(7 h/d; 5 d/wk) increase in lung weights in rats
were the only adverse treatment-
related effects noted.
50-55 (6 h/d, 5 d/wk) Hunched posture and reduced loco-
motion observed in both sexes.
No histopatho logic effects were
observed.
130 (6 h/d, 5 d/wk) No clinical signs of toxiclty or
treatment related mortality
(3/20 exposed and 3/20 control
rats, and 8/30 exposed and U/30
control mice died). No significant
pathologic changes in additional
groups of 60 rats or mice.
T45 (6 h/d, 5 d/wk) Early deaths and decreased body
weight gain were observed
starting at week 1.
130 (6 h/d, 5 d/wk) Normochronic anemia (decreased
red blood cell, hemoglobin, and
hematocrit) indicated in 1 and
suggested in 1 of 3 dogs. No
changes in the 3rd exposed dog,
or in control dogs.
50-55 (6 hr/d, 5 d/wk) Hunched posture in males and re-
duced locomotion in females. No
histopatho logic effects were ob-
served.
127-131 in 180-18H days No adverse effects as Judged by
(7 h/d, 5 d/wk) general appearance, behavior,
Reference
Hollingsworth
et al. , 1956
aiellings et al. ,
198ta
Jacobson et al. ,
1956
Snellings et al. ,
198l|b
Jacobson et al. ,
1956
Sne llings et al. ,
198Ha
Hollingsworth
et al. , 1956
M rabbits (2/sex)
10 mice ( female)
mortality, growth, final body
and organ weights, and gross
or microscopic pathologic
examination.
-------
TAELE 9-2 (cent. )
Number of
Route Species Concentration Exposures
inhalation 210 rats (120/sex) 33 ppm 115 (6 h/d, 5 d/wk)
inhalation 60 mice (30/sex) 10 ppm 50-55 (6 h/d, 5 dA
-------
exposures to this level of ethylene oxide. Knee jerk reflexes became very
weak, pain perception in the hind quarters decreased, the creraasteric reflex
was elicited, and the extensor reflex of the palms of the hind feet was
abolished. Impairment of both sensory and motor function at the lumbar and
sacral level of the spinal cord was indicated. Exposure of monkeys to a lower
level of ethylene oxide (204 ppm for 176-226 days) produced partial paralysis
and some muscular atrophy of the hind legs with moderate suppression of the
leg reflexes. The Babinski reflex was present after this lower level exposure
to ethylene oxide.
In a more recent subchronic study, Snellings et al. (1981*3) exposed
groups of 30 male and 30 female B6C3F1 mice to the vapors of ethylene oxide.
Exposures were for 6 hours/day, 5 days/week for 10-11 weeks to nominal levels
of 1, 10, 50, 100, and 250 ppm. No effects were observed on survival, body
weight or histologic sections of a variety of organs. At the three higher
exposure levels, however, signs of neuromuscular toxicity were observed. In
both sexes of the high exposure group, there was a statistically significant
increase in hunched posture, reduced locomotion, and righting reflex. The
former two were observed in the 100 ppm group and in males and females,
respectively, of the 50 ppm group. The abnormal righting reflex was observed
only during intermediate testing of females in the 100 ppm group. In addi-
tion, reduced toe pinch reflex was reported for females tested at intermediate
periods and reduced tail pinch reflex was reported for males at termination in
the 250 ppm group. Also at termination in the high dose group, hematologic
parameters, red blood cell count, packed cell volume and hemoglobin concentra-
tions were decreased, and some changes in either absolute or relative organ
9-21
-------
weights were observed. In this study, the neu ro muscular effects appeared to
be the most sensitive indicator of exposure to ethylene oxide.
Preliminary results of a chronic inhalation study conducted by NIOSH have
been reported (Lynch et al., 1982a). Male F344 rats (80 per treatment group)
and male Cynomolgus monkeys (12 per treatment group) were exposed to either 50
or 100 ppm ethylene oxide for 7 hours/day, 5 days/week for 24 months. Addi-
tional details of the experimental design are presented in Section 9.5, but it
should be noted that the rats were included primarily for carcinogenicity
evaluation, and the monkeys used to determine target organ toxicity. A number
of indices were evaluated including body weights, hematology, clinical
chemistry, urinalysis, opthalmology, pulmonary function, neurophysiology,
neuropathology, gross and histopathology, sister chromatid exchange rates, and
chromosomal aberrations in peripheral lymphocytes. The results that are
currently available are summarized below.
As detailed in Section 9.5, weight gain throughout most of the exposure
and survival periods were significantly depressed in the rats at both exposure
levels (Lynch et al. , 1982a). Weight gain was significantly depressed in the
treated monkeys beginning at week 25. The liver and spleen of the rats were
the only organs in which histopathological evaluations have been completed,
but the preliminary terminal sacrifice spleen data indicate a dose-related
induction of leukemia (Section 9.5). Hematologic analyses showed no statisti-
cally significant change in red blood cell count in the treated rats, but
white blood cell counts were highly variable and reflected the presence of the
leukemia. There were no differences in the red or white blood cell counts in
either of the monkey groups, although increased frequencies of chromosomal
9-22
-------
aberration and SCE were observed in the peripheral lymphocytes of these
animals.
In another 2-year inhalation toxicity study in rats conducted by
Snellings et al. (1984b), the only non-neoplastic effect reported was a
decrease in body weight gain. As described more fully in Section 9.5, groups
of 120 male and 120 female Fischer 344 rats were exposed 6 hours/day, 5
days/week to ethylene oxide at target levels of 0, 10, 33, and 100 ppm.
Decreased body weight gain was observed in both sexes after 4 weeks in the
high exposure group and in females of the 33 ppm group after 10 weeks. Other
groups were similar to control animals. Early deaths were reported for the
high exposure group and were likely to be tumor-related. The incidence and
type of neoplastic lesions are discussed in Section 9.5.
Significant hematological effects (i.e., anemia) have also been observed
in ethylene oxide-exposed dogs. Jacobson et al. (1956) found decreased red
blood cell counts, hemoglobin, and hematocrit in 2 of 3 beagle dogs that were
exposed to 292 ppm ethylene oxide vapor for 6 hours/day, 5 days/week for 6
weeks. Definite (1 dog) and suggestive (1 dog) hematologic effects of the
same type were also observed in 2 of 3 dogs that were similarly exposed to 100
ppm ethylene oxide for 6 months. Woodward and Woodward (1971) demonstrated a
dose-related increase in anemia in dogs that were administered 6-36 mg/kg
ethylene oxide in 30 daily subcutaneous injections. Pathologic examination
showed hyperplastic bone marrow and ectopic hematopoiesis. Balazs (1976),
however, was unable to repeat these findings in beagle dogs with an ethylene
oxide-glucose solution administered intravenously over the same concentration
range in a 21-day study.
9-23
-------
An oral feeding study using 1Q% ethylene oxide in olive oil was performed
on rats (Hollingsworth et al., 1956). Rats fed 100 rag/kg ethylene oxide in 15
doses over 21 days showed narked weight loss, gastric irritation, and slight
liver damage. Feeding of 30 mg/kg in 22 doses produced no observable adverse
effects.
9.2.3- Summary of Toxicity. The primary effects of acute inhalation exposure
to high concentrations of ethylene oxide gas are respiratory tract irritation
and CNS depression. Headache, vomiting, dyspnea, and diarrhea are common
systemic effects of vapor exposures in humans, and excessive exposures have
produced bronchitis, pulmonary edema, and convulsive movements. Similar
effects have been observed in a variety of animal species, but paralysis (par-
ticularly of the hindquarters) and periodic convulsions frequently preceded
death. Death in ethylene oxide-exposed laboratory animals is usually due to
lung edema or secondary lung infections, and postmortem pathologic findings in
other organs include widespread hypereraia and congestion (liver, kidneys,
spleen) and fatty degeneration (liver).
Derma to logical effects following skin contact with ethylene oxide in
humans from accidental or experimental exposure include edema, erythema, and
vesiculation with possible bleb formation, in that sequence. Vesicle
formation is usually delayed (e.g., 6-12 hours), the magnitude of skin injury
appears to be related to concentration and duration of contact, and the
effects are reversible. Concentrated ethylene oxide evaporates from the skin
resulting in a freezing effect, but more dilute solutions penetrate the skin
more effectively, resulting in chemical burning; weak solutions lack suffi-
cient chemical strength to cause injury except after prolonged contact. Skin
9-24
-------
burns have also been caused by residual ethylene oxide in clothing or footwear
treated or accidentally contaminated with the compound. Sens it izat ion has
also been associated with repeated dermal exposure to ethylene oxide at the
sites of contact. Similar dermal irritative effects have been observed in
experimentally exposed rabbits and guinea pigs, but sensitization was not
demonstrated by topical or intradermal administration in guinea pigs. High
concentrations of ethylene oxide vapors are irritating to the eyes of humans
and animals, and direct ocular contact with liquid ethylene oxide can produce
cornea 1 injury.
Case reports indicate that neurological effects (e.g., headache/vomiting,
sensorimotor neuropathy, seizures) and ocular effects (e.g., cataracts) may be
primary effects of limited repeated exposure to high levels of ethylene oxide,
and hematological effects (reduced hemoglobin and increased number of lympho-
cytes) have been noted in chronically exposed ethylene oxide production plant
workers. Retrospective morbidity and mortality studies of ethylene oxide pro-
duction workers do not, however, suggest chemical-related, non-neoplastic
toxicity. Subchronic exposure of different species of animals to ethylene
oxide by different routes of exposure produced effects similar to those seen
in acute studies; symptoms of poisoning primarily reflect neurotoxic action
(e.g, hindquarter neuropathy) and pathologic changes generally occur in the
lungs, kidney, and liver (e.g., congestion and degenerative changes), although
testicular effects (e.g., tubule degeneration) and hematologic effects (e.g.,
anemia) have also been observed.
9-25
-------
9.3. TERATOGENICITY AND REPRODUCTIVE TOXICITY
9.3.1. Teratogenic Effects. Batelle Pacific Northwest Laboratories
(Hackett et al. , 1982) conducted teratology and reproductive studies for the
National Institute for Occupational Safety and Health investigating the
effects of ethylene oxide produced by inhalation exposure. Pregnant rabbits
and rats were exposed to a single dose of ethylene oxide (150 ppm, Union
Carbide, Linda Lot No. 01901, 99.7? pure) both prior to and during the period
of organogenesis. Thirty New Zealand White rabbits per group were exposed in
three different regimes (filtered air alone, ethylene oxide exposure on days
7-19 of gestation, and ethylene oxide exposure on days 1-19 of gestation).
Forty-one Sprague-Dawley CD rats per group were exposed according to four
different schedules (filtered air alone, ethylene oxide exposure on days 7-16
of gestation, ethylene oxide exposure on days 1-16 of gestation, and ethylene
oxide exposure 3 weeks prior to mating and through days 1-16 of gestation).
In the rabbits, no toxic effects were observed in the mothers (i.e.,
changes in body weight, organ weight, histopathological changes in the
organs). In addition, there were no decreases in the percentage of pregnant
animals nor was there any indication of adverse effect on the fetus (i.e.,
decreases in fetal body weight, crown rump length, sex ratios, or morphologic
alterations).
In the rats, maternal toxicity was observed with sporadic decreases in
food consumption, decreases in body weight, and increases in kidney and spleen
weights, with the increases in spleen weights roughly proportional to the
duration of exposure. Adverse effects were also observed in the developing
conceptus. There was an increase in resorptions in animals exposed both pre-
9-26
-------
and postgestationally with a trend for early midgestational resorptions. In
addition, lowered fetal body weight, decreased crown-rump length and an
increased incidence of incomplete skeletal ossification were observed in all
ethylene oxide exposed offspring, and this was especially pronounced in
animals exposed both pre- and postgestationally. It was concluded from this
study that exposures of 150 ppm in rats caused significant adverse effects in
both the nether and developing fetus; however, since only one dose was used in
this study, it is not known whether these developmental effects would occur in
the absence of maternal toxicity.
Because of concerns over adverse reproductive effects which could occur
as a result of exposure to ethylene oxide or ethylene oxide reaction products
left on improperly degassed surgical supplies, LaBorde and Kimmel (1980) con-
ducted studies on the effects of ethylene oxide administered intravenously.
CD-1 mice in four replicates of three treatment groups (10 animals per group)
were treated with 0, 75, 150 mg/kg ethylene oxide (Eastman Organic Chemicals
Co., purity not stated, ethylene oxide was injected in 5% dextrose solution).
The animals were exposed in the following treatment periods of gestation: days
4-6 (period I), days 6-8 (period II), days 8-10 (period III) and days 10-12
(period IV).
Clinical signs of maternal toxicity (weakness, labored breathing,
tremors, and death) were observed in animals injected with 150 mg/kg ethylene
oxide on gestational days 4-6 (period I), 8-10 (period III) and 10-12 (period
IV) but not in the group injected on days 6-8 (period II). Decreases in mean
maternal body weight gain were observed in animals in period I, period III,
and period IV and were accompanied by decreases in the mean number of live
fetuses in periods III and IV. Embryotoxicity as manifested by significant
9-27
-------
reductions in mean fetal weight was observed in all four periods at the 150
mg/kg dose. There was no significant change in the mean number of implants
per litter, but there was a reduction in the mean number of live fetuses per
litter (and an increase in the number of dead and resorbed offspring) in
periods III and IV at the 150 mg/kg level. An increase in the percent of
malformed fetuses/litter was noted in periods II, III, and IV at 150 mg/kg
level, but in period III the incidence did not achieve statistical signifi-
cance. It was concluded that the ethylene oxide exposure under these condi-
tions was selectively affecting the development of the conceptus (as seen by
skeletal malformations and embryotoxicity) since ethylene oxide exposure in
period II (days 6-8 of gestation) produced malformations and embryonic death
while not affecting the mother (no clinical signs of toxicity) ; however, this
conclusion was tempered somewhat by the observation of maternal deaths in
treatment groups before and after this time period. Although there was no
dose-response relationship in the severity of adverse effects in either the
mother or fetus, the types of malformations seen in periods II and III
appeared to follow a developmental pattern. The authors reported that in
animals treated on days 6-8 cervical and upper thoracic vertebrae malformation
were observed. Animals treated on days 8-10 had defects primarily in the
lower thoracic region.
Another study by the same investigators (Jones-Price et al. , 1983) evalu-
ated the reproductive effects of intravenous injections of ethylene oxide in
rabbits. New Zealand White rabbits were intravenously injected in two treat-
ment regimes; 0, 9, 18, or 36 mg/kg ethylene oxide (source and purity not
reported) on days 6-14 of gestation, or 0, 18, or 36 mg/kg on days 6-9 of
9-28
-------
gestation. Seventeen to 21 animals were examined in the group exposed on days
6-9, 18-24 animals examined in the group exposed on days 6-14.
Maternal toxioity was observed in both exposure groups, with more severe
effects observed in the groups treated on days 6-14 than on days 6-9 of gesta-
tion. Significant decreases in maternal weight gains were observed during the
entire treatment at the 18 and 36 mg/kg level. These decreases included both
decreases in pregnancy weight gains and decreases in absolute weight gains
(weight gained during pregnancy minus uterine weight). No embryotoxic effects
were observed in the day 6-9 treatment groups; however, in the 6-14 day treat-
ment group significant dose-re la ted trends for decreased numbers of live
fetuses/litter and resorptions/litter were observed. At the 36 mg/kg level,
the incidence of resorptions/litter was statistically significantly different
from control levels. The authors concluded that intravenous administration of
ethylene oxide in rabbits produced embryotoxicity, though at doses which also
produced significant maternal toxicity.
LaBorde et al. (1982) presented data at the 1982 Society of Toxicology
meeting regarding the teratogenic effects of ethylene chlorhydrin (ECH), a
reaction product of ethylene oxide, in mice and rabbits. ECH is produced by
the interaction of ethylene oxide and chloride ions, so it is a residue of
ethylene oxide that could be left on medical devices after improper degassing
of ethylene oxide during sterilization. Forty-one to 65 CD-1 mice were
injected intravenously with 60 mg/kg or 120 mg/kg ECH (source not reported; in
5% sterile dextrose) on days 4-6, 6-8, 8-10, or 10-12 of gestation. Seventeen
to 22 New Zealand White rabbits were intravenously injected with 9, 18, or 36
mg/kg ECH on days 6-14 of gestation.
9-29
-------
In this study, no adverse effects were observed in either the mother or
the fetus of the New Zealand White rabbits; however, in CD-1 mice, clinical
signs of toxicity (weight loss of >1 g in 24 hours) were observed in the
mothers in all treatment periods at the 120 rag/kg dose. Maternal weight gain
during the entire treatment period and during pregnancy were significantly
reduced at the 120 mg/kg level in day 4-6, 6-8, and 10-12 groups. There was
also a trend for increased resorptions/litter in animals exposed on days 4-6
and 10-12 at the 120 mg/kg level. At the 120 mg/kg dose for all treatment
periods, there was a significant decrease in mean fetal weight/litter. At the
60 mg/kg level in animals exposed on days 8-10, there was a significant
reduction in fetal weight in the absence of maternal toxicity. The authors
reported a trend for an increase in the number of malformed fetuses treated on
days 8-10; however, the incidence of this effect did not achieve statistical
significance.
The conclusion reached by LaBorde et al. (1982) was that ECH administered
intravenously in mice produced embryo/fetal toxicity and possibly a slight
increase in malformations at maternally toxic doses. At the 60 mg/kg level,
in animals treated on days 8-10, fetal weight reductions occurred in the
absence of maternal toxicity; therefore, it was concluded that ECH may pose a
hazard specific to the developing conceptus.
Verret (1974) investigated the toxic and teratogenic effects of ethylene
chlorohydrin (ECH) in the developing chick embryo. ECH (source and purity not
reported) was administered via the air cell during a p re-incubation period (0
hour) and after 96 hours of incubation at levels equivalent to 10, 25, 50,
100, and 200 mg/kg. The control groups were treated with the vehicle (water)
or left untreated. One hundred eggs were used per group. Ethylene chloro-
9-30
-------
hydrin was found to be toxic in this system with significantly increased mor-
tality (no hatch) at levels >25 mg/kg at the 0 hour exposure, and at levels
>12.5 mg/kg at the 96 hour exposure. Statistically significant increases in
the number of structural anomalies were observed at two dose levels (50 and
100 mg/kg) at the 0 hour exposure, and at four dose levels (12.5, 25, 50, and
100 mg/kg) at the 96 hour exposure. The significance of these observations in
terms of mammalian effects is not known, however, since teratogenic effects in
chick embryos may not be predictive of mammalian effects.
9.3.2. Reproductive Effects. The Carnegie-Me lion Research Institute
(Snellings et al. , 1982a) conducted a one-gene rat ion study evaluating the
reproductive effects of inhalation exposure to ethylene oxide. Thirty male
and female Fischer-344 rats were continually exposed to 10, 33, and 100 ppra
ethylene oxide with the control animals exposed to filtered air. Prior to
cohabitation, all groups were initially exposed to ethylene oxide for 6
hours/day, 5 days/week for 12 weeks. After 1 week of cohabitation, females
with vaginal plugs were removed, and the other females were rotated to a
different male to allow for mating for another week. At the end of 2 weeks
all male and female animals were separated. The males were then exposed to
ethylene oxide for 6 hours/day, 7 days/week for an additional 3 weeks. The
females were exposed for 6 hours/day, 7 days/week from day 1 through day 19 of
gestation. On day 20 of exposure, females not pregnant were sacrificed. The
pregnant females were allowed to deliver and 5 days after parturition were
again exposed to ethylene oxide for 6 hours/day, 7 days/week until day 21
postpar turn.
9-31
-------
The following criteria were used to establish fertility. If a female
produced a litter, or if gross examination revealed implantation sites after
staining, she was considered fertile. Any female not pregnant after two
different matings was considered infertile. If the male impregnated a female
after the first mating, then he was considered fertile; any male failing to
impregnate a female in two different mating periods was considered infertile.
By this criteria, females exposed to 100 ppm had a higher incidence of infer-
tility after mating with a male of proven fertility; however, this incidence
did not achieve statistical significance. In the males there was no decrease
in fertility. In the 100 ppm group, significantly more females had lengthened
gestation periods (time from vaginal plug to birth of litter) than the con-
trol, 10, or 33 ppm groups. The control, 10, and 33 ppm groups had gesta-
tional periods of 22 days, while the 100 ppra group had gestations ranging from
22-31 days (7/14 rats had 22 day gestation, V1U rats had 23 day gestation,
3/1^ rats >25 day gestation). Since most of the animals did not have exten-
sively long gestational delays, it is not clear whether this lengthening of
gestation represented a true adverse biological effect.
In this study (Snellings et al. , 1982a), the number of pups was signifi-
cantly reduced with a decrease in the number of implantation sites at the 100
ppm level. Of the surviving pups, however, there was no effect on survival
after parturition. In the parental generation, there were no adverse effects
on body weight or organ histology (testes, epididymides, accessory sex glands,
cervix, uterus, ovaries, oviducts, mammary tissues). In the F..a generation,
-25% of the animals suffered from sialoacryoadenitis virus infection but this
infection appeared to be unrelated to the ethylene oxide exposure.
9-32
-------
It was concluded from this study (Snellings et al., 1982a) that ethylene
oxide administered to rats by inhalation has the potential to disrupt repro-
duction by causing an increased incidence of embryolethal effects; however,
this embryotoxic effect was only observed when the animals were exposed to the
highest dose (100 ppm) and not at the lower doses (10 or 33 ppm) of ethylene
oxide.
Snellings et al. (1982b) have also examined teratological aspects by
exposing rats by inhalation to 10, 33, or 100 ppm ethylene oxide for 6
hours/day on days 6-15 of gestation. Exposure to 100 ppm caused depression of
fetal weight gain, but did not result in fetal death or abnormalities other
than variations in ossification of sternebrae and distal thoracic vertebral
centra.
9.3-3. Testicular Effects. Hollingsworth et al. (1956) investigated the
acute and chronic toxicity of ethylene oxide in a variety of animal species.
Positive responses related specifically to the male reproductive system were
observed in hamsters and rats. Eight guinea pigs were exposed to 357 ppm
ethylene oxide (commercial grade ethylene oxide, 97-98.6? pure by weight) and
received 123 seven hour exposures over a 176 day period. There was only
moderate growth reduction in the males; however, appreciable degeneration of
testicular histology was noted. In another phase of this experiment, both
rats and hamsters were exposed to 204 ppm ethylene oxide, 7 hours/day, in
122-157 exposures given over an experimental period of 176-226 days. A slight
but not statistically significant decrease in testis weight of rats and guinea
pigs was observed. In rats, there was histological evidence of degeneration
of the testicular tubules.
9-33
-------
A recent study sponsored by NIOSH described the effects of inhaled
ethylene oxide on semen production in Cynomologus (Macaca fasicularis) monkeys
(Lynch et al. , 1983). The monkeys were exposed to 50 and 100 ppm ethylene
oxide, (toion Carbide, 99.7? pure) 7 hours/day, 5 days/week for 2 years. In
the preliminary range-finding study, only two animals per group was used.
Testicular weights were diminished in animals exposed to 100 ppm ethylene
oxide but were only marginally decreased in those exposed at the 50 ppm level.
Similar decreases in epididymal weights were reported. Sperm motility was
significantly reduced at the 50 and 100 ppm level, both in terms of the
percentage of motile sperm and the ability of the sperm to travel a given
distance in a given time (drive range). In the preliminary study, the sperm
concentration was decreased at the 50 and 100 ppm levels. In a subsequent
study with larger numbers of monkeys per group (8 or 9), the same types of
adverse testicular effects were observed. In this study, there was a 30%
decrease in sperm concentration, 30% reduction in motile sperm, and a 3- to 4-
fold decrease in distance traveled when the animals were exposed to 50 and 100
ppm ethylene oxide; however, there was no effect on sperm head morphology
(Lynch et al., 1983).
In another study relevant to the effects of ethylene oxide on the testis,
radiolabeled ethylene oxide was detected in auto radiograms of mouse gonads
(epididymis and testis) 20 minutes after intravenous injection (Appelgren et
al. , 1977). Radioactivity was found in the epididymis up to 24 hours after
injection. The results of the dominant lethal mutagenicity test were
negative, although inadequacies in this study prevent a firm conclusion from
being made (see Section 9.4). This study is relevant to testicular effects
because it establishes that ethylene oxide has access to the gonads.
9-34
-------
9.3.1*. Adverse Reproductive Outcome in Humans. There is little information
on the effects of ethylene oxide on the human reproductive system. In one
study a comparison was made between the health of 37 male employees involved
in ethylene oxide production with 41 men who worked in other production units
(Joyner, 1964). This study evaluated many health endpoints including genito-
urinary problems. The mean exposure period was 10.7 years with a general
level of exposure on the order of 5-10 ppm. The range of exposure levels
varied from 0-55 ppm. The health survey of the workers considered the
following information: 1) the number of sick days taken in a 10 year period
with information on the etiology and duration of the illness, 2) any medical
diagnosis entered into the medical records and confirmed by an outside
physician, 3) any visits to the Medical Division related to respiratory,
gastrointestinal, or genitourinary problems. In this study there was a higher
incidence of chest abnormalities and a higher incidence of absenteeism
attributable to gastrointestinal and genitourinary causes; however, the higher
incidence of absenteeism was attributable to a single individual in each
category. Therefore, it was concluded that long term exposure to ethylene
oxide had no adverse health effect on the men involved in ethylene oxide
production. Since this study did not deal specifically with reproductive
health problems, it is of limited value in determining the potential of
ethylene oxide to cause adverse reproductive effects.
A study by a Russian investigator (Yakubova et al. , 1976) reported that
female workers involved in ethylene oxide production experienced a number of
gynecological and obstetrical problems. Tnese problems included diseases of
the cervix, Inflammation of the uterus, obstetric anamnesis (this word, as
well as others may have been incorrectly translated), hypertonic disease,
9-35
-------
anemia, toxicosis, and shortened pregnancies. These observations were
reported in an anecdotal manner with no presentation of actual data or
description of methodologies. This study is therefore of little value in the
scientific review of adverse reproductive effects.
Holmberg (1979) and Hotaberg and Nurminen (1980) reported case studies of
a mother exposed to a variety of organic solvents. These studies describe an
adverse reproductive outcome in a woman exposed to alkylphenol and dyes as
well as ethylene oxide. It is not clear whether the two articles describe the
same woman or two different women. Both reports describe an infant born with
hydrocephalus and Holmberg (1979) described a child with additional malforma-
tions (cleft palate, double uterus, polydactyly). These reports are not
useful in establishing causal relationship between ethylene oxide exposure and
congenital malformations because ethylene oxide was not the only chemical
involved and because a larger population size would have to be evaluated
before such an association can be established.
An epidemiology study has been conducted concerning the effects of
ethylene oxide exposure on pregnancy outcomes in nursing personnel. This
report is the only one which adequately evaluates the possible causal associa-
tion between ethylene oxide exposure and adverse human reproductive effects,
and has been reviewed in depth by an Environnental Protection Agency epidemi-
ologist (Margosches, 1983). In cooperation with the Finnish investigators,
the data have been critically analyzed and reviewed. The following is the
text of this evaluation:
"In November, 1982, K. Hemminki et al. published a study of
"Spontaneous abortions in hospital staff engaged in sterilizing
instruments with chemical agents" in the British Medical Journal.
This study, encompassing all Finnish sterilizing staff at that time,
9-36
-------
claimed adjusted spontaneous abortion (s. a.) rates of 16.7% for
"exposed" and 5.6/5 for nonexposed pregnancies among these staff.
The report singled out ethylene oxide, glutaraldehyde, and formalde-
hyde use and suggested concentrations as low as 0.1-0.5 ppm ethylene
oxide might have been associated with adverse outcomes. In
particular, among hospital-discharge-corroborated pregnancies, the
ethylene-oxide-exposed s.a. rate (22.6) significantly exceeded the
control s.a. rate (9.2). Also rates among all pregnancies exposed
to ethylene oxide or to glutaraldehyde differed significantly from
rates among pregnancies not exposed.
This study encompassed staff employed in 1979 at hospitals
throughout Finland (including tuberculosis sanitoria and mental
hospitals). It was a cohort study looking at past events;
determination of exposure status was based on the responses to two
questionnaires. The unit for most statistical tabulations and
analyses was the pregnancy; while not uncommon in the literature,
such a basis cannot take into consideration the relatedness of
sibling births or repeated miscarriages of a single woman.
The cohorts, sterilizing staff and controls [hereafter also
called group (1) and (2)], were identified by the head nurses at the
study hospitals. The former were named in response to a first
questionnaire that also queried chemical sterilizing agent use
history at each hospital. The latter were obtained as members of
cluster samples from auxiliary nurses in departments (not including
chemical sterilization, x-ray, or surgery) at the time of distribu-
tion of a second questionnaire that focussed on pregnancy and
employment history. The investigators obtained a very high return
rate (92% among sterilizing staff, 91& among auxiliary nurse
controls) and studied the 645 (63$) ever pregnant women among steri-
lizing staff and the 574 (55%} ever pregnant women among controls.
The 17 male sterilizing staff were not studied.
While the study population was selected on the basis of
hospital employment as sterilizing staff (1) or non-sterilizing-
staff auxiliary nurses (2), Dr. Hemminki classified pregnancies of
each group (1) member according to likelihood of exposure and the
agent(s) present in order to make finer comparisons. He considered
all pregnancies occurring after the first use of ethylene oxide at a
hospital to be exposed to ethylene oxide unless an individual did
not work at the hospital during a particular pregnancy; similarly
for glutaraldehyde and formaldehyde. This was a fairly conservative
classification. This study design precluded the examination of the
questions whether spontaneous abortions were related to an
individual's ever having been exposed to an agent.
Another limiting factor of the study design was the characteri-
zation of individual exposures in purely qualitative terms. Dr.
Hemminki believes that typical exposures have averaged <1 ppm
(measured by gas-tight syringes). He bases this belief on papers
9-37
-------
published by colleagues at the Institute of Occupational Health
covering a 3-year period overlapping the close of the study period,
on the unchanged instrumentation of ethylene oxide use over its 20
or so years in Finland, and on the measurement method's 1 ppm
detection limit. He did not, however, make any unwarranted
inferences regarding possible dose-response relationships.
Nevertheless, he did find sizeable differences in adjusted
spontaneous abortion rates in both nurses and sterilizing
professionals (these are 2 education levels of sterilizing staff)
between ethylene-oxide-ex posed and non exposed, age-adjusting (£30)
the rates among discharge-registry-identified pregnancies, the
ethylene oxide-exposed s. a. rate (16.1) also exceeded the rate (9.^)
in control pregnancies but no longer significantly. (Certain of the
pregnancies occurring during 1973 to 1979 could be cross-identified
through a national hospital discharge register and parallel analyses
were carried out on this set and the total questionnaire-obtained
set. )
On the whole, this study and its report paid close attention to
the possibilities and consequences of such typical epidemic logic
afflictions as reporting and recall bias. Additionally, the
methodology for statistical analysis, based on rates adjusted for
such concomitant variables as age, parity, and decade of pregnancy
by logistic regression, is sound (although there may be good reasons
to investigate a finer categorization of age). While a "per woman"
analysis, the analytic methods for incorporating an individual's
pregnancy history have not yet been perfected or standarized.
Unfortunately, the investigators introduced a possible source
of bias through telling the supervisory nurses (who identified group
(1) and selected group (2)) the purpose of the study, including the
names of the agents of interest. Another shortcoming is the
impreciseness with which hospital exposure history was determined.
Finally, although the authors planned a priori to investigate
relationships between ethylene oxide and spontaneous abortions and,
possibly, other adverse pregnancy outcomes, the underlying
relatedness of multiple pregnancies and of certain of the analyses
(e.g., regroupings of the same pregnancies to look at different
exposures) dilute the strength of any associations perceived in this
study. Notwithstanding these limitations, this work is sufficiently
suggestive to support further study of the possible associations
between ethylene oxide exposure and adverse pregnancy outcomes or
other reproduction effects." (Margosches, 1983).
9.3.5. Summary of Teratogenicity and Reproductive Toxicity. The potential
of ethylene oxide to cause teratogenic or adverse reproductive effects has
been examined in four animal species (mouse, rat, rabbit, monkey) by two
routes of administration (inhalation and intravenous) (Table 9-3).
9-38
-------
TAK.E 9-3
Sumnary of Studies on Tera togenic ity and Reproductive Toxicity
Type of Route of
Study Administration
Te ra to logy iv
Teratology iv
Teratology/ Inhalation
Reproduction
Teratology/ Inhalation
Reproduction
Dose Level and
Species Time of Exposure
CD-1 mouse 0, 75, 150 rag/kg
day t-f>, 6-8, 8-10
or 10-12 of gestation
New Zealand 0, 18, 36 mg/kg
White rabbit day 6-9 of gestation;
0, 9, 18, 36 mg/kg
day 6-1t of gestation
Sprague- 150 ppm, 7 hr/day:
Dawley CD day 7-16 gestation,
rat day 1-16 gestation,
3 weeks pregestatlon
plus day 1 -1 6
gestations
New Zealand 150 ppm, 7 nr/day:
Kiite rabbits day 7-19 gestation,
day 1-19 gestation
Findings
1. Developmental toxic ity
at or near dose level
which produced maternal
toxlcity (150 mg/kg)
1. Developmental toxicity
only at levels which
were maternally toxic
(36 UK/kg, day 6-1 "4)
Te ra to logy :
1 . Retarded fetal
development
Reproduction :
1 . ffetemal toxicity
2. Increase In trau terine
mortality
1. No tera togenic or
reproductive effects
References
LaBorde and
Klmmel, 1980
Jones-Price
et al. , 1983
Battelle Pacific
Northwest Labora-
tories (NI031
210-80-0013)
Hackett et al. ,
1982
Battelle Pacific
Northwest Labora-
tories (NIOSH
contract No.
210-80-0013)
Hackett et al. ,
1982
Comments
Te ra to logy :
1. Inadequacies
a) no maternally
toxic dose
b) dose-response
not determined
Reproduction :
1 . Inadequacies
a) dose -response
not determined
Te ra to logy :
1 . Inadequacies
a) no maternally
toxic doses
b) no develop-
mental toxic
doses
c) dose-response
not determined
Reproduction
1. Inadequacies
a) do se -response
not determined
-------
TABLE 9-3 (oont. )
jr
O
Type of Route of
Study Administration Species
•One genera- Inhalation Fischer 31t
tion reproduc- rats
tion
Teratology Inhalation Fischer 3m
rats
Chronic Inhalation Guinea pigs
tox ic ity
(nale
reproduction)
Dose Level and
Time of Exposure
0, 10, 33, 100 ppm
12 wks prior to Dating,
6 hr/day, 5 day/wk.
During gestation -
days 0 through day 19.
During lactation -
days 5 through 21.
0, 10, 33, 100 ppm
6 hr/day, days 6-15
of gestation
357 ppm, 123 7-hr
exposures In 176
days
Findings
1. No difference In F
fertility. No FQ
toxic ity.
2. No adverse effects on F
survival, growth rate,
or lactation.
3. Adverse reproductive
effects at highest dose,
100 ppm.
a) increased gestatlonal
length
b) decreased litter size
c) decreased implantation
sites (i.e., decreased
fecundity)
d) decreased fetuses/
implantation sites
(embryo lethal)
1 . Depression of fetal body
weight (slight)
1. Tubular degeneration of
test with replacement
fibrosls
References Comments
Carnegie -Me lion
Research
Institute, 1979
(Snelllngs et al. ,
1982a)*
Snelllngs et al. , 1. htatemal body weight
198lb not monitored during
treatment
Hollingsworth
et al. , 1956'-
20M ppm, 122-157
7-hr exposures in
176-226 days
1. Slight decrease in testes
weight, not statistically
significant.
Testicular Inhalation
tox ic ity
Rats 201 ppm, 122-157
7-hr exposures In
176-226 days
Cynomologus 50, 100 ppm,
monkeys 7 hrs/day for 2
years
1.
2.
1 .
2.
3.
1.
Slight decrease in testes
weight, not statistically
significant.
Testes: small, slight
degeneration of tubules.
Decreased testicular Lynch et al. , 1983
weight.
Decreased sperm
concentration.
Decreased sperm raotllity.
No change In sperm
morphology.
•A variety of experimental protocols were utilized; only those which provided positive information on reproduction effects are noted here
-------
TAELE 9-3 (cent. )
I
-Cr
Type of
Study
Medical
survey of
workers
Medical
survey of
workers
Case study
Epidemiology
study
Route of Dose Level and
Administration Species Time of Exposure
Occupational 37 male Mean exposure time:
exposure workers 10.7 years.
General levels:
5-10 ppm
Occupational 282 female <0.2-0.3 mg/m3
exposure production
workers
259 female
management
coworkers
100 females
controls
Occupational Pregnant
exposure women
Occupational Pregnant <1 ppm
exposure women
Findings
1. No observed Increase
In male reproductive
disorders.
1. Gynecological dis-
orders, spontaneous
abortions, toxicosis,
decrease birth weights.
1. Infant with hydro-
cephalus
1. Ethylene oxide exposure
associated with an
Increase in spontan-
eous abortion
References Comments
Joyner, 1961 1. Small sample
size
2. Study did not
evaluate
ffertl lity or
test leu la r
function.
Yakubova et al. , 1. Difficulties in
1976 translated material
2. Little information
provided en experi-
mental design.
3 . Mu It ip le ex posu re s
to noise and high
tempera turea
Holmberg, 1979; 1. Mother exposed to
Halm berg and multiple chemicals
Nurmlnen, 1980 2. Only one Infant
studied
Hemmlnkl et al. , 1. Possible bias
1982 introduced by
supervisors
who categorized
participants in
this study
2. Limited exposure
data
Teratology
In travenous
CD-I mouse
0, 60, 120 mg/kg on
days ")-6, 6-8, 8-10,
or 10-12 of gestation
Maternal toxic Ity at
120 mg/kg fbr all
treatment periods
Embryo toxic Ity at 120
mg/kg for all treatment
periods and at 60 mg/kg
on days 8-10 (fetal
weight reduction).
LaBorde and Kimroel,
1980
-------
TAELE 9-3 (cent. )
I
-t
Type of
Study
Teratology
Teratology
and Toxic Ity
Route of Dose Level and
Administration Species Time of Exposure
Intravenous New Zealand 0, 9, 18, 36 rag/kg on
white rabbits days 6-14 of gestation
Air cell thick enbryo 0, 10, 25, 50, 100 or
Injection 200 ng/kg at 0 hour
incubation ; 0
5, 12.5, 25, 50 or
100 mg/kg at 96
hours incubation
Findings
1 . Ho effect on anther or
fetus
1. Ovo-toxic at levels
^25 rag/kg at 0 hour,
and >12.5 ing /kg at
96 hours.
2. Teratogenic to chick
embryo
References Comments
LaBorde et al. , 1. Inadequacies
1982 a) no maternally
toxic doses
b) no develop-
mentally toxic
doses
Verrett, 197« Uncertainties in
extrapolating avian
developmental effects
to those of manuals
•Original study performed by Carnegie-Me lion Research Institute (Bayes, 1979); later published aa Snelllngs et al. , 1982a
-------
In a teratology study, Hackett et al. (1982) reported that rats exposed
to a single 150 ppm dose of ethylene oxide displayed both maternal toxicity
(decreases in food consumption, decreases in body weight, increases in kidney
and spleen weights) and toxicity to the developing conceptus (increases in
resorptions, decreases in fetal weight, decreases in crown-rump length, and
increases in incomplete skeletal ossification). Similar effects were not
produced in rabbits exposed to 150 ppm ethylene oxide in this study.
LaBorde and Kimmel (1980) administered 75 and 150 mg/kg ethylene oxide to
pregnant CD-1 mice for several gestational intervals. The animals displayed
signs of maternal and fetal toxicity at the highest dose level. There were
maternal deaths with decreases in the number of implants per litter and an
increase in the percentage of malformed fetuses/litter. The malformations
appeared to follow a developmental pattern and in at least one gestational
interval (days 8-10 of gestation) occurred in the absence of significant
observed maternal toxicity.
Similar studies were conducted by Jones-Price et al. (1983) on the
effects of 18 and 36 mg/kg ethylene oxide administered intravenously to New
Zealand rabbits. Significant maternal toxicity (decreased weight gain) was
observed in addition to embryotoxicity observed in the offspring (decreases in
the number of live fetus/litter, increases in the number of
resorptions/litter). No embryotoxicity was observed in the absence of
maternal toxicity.
LaBorde et al. (1982) investigated the teratogenic effect of
intravenously administered ethylene chlorohydrin (ECH), a reaction product of
ethylene oxide, in CD-1 mice and New Zealand rabbits. No adverse maternal or
embryotoxic effects were produced in the rabbits. In the mice at the highest
-------
dose (120 mg/kg), however, severe maternal weight loss with increases in
resorptions/litter and decreases in fetal weight were observed. At the 60
mg/kg level, with exposure on gestational days 8-10, there was significant
fetal weight loss in the absence of maternal toxicity; therefore, the authors
concluded that ECH may be a specific hazard to the developing conceptus at
this dose level. ECH has also been reported to produce adverse effects in
developing chick embryos (Verrett, 1974). Structural abnormalities were
produced by 12.5-100 mg/kg of ECH when the egg was incubated with the chemical
for up to 96 hours.
In a one-gene ration study (Snellings et al. , 1982a), female rats exposed
by inhalation to 100 ppm ethylene oxide had a higher incidence of infertility
with indications of a longer gestational period. There was a decrease in the
number of pups produced by mothers exposed to 100 ppm ethylene oxide, as well
as a decrease in the number of implantation sites. However, there were no
significant signs of toxicity in the mothers (no decreases in body weight or
changes in organ histology). The sane group (Snellings et al. , 1982b)
observed lowered fetal weights, but not a substantial level of malformations
in response to 100 ppm ethylene oxide administered to rats by inhalation on
gestation days 6-15.
Adverse effects on the testis resulting from ethylene oxide exposure have
been reported for the hamster and rat (Hollingsworth et al. , 1956) and
Cynonclogus monkey (Lynch et al. , 1983). Hollingsworth reported testicular
degeneration occurring in hamsters and rats exposed to ethylene oxide by
inhalation (204-357 ppm). Lynch et al. (1983) reported adverse effects on
sperm concentration and notility, but not morphology, in Cynomologus monkeys.
The monkeys in this study were exposed over 2 years to 50 and 100 ppm ethylene
9-44
-------
oxide by inhalation. In mice, radiolabeled ethylene oxide has been found to
persist in the epididymis up to 24 hours after a single injection (Appelgren
et al., 1977).
Very little information exists on the adverse reproductive effects of
ethylene oxide in the human. Medical surveys have described effects ranging
from no adverse reproductive outcome (Joyner, 1964) to a variety of adverse
outcomes (Yakubova et al., 1976). The study by Joyner (1964) is inadequate
because it did not deal specifically with adverse reproductive outcomes. The
report by Yakubova et al. (1976) was presented in an anecdotal manner and
therefore is of little scientific value. A case report described by Holmberg
(1979) and Holmberg and Nurminen (1980) indicated that one women exposed to a
variety of substances, including ethylene oxide, produced an infant with
multiple defects and hydrocephalus. However, because of the multiple chemical
exposures involved, this study is of little value in establishing the
potential of ethylene oxide to cause adverse effects.
A recent epidemiological study has been conducted evaluating the
pregnancy outcome of nursing personnel exposed to ethylene oxide (Hemminki et
al., 1982). Although there were problems in the study design and collection
of data, the data are sufficient to suggest an association between ethylene
oxide exposure and spontaneous abortion, warranting further examination of
adverse pregnancy outcomes. Additional epidemiology studies would be helpful
to more firmly establish the potential of ethylene oxide to cause adverse
reproductive effects in humans.
In conclusion, ethylene oxide produces adverse reproductive and
teratogenic effects in both females (maternal toxicity, depression of fetal
weight gain, fetal death, fetal malformation) and males (reduced sperm numbers
9-45
-------
and sperm motility) if the concentration of the chemical reaching the target
organ is sufficiently high or if exposure at lower levels is sufficiently
long. Thus, the experiments in which ethylene oxide was injected
intravenously have produced more detrimental effects than the short-term
inhalation experiments. Even short-term inhalation experiments, however, have
resulted in suggestive evidence of detrimental effects. The levels needed to
produce the developmental effects approach or equal the levels needed to
produce toxicity in the dams. The effects of ethylene oxide on human
reproduction have not been studied in depth, although one study indicates that
ethylene oxide may be associated with spontaneous abortion (Hemminki et al.,
1982). Future studies are needed to establish this effect in humans.
9.4. MUTAGENICITY
Ethylene oxide has been evaluated for mutagenicity in several different
systems including tests in bacteria, fungi, higher plants, Drosophila,
mammalian cells in vitro, and rodents. Effects in humans are also reported.
The available data concerning the mutagenicity of ethylene oxide are discussed
below and summarized in Tables 9-4 to 9-16. The reader may also wish to refer
to other reviews of the mutagenic potential of ethylene oxide (e.g., Fishbein,
1976, Wolman 1979, Ehrenberg and Hussain, 1981, and NIOSH, 1981).
9-46
-------
TAH.E 9-4
Summary of Mutagenicity Testing of Ethylene Oxide: Gene Mutations in Bacteria
Re f e re nc e
Rannug et al. ,
1976
Test
System Strains
Salmonella/ TA1535
micro some assay
(suspension
assay)
Activation
System
None
Chemical
In ft) r nation
Concentration tested:
0-95.5 mM
Source: Fluka
Results
Strong
positive
response
Comments
1. Ethylene oxide used as a
control.
2. Dose -dependent response.
positive
15-fold
Pfeiffer and Salmonella/ TA98
Dunkelberg, microsome assay TA100
1980 (plate test) TA1535
TA1537
None
Purity: Not given
Solvent: Cold ethanol
Concentration tested:
0-200 (moI/plate
(0-8.8 rag/plate)
Source: J.T. Baker
Chemicals BV
Deventer, The
Netherlands
Purity: 99.7 %
Solvent: Cold acetone
increase in revertants noted at
highest dose compared to negative
controls.
3. Five plates used per dose.
Positive 1. Dose-dependent response for TA1535
and TA100.
2. Concurrent negative control values
not given.
3. Compared to lowest dose (20 (jnol/
plate), revertant count at highest
dose (200 uraol) was elevated
18-fold for TA1535 and 2.25-fold
for TA100.
4. Between 6 and 10 independent runs
were done in duplicate for each
experiment.
-------
TAELB 9-4 (cent. )
Re fe re nc e
Test
System
Activation
Strains System
Chemical
Information Results
Comments
Tanooka, 1979 Bacillus subtills
spores (reversion
to his*
pro to trophy)
CD
HA 101
(his met
leu)
TKJ 5211
(his net
uv rA10)
TKJ 8201
(his net
polA151)
None
Concentration tested:
27.3% atmosphere of
ethylene oxide gas
for times ranging
from 5-50 minutes.
Source: Daicide LS gas
Daldo Oxygen Co.
Tokyo, Japan
Purity: 27.3 J ethylene
oxide 72 .1% Freon
Positivel. Tests conducted in a polyethylene bag;
response *) x 10 spores placed on sterile
filter inside bag.
2. Negative control values not provided.
3. Revertant values expressed as muta-
tion frequency (6 x 10" after
5 minutes exposure and 8 x 10
after 50 minutes exposure of HA 101
and TKJ 5211).
H. Lethal and ma tag en ic effects were
enhanced In the polA strain; TKJ
8201 was 10x no re sensitive than
HA 101 and TKJ 5211.
-------
9.1.1. Gene Mutation Studies.
9.1.1.1. PROKARYOTIC TEST SYSTEMS (Bacteria) — Several investigators
have shown that ethylene oxide causes point mutations in bacteria (Table 9-1).
Ethylene oxide is a very effective sterilant for products that would be
damaged by other sterilization methods. Bacillus subtilis var. niger is
commonly used to monitor the effectiveness of ethylene oxide sterilization.
Jones and Adams (1981) found that treatment of spores of these bacteria with
Pennges (12:88 ethylene oxide-Freon mixture by weight) for 5 minutes increased
the number of colony variants by five fold over the spontaneous level. Forty
aberrant isolates (out of 125 found) were plated five times in succession; of
these, 11 reverted to typical appearance, 12 changed to other atypical
appearances, and 17 remained stable. Although the changed colonies were not
well-defined genotypically these data suggest that ethylene oxide induced
mutations in the surviving spores.
In a study by Rannug et al. (1976), ethylene oxide was chosen as a
positive control chemical in tests of other chemical substances in the
Salmonella/microsome assay. In this study, strain TA1535 was exposed to
concentrations of ethylene oxide (purity not reported) ranging from 0-95.5 mM
in a suspension test without addition of an exogenous mammalian metabolic
activation system (Table 9-1). A statistically significant dose-related
response was observed (Figure 9-1) where the maximum killing was -20%.
In another Salmonella assay, Pfeiffer and Dunkelberg (1980) exposed
strains TA98, TA100, TA1535, and TA1537 to concentrations of ethylene oxide
(99.7% pure diluted in cold acetone) ranging from 0-200 p.M (0-8.8 mg/plate)
(Table 9-4). Between 6 and 10 trials were performed in duplicate. A clear
9-19
-------
108
Q.
C/5
y = 0.996x +6.02
R = 0.9897
p <0.01
24 36 48 60 72
ETHYLENE OXIDE CONCENTRATION, mM
84
96
Figure 9-1. Mutagenic response of Salmonella typhimurium strain TA 1535 exposed
to ethylene oxide.
Source: Rannug et al. (1976)
9-50
-------
dose-dependent response was observed for the base-pair substitution detecting
strains TA100 and TA1535 but not for the frameshift detecting strains TA98 and
TA1537. This result is consistent with responses observed with other alky-
lating agents.
Tanooka (1979) exposed spores from three different his~ Bacillus subtilis
strains to an ethylene oxide gas mixture (Daicide LS, comprised of 27.3%
ethylene oxide and 72.7$ freon gas) in a plastic bag (Table 9-4). Histidine-
independent revertants were selected after treatment; a repair-competent
strain and a uvrA repair-deficient strain were treated for times ranging from
5-50 minutes. Exposure-related revertant frequencies were observed for both
-6 -4
strains (ranging from 3 x 10 after 5 minutes exposure to 2 x 10 after 50
minutes exposure). In a similar experiment conducted with a polA strain a
significantly higher dose-related revertant rate was reported compared to that
obtained with the repair competent and uvrA strains. The revertant
frequencies corresponding to 5 and 40 minutes of exposure were about 8 x 10~
and 3 x 10 , respectively. A similarly elevated sensitivity of the polA
strain was observed for ethylene oxide-induced toxicity. No data were given
for negative controls for any of the strains. The his+ revertants produced in
the repair-competent strain exposed to ethylene oxide gas for 30 minutes were
characterized, and 85/t of them were found to contain suppressor mutations; 15/f
were true revertants as measured by cotransformation of hisB+ with the
neighboring trpC+ marker using DNA extracted from each his+ colony. Although
this study was not conducted using a "standard" assay system, it does indicate
that ethylene oxide is mutagenic in B. subtilis.
The positive responses in these tests show that ethylene oxide causes
genetic damage as evidenced by induction of mutations in bacteria. The
9-51
-------
studies described below show that ethylene oxide causes genetic damage in
higher organisms also.
9.4.1.2. EUKABYOTIC TEST SYSTEMS
9.4.1.2.1. Plants — Kolmark and Kilbey (1968) studied the induction of
ad+ revertants in Neurospora crassa strain K3/17 (macroconidia) after
treatment with ethylene oxide (source and purity not given). Five doses
ranging from 0.0015-0.15M were employed, but the corresponding mutation
frequencies were not reported (Table 9-5). The purpose of the study was to
investigate the kinetics of nutation induction. In this study, ethylene oxide
was found to be 15-21 times more effective as a mutagen than diepoxybutane.
Migliore et al. (1982) tested a series of aliphatic epoxides for their
ability to induce forward mutations in Schizosaccharomyces pombe. Ethylene
oxide treatment in liquid suspension at concentrations from 0.5-15 raM resulted
in dose-related increases in mutation frequency; survival was reduced -60% at
the high dose. One hundred-fold increases in nutation frequency were noted at
the high dose levels compared to the corresponding negative controls both with
and without metabolic activation by phenobarbitone-induced mouse liver S9 mix
(50.28 + 1.76 vs. 0.59 + 0.22 and 66.21 + 29.44 vs. 0.66 + 0.59 mutations/104
survivors, respectively). The ranking of the chemical substances tested with
respect to their relative specific activity was: epichlorohydrin > ethylene
oxide > glycidol > 1 ,2-epoxybutane > 1 ,1 ,1-trichloropropylene oxide >
propylene oxide > 2,3-epoxybutane.
Ethylene oxide is known to be a very effective mutagen in higher plants.
l"kny tests have been performed in which ethylene oxide has been shown to be
9-52
-------
TABLE 9-5
Sunmary of Mitagenicity Testing of Ethylene Oxide: Gene Natation Teats in Lower Plants (Yeast)
Reference
Test System
Chemical Information
Results
Co amenta
Kolmark and Kilbey, ad-3A revertants
1968 In Neuroapora era333
I
Ul
Migllore et al. ,
1982
Forward ma tat ion 3 at
the ade locus In
Sehizocaacharomycea
Concentration tested:
ranged from 0-0.1 M
(0-6.2 g/i) ethylene oxide.
Source: Imperial Chemical
Industries Ltd.
Purity: Not given
Solvent: Distilled water
Source: Montedison (Italy)
Purity: 99.70*
Do ae-re la ted
positive response
1. Objective of work was to study
kinetics of mutation.
2. Revertant values given in Figure of
paper as mutation frequencies (i.e.
ad+/1o6 survivors).
Do se -re la ted
positive response
ponfce
Solvent:
Water and
DM SO
Without S9
Doae
(mM)
0
0.5
1.5
5
15
Survival
71
99
80
35
100
.78
.19
.3
.14
Mutation
.X10-4
0.
1.
4.
18
66
66 ± 0
89 ± 1
17 ± 0
.77 ±
.21 ±
Freq.
.59
.00
.75
0.72
29.44
Survival
100
100
76.64
100
42.87
With
S9
Mutation
.xlO-1*
0.
3.
7.
14
59 ± 0
32+0
15+0
.33 ±
50.28 +
Freq.
.22
.96
.24
7.62
1.76
-------
mutagenic. The results of these studies will not be analyzed in depth. Most
were directed rautagenesis tests conducted to generate desirable traits in food
crops. The results of two tests, in which plants were treated with ethylene
oxide, will be discussed for illustrative purposes (Ehrenberg et al., 1956;
Jana and Roy, 1975). Ehrenberg et al. (1956) administered several chemical
substances including ethylene oxide (purity not given) to dry and presoaked
barley seeds which were subsequently screened for sterility (dependent on
chromosomal aberrations) and chlorophyll mutations (caused by gene mutations,
either chromosomal or ex trachomro so ma 1) in the developing plants (Table 9-6).
The seeds were exposed to ethylene oxide either as a gas (dry seeds receiving
80? ethylene oxide for 6 days) or in solution [seeds were presoaked in 0.12
and 0.03$ (0.27 and 0.07 M) solutions for 2 hours]. Ethylene oxide induced
mutations in a dose-dependent manner as can be seen in Table 9-6. A 5-fold
increase in lethal nutations and a 33-fold increase in chlorophyll mutations
were observed.
Jana and Roy (1975) treated dry seeds of two varieties of rice, IR8 and
Dular, with 0.1-0.6? (0.02-0.14 M; pH 7.0)) ethylene oxide (purity not
reported) solutions at 10°C for 8 hours. The seeds were sown and the plants
grown and harvested. Seeds from single plants were collected and thoroughly
mixed to obtain a random sample of seeds. These were then grown to get at
least 100 plants from treated original seed for the next generation. These
plants were scored for gene mutations affecting chlorophyll expression, and a
dose-related mutation frequency was observed (Table 9-6). Although negative
controls were not reported, and the spontaneous mutation frequency was not
provided, about three times as many mutants were reported in offspring from
plants receiving the highest dose compared to those receiving the lowest dose.
9-54
-------
TABLE 9-6
Summary of Mutagenicity Testing of Ethylene Oxide: Mutation Teats In Higher Plants
Reference
Test System
Chemical Information
Results
Comments
Ehrenberg et al.
1956
Lethal ( chromosomal)
and chlorophyll
(gene) nutations In
barley.
Vhen tested as a gas, resting
seeds exposed to 80J ethylene
oxide for 6 days. Wien tested
In solution, partly preaoaked
seeds exposed to 0.03) and
0.12J (0.27 and 0.07 M)
ethylene oxide fbr 2 h at 20°C.
Postive response 1. Third generation progeny not
available for analysis when report
written; positive response nay be
due to extra chromosomal mutations.
2. Mutagenic response observed after both
types of treatment.
3. Half-life of ethylene oxide in water
solution la around 100h at 20°C.
%
EtO
0
0.03
0.12
80
%
Sterility
t
5.7
9.5
22.1
* 2nd
generation
ch lorophyll
gene
mu ta t Ion s
0.05U
0.20
0.75
1.8
No. spikes
analyzed
15,861
2,510
1,872
989
treatment
condition
None
Solution
Solution
Gas
-------
TABLE 9-6 (oont. )
Reference
Test Systea
Chemical Information
Results
Comments
Jana and Roy,
1975
Chlorophyll gene
nutations In rice
(IR8 and Dular)
Ul
CTv
Concentration tested:
ranged from 0 to 0.6<
ethylene oxide. Seeds
treated fbr 8 hours at
10«C and pH 7.0
Source: Bast man Organic
Chemicals
Purity: Not given
Solvent: Not given
Dose-re la ted
positive response
% 2nd Generation Chlorophyll
Gene filiations
jtEtO Dular IR8
1. Objective of study UBS to study
kinetics of mutation.
2. Revertant values given in Figure
In text as nutation frequencies.
0.1 5.0 + 0.36 5.9 + 0.13
0.3 7.0 + 0.37 7.0 + 0.30
0.5 12.3 + 0.32 12.0 ± 0.19
0.6 11.6 + 0.13 13.1 + 0.16
-------
The positive responses observed in plants are consistent with the bacterial
results and show that ethylene oxide is rautagenic in plants.
9.4.1.2.2. Animals — Ethylene oxide has also been shown to cause both
gene and chromosomal mutations in animals. Bird (1952) injected adult male
Drosophila melanogaster (Oregon K) with 0.5 and 0.8$ (0.11 and 0.18 M)
ethylene oxide to test its ability to induce sex-linked recessive lethal
mutations (Table 9-7). The highest dose level approximated the LD The
exact amount administered and the purity of the sample were not reported.
There were no sex-linked recessive lethals in 494 offspring of untreated
flies. Ten lethals out of 713 offspring (1.4$) and 9 lethals out of 198
offspring (4.5$) were detected after treatment with 0.5$ and 0.8$ ethylene
oxide, respectively. The dose-related positive response reported indicates
ethylene oxide is mutagenic in Drosophila.
Vfatson (1966) fed ethylene oxide to male Oregon K Drosophila melanogaster
to compare the induction of sex-linked recessive mutations with the induction
of heritable translocations. A second objective of this study was to compare
the effect on mutation yield of storing sperm in seminal receptacles after
treatment with alkylating agents. A positive dose-related increase in both
endpoints resulted from ethylene oxide treatment (Table 9-7). For the sex-
linked recessive lethal test, -3% lethals were detected at the low dose (0.4$
ethylene oxide) compared to 1% at the high dose (0.7$ ethylene oxide). For
translocations these values were =0.28 and 0.7$, respectively. Negative
control values were not given. Storage of ethylene oxide-treated sperm in the
seminal receptacles for 6 days had no effect on the frequencies of the two
types of genetic damage.
9-57
-------
TABLE 9-7
Summary of Mutagenicity Testing of Ethylene Oxide: Gene Mutation Tests in Insects
MD
I
CO
Test
Reference System Strain
Bird, 1952 Drosophila Orgeon K:
melanogaster adult males
sex- linked
recessive
lethal test
Watson, 1966 Drosophila Oregon K:
melanogaster adult males
sex-linked
recessive
lethal test
and heritable
translocation
test
Chemical
Information Results
Ethylene oxide administered Dose-related 1.
by feeding, inhalation or positive
injection. (Data not pre- response
sented for first two routes
of of administration.) For 2.
injection experiments 0.5-5%
solutions administered to
20 males. Dosages >0.8J
lethal. 0.8J ethylene oxide
killed 50J of treated flies
while 0.5f ethylene oxide
did not affect viability
Source: Not given t No.
EtO Chromosomes
Purity: Not given
0 191
Solvent: O.W saline 0.5 713
0.8 198
Concentration tested: Positive dose- 1.
0, O.Ot, or 0.7* (0, related
0.09, or 0.16 M) ethylene response
oxide
Source: Not given
2.
Purity: Not given
Solvent: Not given 3.
Comments
Objective of experiment was to find most
effective method of administration for
routine testing.
Cannot determine germ cell stage
specificity.
No. %
Lethals Lethals
0 0
10 1.1|
9 4.5
Objective of experiment was to determine
effect of sperm storage in female seminal
receptacle on mutation frequency after
treatment with monofunctional and
bi functional alkylating agents.
Did not observe storage effect for ethylene
oxide with respect to either endpoint.
Cannot determine germ cell stage
specificity.
Pre-stored
Post-stored
0.7
O.I)
itSIRL
3.3
3.6
7.1
3.3
3.1
0.29
0.39
0.69
0.79
0.37
% Trans.
0.08
0.1
0.1
0.2M
0.12
0.7
6.8
0.60
0.09
-------
TABLE 9-7 (cont. )
U1
Reference
Lee, unpublished
Test
System Strain
Drosophlla
me la nog aster
sex -linked
recessive
lethal test
and gonadal
Chemical
In fo rma t ion He su It s
Source: Not given for
unlabeled ethylene
oxide ^H-ethylene
oxide from New
England Nuclear
sp. act. = 2.8
cl /mmo le
Comments
1. Objective of experiment was
the relation of exposure to
alkylatlon of germ cell DNA
to nutational response.
to determine
level of
Purity: Not given
Exposure
(umole/25 ml vial)
0
0.086
O.U3
(Dose)
Alkylation/
Nucleotlde x 1Q-3
5.58
22.3
% SLRL
0.12
0.35 + 0.07
0.92 + 0.2
-------
Lee (unpublished data) conducted parallel experiments with unlabeled and
^H-labeled ethylene oxide to determine: 1) the relationship between exposure
and the level of alkylation of germ cell DNA, and 2) the relationship of germ
cell DNA alkylation to mutational responses in Drosophila melanogaster males.
For both the dosimetry and genetic test treatments, ethylene oxide was
given to the flies by adding 0.7 m& of cold water solutions to glass fiber
paper in 25 mJl scintillation vials (0.086 or 0.43 (jmole/vial). Immediately
afterward 50 males were added to the vials which were then sealed. Treatment
ill
was continued for 24 hours at 25°C. C-Thymidine was also given to males in
the dosimetry experiment. The number of alkylations per nucleotide of DNA was
calculated based on the 3n/ 1^C ratios in purified sperm DNA (to determine the
number of alkyl groups present) and the 1l*C/sperm cell ratio (to determine the
amount of sperm cell DNA in the extraction product). The genetic data showed
ethylene oxide to be an effective mutagen since dose-re la ted increases in sex-
linked recessive lethals were observed (see Table 9-7). Using the exposure-
dose relationship determined from the dosimetry experiments and the genetic
data, a doubling dose of 2.3 x 10 alkylations/nucleotide was calculated.
These studies show that ethylene oxide is distributed to the gonads of a
higher eukaryote (Drosophila) and causes heritable genetic damage.
9.4.1.2.3. Nfemmalian Cells in Culture — Three tests have been conducted
to ascertain the ability of ethylene oxide to cause gene mutations in
mammalian cells in culture. Brown et al. (1979) reported in an abstract that
polymethacrylate plastic sheets and polypropylene plastic sheets and mesh
sterilized by ethylene oxide gas adsorbed ethylene oxide molecules which could
be released later to exert a mutagenic effect. They placed the ethylene oxide
9-60
-------
treated plastic, of unspecified size, in culture flasks containing L5178Y
TKV- mouse lymphoma cells for 3 days. This was followed by dilution in
ethylene oxide-free media for 3 days prior to selection using BUdR. Poly-
methacrylate sheets treated for 18 hours with pure ethylene oxide were
estimated to release 8-40 |ig ethylene oxide (as measured by GC into the
flasks, while similarly treated polypropylene sheets and meshes released
5-100 |ig ethylene oxide. Although the spontaneous negative control mutation
frequencies were not given, the released ethylene oxide was reported to result
in a 2- to 20-fold increase in induced mutation frequency relative to the
controls (Table 9-8). It was not possible to evaluate this report critically
because it was presented in abstract form.
Tan et al. (1981) administered ethylene oxide (Matheson Co., 99.7? pure,
Dr. R. Gumming, personal communication) to Chinese hamster ovary cells at
concentrations ranging upwards to 10 mM in the medium. Mutations at the HGPRT
locus were selected after 5 hour ethylene oxide treatments both with and
without an exogenous metabolic activation system (39 mix derived from Aroclor
1254-induced rat livers) followed by a 16-18 hour recovery period and subcul-
turing for 1 week. A dose-dependent positive response was obtained at concen-
trations causing between 10? and 90? cell killing (Figure 9-2) both with and
without metabolic activation. The mutation frequency at the highest dose not
resulting in excess toxicity (<80? cell killing) was roughly 10 times greater
than the reported spontaneous frequency (see Table 9-8).
Hatch et al. (1982) and Dr. S. Nesnow (personal communication, 1983)
exposed Chinese hamster V-79 cells to ethylene oxide gas at concentrations up
to 7500 ppm and selected for ouabain- and 6-thioguanine resistant mutants.
Significant numbers of mutants were produced for both genetic markers. There
9-61
-------
TABLE 9-8
Summary of Mutagenicity Testing of Ethylene Oxide: Mammalian Cells in Culture
Reference
Test
System
Activation
System
Chemical
Information
Results Comments
Brown et al., L5178Y TK+/-
1979 mouse lymphoma
gene mutation
assay
None
Polymethacrylate (PUMA)
plastic sheets and
polypropylene (PP)
plastic sheets and meshes
sterilized for 18 h in
pure gaseous ethylene oxide.
PMMA retained ethylene oxide
and established concentra-
tions of 8-40 (ig/20 mi
cultured medium (1-5 x
10-5M ethylene oxide).
PP retained ethylene oxide
and established concentrations
of 5-100 ng/ 20 ml in cultured
medium.
Source: Not given
Purity: Not given
Solvent: None
2 to 20-fold
induced
mutation
frequency
observed
Presented in abstract.
Chemical concentrations measured by gas
chromatography.
Two ethylene oxide metabolites also
tested. At the low, but unspecified,
level tested, ethylene glycol residues
did not produce an effect. Chlorohydrin
produced residues of 15-30 |ig/piece of PP.
Direct addition of this compound to the
medium resulted in a 2-3 x Induced
mutation frequency.
Tan et al.,
1981
HGPRT Chinese
Hamster Ovary
cell gene
mutation
assay
Liver S9 mix
from Aroclor
12511-induced
Sprague-
Dawley rats
Concentrations tested
0 to 10 mM
Dose-related
positive
response with
and without
activation
Concentrations and induced mutants
extrapolated from Figure 9-1 of text.
250-300 mutants/106 cells at high dose
both with and without activation compared
to 0-10 mutants/10" cells in negative
controls.
Direct acting rautagen.
Ethylene oxide both cytotoxic and mutagenic
-------
MUTATION FREQUENCY, HGPRT mutants x 1(r6/clonable cell
i
cri
CO
I
-1
o
CD
Q.
T3
CD
Q.
—h
O
3
O)
3
CO
00
in
CD
c
03
to
OJ
3
-a
o
3
in
CD
O
—h
O
I
o
o
o
CD
fD
(0
O
X
O
X
o
o
z
o
m
30
H
O
z
3
RELATIVE SURVIVAL, percent
-------
was a dose-related increase in mutation frequency. The response at the
highest dose was 20 times greater than negative control rates. The work was
reported in an abstract.
The studies by Brown et al. (1979), Tan et al. (1981), and Hatch et al.
(1982) indicate that ethylene oxide causes gene mutations in cultured
mammalian cells.
9.4.2. Chromosome Aberration Studies. Nkny studies have shown that
heritable chromosome aberrations are induced in plants after ethylene oxide
exposure [e.g., Moutschen et al. (1968) in barley and Mackey (1968) in wheat].
These studies will not be discussed in this report. Nfost were directed
mutagenesis studies designed to obtain desirable variants. The ability of
ethylene oxide to cause such mutations shows it to be an effective clastogen
in plants.
9.4.2.1. DOMINANT LETHAL TESTS — Ethylene oxide causes chromosome
damage in both mammalian germ cells and somatic cells (Tables 9-9 to 9-13).
Ethylene oxide has been tested in dominant lethal tests in both rats and mice
and has yielded a positive response in each (Table 9-9). The precise nature
of the damage causing dominant lethal effects is not known, but there is a
good correlation between chromosome breakage in germ cells and dominant lethal
effects (Nfetter and Jaeger, 1975). When dominant lethal effects are observed
in the offspring of treated males, it can be concluded that the test agent
reached the gonads and likely caused genetic damage. Embree et al. (1977)
conducted a dominant lethal test with Long Evans rats. Twe Ive-week-old males
inhaled 1000 ppm ethylene oxide for 4 hours (Matheson Gas Products, Newark,
9-64
-------
TABLE 9-9
Summary of Mutageniclty Testing of Ethylene Oxide: Dominant Lethal Testa
Reference
Embree et al. ,
1977
Generoso
et al.,
I960
Test
System
Dominant
lethal assay
in Long Evans
rats
Dominant lethal
assay: male
mice T stock
(Experiment
I) and (101 x
C3H)F,
(Experiment
ID
Mating and
Sacrifice
Each male
placed with
2 virgin
females per
week for 10
weeks. Females
sacrificed on
the 17th day
after first
exposure to
male.
Experiment I:
Mated to 2
virgin (SEC
x C57BDF,
females about
12 weeks old.
Females replaced
when vaginal
plug observed.
Sacrificed 12-
15 days later.
Experiment II:
Mated to 2
virgins from
one of the
following stocks
Chemical
Information
12 week old male
animals exposed to
1000 ppm ethylene
oxide via inhalation
for t hours
Source: Not given
Purity: Not given
Single i.p. injection
of 150 mg/kg.
Maximum volume of
1 ml
Source: Eastman
Kodak Co.
Purity: Not given
Solvent: Double-
distilled
water
Results Comments
Positive response. 1. Animals exhibited toxicity but no deaths
Significant Increase resulted.
in postimplantational
fetal deaths during 2. Pattern of positive response indicates
first 5 weeks of the postmeiotic effect.
experiment
% Dead Implants
Week EtO Control
1 12» 2
2 30* 10
3 30» 1
19 8
5 10" 4
10 9 11
•P<0.05
Positive response 1. i.p. route of administration chosen to
observed for days mimic implanation of medical device.
2.5-11.5. Corresponds
to treated spermatozoa
and late spermatids.
During this period 12
to 31> dead implants
in treated group
compared to 3 to 5J
dead implants in
negative control group.
Little or no difference
in the yield of dominant
lethal mutations in male
postmeiotic germ cells
when mated to females
from different stocks.
T, (SEC x C57BDF,,
(101 x C3H)F,, or
(C3H x C57BDF!.
Sacrifice 12-15
days after
observation of
vaginal plug.
-------
TABLE 9-9 (cent.)
Re fe re no e
Appelgren
et al. ,
1977
M3
1
cr>
ON
Test (feting and
System Sacrifice
Dominant lethal teles rated to
assay: mice 3 virgin
females per
week. Females
sacrificed
on 17th day
after first
exposure to a
male.
Chemical
Information Results
Single injection of Negative
either 0, 0.025, response
0.05 , or 0.1 g/kg
of ethylene oxide
given i.v.
Source: Not given
Purity: Not given
Comments
1. Reported data of dominant lethal test
from work by Bateman.
2. Positive controls showed a significant
dose-related positive response.
3. Highest dose is 1/3 that used by Generoso
et al. , 1980; route of administration
different from those used by Generoso
et al. and Embree et al. (1977).
Solvent: Saline
'I. Conducted whole body autoradiography study.
Determined ethylene oxide distribution to
various tissues in the body, Including
gonads after either Inhalation or injection.
-------
TABLE 9-10
Summary of Mutageniclty Testing of Ethylene Oxide: Heritable Tranalocation Test
1
en
— 0
Reference
Generoso
et al.,
1980
Test
System Strains
Heritable T stock males
translocation treated and
mated to
(SEC x
C57BDF,
females
Chemical
Information
Single daily intra-
peritoneal Injection
of 0, 30, or 60
mg/kg of ethylene
oxide weekdays for
5 weeks
Results
Dose-related
positive
response
Comments
1. Shape of response curve consistent with
dose-squared kinetics.
2. Demonstrates capability of
to cause heritable genetic
in vivo.
ethylene oxide
damage in mice
Dose
(mg/kg)
0
30
60
60
Translocation Heterozygotes
Frequency %
0/822
6/1456
38/U06
6/72
0
1.32
9.36
8.33
-------
TABLE 9-1 1
Summary of Mutageniclty Testing of Ethylene Oxide: Chromosome Aberration Teats
I
CTi
OO
Reference
Fomenko and
Strekalova, 1973
Strekalova, 1971
Poirier and
Papadopulo, 1982
Teat
System
Chromosomal
aberrations
in bone marrow
from rata
Chromosome
aberrations in
bone marrow from
random bred
white rata
Chromosomal
aberrations in
the human
amniotic cell
line FL.
Chemical
Information
Concentration tested:
0.001-0.003 and
0.030-0.060 mg/liter
for 2, 4, 8, and 30 days
by inhalation
Source: Not given
Purity: Not given
Solvent: Not given
Concentration tested:
9 rag/kg per os
Source: Matheson Gas
products
Purity: Commercial
Grade
Reaulta Comments
Time-dependent poaitive 1. Method of preparing cells for analysis
response at highest dose not given.
2. Criteria for scoring aberrations not given.
3. Definition of terms not given.
1. Insufficient information for adequate
evaluation of results.
Positive response reported 1. Animals killed 24 and 48 hours after
treatment.
2. Chromosome preparations made from bone marrow
squashes.
3. Criteria for classification of aberrations
not defined.
4. Insufficient information for adequate evaluation
of results.
Dose-related poaitive 1. 1 hour vapor exposure.
response
2. Selected data presented only for cella harvested
72 hours after exposure.
EtO J Chromatid aberrations/ 100 cells
Dose Abnormal J
(mM) Metaphaaes Breaks Exchanges Survival
0 10.8 3.0 5.4 100
5 21.7 15.0 5.0 58
7.5 59.7 37.6 45.5 25
10 77.8 79.2 115.1 9.2
-------
TABLE 9-12
Summary of Mutagenioity Testing of Ethylene Oxide: Micronucleus Tests
Reference
Test
System
Chemical
Information
Results
Comments
Appelgren et al.
1978
Micronucleus
test: NMRI
mice and
Sprague-
Dawley rats
Concentration tested:
0 to 0.3 g/kg (mice)
or 0 to 0.2 g/kg (rats)
via intravenous
injection 30 and 6
hours before the
animals are killed.
Source: Not given
Purity: Not given
Solvent: Cold water
Dose-dependent response in 1.
mice. Increased Incidence
in rats, but severe bone
marrow depression prevented 2.
further characterization.
The animals given the highest doses died after
the first or second injection.
1000 polychromatic erythrocytes screened for
micronuclei per animal.
Conan et al.,
1979
Micronucleus
test: Swiss
mice
Concentration tested:
Two injections. Doses
ranged from 0-200 mg/kg
for i.p. injection, or
0-5 rag adsorbed to
Implanted plastic
devices.
Source: Not given
Purity: Not given
Solvent: Water
Dose-dependent positive
response after i.p.
injection.
Jenssen and
Ramel, 1980
Micronucleus
test: CBA
mice (males)
Concentration tested
0-175 mg/kg
Source: Fluka AC,
Switzerland
Purity: Not given
Solvent: Not given
Positive response
1. Two-fold increase noted in micronucleus
formation (0.33 + 0.10 in controls compared
to 0.93 + 0.31 at 150 mg/kg.
-------
TABLE 9-13
Summary of Mutagenicity Testing of Ethylene Oxide: Chromosome Mutations in Human Populations
Reference
Thiess et al.,
1981
Test
System
Chromsome
aberrations:
peripheral blood
of occupationally
exposed workers
Chemical
Information
Exposure:
1 . Long-term
(>20 years)
2. <20 years
3. Long-term plus
Results
Mutagenic effect Indicated
Aberrations excluding
gaps:
1. a. 3-5
b. 2.7
2. 2.3
3. 2.2
Comments
1. Workers were exposed to other alkylene
oxides besides ethylene oxide. Cannot
assign damage to one agent.
accident
4. Accident
5. Control
1. 1.1
Pero et al.,
1981
Chromosome
aberrations:
peripheral
blood lymphocytes
from ethylene
oxide-exposed
workers
Exposure levels:
0.5-1.0 ppra in air
Suggestive positive
response for aberrations
excluding gaps. Noted
only in comparison
1. Both exposed groups had significantly higher
levels of total aberrations (breaks and gaps)
compared to the control group
-------
California, purity not given). The LC is reported to be 1462 ppm for 4
hours. Erabree et al. (1977) reported signs of toxicity after treatment but
no deaths. Immediately following treatment, each male was mated to two virgin
females per week for 10 weeks. The females were sacrificed 17 days after they
were caged with a treated male. Statistically significant (P<0.05) increases
in the number of post imp Ian tat ion deaths were observed on weeks 1, 2, 3, and 5
after treatment, but not on other weeks, indicating that ethylene oxide exerts
its effects on post-meiotic cells. It should be noted that the statistical
significance of the increases observed for weeks 1 and 5 may have been due to
low negative control values for the corresponding weeks.
Generoso et al. (1980) also observed an increased incidence in postim-
plantation deaths in mice during the first 2 weeks after administration of 150
mg/kg ethylene oxide (Eastman Kodak, purity not reported) by a single intra-
peritoneal injection. One dose of 200 mg/kg ethylene oxide was shown to kill
10 of 12 mice. The testing for dominant lethal effects in this study was done
in two ways. In the first experiment, T stock males treated with ethylene
oxide were mated to two virgin (SEC x C57BL)F.. females. When a female was
impregnated, as evidenced by the observation of a vaginal plug, she was
replaced with another female. These females were also replaced after the
observation of a vaginal plug and so forth for three weeks post-treatment.
The females were sacrificed 12-15 days after the observation of the vaginal
plug and were dissected to determine the frequency of dominant lethal effects.
A significant increase in post implantation deaths was observed in females that
were bred with treated males between days 2.5 and 11.5 post-treatment (from
12-31? dead implants in treated group compared to 3-5? dead implants in
negative control group). This indicates that late spermatids and spermatozoa
9-71
-------
are sensitive to the test compound. In the second experiment (101 x C3H)F
males were injected with ethylene oxide and divided equally into four groups.
Four days post-treatment they were mated either to T stock, (SEC x C57BL)F
(101 x C3H)F1} or (C3H x C57BL)F1 females. The females were checked for
vaginal plugs each morning until the—S-th day post-treatment and were killed
for uterine analysis 12-15 days after the observation of a vaginal plug. The
purpose of this experiment was to determine whether the different stocks of
mice differed with respect to the ability of oocytes to repair genetic damage
induced in the treated male genome. The results of this experiment were
consistent with those of the first experiment in showing an increased inci-
dence of post implantation deaths. However, no significant difference was
observed when (101 x C3H)F.. -treated males were mated to females of different
stocks.
Appelgren et al. (1977) studied the whole-body distribution of radio-
labeled ethylene oxide in mice and reported the results of a dominant lethal
in
test. Nfeile mice were treated with [ C] ethylene oxide (specific activity not
given) by inhalation or intravenous injection. The animals were later sacri-
ficed and auto radiograms of midsagittal sections were prepared. The auto-
radiograms from mice that inhaled ethylene oxide differed qualitatively from
those that received the material intravenously in only one respect: the
mucosal membranes of the respiratory tract of animals that inhaled the com-
pound accumulated ethylene oxide. In experiments conducted using the intra-
venous route of administration, ethylene oxide was present in the gonads
(epididymis and testicle) 20 minutes after administration. Radioactivity was
still present in the epididymis 24 hours after injection.
9-72
-------
These observations that ethylene oxide reaches the gonads are consistent
with the positive dominant lethal responses reported by Erabree et al. (1977)
and Generoso et al. (1980). However, the results of the dominant lethal test
cited by Appelgren et al. (1977) were negative, in that there was no increase
in the incidence of dominant lethal mutations. The highest dose used in this
study was 100 mg/kg, as compared to the 150 mg/kg used by Generoso et al.
(1980). Since the chemical was administered by intravenous injection in the
study by Appelgren et al. (1977) and intraperitoneally by Generoso et al.
(1980), it is not clear whether the apparently negative response in the study
of Appelgren (1977) is attributable to the difference in the dose or to other
factors.
The positive dominant lethal tests reported by Embree et al. (1977) and
Generoso et al. (1980) indicate that ethylene oxide reaches the germinal
tissue in intact mammals and causes genetic damage. Although these tests do
not unambiguously demonstrate heritable effects caused by ethylene oxide, the
positive heritable translocation test reported by Generoso et al. (1980) does.
Mouse-specific locus tests, which measures heritable gene mutations, are now
underway at Oak Ridge National Laboratory and Research Triangle Institute, and
the results should provide additional insight into the ability of ethylene
oxide to cause heritable mutations in intact mammals.
9.4.2.2. HERITABLE TRANSLOCATION TEST — In conjunction with their
study of dominant lethal effects, Generoso et al. (1980) tested ethylene oxide
for its ability to cause heritable translocations in mice (Table 9-10).
T stock male mice were given 0, 30, or 60 mg ethylene oxide per kg once daily,
weekdays, for 5 weeks. Immediately after the last injection each male was
9-73
-------
caged with three (SEC x C57BL)F1 females. After 1 week the treated males were
removed, and the females were separated from each other. In the control
group, each male was left with one of the three females for -5 months after
the first litters were born in order to produce additional progeny. The
incidence of heritable translocations was as follows: negative control, 0%;
30 mg/kg, 1.32$; and 60 mg/kg, 9.36$. These positive results demonstrate that
ethylene oxide causes heritable chromosomal mutations in whole mammals.
9.4.2.3. CHROMOSOME ABERRATION TESTS — The ability of ethylene oxide
to cause well-defined chromosomal aberrations (breaks, rings, inversions,
translocations, etc.) has been studied by several investigators. Some of
these studies have been discussed above, including the positive heritable
translocation tests (Watson, 1966 and Generoso et al., 1980), and work
conducted with plants (e.g., Jana and Roy, 1975). Two additional experimental
studies were evaluated (Table 9-11). One was by Fomenko and Strekalova
(1973), who administered from 0.001-0.003 mg/Jl or from 0.030-0.060 mg/£
ethylene oxide (purity not reported) by inhalation for 2, 4, 8, or 30 days to
white rats (strain unspecified). A time-related increase in total aberrations
in bone marrow cells was noted in the high dose group (7.1-11.6$) compared to
the negative controls (3.0$). The significance of these results cannot be
determined, however, because of deficiencies in reporting how the chromosomes
were prepared and in defining criteria for scoring aberrations.
Similarly, Strekalova (1971) reported that administration of one 9 mg/kg
dose of ethylene oxide per os^ in aqueous solution resulted in an increased
incidence of total aberrations in bone marrow cells scored 21 , and in some-
cases, 48 hours later; the vague manner in which the study is reported,
9-74
-------
however, precludes an independent evaluation of the results. The most notable
problem is that the terms and the criteria for scoring aberrations are not
defined. Furthermore, bone marrow squashes were used to prepare metaphase
chromosomes for analysis. This technique is not suitable, because it does not
yield high quality chromosome spreads compared to chromosome preparations made
by the air-drying technique.
Poirier and Papadopoulo (1982) exposed F1 cells (derived from human
amnios) to ethylene oxide (commercially available from Matheson Gas Products)
at 5, 7.5, and 10 mM for 1 hour. The corresponding number of cells surviving
was 58, 25, and 9.2?. Three separate experiments were performed. After
harvesting (at 48, 72 and 196 hours) and slide preparation, 150 metaphases
were scored for each dose and fixation time (50 from each experiment). Dose-
related increases in chromatid aberration were found. For example, at 48
hours after treatment the frequency of exchanges (triradials, 'dicentric' and
'centric' rings) per 100 cells was 5.9, 10.6, 56.7, and 127.3 for the corres-
ponding treatments of 0, 5, 7.5, and 10 mM ethylene oxide/1 hour exposure
(Table 9-11).
Ethylene oxide at 50 and 100 ppm 7 hours/day, 5 days/week for 104 weeks
also significantly increased the frequency of chromatid/chromosomal aberra-
tions in peripheral lymphocytes of male Cynomologus monkeys (Lynch et al.,
1982a; Dr. D. Lynch, personal communication, 1983). The response was dose-
related; roughly 4-fold increases in cells with one or more chromatid and/or
chromosome aberrations were noted in the high dose animals compared to the
negative controls.
9-75
-------
9.4.2.4. MICRONUCLEUS FORMATION — Three studies addressed the ability
of ethylene oxide to induce micronuclei (Table 9-12). Appelgren et al. (1978)
treated NMRI mice by intravenous injection with two doses of ethylene oxide
ranging from 50-300 rag/kg, 30 and 6 hours before sacrifice, and Sprague-Dawley
rats according to the same regimen with doses up to 200 mg ethylene oxide/kg.
Mice given 300 mg/kg died after the first injection. Rats given 200 rag/kg
died after the second injection. In mice, ethylene oxide caused a highly
significant dose-related increase in micronuclei. At the highest dose there
were 2.48? polychromatic erythrocytes with micronuclei compared to 0.52? poly-
chromatic erythrocytes with micronuclei in the negative control animals
(P<0.001). Rats also exhibited a statistically significant increase in micro-
nuclei, but it was not shown to be dose-related. Toxicity to the bone marrow
confounded the results. The mid-dose level caused 1.08? polychromatic eryth-
rocytes with micronuclei compared to 0.49? polychromatic erythrocytes with
micronuclei in the negative controls (P<0.05).
Using male Swiss mice, Conan et al. (1979) conducted three different
types of experiments to assess the ability of ethylene oxide or its metabo-
lites, ethylene glycol and 2-chloroethanol, to cause micronuclei. Ethylene
glycol and 2-chloroethanol were given to the experimental animals via oral
administration or intraperitoneal injection. Ethylene oxide was administered
by intraperitoneal injection, intravenous injection or intraperitoneal
implantation of gas sterilized medical devices. Implantation of the ethylene
oxide gas sterilized medical devices did not induce elevated numbers of poly-
chromatic erythrocytes with micronuclei. Similarly, when ethylene oxide was
injected intravenously (two injections of 100 mg/kg 24 hours apart) and the
animals killed 6 hours after the second injection, no statistically signifi-
9-76
-------
cant increase in micronucleus formation was observed after treatment.
However, when ethylene oxide was given in traper itonea lly a suggestive positive
response was observed. As the dose of ethylene oxide increased from 0-4000
rag/kg (in traper itoneal), the percentage of polychromatic erythrocytes with
micronuclei increased from 0.23-0.47.
Jenssen and Ramel (1980) used CBA male mice in their assessment of the
ability of ethylene oxide to cause micronuclei. Ethylene oxide was
administered intraperitoneally at dosages up to 175 mg/kg, and micronuclei
were scored in polychromatic erythrocytes 24 hours later. The response was
not clearly dose-related, but a two-fold increase in micronuclei was observed
in the animals at the two highest doses (150 and 175 mg/kg) compared to values
for the negative control animals (0.93 + 0.31? and 0.66 + 0.1955 compared to
0.38+0.10$, respectively).
The positive responses obtained in the micronucleus tests of Appelgren et
al. (1978) and of Jenssen and Ramel (1980) indicate that ethylene oxide
reaches bone marrow and exerts a chromosome damaging (breakage and/or nondis-
junction) effect on heraatopoietic cells of mammals.
9.4.3. Chromosome Mutations in Human Populations. Three studies have been
conducted in which workers exposed to ethylene oxide have been monitored fbr
the presence of chromosome damage in peripheral blood lymphocytes.
Ehrenberg and Hallstrom (1967) monitored eight workers for the presence
of chromosome aberrations in peripheral lymphocytes eighteen months after an
acute exposure to high, but unspecified, concentrations of ethylene oxide.
Ten unexposed persons were selected as controls. The two groups were not
characterized in the report and it is not known how well the control group was
9-77
-------
matched to the exposed group. No analyzable cells were obtained from one
person in the exposed group. All samples were coded and an average of 20
metaphase plates was analyzed per remaining person (range = 6-26). Gross
chromosome aberrations (i.e., chromosome and chromatid breaks and exchanges,
supernumerary chromosomes, and one case of endoreduplication) were elevated in
the exposed subjects (17.5?) compared to the unexposed control subjects
(4.3?). Chromosomal effects such as this are potentially heritable and
represent clear evidence of genetic damage. The addition of chromosome gap
data to these values increased the respective incidences to 30.2 and 16.5/5.
Because of the small size of the study population and the low number of
metaphase spreads analyzed, the discriminating power of the study was not
great, and, thus, the elevated levels of chromosome damage observed in the
exposed population was judged not to be a significant positive effect.
Thiess et al. (1981) monitored 43 humans exposed to ethylene oxide and,
to a lesser extent, other alkylene oxides for the presence of chromosomal
aberrations (Table 9-13). The workers ranged in age from 27-63 years (x =
47.1 years). Individuals were divided into four groups based on the type and
extent of ethylene oxide exposure they had received:
1. Long-term exposure (>20 years), 11 men
2. Less than 20 years of exposure, 6 men
3. Long-term exposure plus accident, 21 men
4. Accident (i.e., short-term high exposure to ethylene oxide), 5 men
Subjects in the first three groups worked in plants where ethylene oxide was
manufactured or processed. Personnel in the fire department or maintenance
workers comprised the fourth group. The negative control group included male
9-78
-------
office and staff workers, none of whom had been exposed to radiation at the
time of testing. The age of individuals in the control group ranged from
24-58 years (x = 38.6). The workplace was monitored for ethylene oxide by
means of spot samples for up to 2-hour periods and for propylene oxide by
personal dosimeters for up to 10 hours over 12-hour shifts. Ethylene oxide
exposures were normally <5 ppm but were found to rise to 1900 ppm for several
minutes during a plant breakdown. Levels of propylene oxide were usually far
below the maximum allowable concentration of 100 ppra, but higher concentra-
tions were measured for brief periods. The percentage of aberrant metaphases,
excluding gaps, in cells cultured from 70-72 hours at 37°C in two control
groups was 1.4 and 1. Based on Fisher exact test analysis of the data, with
Yates correction, significantly increased incidences of chromosomal aberra-
tions were observed in Group I individuals (>20 years exposure) compared to
the control group upon examination in October 1978 (3.5%, P<0.005). An
increased incidence of aberrant metaphases was also noted when these indivi-
duals were examined again in August 1979 (2.7$, P<0.05). No statistically
significant increase was observed for the other groups. The significantly
increased rate of chromosome aberrations (excluding gaps) in workers exposed
to ethylene oxide for >20 years suggests a mutagenic effect. The results do
not conclusively indict ethylene oxide as the causative agent, however,
because the workers were exposed to other substances (such as ethylene chloro-
hydrin, ethyleneimine, propylene oxide, etc.) which may have caused or contri-
buted to the effect. Furthermore, it should be noted that the authors may not
have used an appropriate statistical test in their evaluation of the data. In
performing the Fisher exact test one must assume that one aberration is
independent of another aberration. Within individuals this may not be the
9-79
-------
case. If a person has one aberration he may be more likely to have a second
aberration particularly if the damage was induced in a stem cell. If this
were the case in the study by Thiess et al. (1981) one of the basic assump-
tions of the Fisher-Yates test, that of independence of the observations,
would not be met. A more appropriate statistical test, therefore, and one
which the authors claimed to have used (but did not report) in their analysis,
is the Mann-Whitney test. Use of the Mann-Whitney test to compare Group 1 and
the control group shows an increased (and perhaps biologically significant)
but not statistical difference between the two groups in regard to aberra-
tions.
Pero et al. (1981) also found increased incidences of chromosome aberra-
tions in factory workers exposed to ethylene oxide (Table 9-13). The workers
were divided into three groups. One was an unexposed control group and two
were exposure groups (i.e., sterilizers and packers) exposed to 50% ethylene
oxide and 50% methyl formate gas (0.5-1.0 ppm ethylene oxide) via inhalation.
Chromosome breaks and gaps were scored in the peripheral blood lymphocytes
from these individuals. Cells were cultured for 72 hours and 200 metaphases
were scored per individual. A statistically significant increase in chromo-
some gaps plus chromosome breaks was observed in cells from the sterilizer
ethylene oxide-exposed group (5 workers) compared to the control group (9
workers), 11-1^ in exposed groups compared to 8.5% in controls, (P<0.05).
With respect to breaks alone, however, a nonsignificant (or at best only a
marginally significant) increase was noted in the comparison between steri-
lizers and control groups (8.2 + 1.0% compared to 5.8 + 1.0?, respectively,
P<0.15). The comparison between the packer (12 individuals), 6.2 + 0.9%, and
control groups was not significant.
9-80
-------
The increased incidences of chromosome aberrations in peripheral lympo-
cytes noted in three studies of workers exposed to ethylene oxide are consis-
tent with one another and with the experimental animal data showing ethylene
oxide to be clastogenic. They indicate that similar effects are caused in
humans as well.
9.4.4. Other Studies Indicative of Genetic Damage. Additional studies have
been conducted bearing on the genotoxicity of ethylene oxide (Tables 9-14 to
9-16). These studies do not measure mutagenic events per se in that they do
not demonstrate the induction of heritable genetic alterations, but positive
results in these test systems do show that DNA has been damaged. Such test
systems provide supporting evidence useful for qualitatively assessing genetic
risk.
9.4.4.1. SCE FORMATION IN HUMAN POPULATIONS — Three studies have been
reviewed here concerning the induction of SCEs in humans (Table 9-14).
Lambert and Lindblad (1980) studied peripheral lymphocytes from five female
workers in a German sterilization plant to determine if ethylene oxide
exposure causes genotoxic effects in vivo as measured by SCE formation. A
description of the exposure these workers received was not reported. The
frequency of SCE formation in exposed individuals (19.1/E) was increased
compared to that of the unexposed control group (14.6$). Although the small
sample size and uncharacterized exposure these workers received preclude a
definitive assessment of the ability of ethylene oxide to cause SCEs in
humans, the results are considered to indicate genetic toxicity in somatic
cells of the exposed workers.
9-81
-------
CO
ro
TABLE 9-14
Summary of Mutagenicity Testing of Ethylene Oxide: SCE Formation in Human Populations
Reference
Johnson and
Johnson,
1982
Garry et al. ,
1979
Test
System
Sister chromatic!
exchange
Induction and
chromosome
aberrations:
Industrial
workers
Sister chromatid
exchange induction:
Chemical
Information
Inhalation exposures
estimated to be: Low
relative exposure
(1 ppm), moderate
relative exposure
(1-10 ppm), high
relative exposure
(5-200 ppm).
Months
0
High Potential 33
Exposure
Low Potential 14
Exposure
Inside Controls 12
Outside Controls
Maximum exposures
estimated to be 36 ppm
Results Comments
Dose-response association 1. Levels of SCE remained elevated after
suggested termination of exposure.
2. Environmental exposure to ethylene oxide causes
increased SCE formation.
3. Report based on preliminary data from
relatively small sample population.
SCE Chromosome Aberrations
after Exposure Months after Exposure
606
35 1.5 2
15 1.1 0.9
12 0.6 0.78
8 — 0.5
Statistically significant 1. Air dried fluorescence plus Giemsa
increases in level of SCE chromosome preps.
peripheral blood
lymphocytes
collected from
hospital workers
(from average measure-
ments over one 8 hour
period). Workers
divided into groups
based on known exposures
to ethylene oxide and
symptoms indicative of
exposure.
observed in exposed
Individuals compared to
controls (unexposed
laboratory personnel).
2. 20 metphases scored/individual.
-------
TABLE 9-1M (cont.)
Reference
Test
System
Chemical
Information
Results
Comments
Yager, 1982
and Yager
et al.,
1983
l
CD
OJ
Sister chromatid
exchange induction:
peripheral blood
lymphocytes
collected from
hospital workers
Exposures determined by
individually monitoring
workers. High exposure
group received a cumula-
tive dose >100 mg
while cumulative dose for
low exposure group was
<100 mg.
1. Control group carefully matched to the exposed
group for age, sex and personal habits.
2. Exposure estimates based on breathing zone
measurements and task frequency estimates.
Laurent
et al.,
1982
Sister chromatid No
exchange induction:
peripheral blood
lymphocytes collected
from hospital
workers.
Mean
Group Exposure (mg)
Control 7
Low exposure 13 7
High exposure 501 10
exposure estimates Exposed group had 1.
statistically significant
increase in SCEs compared
to control group. Range
of SCEs for the exposed group
was 9.61 - 17.57 compared to
a range of 7.0U - 8.52 for the
control group.
SCEs/
cell
.56 + 1.01
.76 + 1.05
.69 + 1.92
Control group
age, sex, and
group.
may not have been matched for
personal habits to the exposed
-------
TABLE 9-15
Summary of Mutagenioity Testing of Ethylene Oxide: SCE Formation in Experimental Studies
Reference
Test
System
Chemical
Information Results
Comments
Star, 1980
Sister chromatid
exchanges: Cultured
human fibroblasts
I
CO
-pr
Concentrations tested:
0 to 3600 ppm and
residues from plastic
children's endotracheal
tubes treated with 1100
mg/cm3 of pure ethylene
oxide for 90 minutes
followed by aeration
from 21 to 96 hours
after sterilization.
Source: STERI-Gas cartriges
3M Germany GmbH,
Neuss
Purity: Hot given
Solvent: Dulbecco's Modified
Eagle's Medium
Toxic as well as mutagenic.
Significant increases in SCE
induction at 36 ppm. Cyto-
toxiclty at 180 ppm and
higher
Cultures from skin biopsies used between
fifth and tenth subculture.
Insufficient data presented to evaluate
conclusions.
Yager and
Benz, 1982
Kligerman
et al.,
1983
Sister chromatid
exchange induction:
New Zealand White
rabbits
Sister chromatid
exchange induction:
CDF rats
Concentrations tested:
0, 10, 50, and 250 ppm
by inhalation
Source: Matheson
Dayton, OH
Concentrations tested:
0, 50, 150 and 150 ppm
for 1 or 3 days by
Inhalation
Positive response at 50 and
250 ppm exposures
Dose- and time-dependent
positive response
1 . Increased SCE levels decreased after exposure
ended but still remained above baseline
levels 15 weeks after exposure.
1. Significant increases at 50 ppm show effects
induced at levels to which workers have been
exposed. Until recently TWA was 50 ppm.
Source:
Matheson Gas
Product
2. Data for 3 days exposure groups shown.
Purity: 99.7*
•Significantly different from controls by one-tailed Dunnett's
Concentration
0
50 + 7
110 + 17
111 + 33
SCEs/
Metaphase
7.5 + 0.5
9.1 + 1.3»
10.3 +1.3*
13.6 + 1.3»
-------
TABLE 9-16
Summary of Mutagenioity Testing of Ethylene Oxide: Unscheduled DNA Synthesis
Reference
Gumming
et al.
(in press)
Test
System
Unscheduled DNA
synthesis:
testlcular DNA of
(101 x C3H)F!
mice
Chemical
Information
Concentration tested:
a. 600 and 800 ppm for
2, 14, 6, or 8 hours.
[3H] dThd administered
intratesticularly
immediately after
ethylene oxide
administration
Results Comments
a. Dose-dependent increase
in UDS over lower range
of doses tested (e.g.,
70 dpm/106 cells,
MB dpm/106 cells, and
8 dpm/106 cells for
800 ppm, 600 ppm, and
negative controls at
1 hours)
I
OO
U1
b. Same as above except
[3H) dThd administered
at different times
after termination of
exposure
c. 300 and 500 ppm 8 hr/day
for 5 days. Aliquots
of animals sacrificed
daily
d. 500 ppm for 2, 4, 6, and
8 h. 6 animals given 80
mg/kg 3-methylchloran-
threne, 6 animals drank
water with 1 mg/m{, sodium
phenobarbital for 1 week
prior to exposure, 6
animals uninduced controls
b. UDS peaks 2 hours after end
of exposure period at day
5 for 300 ppm; at day 1 for
500 ppm
Response peaked at day 5 for
300 ppm; at day 1 for 500 ppm
UDS response dramatically
reduced In animals receiving
mixed-function oxidase
inducers
Source:
Matheson Co., East
Rutherford, NJ
Purity: 99.7*
Pero et al.
1981
Unscheduled DNA
synthesis: Human
lymphocyte cultures
Exposure levels: 0.5 to
1.0 ppm in air
Positive response
1. UDS induced by exposure to N-acetoxy
acetyl aminofluorene (NA-AAF).
2. Decreases in NA-AAF-induced UDS measured
biochemically and by autoradiography in
lymphocytes from ethylene oxide-exposed
workers. UDS peaked at 2 mM exposure
NAAAF.
-------
In a preliminary, unpublished report Johnson and Johnson (1982) described
how they monitored workers at three sterilant facilities for the presence of
SCEs and chromosome aberrations in peripheral blood lymphocytes. Based on
environmental sampling the workers were assigned to one of the following cate-
gories depending upon the plant site at which they worked: high relative
exposure (5-200 ppm), moderate relative exposure (1-10 ppm), and low relative
exposure (1 ppm). The numerical exposure values represent the estimated range
of an 8-hour time-weighted-average inhalation exposure. Employees at each
plant were further categorized as to high or low potential for ethylene oxide
exposure based on their job description and other factors. During the course
of the study it was noted that the SCE levels in the control group of presum-
ably unexposed workers at Plant III were higher than those of other control
groups available for comparison at the time (12/metaphase compared to 7/meta-
phase). The study was therefore expanded to include an additional control
group, which was taken from the local community and matched by sex and age to
potentially exposed Plant III employees.
Preliminary analysis of the data indicates a consistent dose-response
trend at Plant III for SCE induction both at the original monitoring and later
after 6 months with no further ethylene oxide exposure (mean values of 12, 14,
and 33 SCEs/metaphase for internal controls, low potential exposure and high
potential exposure groups, respectively, compared to 8 SCEs/raetaphase for the
external control groups). A much less pronounced trend was noted at Plant II,
and the SCE data for Plant I showed no significant difference between poten-
tially exposed and control groups. Analysis of the chromosome aberration data
suggests a dose-related increase in damage, but the magnitude of the dif-
ferences between groups is not great. Thus, it appears that a dose-response
9-86
-------
association exists between exposure to ethylene oxide and SCEs in humans and
that the increased levels of SCEs appears to be stable, perhaps suggesting
long-lived adverse effects caused by human exposure to ethylene oxide. How-
ever, it is important to bear in mind that these conclusions are based on
preliminary data from a relatively small study population.
In a study of 12 ethylene oxide exposed workers from the instruments- and
materials-sterilization areas of a hospital, Garry et al. (1979) reported
increased SCE levels in the peripheral blood lymphocytes. The maximum
exposure concentration sampled 15 feet from the sterilizer was estimated to be
36 ppm based on an infrared spectroscopy measurement over one 8-hour period
during the course of the study. Individuals reporting upper respiratory irri-
tation had statistically significant increases in the incidence of SCEs com-
pared to the control population of 12 unexposed persons working in the
adjacent operating room (10.3 + 1.8 vs. 6.4 + 0.47, P<0.01).
Yager (1982) and Yager et al. (1983) also monitored 14 hospital workers
exposed to ethylene oxide. Thirteen persons not exposed to ethylene oxide
served as matched controls. Cumulative exposure doses during the 6 months
prior to blood sampling were estimated by monitoring air concentrations during
TM
defined tasks (using a Wilkes-Miran 1A Gas Analyzer) and multiplying this
value by the number of sterilizer loads processed. Based on these estimates,
the workers were assigned to the low dose group (13 + 18 mg ethylene oxide) or
the high dose group (501 + 245 mg ethylene oxide). An increased incidence of
SCEs/cell was observed in the high dose group (10.7 + 1.92) compared to the
low dose (7.8 + 1.05) and unexposed control (7.56 + 1.01) groups.
Laurent et al. (1982) also collected peripheral blood from hospital
workers exposed to ethylene oxide. Ten persons in good health and not exposed
9-87
-------
to any known toxicants were selected as the negative control group. It was
not reported whether the controls were matched for sex, smoking habits, etc.
They do not appear to have been matched for age because the age of the control
group ranged between 20 and 35 years while that of the ethylene oxide exposed
workers ranged between 23 and 51 years. No estimate was made of the exposure
received by the sterilizer operators, but they had a significantly elevated
level of SCE compared to the controls (13.02 + 2.294 vs. 7.86 + 0.479).
The increased incidence of SCEs observed in five groups of workers
exposed to ethylene oxide does not demonstrate that mutations occurred but
does indicate that ethylene oxide can cause genotoxic effects in somatic
tissue of humans in vivo.
9.4.4.2. SCE FORMATION IN EXPERIMENTAL STUDIES — Human cells in cul-
ture also exhibited increased SCE levels after exposure to ethylene oxide
(Table 9-15). Star (1980) exposed skin fibroblast cells from normal healthy
human tissue biopsies to 0-3600 ppm ethylene oxide or to plastic children's
endotracheal tubes sterilized with 1400 mg/cmj ethylene oxide at 55°C for 90
minutes followed by aeration in room air for varying times from 24-96 hours.
The cell lines were kept frozen in liquid nitrogen and used between their 5th
and 10th subculture. The placement of the plastic tubes in the culture medium
resulted in ethylene oxide concentrations ranging from 12-800 ppm as estimated
by GC of head space material. Excessive cell killing precluded scoring SCEs
above 600 ppm ethylene oxide for the experiment. No statistically significant
increase in SCEs was noted in the experiment using the endotracheal tubes, but
a consistent, apparently dose-related, rise in SCSs was noted in this part of
the study at doses >217 ppm. In the other set of experiments a statistically
9-88
-------
significant increase in SCE induction was reported at 36 ppm; however, insuf-
ficient data were presented to permit an adequate evaluation of the results.
A membrane dosimetry system was developed by Garry et al. (1982) to
enable the measurement and determination of dose-response relationships for in
vitro exposure to toxic gases. Elevated SCEs were observed in peripheral
lymphocytes cultured from healthy humans at exposures as low 10 (ig/mJ, (in the
media) during a 20-minute exposure period. A dose-related increase was noted
up to ethylene oxide concentrations of 35 [ig/mJl (the highest dose tested). At
this dose there were about 20 SCEs/cell compared to control levels of =5
SCEs/cell.
Yager (1982) and Yager and Benz (1982) administered from 10-250 ppm
ethylene oxide gas to 4-month-old male New Zealand White rabbits via inhala-
tion. Eight animals were placed in each exposure chamber for 6 hours/day, 5
days/week, for 12 weeks. Blood samples were taken from the marginal ear vein
at 1, 7, and 12 weeks of exposure and 2, 7, and 15 weeks after exposure.
Three animals per chamber were used for serial blood sampling for SCE and
hematological assays [i.e., red cell count (total and differential), white
cell count, hematocrit, and hemoglobin concentration]. One animal was held in
reserve and four animals were sacrificed immediately at the end of the 12-week
exposure period for analysis of reduced glutathione (GSH) in liver and blood.
Positive and negative controls were performed using intraperitioneal injec-
tions of mitomycin C and Hanks balanced salt solution, respectively, at each
time point. Exposure to 10 ppm did not cause a detectable increase in the
incidence of SCEs; however, exposure to 50 and 250 ppm did cause an increase
in SCEs (9.^7 + 0.26 and 13.17 + 0.32, respectively) that decreased after
exposure ended, but still remained above baseline levels (7.8 + 0.23) 15 weeks
9-89
-------
after exposure (8.45 + 0.30). Heraatological and GSH measurements from the
animals did not differ from controls.
After exposures to ethylene oxide of 0, 50, 150, or 450 ppra for 6
hours/day for 1 or 3 days, blood was removed from male CDF rats by cardiac
puncture, cultured in the presence of 5-bromodeoxyuridine and scored for SCEs
and chromosome breakage (Kligerman et al., 1983). No significant dose-
dependent increase in chromosome breakage was observed, but there was a
concentration dependent increase in SCEs. Animals in the highest dose group
exposed for 3 days had 13.6 + 1.3 SCEs/cell compared to the control value of
7.8 + 0.5 SCEs/cell. SCE induction was also significantly elevated after 3
days of exposure to 50 ppm ethylene oxide (9.1 + 1.3) showing effects in rats
at levels to which workers have been exposed. There was no significant
reduction in mitotic activity or slowing of cell kinetics.
9.4.4.3. UNSCHEDULED DNA SYNTHESIS — Gumming et al. (in press) tested
ethylene oxide for its ability to cause unscheduled DNA synthesis in germ
cells of male mice after inhalation exposures. Four experiments were
performed in which hybrid mice (101 x C3H)F.. were treated with 99.7% pure
ethylene oxide (Matheson Co.). In the first experiment, the effect of differ-
ential time exposures on unscheduled DNA synthesis induction was assessed.
Animals were treated with 600 and 800 ppm ethylene oxide for 2-8 hours, after
which exposed animals were anesthetized with metofane and injected intrates-
ticularly with [ H]thymidylic acid (dThd). A dose-dependent increase in
unscheduled DNA synthesis was found over the lower end of the dose range for
the first 4 hours of exposure in that a higher response was seen at 800 ppm
5
than at 600 ppm (e.g., 70 dpm/10 cells for 4-hour exposure at 800 ppm
9-90
-------
compared to 48 dpm/10 for 4-hour exposure at 600 ppm; controls incorporated 8
dpm/10 cells). Due to the toxioity of ethylene oxide at 800 ppm it was only
possible to measure up to 6 hours exposure for this concentration. In a
second experiment, ethylene oxide administration was the same as above, but
•3
[ HjdThd was administered to the animals at different times after the end of
their ethylene oxide exposure to characterize the unscheduled DMA synthesis
response at different times after treatment. Unscheduled DNA synthesis was
found to increase with time to a peak 2 hours after the end of the exposure
period, and to decrease subsequently. Two additional sets of experiments were
performed. The first was a workweek exposure regimen of 300 or 500 ppm for 5
hours/day for 5 days, and the second involved pretreatment of the animals with
mixed-function oxidase inducers (either a single intraperitoneal injection of
80 mg/kg 3-methylcholanthrene or administration of drinking water containing 1
mg/mfl, phenobarbital for 1 week prior to ethylene oxide treatment). Concerning
the workweek exposures, little effect was noted after the first two exposure
periods at 300 ppm. An effect was subsequently noted which rose to a maximum
after the 5th exposure period. At 500 ppm the maximum effect was seen after
the first exposure period. Apparently, increased levels of DNA damage
occurred throughout the week, but after the third exposure period the capacity
to respond to this damage appeared to be limited.
Pero et al. (1981, 1982) treated peripheral lymphocytes taken from
ethylene oxide exposed workers with 10 mM N-ace toxy-2-ace ty lam inof luorene (NA-
AAF) for 1 hour and subsequently measured the incorporation of [ H] thymidylic
acid into DNA to detect unscheduled DNA synthesis (Table 9-16). NA-AAF-
induced unscheduled DNA synthesis was found to be inversely related to the
duration of worker exposure to ethylene oxide and to the number of
9-91
-------
chromosome breaks observed. This suggests an inhibition of the cellular DNA-
repair capacity by ethylene oxide. Biochemical and auto radiography studies
were consistent with this response. When NA-AAF-treated lymphocytes were ex-
posed to ethylene oxide, it was found that concentrations above 2 mM resulted
in inhibition of unscheduled DNA synthesis.
As was the case for the studies of SCE induction, these results do not
show that ethylene oxide is mutagenic but do indicate it causes damage to DNA
and are consistent with the results showing that ethylene oxide causes
nutations.
9.^.5. Summary and Conclusions of the Mutagenicity of Ethylene Oxide
Ethylene oxide has been shown to induce gene mutations in bacteria, fungi,
higher plants, Drosophila, and cultured mammalian cells in tests conducted
without the use of exogenous hepatic metabolic activation systems. It is
therefore a direct-acting mutagen. Ethylene oxide has also been shown to
induce dominant lethal effects in mice and rats; chromosomal aberrations in
higher plants, Drosophila, mice, and rats; and micronuclei in mice and rats.
Based on these positive findings in different test systems, ethylene oxide is
judged to be capable of causing chromosomal aberrations. It has also been
shown to induce SCE in rabbits, rats and humans.
Tissue distribution studies have shown that ethylene oxide reaches the
gonads. This result is consistent with evidence that ethylene oxide causes
unscheduled DNA synthesis in germ cells of male mice and heritable mutations
in insects and rodents (i.e., sex-linked recessive lethals and heritable
translocations in Drosophila, dominant lethals in rats and mice and heritable
9-92
-------
translocations in mice). Ethylene oxide can therefore be regarded as rnuta-
genic both in somatic cells and in germ cells.
Based on the available data, there is overwhelming evidence that ethylehe
oxide is a direct-acting mutagen that has the potential to cause mutations in
the cells of exposed human tissue. The observations that ethylene oxide
reaches and reacts with mammalian gonadal DNA, and causes heritable mutations
in intact mammals, indicate that it may be capable of causing heritable
mutations in man provided that the pharmacokinetics of ethylene oxide in
humans also results in its distribution to the DNA of germ cells. Thus,
ethylene oxide should be considered a potential human mutagen.
9-93
-------
9.5. CARCINOGENICITY
The purpose of this section is to evaluate the likelihood that ethylene
oxide is a human carcinogen and, on the assumption that it is a human carcino-
gen, to provide a basis for estimating its public health impact and evaluating
its potency in relation to other carcinogens. The evaluation of carcinogeni-
city depends heavily on animal bioassays and epidemiologic evidence. However,
other factors, including mutagenicity, metabolism (particularly in relation to
interaction with DNA), and pharmacokinetic behavior, have an important bearing
on both the qualitative and the quantitative assessment of carcinogenicity.
The available information on these subjects is reviewed in other sections of
this document. The carcinogenicity of ethylene oxide has also been evaluated
by the International Agency for Research on Cancer (1976). This section pre-
sents an evaluation of the animal bioassays, the human epidemiologic evidence,
the quantitative aspects of assessment, and, finally, a summary and conclusions
dealing with all of the relevant aspects of the carcinogenicity of ethylene
oxide.
9.5.1. Animal Studies. Only a few studies have been conducted to assess the
carcinogenicity of ethylene oxide. Most of the reported studies have dealt
with subcutaneous administration and skin painting of the compound in mice and
intragastric administration in rats. These studies are discussed briefly
herein. Two lifetime inhalation studies in rats have been performed (Snellings
et al., 1981 and Lynch et al., 1982), and they will be described in detail.
9.5.1.1. MICE — Reyniers et al. (1964) conducted a study of female germ-
free mice that developed tumors (63/83) after being accidentally exposed to
9-94
-------
ethylene oxide-treated ground-corncob bedding for 150 days, and were moved to
untreated bedding for the rest of their lifespans. These animals developed
ovarian, lymphoid, and pulmonary tumors. Colony mates maintained on untreated
bedding did not develop tumors. All males exposed to ethylene oxide-treated
bedding died, with necropsy showing massive hemorrhage. The causative agent
was not identified, since chemical analysis of the bedding was not done. The
high number of tumors could have been due to other chemicals (such as ethylene
glycol or 2-chloroethanol, both derived from ethylene oxide) or to a viral
agent, although the authors believed that a viral agent was unlikely. High
toxicity is indicated by these findings in male mice. Because germ-free mice
are T-lymphocyte deficient, they may be more susceptible than normal animals to
tumor development, or the tumor development may be due to immune suppression.
At present, however, there is no evidence to support these hypotheses.
Dunkelberg (1979) studied the oncogenic activity of ethylene oxide dis-
solved in tricaprylin and administered subcutaneously to the interscapular area
of groups of 100 female NMRI mice in weekly dosages of 0.1, 0.3, and 1.0 mg.
The incidence of spontaneous subcutaneous tumors in these mice was between 0
and 2%. Preliminary results up to the 91st week of treatment showed that 6,
8, and 12 local tumors (sarcomas) occurred in mice receiving total ethylene
oxide doses of 9.1, 27.3, and 91.0 mg, respectively. No local tumors occurred
in mice receiving no treatment or tricaprylin alone. The number of tumors at
sites distant from the injection area was not significantly greater in the
group treated with ethylene oxide than in the two control groups. The final
report of this study (Dunkelberg, 1981) covers the period from the start of
the study to 106 weeks, at which time all of the animals were sacrificed. No
increase in tumors at remote sites was observed.
Lifetime skin painting studies with 10% ethylene oxide in acetone (three
9-95
-------
times weekly) were performed on 30 female mice by Van Duuren et al. (1965).
Application of 0.1 mL of ethylene oxide solution to the clipped dorsal skin
produced no tumors. Median survival time for the mice was 493 days. The
investigators indicated that rapid evaporation of the compound from the skin
was responsible for the negative results observed.
9.5.1.2. RATS — Walpole (1958) injected 12 rats subcutaneously with a
maximum total ethylene oxide dose of 1 g/kg (dissolved in arachis oil) over 94
days (dosing schedule not specified). Rats were observed for their lifetimes
following treatment, and no tumors were observed. Since the total amount of
ethylene oxide administered and the frequency of injection were not specified,
it is difficult to evaluate this negative result.
Dunkelberg (1982) administered ethylene oxide intragastrically by gavage
at two dosages, 30 and 7.5 mg/kg body weight, to two groups of 50 female
Sprague-Dawley rats with empty stomachs twice weekly for a period of nearly 3
years, using salad oil as the solvent. One group was treated with the solvent
alone, and the other group was left untreated. A positive control group was
treated with3 -propiolactone. The test substances were dissolved in 1 mL of
oil immediately before treatment. The design of the experiment is summarized
in Table 9-17 and the results are summarized in Table 9-18. Ethylene oxide
induced local tumors, mainly squamous cell carcinomas of the forestomach. The
first tumor occurred in the 79th week. The tumor rates were 62% in the 30
mg/kg group and 16% in the 7.5 mg/kg group. In addition, carcinomas in situ,
papillomas, and reactive changes of the squamous epithelium of the forestomach
were observed in other animals. An unspecified number of tumors occurred in
the glandular stomach. Ethylene oxide did not induce tumors at sites away from
the point of administration. Survival decreased in the positive control group.
9-96
-------
TABLE 9-17. DESIGN SUMMARY FOR CARCINOGENICITY TESTING
OF ETHYLENE OXIDE BY INTRAGASTRIC ADMINISTRATION
TO SPRAGUE-DAWLEY RATS
Group
Ethylene oxide I
Ethylene oxide II
Oil (vehicle)
Untreated
g -Propiolactone
Single
dose (mg/kg body wt)
(2x weekly)
30.0
7.5
1.0 mL
-
30.0
Average total
dose (mg/kg body wt)
5112
1186
-
-
2868
Number of
animals
50
50
50
50
50
SOURCE: Adapted from Dunkelberg, 1982.
TABLE 9-18. TUMOR INDUCTION BY INTRAGASTRIC ADMINISTRATION
OF ETHYLENE OXIDE IN FEMALE SPRAGUE-DAWLEY RATS
Number of rats with stomach lesions
Dose
7.5
30. Ob
Reactive
changes3
9
11
Carcinoma
in situ
4
4
Fibrosarcoma
0
2
Squamous cell
carcinoma
8
29
No stomach tumors were seen in either vehicle-controls or untreated controls.
aReactive changes of the squamous epithelium of the stomach comprised hyper-
keratosis, hyperplasia, and papillomas.
'-'Fifteen animals from the ethylene oxide I group developed stomach tumors, of
which 10 exhibited metastasis and invasive growth into neighboring organs.
SOURCE: Adapted from Dunkelberg, 1982.
9-97
-------
Two other studies designed to test for chronic toxicity of ethylene oxide
reported no tumors; however, the exposure and observation periods were too
short to adequately test the carcinogenicity of ethylene oxide in rats, mice,
monkeys, guinea pigs, and rabbits (Hollingsworth et al., 1956; Jacobson et al.,
1956),
9.5.1.2.1. Snellings et al. (1981) Inhalation Study — A 2-year inhala-
tion study (unpublished) was performed by Bushy Run Research Center, Pitts-
burgh, Pennsylvania (Snellings et al., 1981). Fischer 344 rats were exposed
to 100, 33, and 10 ppm of ethylene oxide vapor by the inhalation route, 6
hours/day, 5 days/ week, for approximately 2 years. Two groups were exposed
to untreated air under similar conditions. Whole-body exposures were conducted
in a dynamic exposure system in which the vapor concentration levels were
determined by gas chromatography. Initially, 120 rats per sex per group were
exposed, with interim sacrifices of 10 animals each at 6 and 12 months and 20
animals at 18 months to determine possible treatment-related effects. Interim
and terminal evaluation included hematology, serum clinical chemistry, urinaly-
sis, body weight, organ weight, bone marrow cytogenetic studies, and gross and
histologic examinations.
In the cytogenetic studies, no statistically significant differences were
noted for the "percentage of abnormal cells," the "average number of chromo-
somal aberrations per cell," or the "total number of chromosomal aberrations
(per rat)" for either males_or females exposed to ethylene oxide at 100 ppm
when compared with values obtained for the air-control groups. However, sta-
tistically significant chromosomal aberrations have been found in other ethy-
lene oxide studies (see section on mutagenicity).
Histopathologic examination was performed on all tissues of each air-con-
trol group and the 100 ppm group at 6 months and at 12- and 18-month necropsy
9-98
-------
intervals. At 6, 12, and 18 months, for the two lower groups (10 and 33 ppm),
this histopathologic examination was performed only when the tissue had gross
lesions. At the 24-month necropsy interval, the histopathologic examination
was performed on all tissues of rats in the 100 ppm group and both control
groups, and on potential target tissues, selected tissues, and tissues with
gross lesions in the two lower-dose groups (10 and 33 ppm).
During the 15th exposure month, all rats became infected with sialodacryo-
adenitis (SDA) virus infection. Clinical signs of infection were noted during
the 62nd and 63rd exposure weeks. After the 64th exposure week, the exposures
were temporarily terminated to permit recovery from the viral infection. Very
low mortality had been observed prior to the infection of the initial 120 rats
per sex per exposure group; no more than five in any group of one sex had died
or were sacrificed because of a moribund condition. During the 64th and 65th
exposure weeks, a total of 24 rats died. There was a higher rate of mortality
among female rats in the 100 ppm exposure group than in any other group. Gross
and microscopic examination of tissues of the animals that died during this
infection period revealed no pathologic findings sufficient to explain the
cause of death. Most of the clinical signs associated with the infection
subsided after 2 weeks of no exposure, as the mortality rate and body weights
returned to preinfection values. As a result, the exposure was restarted. No
increase in mortality in association with this disease had been reported in the
literature.
According to Snellings et al. (1981), the total numbers of rats that died
or were sacrificed in a moribund condition were 49, 39, 28, 31, and 29 for the
males and 53, 31, 25, 19, and 20 for the females in the 100 ppm, 33 ppm, 10
ppm, Air Control 1, and Air Control II groups, respectively. One additional
male in the 33 ppm group and one female in Air Control Group I were acciden-
9-99
-------
tally killed.
The cumulative mortality data and statistical significances for male and
female rats are shown in Tables 9-19 and 9-20, respectively. The cumulative
percentage dying in the 100 ppm group for both sexes was significantly higher
than that of controls for at least the last four exposure months of the study.
Very few significant differences were observed in males of the 33 ppm group.
During the 15th exposure month, the mortality rate of females in the 100
ppm group increased significantly. This increase was also noted for males in
the 100 ppm group and females in the 33 ppm group, but to a lesser degree.
Since the SDA virus may have contributed significantly to this mortality, the
data were re-evaluated by Snellings et al. (1981), using the number of rats
alive at the beginning of month 17 as the starting point. This re-evaluation
eliminated the immediate effects of the SDA virus infection. The results of
these calculations, presented in Tables 9-21 and 9-22, indicate a significant
increase in mortality in the 100 ppm group versus the controls for both males
and females, but the increased mortality was not significant until month 23 for
the males and month 22 for the females. In no time interval was the cumulative
percentage mortality value for either sex in the 33 ppm group significantly
different from that of combined controls. However, from the 21st month on, the
values for both sexes in the 33 ppm group were higher than those for both
control groups. At no time were significant increases in mortality observed in
the 10 ppm exposure group of either sex.
Of the many tumor types occurring in the Snellings et al. (1981) study,
six types, which may be treatment related, are reviewed here: subcutaneous
fibroma, peritoneal mesothelioma, pancreatic adenoma, pituitary adenoma, brain
neoplasm, and mononuclear cell leukemia. The authors presented no evidence
that the SDA viral infection increased the tumor incidence in the experimental
9-100
-------
TABLE 9-19. CUMULATIVE PERCENTAGES OF MALE FISCHER 344 RATS THAT DIED OR
WERE SACRIFICED IN A MORIBUND CONDITION AFTER EXPOSURE TO ETHYLENE OXIDE VAPOR3
Exposure
month
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
24.5
25.0
100 ppmb
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
1.0
3.0
7.0
7.0
7.0
10.4
11.7
18.2
24.7(~,~ja)
27.3(~>~>~)
44.2(a»c»c)
50.7(a,c,c)
55.9(3, b,c)
65'2(a'"'b)
Exposure concentration
Air
33 ppmb 10 ppm Control I
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.9
1.8
2.8
2.8
4.8
4.8
6.8
8.8
9.8
9.8
12.5
15.1
20.3
29.4(~>a>a)
36.o(~»b»b)
39>9(-,a,-)
42.5
54.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.9
0.9
0.9
0.9
0.9
0.9
0.9
1.9
2.9
2.9
2.9
8.0
10.6
14.4
18.3
25.9
31.0
38.3
0.0
0.0
0.0
0.0
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
1.8
1.8
1.8
2.9
5.1
5.1
9.0
11.5
17.9
21.8
12.9
34.6
41.9
Air
Control II
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.0
1.0
4.1
5.2
5.2
6.5
10.4
11.7
13.0
20.8
28.6
42.6
Combined
controls
0.0
0.0
0.0
0.0
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.9
1.9
1.9
3.5
5.2
5.2
7.8
11.0
14.8
17.4
25.2
31.6
42.3
aLife table analysis, adjusted for scheduled interim sacrifices,
Superscripts in parentheses denote values significantly higher than those of control
groups. First letter denotes degree of significance vs. Control I group; second
letter denotes degree of significance vs. Control II group; third letter denotes
degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001 - = not significant
SOURCE: Adapted from Snellings et al., 1981.
9-101
-------
TABLE 9-20. CUMULATIVE PERCENTAGES OF FEMALE FISCHER 344 RATS THAT DIED OR
WERE SACRIFICED IN A MORIBUND CONDITION AFTER EXPOSURE TO ETHYLENE OXIDE VAPORa
Exposure
month 100 ppmb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
24.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.8
1.8
1.8
2.8
3.9
16.o(b»b»b)
18.o(b»b»b)
21.l(c,c,c)
22.2(b»c>c)
25.o(a»c»c)
30.4(b>a»b)
34.4(b»a»b)
41.3(c,b,c)
49.5(c>b>c)
63.3(c»c»c)
70.0(c,c,c)
33 ppm
0.0
0.0
0.0
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
1.8
1.8
5.9
5.9
6.9
10.3
15.5
16.8
22.0
24.6
32.4
35.2
41.1
Exposure concentration
Air Air Combined
10 ppm Control I Control II controls
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.0
3.0
5.0
6.2
11.3
11.3
12.6
13.9
24.2
28.5
34.7
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
2.8
3.8
3.8
6.1
8.6
9.9
9.9
9.9
18.8
22.9
25.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.1
3.1
3.1
4.3
5.6
12.3
16.3
18.9
22.9
25.8
25.8
0.4
0.0
0.0
0.0
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
3.0
3.5
3.5
3.5
5.2
7.1
11.0
13.0
14.3
20.8
24.3
25.9
aLife table analysis, adjusted for scheduled interim sacrifices.
bSuperscripts in parentheses denote values significantly higher than those of control
groups. First letter denotes degree of significance vs. Control I group; second
letter denotes degree of significance vs. Control II group; third letter denotes
degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001
SOURCE: Adapted from Snellings et al., 1981.
not significant
9-102
-------
TABLE 9-21. CUMULATIVE PERCENTAGES OF MALE FISCHER 344 RATS THAT WERE ALIVE AT THE
BEGINNING OF MONTH 17, BUT DIED OR WERE SACRIFICED IN A MORIBUND
CONDITION AFTER SUBSEQUENT EXPOSURE TO ETHYLENE OXIDE VAPOR3
Exposure concentration
Exposure Air
month 100 ppmb 33 ppmb 10 ppm Control I
17
18
19
20
21
22
23
24
24.5
25.0
0.0 1.1
3.7 1.1
5.0 4.0
12.0 6.8
19.0 12.6
21.8 22.6
40.o(a,c,0 29.8(->a>-)
46.9<-,c,b) 34.1
52.5(~>b»b) 36.9
62.5(~>~>a) 49.8
1.0
1.0
1.0
6.2
8.8
12.8
16.7
24.5
29.7
37.1
1.0
3.3
3.3
7.2
9.9
16.4
20.3
28.2
33.4
40.8
Air
Control II
2.1
3.2
3.2
4.6
8.6
9.9
11.2
19.2
27.1
41.4
Combined
controls
1.6
3.3
3.3
5.9
9.2
13.2
15.8
23.7
30.3
41.2
aLife table analysis, adjusted for scheduled interim sacrifices.
Superscripts in parentheses denote values significantly higher than those of control
groups. First letter denotes degree of significance vs. Control I group; second
letter denotes degree of significance vs. Control II group; third letter denotes
degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001 - = not significant
SOURCE: Adapted from Snellings et al., 1981.
9-103
-------
TABLE 9-22. CUMULATIVE PERCENTAGES OF FEMALE FISCHER 344 RATS THAT WERE ALIVE AT
THE BEGINNING OF MONTH 17, BUT DIED OR WERE SACRIFICED IN A MORIBUND
CONDITION AFTER SUBSEQUENT EXPOSURE TO ETHYLENE OXIDE VAPOR3
Exposure
month 100 ppmb
17
18
19
20
21
22
23
24
24.5
3.7
5.1
8.4
15.1
20.1
28.4(b,-,a)
38.4 p > 0.01
0.01 > p > 0.001 c = p < 0.001 - = not significant
cControl I group differed significantly from Control II group at the p < 0.05
level only for the 22-month mortality count.
SOURCE: Adapted from Snellings et al., 1981.
9-104
-------
groups. The time to first tumor for some neoplasms (but not for mononuclear
cell leukemias) was decreased in the high-dose group as compared to controls,
as shown in Table 9-23. Median time-to-tumor was not reduced.
Histopathologic examinations were performed on tissues of all the rats in
the 100 ppm group and both control groups. In the 33 and 10 ppm groups, only
those tissues that had gross lesions were examined. Therefore, some small
tumors in these two groups may have been missed, yielding an erroneously low
estimate of tumors.
In male rats sacrificed at 24 months, a statistically significant increase
in subcutaneous fibromas (10/28, 35.7%) was observed in the group exposed to
100 ppm ethylene oxide as compared with combined controls (3/91, 3.3%) (Table
9-24). An increased prevalence of these tumors was also observed in the 10 ppm
group (8/48, 17%); however, this increase was not significant. No increase in
subcutaneous fibromas was observed in the 33 ppm group. The authors concluded
that the increased prevalence of subcutaneous fibromas in the 100 ppm group
represented an effect of treatment. It should be noted, however, that histo-
logic examinations were performed only on skin sections that showed gross
lesions; therefore, many tumors too small for gross detection were probably
missed. When the incidences of this tumor type were added to those for animals
that died spontaneously or were euthanized when moribund, the totals were even
higher in both the 100 and 10 ppm groups than in the controls (Table 9-25).
An increase in the frequency of peritoneal mesothelioma was observed in
all of the male treatment groups sacrificed at 24 months (4/30 at 100 ppm,
4/39 at 33 ppm, 2/51 at 10 ppm vs. 1/48 for the Control I group and 2/84 for
the Control II group) (Table 9-24). Although the increase was riot significant
at any dose level, this enhanced prevalence in the 100 and 33 ppm groups is
considered a treatment-related effect. This tumor was also found in a large
9-105
-------
TABLE 9-23. SUMMARY OF SELECTED TUMOR INCIDENCE COMPARISONS FOR MALE AND FEMALE
FISCHER 344 RATS EXPOSED TO ETHYLENE OXIDE FOR 2 YEARS
Ethylene oxide
concentration
ppm
100
33C
10C
0-1
O-II
100
33C
10C
0-1
O-II
100
33c
10C
0-1
O-II
100
33c
10C
0-1
O-II
100
33C
10C
0-1
O-II
Time in months to:
Total number of rats
With tissues examined With tumora
Mononuclear cell leukemia - Males
119 26
81
79
116
118
25
21
20
18
Mononuclear cell leukemia - Females
113 28(c»b>c)
79
77
118
117
Peritoneal mesothelioma
119
91
89
114
116
Pituitary adenoma -
117
79
80
117
117
Pituitary adenoma -
117
90
90
119
116
24(c,c,c)
14
9
13
- Males
22(c,c,c)
y(a, a, a)
3
2
2
Males
27
16
27
28
22
Females
32
38
39
38
38
First
tumor
19
13
20
18
21
18
18
19
19
18
15
18
20
18
20
15
15
18
17
18
10
17
16
15
18
Median
tumorb
24
25
25
23
25
24
24
25
24
23
23
25
—
—
—
25
25
25
25
25
24
25
24
25
25
Superscripts in parentheses denote values significantly higher than those of
control groups. First letter denotes degree of significance vs. Control I
group; second letter denotes degree of significance vs. Control II group; third
letter denotes degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001 - = not significant
"Medians were not presented if the total number of a particular tumor was
three or less.
C0nly organs with gross lesions were histologically examined from this exposure
level at the 6-, 12-, and 18-month sacrifice intervals.
SOURCE: Adapted from Snellings et al., 1981.
9-106
-------
TABLE 9-24. ETHYLENE OXIDE 2-YEAR VAPOR INHALATION STUDY:
24-MONTH FINAL SACRIFICE FREQUENCY OF EXPOSURE-RELATED NEOPLASMS FOR
110- TO 116-WEEK-OLD FISCHER 344 RATS
ppm of Ethylene Oxide
Organs/Findings/Sex 100a 33a 10a Control I Control II
Total number
examined grossly
Male 30 39 51 48 49
Female 26 48 54 60 56
Pituitary
Adenomas
Male 12/29b 13/39 15/51 16/48 13/49
Pancreas0
Adenomas
Male 5/30 1/2 2/3 2/48 5/49
Subcutisd
Fibromas
Male 10/28(c>c>c) 1/34 8/48<-a»a»b) 1/44 2/47
Peritoneum
Mesotheliomas
Male 4/30 4/39 2/51 1/48 1/49
Spleen
Mononuclear
cell leukemias
Male 8/30 10/39 9/51 5/48 8/49
Female
14/48(b>b>b) ll/54(-.~»a) 5/60 6/55
Superscripts in parentheses denote values significantly higher than those of control
groups. First letter denotes degree of significance vs. Control I group; second
letter denotes degree of significance vs. Control II group; third letter denotes
degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001 - = not significant
bNumerator equals number of rats with specified finding. Denominator equals
number of rats for which specified tissues were examined.
"-Tissues from 33- and 10-ppm groups examined only if gross lesions were present.
Since tissues were not examined from all rats, data from the 33- and 10—ppm
groups were not statistically compared with data from other groups.
^Examined only if gross lesions were present (except flank region skin and
subcutis, which was routinely examined microscopically).
SOURCE: Adapted from Snellings et al., 1981.
9-107
-------
TABLE 9-25. ETHYLENE OXIDE 2-YEAR VAPOR INHALATION STUDY: FREQUENCY OF
EXPOSURE-RELATED NEOPLASMS AT 24-MONTH FINAL SACRIFICE AND IN
FISCHER 344 RATS DYING SPONTANEOUSLY OR EUTHANIZED WHEN MORIBUND3
Organs/Findings/Sex 100b
ppm of Ethylene Oxide
33b 10b Control I Control II
Pituitary
Adenomas
Male
Pancreas^
Adenomas
Male
Subcutis6
Fibromas
Male
Peritoneum
Mesotheliomas
Male
Spleen
Mononuclear
cell leukemias
Male
Female
24/79c
16/79
26/79
2/32
24/79
2/80
ll/80(b,-,a) 1/43
15/78(c,b,c) 3/75 10/77(b,a,b) !/76
21/80(c.c»c) 6/80(~»-»a) 3/80 1/80
25/80
23/80
21/80
14/80
20/80
9/80
19/78
5/80
3/78
2/80
18/80
13/76
aConcerning the animals that died spontaneously or were euthanized when moribund,
it was not specified whether tissues were examined microscopically only when
gross lesions were present, or if all tissues were reviewed in this way.
It is therefore assumed that all of the tissues from these animals were studied
histologically, whether or not gross lesions were observed. Not to have per-
formed such studies would have yielded erroneously low frequencies of exposure-
related neoplasms.
Superscripts in parentheses denote values significantly higher than those of control
groups. First letter denotes degree of significance vs. Control I group; second
letter denotes degree of significance vs. Control II group; third letter denotes
degree of significance vs. combined controls (C-I plus C-II).
a = 0.05 > p > 0.01 b = 0.01 > p > 0.001 c = p < 0.001 - = not significant
cNumerator equals number of rats with specified finding. Denominator equals
number of rats for which specified tissues were examined.
"^Tissues from 33- and 10-ppm groups were examined only if gross lesions were
present. Since tissues were not examined from all rats, data from the 33- and
10-ppm groups were not statistically compared with data from other groups.
eExamined only if gross lesions were present (except flank region skin and
subcutis, which was routinely examined microscopically).
SOURCE: Adapted from Snellings et al., 1981.
9-108
-------
number of treated animals that died spontaneously or were euthanized when
moribund. When the tumor incidence in this latter group was added to that for
animals sacrificed at 24 months, the numbers were much higher than controls and
were statistically significant for the high-dose group versus controls (21/80
at 100 ppm, 6/80 at 33 ppm, 3/80 at 10 ppm vs. 1/80 for the Control I group and
2/80 for the Control 11 group) (Table 9-25).
Pancreatic adenomas were statistically significant for the male high-dose
group sacrificed at 24 months and the animals that died spontaneously or were
euthanized when moribund (11/80 at 100 ppm, 1/43 at 33 ppm, 2/32 at 10 ppm vs.
2/80 in the Control I group and 5/80 in the Control II group) (Table 9-25).
Tissues from the 33 and 10 ppm groups were examined only if gross lesions were
present in the 24-month sacrifice group, which may explain the paucity of
tumors in these groups (Table 9-24). The denominator in Table 9-25, the number
of rats for which the specified tissue was examined, may be erroneously high
for the data combining the 24-month sacrifice with the animals that died
spontaneously or were euthanized when moribund.
While Tables 9-23 and 9-24 show no significant increase in the frequency
of pituitary adenomas in the groups of treated males, Table 9-23 shows some
indication of a decreased time-to-tumor. In males, the first pituitary adenomas
appeared at 15 months in the 100 and 33 ppm groups, and in the 17th or 18th
month in all other groups; in females, the corresponding times were 10 months
for the 100 ppm group versus at least 15 months for all other groups. The
time-to-tumor decreased significantly with increasing dose (p < 0.01 for males,
p < 0.0001 for females), suggesting that the normal incidence of pituitary
adenomas was accelerated by exposure to ethylene oxide.
An increased frequency of mononuclear cell leukemia was observed in the
ethylene oxide-treated animals at the 24-month sacrifice interval (Table 9-24).
9-109
-------
Statistical significance was observed in females in both the 100 and 33 ppm
groups versus combined controls (p < 0.01). The responses for the 24-month
sacrifice were 15/26 (58%), 14/48 (29%), and 11/115 (10%) for the 100, 33, and
10 ppm groups and combined controls, respectively. The frequencies for male
rats were not significantly increased in the treated versus the control groups.
In females, the results for animals dying spontaneously or euthanized when
moribund and for those sacrificed at 24 months remained statistically signifi-
cant for the two higher-dose groups versus combined controls. The frequencies
for females (Table 9-25) were 27/80 (34%), 24/80 (30%), 14/80 (18%), and 22/156
(14%) for the 100, 33, and 10 ppm groups and combined controls, respectively,
with statistically significant differences in the two higher-dose groups versus
combined controls (p < 0.01) and a significantly positive linear dose-response
trend (p < 0.01). The trend became even stronger (p < 0.00001) when the pro-
portions were adjusted for early mortality. These data suggest that exposure
to ethylene oxide not only increased the total incidence of leukemia but also
accelerated its rate of development (Figure 9-3). The authors also reported
that the number of female rats with three or more tumors was significantly
(p < 0.001) increased in the 100 ppm group as compared to the controls.
A letter to the U.S. Environmental Protection Agency (Browning, 1982)
stated that a recent histologic examination of all brain tissue from the
Snellings et al. (1981) study revealed the presence of primary brain neoplasms
(see Tables 9-26 and 9-27, and Table 9-33 in Section 9.5.3.3.2). These tumors
were shown to be statistically significant by the Fisher Exact Test in both
males and females.
In summary, ethylene oxide has produced significant increases of several
tumor types in rats. A dose-related increase in mononuclear cell leukemia
occurred in female rats. The occurrence of pituitary adenoma appeared to be
9-110
-------
60
50
til 40
o
z
UJ
o
(E
MALE
I
I
I
20 40 60 80 100
CONCENTRATION OF ETHVLENE OXIDE, ppm
Figure 9-3. Percentages of male and female Fischer 344 rats with
histologically confirmed mononuclear cell leukemia at 24-month
sacrifice.
Source: Snellings et al. (1981).
9-111
-------
TABLE 9-26. ETHYLENE OXIDE 2-YEAR VAPOR INHALATION STUDY: FREQUENCY OF
PRIMARY BRAIN NEOPLASMS IN FISCHER 344 RATS
Exposure level (ppm)
Sex 100 33 10 0 (CI) 0 (CII)
18-month sacrifice8
Male 0/20 1/20 0/20 0/20 0/20
Female 1/20 0/20 0/20 1/20 0/20
24-month sacrifice5
Male 3/30 1/39 0/51 1/48 0/49
Female 2/26 2/48 0/51 0/60 0/56
Dead/euthanized moribund3
Male 4/49 3/39 1/28 0/30 0/29
Female 1/53 1/31 1/24 0/18 0/20
18- and 24-month sacrifices and dead/euthanized moribund5
(Combined from above)
Male 7/99 5/98 1/99 1/98 0/98
Female 4/99 3/99 1/95 1/98 0/96
2-year study*3
(Combined 6-, 12-, 18-, and 24-month sacrifices and dead/euthanized moribund animals)
Male 7/119 5/118 1/119 1/118 0/118
P=0.002C p=0.017c
Female 4/119 3/119 1/115 1/118 0/116
p=0.045c
^Numerator equals the number of brains with primary neoplasms. Denominator
equals total number of brains examined microscopically.
^Numerator equals the number of brains with neoplasms. Denominator equals total
number of brains examined microscopically. Although animals sacrificed at
6 and 12 months are included, no brain neoplasms were discovered in these
groups. The 6- and 12-month animals can be eliminated by subtracting 20 from
each denominator.
GFisher Exact Test.
SOURCE: Adapted from Snellings et al., 1981.
9-112
-------
TABLE 9-27. ETO 2-YEAR VAPOR INHALATION STUDY: FREQUENCY OF
PRIMARY BRAIN NEOPLASM TYPES IN FISCHER 344 RATS
(Combined data for 6-, 12-, 18-, and 24-month sacrifices, and
dead/euthanized moribund animals)
Neoplasm type
100
Exposure level (ppm)
33
10
0 (CI)
0 (CII)
Granular cell tumor
Astrocytoma/oligodendro-
glioma/mixed glioma
Malignant reticulosis-
microglioma
Males3
1/119 1/118 1/119 0/118 0/118
5/119 2/118 0/119 1/118 0/118
1/119 2/118 0/119 0/118 0/118
Granular cell tumor
Females3
1/119 1/119 0/118 1/118
0/116
Astrocytoma/oligodendro-
glioma/mixed glioma 2/119 2/119 1/118 0/118 0/116
Malignant reticulosis-
microglioma 1/119 0/119 0/118 0/118 0/116
aNumerator equals the number of brains with primary neoplasms. Denominator equals
total number of brains examined microscopically. Although animals sacrificed at
6 and 12 months are included, no brain neoplasms were discovered in these groups.
The 6- and 12-month animals can be eliminated by subtracting 20 from each
denominator.
SOURCE: Adapted from Snellings et al., 1981.
9-113
-------
accelerated in female rats exposed to 100 ppm, although there was no statisti-
cally increased incidence of these tumors. The frequency of peritoneal meso-
thelioma was treatment-related in the male rats exposed to 100 and 33 ppm.
Further, a significant increase occurred in subcutaneous fibromas in male rats.
Increases in brain neoplasms were also observed in both sexes.
9.5.1.2.2. Lynch et al. (1982) Inhalation Study (NIOSH) — Another chronic
inhalation study (unpublished draft) on ethylene oxide and propylene oxide was
performed by the National Institute for Occupational Safety and Health (NIOSH)
(Lynch et al., 1982). In the present report, only the preliminary findings of
the ethylene oxide section of the study will be discussed. Male Fischer 344
rats (80 in each group) and 12 male cynomolgus monkeys were exposed to ethylene
oxide at either 50 or 100 ppm for 7 hours/day, 5 days/week, for 24 months.
Each treatment group consisted of 80 rats and 12 monkeys at the start of the
study. Rats and monkeys were housed together in the same chambers during the
7-hour exposure period. Food and water were available ad libitum except during
the exposure periods. In analyzing for carcinogenicity, only limited data were
available for monkeys because of their longer lifespans; however, the authors
reported that there was no evidence of leukemia in any of the exposed monkeys.
An overall statistically significant depression in weight gain was noted
for ethylene oxide-exposed rats. This development, which appeared to begin at
about week 7 for the 100 ppm group and at week 15 for the 50 ppm group, con-
tinued throughout the study. Survival was also adversely affected by exposure
to ethylene oxide, with estimated mean survival times of greater than 720 days
for the controls, 690 days for the 50 ppm group, and 653 days for the 100 ppm
group. An outbreak of mycoplasma infection also caused an abrupt decline in
survival at about 480 days into the study.
9-114
-------
With respect to pathology, the authors reported that the livers and
spleens of the ethylene oxide-exposed rats were the only organs for which
histopathologic evaluations were completed. While the results are prelimi-
nary (Table 9-28), the data obtained at terminal sacrifice indicate that the
incidence of leukemia followed a dose-response pattern ranging from 33.3% in
controls to 64.3% in the 100 ppm group (p = 0.07, Table 9-28). The one-tailed
test for linear trend at terminal sacrifice was significant at the p < 0.05
level. Using a two-tailed test, the significance level was p = 0.08. These
preliminary data, therefore, do provide some evidence of ethylene oxide-induced
leukemia. The data from moribund sacrifice and deaths (Table 9-28) merely
accentuate both the early toxicity and the mortality in the 100 ppm group as
compared with the other groups, and the relatively high leukemia rates in these
rats. Neither these rats nor the total was significantly higher than controls.
TABLE 9-28. LEUKEMIA INCIDENCE IN MALE FISCHER 344
RATS EXPOSED TO ETHYLENE OXIDE FOR 2 YEARS3
Treatment group
Terminal
sacrifice
only (%)
Leukemia incidence
Moribund
sacrifice
and death (%)
Terminal
sacrifice
plus moribund
sacrifice and
death (total)
Control
Ethylene oxide, 50 ppm
Ethylene oxide, 100 ppm
7/21 (33.3%)
12/27 (44.4%)
9/14 (64.3%)b
5/18 (27.9%)
26/52 (50.0%)
21/62 (33.9%)
12/39 (30.8%)
38/79 (48.1%)
30/76 (39.5%)
aBased on histopathologic evaluation of spleens.
bp = 0.07 based on the one-tailed Fisher Exact Test.
SOURCE: Lynch et al., 1982.
9-115
-------
Lynch et al. (1982) also reported that exposure to ethylene oxide signifi-
cantly increased the incidence of peritoneal mesotheliomas. These tumors were
present on the tunica vaginalis surrounding the testes and epididymis, and
occasionally spread to the peritoneal cavity. A nonsignificant increase in
pheochromocytoraas was observed in exposed groups (Table 9-29).
Lynch et al. (1982) reported the following incidences of mixed-cell gliomas
in male rats: 0/76 in controls, 2/77 in the 50 ppm group, and 5/79 in the 100
ppm group. The term "glioma" was used because the tumors contained both astro-
cyte and oligodendroglia cells within the tumor. These findings are significant
because the above-described tumors are unusual in Fischer 344 rats. Additional
data collected from this study are currently being evaluated, and a final com-
prehensive report is scheduled to be published within a year.
9.5.1.2.3. Summary of Animal Studies — The Snellings et al. (1981)
study, which showed an increase in leukemia in Fischer 344 rats, is also
supported by a preliminary NIOSH study (Lynch et al., 1982, Table 9-28) in
which an increase in leukemia appeared in rats of the same strain but of a
different sex and with mycoplasma instead of SDA viral infections. Increases
in peritoneal mesotheliomas were observed in both studies (Snellings et al.,
1981 and Lynch et al., 1982), and significant increases in subcutaneous fibro-
mas in the males were observed in the Snellings study. Snellings et al. (1981)
also concluded that the frequencies among female rats with more than two neo-
plasms were significantly greater for all three groups when compared to com-
bined controls.
Further, both studies found significant increases in brain neoplasms, a
development that requires further review in terms of its possible value for
risk evaluation. Like the finding of gliomas in male rats reported previously,
9-116
-------
these studies are significant because brain neoplasms are unusual in the Fischer
344 strain of rats.
In 1980, the National Toxicology Program (NTP) began a cancer bioassay in
B6C3F1 mice (inhalation exposure). Exposure to ethylene oxide at 0, 5, and 100
ppm for 6 hours/day, 5 days/week began in August 1981. The final report is
expected in mid-1985.
TABLE 9-29. INCIDENCE OF NEOPLASTIC LESIONS IN MALE FISCHER 344
RATS EXPOSED TO ETHYLENE OXIDE FOR 2 YEARS3
Organs/Findings
Exposure level (ppm)
Control
50
100
Adrenal
Pheochromocytomas
Brain
Gliomas (mixed-cell)
8/78
0/76
14/77
2/77
13/78
5/79
Body cavity
Peritoneal mesotheliomas
Spleen
Mononuclear cell leukemia
3/78
24/77
(p = 0.032)b
9/79 21/79
(p = 4.95 x 10~5)b
38/79
(p = 0.22)b
30/76
aEach group consisted of 80 male rats. Denominators of less than 80 reflect
tissues accidentally lost or tissues that could not be examined histologically
due to autolysis.
bFisher Exact Test.
SOURCE: Lynch et al. , 1982.
9-117
-------
9.5.2. Epidemiologic Studies
9.5.2.1. JOYNER (1964) -- Joyner (1964) conducted a health evaluation of
employees at an ethylene oxide plant in Texas. The evaluation included a phy-
sical examination of 37 male ethylene oxide operators, aged 29 to 56, and a
similar number of age-matched controls. The operators were reported to have
been exposed to ethylene oxide at approximately 5 to 10 ppm for the durations
of their service. The controls, who were chosen from operators assigned to
other production units, had been exposed to many different agents encountered
in the petrochemical industry. The author stated that the mean length of
service for the control group was 11 2/3 years, as compared with 10 2/3 years
for the exposed group. The author used company medical records for the period
1952-1963 to compare the exposed group and controls with respect to days lost
for illness, specific diagnoses, and initial visits for respiratory, gastro-
intestinal, or genitourinary complaints. The author found that the ethylene
oxide operators who were currently employed exhibited less absenteeism, fewer
symptoms, and fewer diagnosed illnesses (including malignant neoplasms) than
the controls.
The author also reviewed the medical records of nine operators who had
experienced accidental exposures in the previous 10 years, and seven workers
other than operators who had experienced accidental exposures in the previous
8 years. Twelve of the accidental exposures were reported to be dermal expo-
sures, while three were reported to be inhalation exposures; one exposure was
reportedly to "vapor." Most of the dermal exposures produced burns. The vapor
exposure produced conjunctivitis. Two of the persons with inhalation exposure
suffered no symptoms; the third developed nausea and vomiting, which lasted
several hours. The authors reported that the persons identified as having had
9-118
-------
accidental exposures did not exhibit any recurring medical problems. The one
person who had suffered symptoms from the inhalation exposure was no longer
with the company and could not be traced.
Additionally, the author reviewed the medical records of eight persons who
had previously worked as ethylene oxide operators for 100 months or more but
who had since been transferred to another division. Among persons formerly
employed as ethylene oxide operators for 100 months or more, no significant
differences were found in the incidence of illness, symptoms, complaints, or
absenteeism when compared to the study cohort or to controls; very few data
were presented in this regard, however.
This study is inadequate for use in evaluating the carcinogenicity of
ethylene oxide for several reasons. First, it is primarily a cross-sectional
study of ethylene oxide operators who were employed as such at the time of the
study. Workers who had developed cancer would probably no longer have been
employed at the plant. Secondly, the period of observation, which in this
study is the same as the duration of exposure for the current operators, may
have been too short to allow adequate assessment of a carcinogenic effect.
Cancer latency may be as long as 20 to 30 years; the longest observation period
among current operators in this study was 16 1/3 years. The mean exposure for
current operators was 10 2/3 years. For those with accidental exposures, the
longest follow-up was 10 years. For the eight workers with over 100 months (8
1/3 years) of exposure, the length of follow-up was not indicated. Third, the
sample sizes studied were so small that only an extremely large carcinogenic
effect could be detected.
9.5.2.2. EHRENBERG AND HALLSTROM (1967) — Ehrenberg and Hallstrom (1967)
conducted a hematologic investigation of workers at a factory that manufactured
9-119
-------
and used ethylene oxide. A preliminary investigation in 1960 revealed certain
hematologic differences between 28 exposed persons who worked in an area of
the factory "where leakage of ethylene oxide from tube joints, pumps, etc. was
possible (and at least occasionally occurred)," and 26 controls in other
departments not working in contact with ethylene oxide. The sex of the study
subjects was not reported. The ages of persons in the exposed group were
reported to be about the same as those in the control group. The exposed
persons were reported to have been active in the ethylene oxide department for
2 to 20 years, with an average of 15 years. One case of leukemia (chronic
lymphatic type) was observed in the exposed group; the expected number of
leukemia cases in the exposed group was not reported. No cases of leukemia
were found in the controls. Three cases of anisocytosis were found in the
exposed group and none in the controls, a finding which the authors suggested
may indicate a disturbed bone marrow function. Hemoglobin values were reported
to be significantly (p < 0.05) lower in the exposed group than in the controls,
and lymphocytes per mm^ were reported to be significantly (p < 0.01) higher in
the 27 exposed healthy persons than in the 20 healthy controls (the presence of
disease may affect the white blood cell count; thus, only "healthy" persons
were considered in the latter comparison). It should be noted that three
persons who were reported to have been accidentally exposed to high levels of
ethylene oxide were added to the exposed group for the lymphocyte/mm^ compari-
son (for a total of 31 persons in the exposed group). The authors did not
state where these three persons worked or even whether they worked in the
factory.
Because of these differences relating to hemoglobin and lymphocytes, and
because ventilation was improved in the plant, the authors did a second study
of the factory workers in 1961. The second study was expanded to include all
9-120
-------
of the workers in the plant. Workers were divided into four categories: "66
persons not working with ethylene oxide (including the 1960 control group);
86 persons intermittently working in ethylene oxide premises; 54 persons who
had once been working in contact with ethylene oxide for some period of time;
and 37 persons permanently working in the ethylene oxide area (including the
1960 exposed group)." The only hematologic analysis in the second study was
for lymphocytes. The authors found an elevated lymphocyte count in the ex-
posed group as compared with controls, but this difference was not significant
(p > 0.05) for either healthy individuals or the total group. The authors
suggested that this lack of a significant difference could possibly be attri-
buted to improved ventilation and safety control in the factory, the small
number (17) of healthy persons in the group permanently exposed (versus 27
healthy exposed individuals in the 1960 investigation), and/or the average age
difference between the exposed and control groups. The average age of the
enlarged control group was reported to be "significantly" lower than that of
the exposed group, and, in general, a decrease in lymphocyte count with age was
found. A significant age difference between the exposed group and the controls
was not present in the 1960 examination. It should be noted that for those
persons examined in the 1960 investigation, a significant difference in average
lymphocyte count between the exposed group and the controls occurred again when
the two groups were examined in 1961.
The authors also compared the number of chromosome aberrations in eight
persons accidentally exposed to ethylene oxide with that in a control group of
10 persons, and found that chromosome aberrations were significantly elevated
in the exposed group. Details of the statistical analysis were not given.
In conclusion, Ehrenberg and Hallstrom (1967) found one leukemia case
among 28 workers exposed to ethylene oxide. The authors indicated that the
9-121
-------
probability of such an occurrence was small, but its statistical significance
was not calculated. The results of the study also suggest that ethylene oxide
may elevate lymphocyte counts and reduce hemoglobin values.
9.5.2.3. HOGSTEDT ET AL. (1979a, 1984) — A follow-up study for the years
1961 through 1977 of these same workers with regard to mortality and cancer
incidence was done by Hogstedt et al. (1979a). A subsequent study by Hogstedt
et al. (1984) followed the mortality of the cohort for the years 1978 through
1982 and the incidence for the years 1978 through 1981. A period of at least
10 years of follow-up from date of first employment was required in order for a
member of the cohort to be considered at risk. The authors reported that the
workers in this factory were exposed to various chemicals. During the period
from 1941 to 1947, it was estimated that the air concentrations were 5 mg/m-^
ethylene chlorohydrin, 100 mg/m^ ethylene dichloride, 0.05 mg/m^ bis(2-chloro-
ethyl) ether, and 600 mg/rn^ ethylene. The authors also cited the possibility
that concentrations up to 1,199 times greater than those reported may have
occurred for short periods of time. For ethylene oxide, the exposure was
reported to be probably < 25 mg/m^, although there were occasional exposures to
the chemical at 1300 mg/m^ (odor threshold). During the 1950s and until 1963,
the authors reported that the average air concentration of ethylene oxide in
the factory was probably 10 to 50 mg/m , although peaks above the odor thresh-
old still occurred. Random samples in the 1970s showed a range of 1 to 10
mg/m^ for ethylene oxide and 10 to 25 mg/m^ for propylene oxide, with the latter
concentrations occasionally being as high as 120 to 150 mg/m^.
Hogstedt et al. (1979a) reported that the study included three subcohorts:
66 men who had never taken part in work involving exposure to ethylene oxide,
86 "intermittently" exposed men (maintenance workers), and 89 men whose work
9-122
-------
involved full-time exposure. The same number of men were reported by Hog-
stedt et al. (1984) for the non-exposed and full-time exposed groups, but they
reported only 79 men in the "intermittently" exposed group. No explanation
for this difference was provided by the authors.
By the end of the second follow-up of these workers (Hogstedt et al.,
1984), a total of 12 cancer deaths had been observed in the full-time exposed
group, while only 4.8 were expected. The Carcinogen Assessment Group (GAG)
calculated that the probability of this occurring was less than 0.01. There
were no statistically significant differences between the observed and expec-
ted number of cancer deaths in the other two exposure groups. Seven of the 12
cancer deaths seen in the fulltime exposed cohort were either from cancer of
the stomach (5 deaths) or from leukemia (2 deaths). Deaths from both causes
were significantly elevated in comparison with the numbers expected (5 observed
versus 0.6 expected for stomach cancer, p < 0.01, as calculated by the GAG; and
2 observed versus 0.15 expected for leukemia deaths, p < 0.05, as calculated by
the GAG). One of the leukemia deaths was from chronic lymphatic leukemia, and
the other was from acute myeloid leukemia. The death from chronic lymphatic
leukemia may well have been the same case that was reported in the Ehrenberg
and Hallstrom (1967) study. Excess mortality for cerebrovascular disease
during the period 1961 through 1982 was also statistically significant (5
observed, 1.5 expected; p < 0.05, as calculated by the GAG) among the full-
time operators. Although the maintenance group showed no overall excess can-
cer mortality, the cancer deaths that occurred in this group were restricted
to cancers of the esophagus and stomach and to leukemia. The leukemia death
was from chronic lymphatic leukemia.
Hogstedt et al. (1984) examined the observed and expected numbers of
deaths from various causes, including different cancer sites, by 1-4, 5-9,
9-123
-------
and 10+ years of exposure. No response by length-of-exposure was found for
any of the causes of death, but the data were rather limited for this type of
analysis.
Cases of cancer in the study group were determined using the Swedish
Cancer Registry. No indication of the completeness of ascertainment of the
Registry was given. Seventeen cases versus 7.9 expected (p < 0.01) were
identified among full-time exposed workers for the period 1961 through 1981.
These included three cases of leukemia versus 0.24 expected (p < 0.01, as
calculated by the GAG). Included in the three leukemia cases were the indivi-
dual with acute myeloid leukemia and the individual with chronic lymphatic
leukemia, both of whom had died, as well as an individual with chronic myeloid
leukemia. One case of stomach cancer was reported in addition to the five
cases in which individuals had died. The expected number of stomach cancers
was not indicated.
In summary, deaths from cancer of all sites, deaths from stomach cancer,
and deaths from leukemia were each significantly (p < 0.05) elevated among the
full-time exposed cohort. The total number of malignancies and the number
of leukemia cases were also significantly (p < 0.01) elevated in this group.
Workers in the full-time exposed cohort were exposed to several chemical
agents, however, and the excess cancer incidence and mortality in this cohort
cannot necessarily be ascribed to the ethylene oxide exposure.
9.5.2.4. HOGSTEDT ET AL. (1979b, 1984) — Hogstedt et al. (1979b, 1984)
reported on the morbidity and mortality of a group of sterilizing operators
exposed to 50% ethylene oxide and 50% methyl formate. The 1979b report indi-
cated that only seven persons worked with the sterilization process. However,
treated boxes (supposedly containing the sterilized equipment) were stored in a
9-124
-------
hall where 30 women worked, and because of leakage from the boxes, the average
exposure in the storage hall was reportedly higher than in the sterilization
room. Exposure measurements made in 1977 showed storage hall concentrations of
2 to 70 ppm with 8-hour time-weighted average concentrations being calculated
at 20± 10 ppm. The concentration was 1,500 ppm inside newly sterilized boxes
and 150 ppm on the floor outside the boxes.
During the period from 1968 to 1977, 70 persons had been employed at some
time in the storage hall, and another 160 had been employed in the neighboring
rooms or as sterilizing operators. Of these, 69 (63 women, 6 men) and 134
(90 women, 44 men) had worked for a year or more, respectively (Hogstedt et
al., 1984). In the Hogstedt et al. (1979b) report, the authors indicated that
among this group of workers three cases of leukemia had occurred. One of these
reported cases was actually a Waldenstrom's macroglobulinemia, however, and
it was subsequently reported (Hogstedt et al., 1984) that this case should have
been classified as a non-Hodgkin's lymphoma. However, according to the Eighth
Revision of the International Classification of Diseases (ICD), which is the
revision used in the Hogstedt et al. (1984) report, Waldenstrom's macro-
globulinemia is classified as a plasma protein abnormality and not as a neoplasm.
The two leukemia cases were among women who worked in the storage hall.
One of the cases was a woman who began working in the storage hall in 1966,
was diagnosed with chronic myeloid leukemia in early 1972 at the age of 51,
and died in 1977. The other case was a woman who began working in the storage
hall in 1968, was diagnosed with acute myelogenetic leukemia in early 1977
at the age of 37, and as of July 1978 was reported to be in complete remis-
sion. Hogstedt et al. (1984), however, reported that she subsequently died
during the extended follow-up period (1978-1982). A third leukemia death was
reported by Hogstedt et al. (1984). This case was a woman who had had inter-
9-125
-------
mittent exposure, passing through the storage hall 2 to 4 times per day. She
was employed during 1969-1972, and in 1979 was diagnosed with blast leukemia
and died the same year.
Two males were reported to have died from cancer in the period 1968-1982.
One was the case of Waldenstrom's macroglobulinemia mentioned earlier.
As previously indicated, this death would not be considered a cancer death by
the Eighth Revision of the ICD. The other death from cancer was not specified
by site. Observed and expected cancer mortality, total mortality, and mortal-
ity from cancer of the lymphatic and hematopoietic systems for both the earlier
(Hogstedt et al., 1979b) and later (Hogstedt et al., 1984) observation periods
as well as the total observation period is reported in Table 9-30.
With regard to morbidity, ten cases of cancer among the female sterilizer
workers had been reported to the Cancer Registry versus 5.2 expected (p < 0.05,
as calculated by the GAG) during the period 1961-1981; there were three cases
of cancer among male sterilizer workers (excluding the case of Waldenstrom's
macroglobulinemia) versus 1.8 expected. The excess morbidity among females was
mainly due to the three cases of leukemia (0.1 expected, p < 0.01, as calcula-
ted by the GAG) and two cases of malignant cancer of the cervix (0.4 expected).
The cases among men were due to tumors of the stomach, colon, and rectum.
Hogstedt et al. (1979b) suggested that the combination of ethylene oxide
and methyl formate may produce a special carcinogenic risk, since methyl for-
mate, the authors indicated, exhibits its antibacterial effect by affecting DNA
structure. No literature reference was cited by the authors as to this point,
however. A literature search conducted for the Carcinogen Assessment Group by
the Environmental Mutagen Information Center at the Oak Ridge National Labora-
tory (Francis, 1985) failed to find any literature citations for mutagenicity
studies of methyl formate.
9-126
-------
TABLE 9-30. OBSERVED AND EXPECTED NUMBER OF DECEASED
AMONG 153 WOMEN AND 50 MEN WITH CONTINUOUS
OR INTERMITTENT EXPOSURE TO ETHYLENE OXIDE3
ICDb
Causes
of death
Women
1968-77 1978-82
Obs Exp Obs Exp
Men
1968-77 1978-82
Obs Exp Obs Exp
Women + Men
1968-82
Obs Exp
1- Total 2 2.9 3 2.7 4 2.6 2 2.2 11 10.4
999
140- All tumors 2 1.2 3 1.1 0 0.6 1 0.5 6 3.4
209
200- Lymphatic 1 0.1 2 0.08C 0 0.06 0 0.05 3 0.3C
207 and hemato-
poietic tissue
aThe observed case of Waldenstrom's macroglobulinemia has been deleted from
this table. It had appeared in the table by Hogstedt et al. (1984).
^International Classification of Diseases, Eighth Revision.
cp < 0.01, calculated by the GAG.
SOURCE: Adapted from Hogstedt et al. , 1984.
9-127
-------
9.5.2.5. MORGAN ET AL. (1981) — Morgan et al. (1981) conducted a retro-
spective study of 767 workers potentially exposed to ethylene oxide who had
worked for at least 5 years at a Texaco Chemical Company plant in Port Neches,
Texas, between January 1955 and December 31, 1977. The authors provided no
analysis of the cohort with respect to length of follow-up. An industrial sur-
vey of the plant (performed in July 1977) showed that the 8-hour time-weighted
average exposure to ethylene oxide was "well below" 50 ppm, except in the area
around the tank car loading operations, where readings were as high as 6,000
ppm. Among the 767 male workers potentially exposed to ethylene oxide in the
study cohort, there were 11 deaths from malignant neoplasms, where 15.24 would
have been expected on the basis of U.S. vital statistics.
There were more deaths than expected from pancreatic cancer (SMR* = 377,
3 observed versus 0.8 expected), bladder cancer (SMR = 322, 1 observed versus
0.31 expected), brain and central nervous system cancer (SMR = 285, 2 observed
versus 0.7 expected), and Hodgkin's disease (SMR = 570, 2 observed versus 0.35
expected). Although the 95% lower confidence limits for these SMRs were all
less than 100, the number of deaths from pancreatic cancer and the number of
deaths from Hodgkin's disease are each significantly (p < 0.05) more than
expected by hypothesis testing using the Poisson test. Excess mortality from
leukemia was not found. Because their study cohort was small and because
excess cases of leukemia following exposure to ethylene oxide were found in the
studies by Hogstedt et al. (1979a, b), the authors calculated the magnitude of
the relative risk of mortality from leukemia, given the sample size of the
cohort, that could be detected at the 95% confidence level with a power of 80%.
This relative risk was calculated to be 10.5 (SMR of 1050). In conclusion, it
*Standardized mortality ratio.
9-128
-------
should be stated that the observed mortalities from pancreatic cancer and from
Hodgkin's disease were each significantly elevated among the study cohort, and
that the study cohort may have been too small for an adequate evaluation of the
risk of mortality from leukemia or other cancer types. Furthermore, there was
no indication by the authors that sufficient allowance had been made for a
cancer latency period.
9.5.2.6. THIESS ET AL. (1982) — Thiess et al. (1982) conducted a cohort
mortality study of 602 persons who had been employed for 6 months or longer in
the alkylene oxide (ethylene oxide/propylene oxide) production or processing
areas of nine BASF Aktiengesellschaft, Ludwigshafen plants in West Germany
during the period from 1928 to 1980. Vital status was ascertained for 523 of
the 536 German employees in the cohort, while that of only 30 of the 66 non-
German employees could be determined. Thus, the percentage of overall follow-
up in this study was 92% (553 of 602). In addition to alkylene oxides, the
workers were reported to have been exposed to a variety of other compounds.
The expected mortality for the total cohort and for those within the co-
hort who were observed for a minimum of 10-years was calculated using mortality
data for Ludwigshafen, Rhinehessia-Palatinate, and the Federal Republic of
Germany. The observed and expected numbers of cancer deaths for those persons
observed for at least 10 years are reported in Table 9-31. The observed number
of deaths from cancer of any site was not significantly (p < 0.05) higher than
that expected based on mortality data for Ludwigshafen, Rhinehessia-Palatinate,
or the Federal Republic of Germany. Deaths from cancer of the brain among
alkylene oxide workers who were followed for at least 10 years did approach
statistical significance (p < 0.07), however, in comparison with those expected
based on Ludwigshafen or Rhinehessia-Palatinate mortality data.
9-129
-------
TABLE 9-31. COMPARISON OF OBSERVED NUMBERS OF CANCER DEATHS IN BASF-AKTIENGESELLSCHAFT, LUDWIGSHAFEN PLANTS
1928-80 FOR PERSONS HAVING 10 YEARS OF OBSERVATION FOLLOWING EXPOSURE TO ALKYLENE OXIDE WITH THAT
EXPECTED BASED ON MORTALITY STATISTICS FOR RHINEHESSIA-PALATINATE 1970-75, LUDWIGSHAFEN 1970-75,
AND THE FEDERAL REPUBLIC OF GERMANY 1971-74, BY ICD CODE AND CAUSE OF DEATH
UJ
o
ICD No.a
151
156
162
188
191
193-199
Total of
140-199c
205
230-239
Cause of death
Malignant tumor
of the stomach
Malignant tumor of
the gall bladder
Malignant tumor
of the bronchi!
Malignant tumor of
the urinary bladder
Malignant tumor
of the brain
Squamous cell
carcinoma of unknown
primary site
malignant tumors in ICD
Myeloid leukemia
Tumor of unknown
character
Rhinehessia-
Palatinate
1970-75
Observed
deaths No. P-value
2 1.852 0.552
1 0.201 0.182
4 3.769 0.520
1 0.469 0.374
1 0.071 0.068
1 0.743 0.525
!0 — d — d
1 0.148 0.138
1 0.454 0.365
Ludwigshaf en
1970-75
No.
1.765
0.243
3.956
0.532
0.066
1.047
__d
0.145
0.426
P-value
0.527
0.216
0.568
0.413
0.064
— c
__d
0.135
0.347
Federal
Republic of
Germany
1971-74
No. P-value
2.033 — b
C C
C C
C C
C C
C C
11.816 — d
0.756 0.531
c c
International Classification of Diseases, Eighth Revision.
bThe probability of observed deaths occurring by chance was not provided by the authors because the observed
deaths were fewer than expected.
cThe authors did not report the number of deaths that would be expected in the cohort based on Federal Republic
of Germany mortality rates for individual tumor sites other than stomach and myeloid leukemia.
dThe authors did not report the number of deaths from tumor sites, ICD 140-199, that would be expected based on
Rhinehessia-Palatinate or Ludwigshafen mortality data.
SOURCE: Adapted from Thiess et al. , 1982.
-------
The authors also compared the observed number of cancer deaths with that
expected, using an internal cohort of 1,662 styrene workers. The minimum obser-
vation period of 10 years required for the comparison in Table 9-31 was not
used for this analysis. Thus, in Table 9-32, there were 14 total observed
cancer deaths, as opposed to 12 observed deaths in Table 9-31. These results
are reported in Table 9-32. The relative risk of death from cancer of all
sites in the alkylene oxide cohort in comparison to what would be expected
based on cancer mortality in the styrene cohort was 1.48. Assuming that the
numbers of observed and expected deaths (14 and 9.44, respectively) are both
Poisson variables, the difference between the two is not statistically signi-
ficant (p < 0.05). In the 65- to 74-year-old age group, the relative risk was
2.78. If it is assumed that both the observed and expected deaths are Poisson
variables, the difference between the two is statistically significant at p <
0.05. It should be noted that although the authors reported in tabular form
that 10 cancer deaths had occurred in the 65- to 74-year-old age group, the
text indicated that 11 had occurred—a difference that obviously would function
to lower the probability of the number of cancer deaths that was observed. A
major problem in evaluating this result, however, is that the workers in the
alkylene oxide cohort were exposed to a variety of chemicals in addition to
ethylene oxide, some of which are known or suspected carcinogens. The authors
did not compare the alkylene oxide and styrene cohorts with regard to the
number of deaths by individual tumor site.
The authors also analyzed the cancer deaths by length of exposure, and did
not find a dose-response. However, they gave no indication that the mortality
analysis by length of employment had been adjusted for length of follow-up.
In summary, this study is inconclusive as to whether persons exposed to
ethylene oxide are at an excess risk of death from cancer. There was a signi-
9-131
-------
TABLE 9-32. RELATIVE RISKS OF DEATH FROM CANCER IN THE ALKYLENE OXIDE COHORT
AS COMPARED WITH THE STYRENE COHORT, BY AGEa
Observed
Age group deaths
15-24
25-34
35-44
45-54
55-64 4
65-74 10
75-84
Total 14
Expected
deaths Relative risk
—
0.35
0.47
1.61
3.41 1.17
3.60 2.78
—
9.44 1.48
aln this analysis, a minimum observation period of 10 years was not made a
requirement.
SOURCE: Adapted from Thiess et al., 1982.
9-132
-------
ficant excess number of cancer deaths in the 65- to 74-year-old age group in
the alkylene oxide cohort, as compared to that expected based on the mortality
data for a group of styrene workers. A fact that may have confounded this re-
sult is that the alkylene oxide workers were exposed to a variety of chemicals
in addition to ethylene oxide, some of which are known or suspected carcino-
gens. Deaths from cancer of any particular site were not found to be signifi-
cantly (p < 0.05) in excess when the expected numbers of deaths for those sites
were derived using mortality data for Ludwigshafen or Rhinehessia-Palatinate.
Two problems with this study are the small sample size and the fact that only
a little more than half of the cohort was observed for 10 years or more. In
regard to leukemia mortality, for which Hogstedt et al. (1979a, 1984) had found
an association with ethylene oxide exposure, the authors found that for those
persons who had had more than 10 years of exposure, one case of myeloid leuke-
mia occurred where only about 0.15 would have been expected based on local mor-
tality data, but this difference was not statistically significant (p < 0.05).
9.5.2.7. SCHNORR (1982) — A proportionate mortality study by Schnorr
(1982) of decedents who had been members of District 1199 of the National
Hospital and Health Care Workers Union found that the proportionate mortality
ratio (PMR) for neoplasms of lymphatic and hematopoietic tissue (1CD code
200-209, Eighth Revision), as well as for other types of tumors, was signifi-
cantly elevated for certain job categories (e.g., "service" and "nursing")
that included job titles of personnel exposed to ethylene oxide (e.g., hospi-
tal central service employees, registered nurses, licensed practical nurses,
and nurse's aides). Such job categories were relatively broad in their inclu-
sion of job titles, however, and the results of the study with regard to a
possible association of cancer risk with ethylene oxide exposure must therefore
9-133
-------
be judged inconclusive.
9.5.2.8. STUDIES IN PROGRESS — Several cohort or case-control studies
testing the association of ethylene oxide exposure and the risk of cancer are
currently in progress or about to begin. A cohort mortality study of approxi-
mately 1,200 workers who were engaged in ethylene oxide production during
their work history in the chemical industry in the Kanawha Valley, West
Virginia, is currently being conducted by the National Institute for Occupa-
tional Safety and Health (NIOSH) and the Union Carbide Corporation. Nested
case-control studies within the cohort will be done for certain kinds of
deaths (e.g., leukemia). The results of the cohort study will not be avail-
able until early 1986. The results of the case-control study will not be
available until some time later (Rinsky, personal communication).
NIOSH and the Health Industry Manufacturing Association are currently
conducting a cohort mortality study of approximately 10,000 persons, consist-
ing primarily of medical equipment manufacturing personnel who use ethylene
oxide as a sterilant. Some exposure information on this cohort is available,
but only for 1978 onward. The results will not be available until at least
1987, and a published report is expected about a year later.
The U.S. Environmental Protection Agency funded a case-control study of
70 cases of cancer of the lymphatic and hematopoietic tissue and 140 controls
in District 1199 of the National Hospital and Health Care Workers Union to
determine if an association existed between such cancers and occupational
exposure to ethylene oxide and/or other substances. The study, conducted by
Dr. Jeanne Stellman of Columbia University, has been completed but has not
yet been published as of the date of this writing. No association between
the cases and exposure to ethylene oxide was reported to be found; however,
9-134
-------
some problems with regard to ascertainment of exposure did exist (Schnorr,
personal communication).
9.5.2.9. SUMMARY OF EPIDEMIOLOGIC STUDIES — Three epidemiologic studies
of persons occupationally exposed to ethylene oxide demonstrated a significant
association between ethylene oxide exposure and cancer incidence or mortality.
A study by Hogstedt et al. (1979a, 1984) found significantly (p < 0.05) in-
creased mortality for stomach cancer and leukemia and significantly (p < 0.01)
increased incidences of cancer of all sites and of leukemia among ethylene
oxide production workers. Hogstedt et al. (1979b, 1984) found significantly
(p < 0.05) increased incidences of leukemia and cancer of all sites and sig-
nificantly (p < 0.01) increased mortality from leukemia among workers exposed
to ethylene oxide used as a sterilant. The study by Morgan et al. (1981)
found increased mortality from pancreatic cancer and Hodgkin's disease that
is statistically significant (p < 0.05) by hypothesis testing.
Excess mortality from leukemia in the Hogstedt et al. (1979a, 1984)
study and excess incidences of leukemia in the Hogstedt et al. (1979b) study
were not limited to any particular types of leukemia. Excess deaths from
leukemia in the Hogstedt et al. (1979a) study included one case of acute
myeloid leukemia and two cases of chronic lymphatic leukemia. Excess cases
of leukemia in the Hogstedt et al. (1979b) study included one case of acute
myeloid leukemia, one case of chronic myeloid leukemia, and one case of
"blast" leukemia. The expected numbers of deaths or cases by type of leukemia
were not calculated in either study.
It should be noted that in all three of the above-referenced epidemio-
logic studies, exposure of the cohort to other chemicals besides ethylene
oxide was reported to have occurred or probably occurred. In the Hogstedt et
9-135
-------
al. (1979a) study, reports were made of exposure to several chemicals, of
which two, ethylene dichloride and bis(2-chloroethyl)ether, are recognized
carcinogens. In the Hogstedt et al. (1979b) study, ethylene oxide-exposed
workers experienced concurrent exposure to methyl formate. In the Morgan et
al. (1981) study, there was no mention of exposure to chemicals other than
ethylene oxide, but the fact that the study was conducted at a chemical plant
would suggest that exposure to other chemicals did occur.
9.5.3. Quantitative Estimation. This quantitative section deals with the
incremental unit risk for ethylene oxide in air and the potency of ethylene
oxide relative to other carcinogens that the GAG has evaluated. The incre-
mental unit risk estimate for an air pollutant is defined as the increased
life-time cancer risk occurring in a hypothetical population in which all in-
dividuals are exposed continuously from birth throughout their lifetimes to a
concentration of 1y g/m^ of the agent in the air they breathe. These cal-
culations are done to estimate in quantitative terms the impact of the agent
as a carcinogen. Incremental unit risk estimates are used for two purposes:
1) to compare the carcinogenic potencies of several agents with each other, and
2) to give a crude indication of the population risk that would be associated
with air or water exposure to these agents, if the actual exposures were known.
Hereinafter, the term "unit risk" will always refer to incremental unit risk.
In the sections that follow, the general assessment procedures used by
the GAG are discussed. These include animal-to-human extrapolation modeling,
data selection, calculation of human equivalent doses, extrapolation modeling
from human epidemiologic studies, and interpretation of the resulting esti-
mates. Following this discussion, the CAG's unit risk calculations and rela-
tive potency estimates are presented.
9-136
-------
9.5.3.1. PROCEDURES FOR THE DETERMINATION OF UNIT RISK FROM ANIMAL DATA
— In developing quantitative estimates of carcinogenic risks, one or both of
two types of data are utilized: 1) lifetime animal studies, and 2) human
studies where excess cancer risk has been associated with exposure to the
agent. In animal studies it is assumed, unless evidence exists to the con-
trary, that if a carcinogenic response occurs at the dose levels used in the
study, then responses will also occur at all lower doses, at incidences
determined by an extrapolation model.
There is, however, no solid scientific basis for any mathematical extra-
polation model that relates carcinogen exposure to cancer risks at the ex-
tremely low concentrations that must be dealt with in evaluating environ-
mental hazards. For practical reasons, such low levels of risk cannot be
measured directly either by animal experiments or by epidemiologic studies.
We must, therefore, depend on our current understanding of the mechanisms of
carcinogenesis for guidance as to which risk model to use. At the present
time, the dominant view of the carcinogenic process involves the concept that
most cancer-causing agents also cause irreversible damage to DNA. This
position is reflected by the fact that a very large proportion of agents that
cause cancer are also mutagenic. There is reason to expect that the quantal
type of biological response, which is characteristic of mutagenesis, is
associated with a linear nonthreshold dose-reponse relationship. Indeed,
there is substantial evidence from mutagenicity studies with both ionizing
radiation and a wide variety of chemicals that this type of dose-response
model is the appropriate one to use. This is particularly true at the lower
end of the dose-response curve; at higher doses, there can be an upward
curvature, probably reflecting the effects of multistage processes on the
mutagenic response. The linear nonthreshold dose-response relationship is
9-137
-------
also consistent with the relatively few epidemiologic studies of cancer
responses to specific agents that contain enough information to make the
evaluation possible (e.g., radiation-induced leukemia, breast and thyroid
cancer, skin cancer induced by arsenic in drinking water, liver cancer induced
by aflatoxins in the diet). There is also some evidence from animal experi-
ments that is consistent with the linear nonthreshold model (e.g., liver
tumors induced in mice by 2-acetylaminofluorene in the large-scale EDgi study
at the National Center for lexicological Research, and the initiation stage
of the two-stage carcinogenesis model in rat liver and mouse skin).
Based on the above evidence of low-dose linearity, and because very few
compounds exhibit low-dose responses that are superlinear, the linear non-
threshold model has been adopted as the primary basis for risk extrapolation
in the low-dose region of the dose-response relationship. The risk estimates
made with this model should be regarded as conservative, representing the
most plausible upper limit for the risk; i.e., the true risk is not likely to
be higher than the estimate, but it could be lower.
The mathematical formulation chosen to describe the linear nonthreshold
dose-response relationship at low doses is the linearized multistage model.
The multistage model employs enough arbitrary constants to be able to fit al-
most any monotonically increasing dose—response data, and it incorporates a
procedure for estimating the largest possible linear slope (in the 95% confi-
dence limit sense) at low extrapolated doses that is consistent with the data
at all dose levels of the experiment.
9.5.3.1.1. Description of the Low-Dose Animal Extrapolation Model — Let
P(d) represent the lifetime risk (probability) of cancer at dose d. The multi-
stage model has the form
9-138
-------
P(d) = 1 - exp [-(q0 + qLd + q2d2 + ... + qkdk)]
where
qt >_ 0, i = 0, 1, 2, ..., k
Equivalently,
Pt(d) = 1 - exp [(qLd + q2d2
where
Pt(d) = P(d) - P(0)
1 - P(0)
is the extra risk over background rate at dose d.
The point estimate of the coefficients q^, i = 0, 1,2, ..., k, and
consequently, the extra risk function, P(-(d), at any given dose d, is calcu-
lated by maximizing the likelihood function of the data.
The point estimate and the 95% upper confidence limit of the extra risk,
Pt(d), are calculated by using the computer program GLOBALB3, developed by
Howe (1983). At low doses, upper 95% confidence limits on the extra risk and
lower 95% confidence limits on the dose producing a given risk are determined
from a 95% upper confidence limit, q*, on parameter q^. Whenever q^ > 0, at
low doses the extra risk Pj-(d) has approximately the form Pj-(d) = qj x d.
Therefore, q^ x d is a 95% upper confidence limit on the extra risk, and
R/q^ is a 95% lower confidence limit on the dose producing an extra risk of
R. Let LQ be the maximum value of the log-likelihood function. The upper
& £
limit, q|, is calculated by increasing q^ to a value q^ such that when the
log-Likelihood is remaximized subject to this fixed value qi for the linear
coefficient, the resulting maximum value of the log-likelihood LI satisfies
9-139
-------
the equation
2 (L0 - L!> = 2.70554
where 2.70554 is the cumulative 90% point of the chi-square distribution with
one degree of freedom, which corresponds to a 95% upper limit (one-sided).
This approach of computing the upper confidence limit for the extra risk,
Pt(d), is an improvement on the Crump et al. (1977) model. The upper confi-
dence limit for the extra risk calculated at low doses is always linear. This
is conceptually consistent with the linear nonthreshold concept discussed
earlier. The slope, qn, is taken as an upper bound of the potency of the
chemical in inducing cancer at low doses. (In the section calculating the
risk estimates, Pfc(d) will be abbreviated as P.)
In fitting the dose-response model, the number of terms in the polynomial
is chosen equal to (h-1), where h is the number of dose groups in the experi-
ment, including the control group.
Whenever the multistage model does not fit the data sufficiently well,
data at the highest dose are deleted and the model is refit to the rest of
the data. This is continued until an acceptable fit to the data is obtained.
To determine whether or not a fit is acceptable, the chi-square statistic
x -
is calculated where N. is the number of animals in the i dose group, X. is
the number of animals in the i dose group with a tumor response, P^ is the
probability of a response in the i^ dose group estimated by fitting the
multistage model to the data, and h is the number of remaining groups. The
9-1^0
-------
fit is determined to be unacceptable whenever ^ ^ is larger than the cumu-
lative 99% point of the chi-square distribution with f degrees of freedom,
where f equals the number of dose groups minus the number of non-zero multi-
stage coefficients.
9.5.3.1.2. Selection of Data — For some chemicals, a number of studies
in different animal species, strains, and sexes, each run at varying doses
and routes of exposure, are available. In such cases, choices must be made
as to which of several data sets are appropriate for use with the chosen
model. The following are the procedures used by the GAG in evaluating these
data for the purpose of risk estimation:
1. The data on tumor incidence are separated according to organ sites
or tumor types. The dose and tumor incidence data set used in the model is
the set in which tumor incidence is statistically significantly higher than
in controls for at least one test dose level, and/or where the tumor inci-
dence rate shows a statistically significant trend with respect to dose level.
The data set that gives the highest estimate of the lifetime carcinogenic
risk, q^, is selected in most cases. However, efforts are made to exclude
data sets that produce spuriously high risk estimates because of a small
number of animals. That is, if two sets of data show a similar dose-
response relationship, and one has a very small sample size, the data set
having the larger sample size is selected for calculating carcinogenic
potency.
2. If there are two or more data sets of comparable size that are
identical with respect to species, strain, sex, and tumor sites, the geo-
metric mean of q^, estimated from each of these data sets, is used for risk
assessment.
9-1U1
-------
3. If two or more significant tumor sites are observed in the same
study, and if the data are available, the number of animals with at least
one of the specific tumor sites under consideration is used as incidence data
in the model.
9.5.3.1.3. Calculation of Human Equivalent Dosages from Animal Data — In
calculating human equivalent dosages, it is necessary to correct for differen-
ces in metabolism among species and for the variations in absorption factors
involved in different routes of administration.
Following the suggestion of Mantel and Schneiderman (1975), it is
assumed that mg/surface area/day is an equivalent dose between species.
Since, to a close approximation, the surface area is proportional to the
two-thirds power of the weight, as would be the case for a perfect sphere,
the exposure in mg/day per two-thirds power of the weight is also considered
to be equivalent exposure. In an animal experiment, this equivalent dose is
computed in the following manner:
Let
Le = duration of experiment
le = duration of exposure
m = average dose per day in mg during administration of the agent
(i.e., during le) and
W = average weight of the experimental animal
The lifetime average exposure is then
d -
L x W2/3
e
When exposure is given in terms of mg/kg/day = m/Wr = s, the conversion is
9-142
-------
simply
m = s x W1/3
2/3
rw
where r is the absorption rate for ethylene oxide (assumed to be 1).
When exposure is via inhalation, as with ethylene oxide, dose calcula-
tions at experimental exposures of up to 100 ppm are performed under the
assumption that the compound is a completely water-soluble gas absorbed
proportionally to the amount of air breathed in. While the GAG has previously
used an existing methodology to determine dose equivalency in such cases, for
ethylene oxide the total body dose resulting from exposure of male Fischer
344 rats to air concentrations of 100 ppm for 6 hours has been measured as
20.24 mg/kg (Tyler and McKelvey, 1980). At 10 ppm exposures under similar
conditions, the measured dose was 2.7 mg/kg. Since daily exposures in the
Snellings et al. (1981) study included 10 ppm and 100 ppm, the human equiva-
lent dosage for the above exposure is estimated as
dh = 20.24 x 5/7* (70/0.42)1/3 = 2.63 mg/kg/day for 100
ppm
and
d, = 2.7 x 5/7* (70/0.42)1/3 = 0.35 mg/kg/day for 10
ppm
where 0.42 kg is the average weight of the male rat in the Snellings et al.
(1981) study, 70 kg is the average weight of the adult human, and 5/7 is the
fraction of days exposed. By interpolation, the 33 ppm exposure is estimated
as 0.94 mg/kg/day in human equivalent doses.
9-TI3
-------
9.5.3.1.4. Calculation of the Unit Risk from Animal Studies — The risk
associated with d mg/kg2/3/day is obtained from GLOBAL83, and for most cases
of interest to risk assessment, the 95% upper-limit risk can be adequately
approximated by P(d) = 1 - exp (-q^d). A unit risk in units X is simply
the risk corresponding to an exposure of X = 1. To estimate this value, it
is simply necessary to find the number of mg/kg^'-^/day that corresponds to
one unit of X, and substitute this number into the above relationship. For
ethylene oxide, human equivalent doses will first be calculated and then
fitted, together with the observed responses, to the linearized multistage
model. An equivalent method of calculating unit risk would be to use mg/kg/
day for the animal exposures and then to increase the j*-*1 polynomial coeffi-
cient by an amount
(Wh/Wa)J/3 j = 1, 2, ..., k
and use the mg/kg/day equivalents for the unit risk values. In the section
of this document that presents unit risk calculations from animal data, the
final qi will always represent the upper-limit potency estimate for humans.
9.5.3.1.5. Interpretation of Quantitative Estimates — Unit risk esti-
mates based on animal bioassays are only approximate indications of absolute
risk in populations exposed to known carcinogen concentrations. This is true
for several reasons. First, there are important species differences in up-
take, metabolism, and organ distribution of carcinogens, as well as in target
site susceptibility, immunological responses, hormone function, dietary
factors, and disease. Second, the concept of equivalent doses for humans as
compared to animals based on the relationship of weight to surface area is
virtually without experimental verification as regards carcinogenic response.
9-1
-------
Finally, human populations are variable with respect to genetic constitution
and diet, living environment, activity patterns, and other cultural factors.
Unit risk estimates can give rough indications of the relative potencies
of given agents as compared with other carcinogens. Such comparisons are, of
course, most reliable when based on studies in which the test species,
strain, sex, and route of exposure are the same.
The quantitative aspects of assessing carcinogenic risks are discussed
here because of the possible usefulness of this information in the regulatory
decision-making process, e.g., in setting regulatory priorities, evaluating
the adequacy of technology-based controls, etc. However, the uncertainty of
present estimations of cancer risks to humans at low levels of exposure
should be recognized. The GAG feels that, given the limited data available
from animal bioassays, especially at the high dosage levels required for
testing, almost nothing can be known about the true shape of the dose-response
curve at low environmental levels. At best, the linear extrapolation model
used here provides a rough but plausible estimate of the upper limit of
risk; i.e., it is not likely that the true risk is appreciably higher than
the estimated risk, but it could very well be considerably lower. The risk
estimates presented in this document should not, therefore, be regarded as
accurate representations of the true cancer risks even when the exposures are
accurately defined. These estimates may, however, be factored into regula-
tory decisions to the extent that the concept of upper risk limits is found
to be useful.
9.5.3.1.6. Alternative Methodological Approaches — The methods used by
the GAG for quantitative assessment are consistently conservative in that they
tend to result in high estimates of risk. This conservatism is primarily due
9-745
-------
to the CAG's use of the linear nonthreshold extrapolation model in prefer-
ence to any one of a variety of other extrapolation models that would give
lower risk estimates. For purposes of comparison, descriptions of these
alternative models (the one-hit, the probit, and the Weibull models) are
presented in Appendix 9A.
Another method of risk estimation employed by the GAG involves the use
of animal bioassay data as the basis for extrapolation. At present, the
CAG's approach is to utilize data corresponding to the most sensitive animal
responses in these studies. An alternative approach would be to use the
average responses of all adequately tested bioassay animals.
Extrapolations from animals to humans can also be made on the basis of
either relative weight or surface area. The latter approach, which is used
by the GAG, has more of a basis in human pharmacological responses; however,
at the present time there is some question as to which of the two approaches
is more appropriate for use with carcinogens. Given this uncertainty, the
GAG has chosen the most generally employed method, which is also the more
conservative of the two. In the case of ethylene oxide inhalation studies,
the use of extrapolation based on surface area rather than weight increases
the unit risk estimates by a factor of 5.5 for the males and 6.8 for the
females.
9.5.3.2. HUMANS—MODEL FOR ESTIMATION OF UNIT RISK BASED ON HUMAN DATA
— Whenever possible, the GAG utilizes data from human epidemiologic studies
in preference to animal bioassay data. If sufficiently valid exposure infor-
mation is available for a given compound, this information is always used by
the GAG in its assessment. If the results of such studies show carcinogenic
effects, the data are analyzed to give estimates of the linear dependence of
9-146
-------
cancer rates on lifetime average doses (equivalent to the factor By in the
equation below). If human epidemiologic studies show no carcinogenic effects
when positive animal evidence is available, then it is assumed that a risk
does exist, but that the risk is smaller than could have been observed in an
epidemiologic study. In such cases it is assumed that the true incidence is
just below the level of detection in the cohort studied, and calculations are
then made to estimate an upper limit of cancer incidence, as determined
largely by the size of the cohort.
Very little information exists that can support extrapolation from high-
exposure occupational studies to situations in which contamination is at low
environmental levels. However, if a number of simplifying assumptions are
made, it is possible to construct a crude dose-response model whose para-
meters can be estimated using vital statistics, epidemiologic studies, and
estimates of worker exposures.
In human studies, responses are measured in terms of the relative risk
of an exposed cohort as compared to a control group. The mathematical model
employed by the GAG assumes that for low exposures the lifetime probability
of death from lung cancer (or any cancer), PQ, may be represented by the
linear equation
P0 = A + bHX
where A is the lifetime probability of death from cancer in the absence of
the agent, and X is the average lifetime exposure to environmental levels in
units such as ppm. The factor, by, is the increased probability of cancer
associated with each unit increase of the agent in air.
If it is assumed that R, the relative risk of lung cancer for exposed
workers as compared to the general population, is independent of the length
9-147
-------
or age of exposure and depends only on average lifetime exposure, it follows
that
R = P = A + bH
P0 A + bH
or
RPO = A + bH (K! + x2)
where Xj = lifetime average daily exposure to the agent for the general
population, X2 = lifetime average daily exposure to the agent in the occupa-
tional setting, and P0 = lifetime probability of dying of cancer with no or
negligible ethylene oxide exposure. Substituting P0 = A + b^ X^ and rearrang-
ing gives
t^ = P0 (R - 1)/X2
To use the above model, estimates of R and X2 must be obtained from ap-
propriate epidemiologic studies. The value of PQ is derived by means of life-
table methodology from 1976 U.S. vital statistics records of age- and cause-
specific death rates for males. For leukemia, the estimate of PQ is 0.0091.
This methodology is utilized by the GAG in the present document, in the sec-
tion on unit risk based on human data.
9.5.3.3. UNIT RISK ESTIMATES FOR ETHYLENE OXIDE
9.5.3.3.1. Unit Risk Estimate Based On Animal Studies — The two long-
term animal inhalation studies presented in the qualitative carcinogenicity
section of this document showed similar results, both qualitatively and
9-148
-------
quantitatively, for the males. Both studies had significantly increased
dose-related incidences of peritoneal mesothelioraas and gliomas, and some
increase in mononuclear cell leukemias. These studies will be analyzed
separately and then compared.
9.5.3.3.1.1. Snellings et al. (1981) (Bushy Run). This study exposed
120 Fischer 344 rats of each sex to three different doses (100 ppm, 33 ppm,
and 10 ppm) of ethylene oxide vapor via inhalation for 6 hours/day, 5 days/
week, for approximately 2 years. Comparable untreated (air) control groups
were also used. Interim sacrifices were conducted to evaluate the time
development of treatment-related effects.
The results of the study show statistically significant increases in
brain gliomas (highest dose group) and in mononuclear cell leukemias in
females in the two highest dose groups, and peritoneal mesotheliomas and
brain gliomas in males in the two highest dose groups. The above tumors all
exhibited dose-response trends. Table 9-33 summarizes the pertinent data
from this study which the GAG has used in calculating potency estimates for
ethylene oxide. In connection with these data, it should be noted that the
brain gliomas were not examined histopathologically until after the results
of the NIOSH study (Lynch et al., 1982) had alerted the Bushy Run researchers
to the possibility of the occurrence of brain neoplasias. For this reason,
only 18-month, 24-month, and dead/euthanized moribund denominator figures
were available for gliomas. For the male peritoneal mesotheliomas and the
female mononuclear cell leukemias, the denominators in Table 9-33 correspond
to the number of animals alive when the first tumor of that type was found.
In the males, the first peritoneal mesothelioma was found at 15 months; in
the females, the first mononuclear cell leukemia was found at 18 months (see
9-149
-------
TABLE 9-33. BUSHY RUN ETHYLENE OXIDE INHALATION STUDY IN FISCHER 344 RATS.
INCIDENCE OF PERITONEAL MESOTHELIOMA AND BRAIN GLIOMA IN MALES, AND MONONUCLEAR CELL
LEUKEMIA AND BRAIN GLIOMA3 IN FEMALES BY DOSE AMONG SURVIVORS TO FIRST TUMOR.
MAXIMUM LIKELIHOOD ESTIMATES OF LINEAR TERM AND 95% UPPER-LIMIT q*
Ul
O
Group
Males
Peritoneal meso./No. examined (X)c
p-values^
Brain glioraas/No. examined (%)e
p-values
Total
p-values
Human equivalent dose (mg/kg/day)'
Females
Mon. leukem/No. examined (%)S
p-values
Brain gliomas/No. examined (%)e
p-values
Total
p-values
Human equivalent dose (mg/kg/day )f
0 (combined)
4/187(2)
<0. 00001
1/196(0.5)
=0.0003
5/187(3)
<0. 00001
0
22/186(12)
<0. 00001
1/194(0.5)
=0.014
23/186(12)
<0. 00001
0
Exposure in air (ppm)
10 33
3/88(3) 7/82(8)
=0.02
1/99(1) 5/98(5)
=0.02
4/88(5) 12/82(15)
=0.0005
0.35 0.94
14/71(20) 24/72(33)
=0.08 0.0001
1/95(1) 3/99(3)
15/71(21) 27/72(38)
<0.0001
0.28 0.75
Linear term
estimates
q* b
MLE 1
100 qj (mg/kg/day)"1
22/96(22) 5.1xlO-2
<0.0001
7/99(7) 3.1x10-2
=0.002
29/96(30) l.lxlQ-1
<0.0001
2.63
28/73(38) 2.0x10-!
<0.0001
4/99(4) 2.0x10-2
=0.05
32/73(44) 2.5x10-!
<0.0001
2.11
1.1x10-1
5.0x10-2
1.7x10-1
2.9x10-!
4.0x10-2
3.5x10-1
Table 9-27.
b95% upper-limit unit risk estimate.
cNumber alive at 15 months.
dFisher Exact Test vs. combined controls (one tailed). P-value under controls is a one-sided Cochran-Armitage
test for a dose-response trend.
eTotal number examined less 6- and 12-month sacrifices.
fBased on measured doses in males of 20.24 and 2.7 mg/kg b.w. following 6 hours' exposure to ethylene oxide at 100 ppm
and 10 pnm, respectively. The animal-to-human dose equivalences are based on a dose per surface area factor of
(70/W ) ' , which increases unit risk estimates by factors of 5.5 for the males and 6.8 for the females over
dose per body weight equivalences.
SNumber alive at 18 months.
SOURCE: Adapted from Snellings et al., 1981.
-------
also Table 9-23).
As reported earlier, a dose of 20.24 mg/kg of body weight has been
measured for male Fischer 344 rats exposed to ethylene oxide at 100 ppm under
conditions similar to those of the Snellings et al. (1981) study. For this
document, dose is assumed to be equivalent between species on the basis of
mg/surface area, or mg/body weights/3. This means that a dose of 2.63 mg/kg
body weight given to a 70 kg human is assumed to produce an equivalent res-
ponse to that produced by 20.24 mg/kg in the male rat. As discussed above
and as shown in Table 9-33, this method of determining dose equivalence
increases the unit risk estimates by factors of 5.5 for females and 6.8 for
males over estimates obtained on the basis of mg/kg of body weight.
Table 9-33, in presenting the total number of significant tumors by sex,
sums the total number of significant tumors over the smallest denominator.
This is done because time-to-tumor data on the gliomas are unavailable. Com-
pared with the usual GAG procedure of counting the total number of animals
with significant tumors, the addition of total significant tumors, as is done
here, increases the risk estimate very slightly.
Calculations of the 95% upper-limit unit risk estimate, based on the
linearized multistage model fitted to the data in Table 9-33, yield a high
value of q^ = 3.5 x 10 (mg/kg/day) , based on total mononuclear cell leu-
kemias and brain gliomas in the female rats. The responses of the males,
based on total peritoneal mesotheliomas and brain gliomas, yield a value of
50% less, qh = 1.7 x 10" (mg/kg/day)"1. The higher estimate is chosen for
safety purposes.
To convert the above estimate to units of y g/nH for humans, the follow-
ing formula is used:
9-151
-------
1 mg/kg/day = 1 mg/kg/day x 70 kg x 1000p g/mg x day/20 m3 = 3.5 x 103y g/m3
or
ly g/m3 = 2.86 x 10~4 mg/kg/day.
The 95% upper-limit slope estimate in terms of y g/m3 is thus calculated as
q* = 3.5 x lO-iCmg/kg/day)'1 x 2.86 x 10~4(mg/kg/day) = 1.0 x 10~4(u g/m3)-1
h — o—
y g/m
To convert from y g/m3 to ppm, the formula is
1 ppm = l'2 g x 44>1 m.w. ethylene oxide x 10 y g x 1Q-6
10~3 m3 28'2 m
-------
study did not show a significant increase in these leukemias, only peritoneal
mesotheliomas and brain gliomas were used for risk assessment. The results
of the potency calculations, shown in Table 9-34, are quantitatively nearly
identical to those in Table 9-33. Based on the above analyses, the maximum
animal 95% upper-limit slope potency value is still q^ = 3.5 x 10~ (rag/kg/
based on the total mononuclear cell leukemias and brain gliomas in
female rats in the Snellings et al. study.
9.5.3.3.1.3. Effects of Results on Different Dose Equivalence Assumptions
— OSHA Versus EPA Assessments. The results of the above assessments depend,
to some extent, on the dose equivalence assumptions. Dose equivalence in the
following discussion means the dose that will cause an equivalent response,
quantitatively, in both species. The GAG has assumed that doses are equiva-
lent on the basis of mg per surface area, an assumption for which there is
some experimental evidence when first-order kinetics apply; for ethylene
oxide, first-order kinetics appear to apply at exposures up to 100 ppm (Tyler
and McKelvey, 1980). As explained in an earlier section, use of the surface
area correction increases the 95% upper-limit unit risk estimate by factors
of 5.5 for the males and 6.8 for the females over estimates obtained on the
basis of mg/kg/body weight.* OSHA, which assumes equivalence on a mg/kg/
body weight basis, calculated exposures of 19.30 mg/kg/day for males and
23.94 mg/kg/day for females exposed to ethylene oxide at 100 ppm in the
Snellings et al. (1981) study, using EPA methodology (Federal Register 48[78]:
*Equivalence could also have been calculated directly on a ppm basis; this
would have yielded a 95% upper-limit estimate approximately 1.8 times as
high as that obtained on the basis of mg/kg/body weight. EPA uses direct ppm
equivalence for partially soluble gases and particulates. Ethylene oxide
can be considered a completely soluble gas.
9-153
-------
TABLE 9-34. NIOSH ETHYLENE OXIDE INHALATION STUDY IN MALE FISCHER 344 RATS.
INCIDENCE OF PERITONEAL MESOTHELIOMA AND BRAIN GLIOMAa BY DOSE,
AMONG TOTAL EXAMINED.
ESTIMATES OF 95% UPPER-LIMIT RISK BASED ON HUMAN EQUIVALENT DOSE (mg/kg/day)
Exposure in air (ppm)
1
50 100 (mg/kg/day)-1
Peritoneal mesothelioma/No.
examined
Brain glioma/No. examined
Total
Human equivalent dose
(mg/kg/day)f
3/78b
0/76C
3/78b
0
9/79
2/77
ll/79b
1.59
21/79b
5/79d
26/79b
3.06
l.OxlO-1
3.4xlO-2
1.3X10-1
—
aSee Table 9-29.
bp < 0.001.
cp < 0.01.
dp < 0.05.
ep-values noted beside control incidences represent values associated with a
one-sided Cochran-Armitage test for a dose-response trend.
fHuman equivalent dose based on transforming ppm to mg/kg/day as in Table 9-33,
except for an adjustment for 7 hours' exposure.
SOURCE: Lynch et al., 1982.
9-154
-------
172-193) for a completely soluble gas. While the results for the males,
19.30 mg/kg/day, are within 5% of the dose measured by Tyler and McKelvey
(1980) (20.24 mg/kg/day), EPA used the more accurate measured dose in this
case. EPA then used the surface area correction factors for animal-to-man
equivalence. Thus, on the basis of the difference in assumptions of equiva-
lent dose alone, the EPA risk numbers are larger than OSHA's by a factor of
about 6.
One other difference between the OSHA and EPA assessments, both based on
the Bushy Run data, is that the EPA added total significant tumors (mononuc-
lear cell leukemias and brain gliomas for the females), while OSHA used the
total number of malignant tumor-bearing animals. For EPA, this led to factors
higher by 50% for the males and 20% for the females. The result, based on
animal data, is that the EPA 95% upper-limit unit risk factor is larger than
that of OSHA by a factor of about 8.
9.5.3.3.2. Comparison of Animal and Human Inhalation Studies — The pur-
pose of this section is to determine whether or not the extrapolated risks
from the animal data can reasonably predict the observed human results. As
presented in Sections 9.5.1 and 9.5.2, there is strong evidence for the carci-
nogenicity of ethylene oxide in rats, while the evidence in humans is limited,
but suggestive of a leukemia effect. Table 9-35 summarizes the leukemia evi-
dence in humans. As can be seen, two of the four mortality studies reviewed
had statistically significant increases in the relative risks of leukemia.
None of the four studies was particularly revealing, however, since both the
observed and expected numbers were quite small. Because the expected numbers
were so small, the 95% confidence limits around the relative risk are quite
large. Even with statistically significant increased cancers, estimates of
9-155
-------
TABLE 9-35. LEUKEMIA (ICD 204-207) INCIDENCE AND MORTALITY: ETHYLENE OXIDE EPIDEMIOLOGY.
INCLUDED ARE RELATIVE RISKS, 95% CONFIDENCE LIMITS, NOMINAL EXPOSURE ESTIMATES,
AND 95% CONFIDENCE LIMITS ON UNIT RISK
Study
Hogstedt (sterllant plant)
(1968-1977)
Storage hall
Adjacent area
Total (original, 1968-77)
(New data, 1968-1982)
N
(70)
(160)
(230)
(203)t>
Obs.
2
0
2
4
Exp.
0.03
0.07
0.10
0.3
Rela-
tive
risk.
66.7
0
20
13.3
(95%
confidence
limits)
8-240
0-52
2.4-72
3.6-34.2
Exposure
(ppm)
20 + 10 (8 hr TWA)
"exposed occasionally
on passing through"
95% confidence
limits on unit
risk3 (ppm~* )
0.07 - 2.5
No other
estimates
possible
ICD 200-207
Hogstedt (prod, plant)
a. (1961-1982)
>_ 10 yr latency
>^ 1 yr exposure
Direct exposure (89) 2 0.18 10.0
Intermittent exposure (79) 1 0.16 6.2
b. > 10 yr exposure
7 20 yr Induction-
latency (1961-1977)
Direct exposure 1 0.04 25
Intermittent exposure 1 0.1 10
Morgan et al.
(1955-1977) (767) 0 0.7 0
Thiess et al. (351) 1 0.15 6.8
(1928-1980)
> 10 yr exposure
1.3-40.1
0.1-42.8
0.3-139
0.1-55.6
0-5.2
0.1-37
< 14
"less exposure"
< 10
mostly < 5 ppm
(a few large variations)
aSee text. Estimates for most groups not possible, due to lack of information about exposure duration.
blncludes those employed at least one year. Hogstedt et al. (1984) subcohorts constructed differently than
earlier Hogstedt et al. (1979b) subcohorts.
-------
hazard due to ethylene oxide are difficult because of the small sample size.
An even greater problem associated with determining potency estimation
from human studies is the general lack of exposure information. Table 9-35
presents some ethylene oxide measurements based on either 8-hour time-weighted
averages or spot measurements, but other important data are generally missing.
Such necessary information as average length of exposure, average age at
exposure, and average length of follow-up cannot even be estimated from three
of the four studies in the literature. Only the subcohort of storage hall
workers in the sterilant plant (Hogstedt et al., 1979b, 1984) is consistent
enough in terms of exposure conditions, duration, and follow-up to estimate
95% confidence limits of ethylene oxide potency. This is presented in the
following paragraphs. The simplifying assumptions add to the uncertainty of
the estimate.
The risk assessment done on the basis of the Hogstedt et al. (1979b)
study probably underestimates the carcinogenic potency of ethylene oxide
because of two factors: 1) In this study, exposure started in 1968 and ended
in 1977—giving a maximum latency period of only 9 years, whereas cancer
usually involves a relatively long latency period. (However, the author
states that leukemia incidence in Hiroshima and Nagasaki due to the atomic
bomb irradiation showed a rapid increase that began shortly after exposure
and reached a peak after 6 years.) 2) Since the study did not report the
number of person-years of exposure, it is assumed for present purposes that
all of the 230 workers were exposed for the full 9 years, an assumption which
tends to underestimate the risk. Another problem with this study is that the
gas used for sterilization was 50% ethylene oxide and 50% methyl formate.
Little is known about the biological effects of methyl formate or of the
combination of methyl formate with ethylene oxide. However, methyl formate
9-157
-------
is known to metabolize to formic acid, which is a normal body metabolite. It
is assumed for present purposes that ethylene oxide was the only leukeraogen
in this study, although one of the cases (the man) had reported some contact
with benzene in laboratory work.
Hogstedt et al. (1979b) states, in connection with exposures in the
factory studied, that infrared spectrophotometry and gas chromatography
measurements in 1977 showed values ranging from 2 to 70 ppm in the factory's
storage hall area. The study also reports that the calculated 8-hour time-
weighted average ethylene oxide concentration in the breathing zone was
20 + 10 ppm, and that the concentration in the storage hall was higher than
in the sterilization room. The accompanying table described the 70 storage
hall employees as having had 8-hour exposures, while all but seven of the
remaining employees were described as "occasionally exposed."
Of the two leukemia cases (acute myeloid and chronic myeloid), both
people worked in the storage hall area, and neither had reported exposure to
benzene. Because the two cases worked in the storage hall, the GAG has
chosen to estimate the expected number of leukemia cases for the persons who
worked only in that area rather than in the entire factory. Based on the
reported expected leukemia incidence of 0.1 cases for the 230 exposed em-
ployees, we can estimate approximately (70/230) x 0.1 = 0.03 cases for the
group exposed in the storage hall. Compared with the two observed cases,
this yields a ratio of observed to expected cases of (2/0.03) = 66.7.
The estimated average exposure to ethylene oxide over the lifetime of
the workers is calculated as follows:
20 ppm x 8/24 hr x 240/365 days x 9/45.6 yr
exposure = 0.865 ppm
9-158
-------
where 45.6 years is the mean age of the 70 storage hall employees at the end
of the study period.
The slope by of the lifetime probability of dying from leukemia due to a
lifetime of breathing ethylene oxide at 1 ppm (Section 9.5.3.2.) is given by
b = P (R - 1) Xl
X2
where PQ is the lifetime probability of dying* from leukemia in the United
States in the absence of ethylene oxide exposure, R is the relative risk, Xj
is the exposure of 1 ppm, and X2 is the exposure experienced by the factory
workers. The relative risk R estimated above is 66.7; the exposure X2 is
given as 0.865 ppm. The lifetime probability of death from leukemia in the
U.S. population is 0.0091. Substituting these values in the above equation
gives
b = 0.0091 (66.7 - 1) = o^Cppm)-1
H 0.865 ppm
The probability associated with breathing ethylene oxide at 1 ppm for a
lifetime is
P = 1 - e~bH (1 ppm) = 0.50
To convert ppm to y g/m , the formula is
*PQ employs both leukemia incidence cases and leukemia mortality rates.
While leukemia mortality in the younger ages « 55) can be closely equated
with incidence, in the older age groups chronic forms predominate in inci-
dence, with death often occurring from other causes. Nevertheless, for this
assessment it is assumed that although ethylene oxide would cause all types
of leukemias, death will result from each case. In this study, the leukemias
in the two women were of the acute form.
9-159
-------
1 ppm = 1'2 8 x 44.1 tn.w. chemical x 10% g x 1Q-6
IQ-3 m3 28.2 m.w. air g
= 1.9 x 103U g/m3
Thus the unit risk estimate in terms of y g/m3 is
bR = O.egCppm)'1 x 1 ppm = 3.6 x 10~4(u g/m3)"1
1.9 x 103y g/m3
Similarly, if the 95% limits on the relative risk from Table 9-35 are sub-
stituted for the point estimate, the 95% confidence limits on the unit risk
become 7.0 x 10~^(ppm)~^ to 2.5(ppm)"~ . This represents a range of 36 and
encompasses the 95% upper-limit incremental unit risk, q^ = 1.9 x 10~
ppm~l, extrapolated from the Snellings et al. (1981) study.
Based on the above analysis, we conclude that the carcinogenic
potency estimates for ethylene oxide derived from human data do not contradict
the estimate based on the rat inhalation studies. Because of the uncertain-
ties in the epidemiologic study, however, the animal inhalation study is
chosen for the 95% upper-limit incremental unit risk estimate for ethylene
oxide.
9.5.3.4. RELATIVE POTENCY — One of the uses of the concept of unit risk
is to compare the relative potencies of carcinogens. For the purposes of the
present analysis, potency is defined as the linear portion of the dose-response
curve, and is used to calculate the required unit risk factors. To estimate
relative potency on a per-mole basis, the unit risk slope factor is multiplied
by the molecular weight of the compound, and the resulting number, expressed
in terms of (mmol/kg/day)~l, is called the "relative potency index."
Figure 9-4 is a histogram representing the frequency distribution of rel-
tive potency indices for 54 chemicals that have been evaluated by the GAG as
9-160
-------
20
18
16
14
>- 12
CJ
:REQUEI^
o
8
6
4
2
n
—
«.
—
_
—
- I
—
"
-1
II
l«
0
4th 3rd 2nd 1st
QUARTILE QUARTILE QUARTILE QUARTILE
1 x 10+ 4 x 10+ 2 x 10
—
PH
IS
vXvX;
"***•**"•***
«***•***•*•*
B
M
^ivvV
•
^^^
—
H
>x'&x*
XvX'X'
:•••:>$•$
:£:;!::|:x'
•x'x'x'i
—
—
—
B
|:::;:;X;S; —
X'X'X'X *X* • "t't* P"^
mm ETTEI s?s R EHEI
12345678
LOG OF POTENCY INDEX
Figure 9-4. Histogram representing the frequency distribution of the potency indices
of 54 suspect carcinogens evaluated by the Carcinogen Assessment Group.
9-161
-------
suspect carcinogens. The data summarized by the histogram are presented in
Table 9-36. Where human data have been available for a compound, such data
have been used to calculate these indices. Where no human data have been
available, data from animal oral studies have been used rather than data from
animal inhalation studies, since animal oral studies have been conducted for
most of these compounds, and their use allows potency comparisons by route.
On the basis of mononuclear cell leukemias and gliomas in female rats in
the Snellings et al. (1981) inhalation study, the relative potency index for
ethylene oxide has been calculated as 1.54 x 10+^. This number was derived
by multiplying the slope in units of (ing/kg/day)"* by the molecular weight of
ethylene oxide, which is 44.1. For the rat study, this slope is 3.5 x 10~1
(mg/kg/day)+1.
The potency index for ethylene oxide is thus 3.5 x 10"~1 x 44.1 =
1.54 x 10~1, putting ethylene oxide at the bottom of the third quartile of the
54 chemicals which the GAG has evaluated as suspect carcinogens. It should
be noted that the ranking of these relative potency indices is subject to the
uncertainties involved in comparing a number of potency estimates for differ-
ent chemicals on the basis of varying routes of exposure in different species,
using studies whose quality varies widely. Furthermore, all of these indices
are based on estimates of low-dose risk that have been calculated by means of
linear extrapolation from the observational range. The indices are, therefore,
not valid for the comparison of potencies in the experimental or observational
range if linearity does not exist there.
9.5.4. Summary. Ethylene oxide has been shown to be carcinogenic in animals
in long-term studies by three different routes of administration (inhalation,
subcutaneous injection, and gavage). The most relevant route for human expo-
9-162
-------
TABLE 9-36. RELATIVE CARCINOGENIC POTENCIES AMONG 54 CHEMICALS EVALUATED BY THE CARCINOGEN ASSESSMENT GROUP
AS SUSPECT HUMAN CARCINOGENS
Level
of evidence3
Compounds
Acrylonitrile
Aflatoxin B^
Aldrin
Allyl chloride
Arsenic
B[a]P
Benzene
Benzidene
Beryllium
1 ,3-Butadiene
Cadmium
Carbon tetrachloride
Chlordane
CAS Number
107-13-1
1162-65-8
309-00-2
107-05-1
7440-38-2
50-32-8
71-43-2
92-87-5
7440-41-7
106-99-0
7440-43-9
56-23-5
57-74-9
Humans
L
L
I
S
I
S
S
L
I
L
I
I
Animals
S
S
L
I
S
S
S
S
S
S
S
L
Grouping
based on
IARC
criteria
2A
2A
2B
1
2B
1
1
2A
2B
2A
2B
3
Slope
(mg/kg/day)"1
0.24(W)
2900
11.4
1.19x10-2
15(H)
11.5
2.9xlO-2(W)
234(W)
2.6
l.OxlO-id)
7.8(W)
l.SOxlQ-1
1.61
Molecular
weight
53.1
312.3
369.4
76.5
149.8
252.3
78
184.2
9
54.1
112.4
153.8
409.8
Potency
index
1x10+1
9xlO+5
4x10+3
9x10-!
2xlO+3
3x10+3
2x10°
4xlO+4
2xlO+1
5x10°
9xlO+2
2xlO+1
7x10+2
Order of
magnitude
(Iog10
index)
+1
+6
+4
0
+3
+3
0
+5
+1
+1
+3
+1
+3
aS = Sufficient evidence; L = Limited evidence; I = Inadequate evidence.
(continued on the following page)
-------
TABLE 9-36. (continued)
Level
of evidence3
Compounds
Chlorinated ethanes
1 , 2-Dichloroethane
hexachloroe thane
CAS Number
107-06-2
67-72-1
1,1,2, 2-Tetrachloroethane 79-34-5
1,1, 2-Trichloroethane
Chloroform
Chromium VI
DDT
Dichlorobenzidine
1 , 1-Dlchloroethylene
(Vinylidene chloride)
Dichlorome thane
(Methylene chloride)
Dieldrin
2,4-Dinitrotoluene
Diphenylhydrazine
Epichlorohydrin
Bis(2-chloroethyl)ether
79-00-5
67-66-3
7440-47-3
50-29-3
91-94-1
75-35-4
75-09-2
60-57-1
121-14-2
122-66-7
106-89-8
111-44-4
Humans
I
I
I
I
I
S
I
I
I
I
I
I
I
I
I
Animals
S
L
L
L
S
S
S
S
L
L
S
S
S
S
S
Grouping
based on
IARC
criteria
2B
3
3
3
2B
1
2B
2B
3
3
2B
2B
2B
2B
2B
Slope
(mg/kg/day)"1
6.9xlO-2
1.42xlO-2
0.20
5.73x10-2
7xlO-2
41(W)
0.34
1.69
1.17(1)
6.3xlO-4(I)
30.4
0.31
0.77
9.9x10-3
1.14
Molecular
weight
98.9
236.7
167.9
133.4
119.4
100
354.5
253.1
97
84.9
380.9
182
180
92.5
143
Potency
index
7x10°
3x10°
3xlO+1
8x10°
8x10°
4xlO+3
1x10+2
4x10+2
1x10+2
5xlO-2
1x10+4
6xlO+1
lxlO+2
9x10-!
2x10+2
Order of
magnitude
(Iog10
index)
+1
0
+1
+1
+1
+4
+2
+3
+2
-1
+4
+2
+2
0
+2
aS = Sufficient evidence; L = Limited evidence; I = Inadequate evidence.
(continued on the following page)
-------
TABLE 9-36. (continued)
Level
of evidence3
Compounds
Bi s ( ch lor ome thy 1) ether
Ethylene dibromide (EDB)
Ethylene oxide
Heptachlor
^ Hexachlorobenzene
0! Hexachlorobutadiene
Hexachlorocyclohexane
technical grade
alpha isomer
beta isomer
gamma isomer
Hexachlorodibenzodioxin
Nickel
Nitrosamines
Dimethvlnitrosamine
Die thy Initrosamine
Dibutylnitrosamine
CAS Number
542-88-1
106-93-4
75-21-8
76-44-8
118-74-1
87-68-3
319-84-6
319-85-7
58-89-9
34465-46-8
7440-02-0
62-75-9
55-18-5
924-16-3
Humans
S
I
L
I
I
I
I
I
I
I
L
I
I
I
Animals
S
S
S
S
S
L
S
L
L
S
S
S
S
S
Grouping
based on
IARC
criteria
1
2B
2A
2B
2B
3
2B
3
2B
2B
2A
2B
2B
2B
Slope Molecular
(mg/kg/day)"1 weight
9300(1)
41
3.5x10-1(1)
3.37
1.67
7.75x10-2
4.75
11.12
1.84
1.33
6.2xlO+3
1 .15(W)
25.9(not by qf)
43.5(not by q|)
5.43
115
187.9
44.1
373.3
284.4
261
290.9
290.9
290.9
290.9
391
58.7
74.1
102.1
158.2
Potency
index
lxlO+6
8xlO+3
2xlO+1
lxlO+3
5x10+2
2xlO+1
lxlO+3
3xlO+3
5x10+2
4x10+2
2xlO+6
7xlO+1
2xlO+3
4xlO+3
9x10+2
Order of
magnitude
(Iog10
index)
+6
+4
+1
+3
+3
+1
+3
+3
-1-3
+3
+6
+2
+3
+4
+3
aS = Sufficient evidence; L = Limited evidence; I = Inadequate evidence.
(continued on the following page)
-------
TABLE 9-36. (continued)
Level
of evidence3
Compounds CAS Number
N-nit rosopyrrolidine
N-nitroso-N-ethylurea
N-nit roso-N-methylurea
N-nitroso-diphenylamine
PCBs
Phenols
2,4, 6-Trichlorophenol
Tetrachlorodibenzo-
p-dioxin (TCDD)
Tetrachloroethylene
Toxaphene
Trichloroethylene
Vinyl chloride
930-55-2
759-73-9
684-93-5
86-30-6
1336-36-3
88-06-2
1746-01-6
127-18-4
8001-35-2
79-01-6
75-01-4
Humans
1
I
I
I
I
I
I
I
I
I
S
Animals
S
S
S
S
S
S
S
L
S
L/S
S
Grouping
based on
IARC
criteria
2B
2B
2B
2B
2B
2B
2B
3
2B
3/2B
1
Slope
(mg/kg/day)"1
2.13
32.9
302.6
4.92xlO-3
4.34
1.99x10-2
1.56x10+5
6.0x10-2
1.13
1.2x10-2
1.75x10-2(1)
Molecular
weight
100.2
117.1
103.1
198
324
197.4
322
165.8
414
131.4
62.5
Potency
index
2x10+2
4x10+3
3xlO+4
1x10°
lxlO+3
4x10°
5x1 0+7
IxlO1
5xlO+2
2x10°
1x10°
Order of
magnitude
(Iog10
index)
+2
+4
+4
0
+3
+1
+8
+1
+3
0
0
aS = Sufficient evidence; L = Limited evidence; I = Inadequate evidence.
Remarks:
1. Animal slopes are 95% upper-limit slopes based on the linearized multistage model. They are calculated based on
animal oral studies, except for those indicated by I (animal inhalation), W (human occupational exposure), and H
(human drinking water exposure). Human slopes are point estimates based on the linear nonthreshold model.
2. The potency index is a rounded-off slope in (mraol/kg/day)"^- and is calculated by multiplying the slopes in
(mg/kg/day)-'- by the molecular weight of the compound.
3. Not all of the carcinogenic potencies presented in this table represent the same degree of certainty. All are
subject to change as new evidence becomes available.
-------
sure is inhalation. Two long-term inhalation studies in rats were performed
that adequately tested the carcinogenic potential of ethylene oxide by inhala-
tion: the Bushy Run study (Snellings et al., 1981) and the NIOSH study (Lynch
et al., 1982). Snellings et al. (1981) found that ethylene oxide exposure
resulted in an increased incidence of mononuclear cell leukemia in females in
the two highest dose groups; this increase was dose related. The test for
linear trend was highly significant (p < 0.0001). There was also a signifi-
cant (p = 0.045) increase in gliomas at the highest dose, and the test for
linear trend was highly significant (p = 0.014). In males, incidences of
primary brain neoplasm, peritoneal mesothelioma, and subcutaneous fibroma
were significantly elevated in at least two exposed groups. The trend analy-
sis was significant for both mesotheliomas (p < 0.00001) and gliomas (p =
0.003) in males. In the NIOSH (Lynch et al., 1982) study, which involved
only male rats, leukemia incidence was significantly increased at low doses
only, while gliomas (mixed-cell) and peritoneal mesotheliomas were increased
significantly in the high-dose groups. For these latter two sites, the dose-
response trend tests were also statistically significant (p < 0.01). Other
positive results for the carcinogenicity of ethylene oxide were demonstrated
by subcutaneous injection in mice and intragastric administration in rats.
Three epidemiologic studies of workers exposed to ethylene oxide demon-
strated significant (p < 0.05) association between ethylene oxide exposure
and the occurrence of cancer. Two of the studies (Hogstedt et al., 1979a, b)
found an association between ethylene oxide exposure and the incidence of
leukemia. Ethylene oxide was not found to be associated with any particular
type of leukemia, however. Other sites or types of cancer found to be signi-
ficantly (p < 0.05) associated with ethylene oxide exposure in an individual
study include pancreatic cancer and Hodgkin's disease in the Morgan et al.
9-167
-------
(1981) study and stomach cancer in the Hogstedt et al. (1979a) study. The
possibility of confounding due to other chemical agents cannot be excluded in
any of the studies, however.
An upper-limit incremental unit risk estimate of 1.0 x 10~^( p g/m^)""!
for ethylene oxide has been calculated, using a linearized multistage model,
on total mononuclear cell leukemias and brain gliomas in female Fischer 344
rats from the Bushy Run (Snellings et al., 1981) study. Extrapolation
from the human leukemia data results in a highly uncertain risk estimate due
to the small numbers of leukemia cases that were observed and expected.
Quantitative comparisons of human and animal inhalation studies do, to the
extent possible, support each other.
9.5.5. Conclusions. Ethylene oxide has been shown to be carcinogenic in
animals by intragastric, subcutaneous injection, and inhalation routes of
exposure. Three human studies show an association between ethylene oxide
exposure and an excess risk of cancer, but each of these studies has some
limitations. Other evidence, which is in the mutagenicity section of this
document, supports the conclusions for carcinogenicity in that ethylene
oxide is a direct-acting alkylating agent, it reacts with mammalian DNA, it
induces base-pair substitutions in the Ames test and gene mutations in
plants and animals, and it breaks chromosomes of plants, animals, and humans
and causes DNA damage in the spermatids of mice.
Using the weight-of-evidence criteria of EPA's Proposed Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 1984), the Carcinogen Assessment Group
considers the animal evidence for carcinogenicity to be "sufficient" and
the human evidence to be "limited" bordering on inadequate. Based on both
the animal and human findings, the overall EPA classification for ethylene
9-168
-------
oxide is Group Bl, meaning that ethylene oxide should be considered as probably
carcinogenic to humans. This Bl classification is qualified as bordering on
Group B2 because of the limitations in the human evidence.
According to the IARC guidelines for evaluating carcinogen evidence (See
Appendix 9B), ethylene oxide would be classified in Group 2A. This classifi-
cation is similarly qualified as bordering on Group 2B because of limitations
in the human evidence. A Group 2 classification, whether 2A or 2B, means
that ethylene oxide should be considered as probably carcinogenic in humans.
An upper-limit carcinogenic potency value of 3.5 x 10" 1 (mg/kg/day)"*
has been calculated based on total mononuclear cell leukemias and brain
gliomas in female Fischer 344 rats in the Snellings et al. (1981) study. An
upper-limit incremental unit risk of 1.0 x 10"^ (y g/ra^)~^ has also been
estimated using a linearized multistage model.
9-169
-------
APPENDIX 9A
COMPARISON OF RESULTS BY VARIOUS EXTRAPOLATION MODELS
The estimate of unit risk from animals presented in the body of this
document was calculated by use of the linearized multistage model. This non-
threshold model is part of a methodology for estimating a conservative linear
slope at low extrapolation doses that is usually consistent with the data at
all dose levels in an experiment. The model holds that the most plausible
upper limits of risk are those predicted by linear extrapolations to low
levels of the dose-response relationship.
Other nonthreshold models that have been used for risk extrapolation
are the one-hit, the log-Probit, and the Weibull models. The one-hit
model is characterized by a continuous downward curvature, but is linear
at low doses. Because of its functional form, the one-hit model can be con-
sidered the linear form or first stage of the multistage model. This fact,
together with the downward curvature of the one-hit model, means that the
model will always yield low-level risk estimates that are at least as large
as those obtained with the multistage model. In addition, whenever the data
can be fitted adequately to the one-hit model, estimates based on the one-hit
model and the multistage model will be comparable.
The log-Probit and the Weibull models, because of their general "S" curv-
ature, are often used for the interpretation of toxicological data in the
observable range. The low-dose upward curvatures of these two models usually
yield lower low-dose risk estimates than those of the one-hit or multistage
models.
The log-Probit model was originally used in biological assay problems
such as potency assessments of toxicants and drugs, and is most often used to
A-l
-------
estimate such values as percentile lethal dose or percentile effective dose.
The log-Probit model was developed along strictly empirical lines, in studies
where it was observed that several log dose-response relationships followed
the cumulative normal probability distribution function, $ . In fitting the
log-Probit model to cancer bioassay data, assuming an independent background,
this relationship becomes
P(D;a,b,c) = c + (1-c) $ (a+blog10 D) a,b > 0< c < I
where P is the proportion responding at dose D, c is an estimate of the back-
ground rate, a is an estimate of the standardized mean of individual toler-
ances, and b is an estimate of the log-Probit dose-response slope.
The one-hit model arises from the theory that a single molecule of a
carcinogen has a quantifiable probability of transforming a single normal cell
into a cancer cell. In this model, the probability distribution function is
P(D;a,b) = l-exp-(a+bd) a,b > 0
where a and b are the parameter estimates (a = the background or zero dose
rate, and b = the linear component or slope of the dose-response model). In
considering the added risk over background, incorporation of Abbott's correc-
tion leads to
P(D;b) = l-exp-(bd) b > 0
Finally, a model from the theory of carcinogenesis arises from the multihit
model applied to multiple target cells. This model, known as the Weibull
model, is of the form
P(D;b,k) = l-exp-(bdk) b,k > 0
A-2
-------
For the power of dose only, the restriction k > 0 has been placed on this
model. When k > 1, the model yields low-dose estimates of risks that are
usually significantly lower than either the multistage or one-hit models,
both of which are linear at low doses. All three of these models—the multi-
stage, the one-hit, and the Weibull—usually project risk estimates that are
significantly higher at low exposure levels than those projected by the
log-Probit model.
The results of both the male and female rat data sets from the Bushy
Run (Snellings et al., 1981) study are presented in Table 9A-1. Surprisingly,
for the female rats, both the Weibull and log-Probit models yielded larger
estimates of risk than the multistage model, which, in this case, produced
results identical to those produced by the one-hit model. For the males, the
one-hit model produced the highest estimates and the log-Probit model produced
the lowest; in this case, the multistage, one-hit, and Weibull all produced
similar results.
A-3
-------
TABLE A-l. ESTIMATES OF HUMAN LOW-DOSE RISK BASED ON DATA FROM MALE AND FEMALE FISCHER 344 RATS
IN THE BUSHY RUN ETO INHALATION STUDY, AS DERIVED FROM FOUR DIFFERENT MODELS.
ALL ESTIMATES INCORPORATE ABBOTT'S CORRECTION FOR INDEPENDENT BACKGROUND RATE
Maximum likelihood estimates of
additional risks
Continuous
human
exposure Multistage One-hit Weibull
ppm model model model
Males
.001 3.1x10-4 1.3x10-4 2.5x10-5
0.01 3.1x10-3 1.3x10-3 3.3x10-4
0.1 3.1x10-2 1.3x10-2 4.3x10-3
1 8.4x10-2 1.2x10-1 5.5x10-2
Females
.001 1.4x10-5 3.6x10-3
0.01 1.4x10-3 1.4x10-2
0.1 1.4x10-2 5.0x10-2
1 1.3x10-1 1.7x10-1
Animal exposure 0, 33 ppm, 100 ppm 6 hours/day, 5 days/week.
DATA
Human eq. dose - mg/kg/day
Males 0 0.35 0.94 2.63 Females
Log-Probit Multistage
model model
3.1x10-1 9.2x10-5
1.6x10-6 9.2x10-4
8.6x10-4 9.2x10-3
5.4x10-2 8.8xlO-2
1.2x10-4 1.9x10-4
3.0x10-3 1.9x10-3
3.3x10-2 1.9x10-2
1.7x10-! 1.7x10-1
Human eq. dose - mg/kg/day
0 0.28 0.75 2.11
957. upper confidence limit
of additional risks
One-hit Weibull Log-Probit
model model model
1.6x10-4 1.3x10-4 5.0x10-'-*
1.6x10-3 1.4x10-3 1.5x10-5
1.6x10-2 1.3x10-2 3.9xUr-l
1.6x10-1 9.7x10-2 9.7x1(1-2
1.5xl()-2 1.0x10-'
4.3x10-2 1.5x10-2
l.lxlO"! 9.Jxl()-2
2.5xH)-l 2.5x11)-'
No. tumors/No, examined 5/187 4/88 12/82 29/96
23/186 15/71 27/72 32/73
Conversions for low doses: Humans 1 mg/kg/day = 1.84 ppm in air
or 1 ppm air = .543 mg/kg/day
Multistage and one-hit models gave identical results in females.
-------
APPENDIX 9B
INTERNATIONAL AGENCY FOR RESEARCH ON CANCER CLASSIFICATION SYSTEM
FOR THE EVALUATION OF THE CARCINOGENIC RISK
OF CHEMICALS TO HUMANS*
ASSESSMENT OF EVIDENCE FOR CARCINOGENICITY FROM STUDIES IN HUMANS
Evidence of carcinogenicity from human studies comes from three main
sources:
1. Case reports of individual cancer patients who were exposed to the
chemical or process.
2. Descriptive epidemiological studies in which the incidence of cancer
in human populations was found to vary in space or time with exposure
to the agents.
3. Analytical epidemiological (case-control and cohort) studies in which
individual exposure to the chemical or group of chemicals was found
to be associated with an increased risk of cancer.
Three criteria must be met before a causal association can be inferred
between exposure and cancer in humans:
1. There is no identified bias which could explain the association.
2. The possibility of confounding has been considered and ruled out as
explaining the association.
3. The association is unlikely to be due to chance.
In general, although a single study may be indicative of a cause-effect
relationship, confidence in inferring a causal association is increased when
several independent studies are concordant in showing the association, when
*Adapted from International Agency for Research on Cancer. Monographs Sup-
plement 4, Evaluation of the Carcinogenic Risk of Chemicals to Humans, 1982.
pp. 11-14. '
3-1
-------
the association is strong, when there is a dose-response relationship, or
when a reduction in exposure is followed by a reduction in the incidence of
cancer.
The degrees of evidence for carcinogenicity from studies in humans were
categorized as:
1. Sufficient evidence of carcinogenicity, which indicates that there
is a causal relationship between the agent and human cancer.
2. Limited evidence of carcinogenicity, which indicates that a causal
interpretation is credible, but that alternative explanations, such as chance,
bias, or confounding, could not adequately be excluded.
3. Inadequate evidence, which indicates that one of three conditions
prevailed: (a) there were few pertinent data; (b) the available studies,
while showing evidence of association, did not exclude chance, bias, or
confounding; (c) studies were available which do not show evidence of carcin-
ogenicity.
ASSESSMENT OF EVIDENCE FOR CARCINOGENICITY FROM STUDIES IN EXPERIMENTAL
ANIMALS
These assessments were classified into four groups:
1. Sufficient evidence of carcinogenicity, which indicates that there
is an increased incidence of malignant tumors: (a) in multiple species or
strains; or (b) in multiple experiments (preferably with different routes of
administration or using different dose levels); or (c) to an unusual degree
with regard to incidence, site or type of tumor, or age at onset. Additional
evidence may be provided by data on dose-response effects, as well as infor-
mation from short-term tests or on chemical structure.
B-2
-------
2. Limited evidence of carcinogenicity, which means that the data sug-
gest a carcinogenic effect but are limited because: (a) the studies involve
a single species, strain, or experiment; (b) the experiments are restricted
by inadequate dosage levels, inadequate duration of exposure to the agent,
inadequate period of follow-up, poor survival, too few animals, or inadequate
reporting; or (c) the neoplasms produced often occur spontaneously and, in the
past, have been difficult to classify as malignant by histological criteria
alone (e.g., lung and liver tumors in mice).
3. Inadequate evidence, which indicates that because of major qualita-
tive or quantitative limitations, the studies cannot be interpreted as showing
either the presence or absence of a carcinogenic effect; or that within the
limits of the tests used, the chemical is not carcinogenic. The number of
negative studies is small, since, in general, studies that show no effect are
less likely to be published than those suggesting carcinogenicity.
4. No data indicates that data were not available to the Working Group.
The categories sufficient evidence and limited evidence refer only to the
strength of the experimental evidence that these chemicals are carcinogenic
and not to the extent of their carcinogenic activity nor to the mechanism
involved. The classification of any chemical may change as new information
becomes available.
EVALUATION OF CARCINOGENIC RISK TO HUMANS
At present, no objective criteria exist to interpret data from studies
in experimental animals or from short-term tests directly in terms of human
risk. Thus, in the absence of sufficient evidence from human studies, evalua-
tion of the carcinogenic risk to humans was based on consideration of both
the epidemiological and experimental evidence. The breadth of the categories
B-3
-------
of evidence defined above allows substantial variation within each. The de-
cisions reached by the Working Group regarding overall risk incorporated these
differences, even though they could not always be reflected adequately in the
placement of an exposure into a particular category.
The chemicals, groups of chemicals, industrial processes, or occupational
exposures were thus put into one of three groups:
Group 1
The chemical, group of chemicals, industrial process, or occupational
exposure is carcinogenic to humans. This category was used only when there
was sufficient evidence from epidemiological studies to support a causal
association between the exposure and cancer.
Group 2
The chemical, group of chemicals, industrial process, or occupational
exposure is probably carcinogenic to humans. This category includes exposures
for which, at one extreme, the evidence of human carcinogenicity is almost
"sufficient," as well as exposures for which, at the other extreme, it is
inadequate. To reflect this range, the category was divided into higher
(Group A) and lower (Group B) degrees of evidence. Usually, category 2A was
reserved for exposures for which there was at least limited evidence of
carcinogenicity to humans. The data from studies in experimental animals
played an important role in assigning studies to category 2, and particularly
those in Group B; thus, the combination of sufficient evidence in animals and
inadequate data in humans usually resulted in a classification of 2B.
In some cases, the Working Group considered that the known chemical prop-
erties of a compound and the results from short-term tests allowed its trans-
fer from Group 3 to 2B or from Group 2B to 2A.
B-4
-------
Group 3
The chemical, group of chemicals, industrial process, or occupational
exposure cannot be classified as to its carcinogenicity to humans.
B-5
-------
10. REFERENCES
Alguire, D.E. 1973. EthyLsne oxide gas. Food Technology, pp. 64-6?.
Altshuller, A.P. and J.J. Bufalini. 1965. Photochemical aspects of air
pollution: A review. Photochern. Photobiol. 4: 97-146.
Anderson, S.R. 1971. Ethylene oxide toxicity: A study of tissue reactions
to retained ethylene oxide. J. Lab. Clin. Med. 77: 346-356.
Ando, A. 1968. Mutation induction in rice by radiation combined with
chemical protectants and mutagens. International Atomic Energy Agency,
Vienna, Tech. Rep. Ser. 86, pp. 7-15. (Cited in NIOSH, 1977).
Anonymous. 1947. Ethylene oxide poisoning. Ind. Hyg. Newsletter. 7: 3-6.
Anonymous. 1977. New process for preserving wood. Adhesives Age. 5: 56.
Anonymous. 1981 a. Chemical Profiles. Ethylene Oxide. Chemical Marketing
Reporter. June 8, 1981. p. 9.
Anonymous. 198lb. Chemical Profiles. Ethylene GLycol. Chemical Marketing
Reporter. June 15, 1981. p. 9.
Anonymous. 198lc. Key Chemicals. Ethylene Oxide. Chem. Eng. News. July 6,
1981. p. 10.
Anonymous. 1982. Aliphatic Organics Forecast. Chemical Marketing Reporter.
January 4, 1982. p. 33.
Appelgren, L., G. Eneroth and C. Grant. 1977. Studies on ethylene oxide:
Miole body auto radiography and dominant lethal test in mice. Proc. Eur. Soc.
Toxicol. 18: 315-317.
Appelgren, L.E., G. Eneroth, C. Grant, L .E. Lanstrom and K. Tenshasen. 1978.
Testing of ethylene oxide for mutagenicity using the micronucleus test in mice
and rats. Acta. Pharmacol. Toxicol. 43(1): 69-71.
10-1
-------
Balazs, T. 1976. Toxicity of ethylene oxide and chloroethanol. Food and
Drug Administration Bylines. 7: 150-155.
Barnard, J.A. and R.K.Y. Lee. 1972. Combustion of n-pentane in a shock tube.
Combust. Sci. Techno 1. 6(3): 143-150.
Bartnicki, E.W. and C.E. Castro. 1969. Biodehalogenation. Pathway for
transhalogenation and the stereochemistry of epoxide formation from halo-
hydrins. Biochemistry. 8(12): 1677-4680.
BASF Wyandotte Corporation. 1972. Ethylene oxide handling manual. Report
No. 21. BASF Wyandotte Corporation, Wyandotte, Michigan.
Bayes, A.L. 1979. Ethylene oxide one-gene rat ion reproduction study.
(Carnegie -Me lion Institute of Research). Submission to the Environmental
Protection Agency under the Toxic Substances Control Act. Citation number
8EHQ-0379-0248. (Later published as Snelling et al. , 1982).
Ben-Yehoshua, S. and P. Krinsky. 1968. Gas chromatography of ethylene oxide
and its toxic residues. J. Gas Chroma tog. 6: 350-351.
Ben-Yehoshua, S. , S. Staler, A. Goldner and P. Krinsky. 1971. Toxic residues
in Hayani dates after fumigation with ethylene oxide. Pest. Chem. Proc. Intl.
Congr. Pest. Chem. 2nd. 4: 593-605.
Bertsch, W. , R.C. Chang and A. Zlatkis. 1974. The determination of organic
volatiles in air pollution studies. Characterization of profiles. J.
Chromatog. Sci. 12: 175-182.
Binder, H. 1974. Ethylene oxide and chlorohydrin in tobacco and its smoke.
Fachliche Mitt. Oesterr. Tabakregie 1974. 15: 294-301.
Binder, H. and W. Lindner. 1972. Determination of ethylene oxide in the
smoke of treated and untreated cigarettes. Fachliche Mitt. Oesterr.
Tabakregie. 13: 215-220.
Bird, M. 1952. Chemical production of mutations in Drosophila: Comparison
of techniques. J. Genet. 50: 480-485.
10-2
-------
Biro, L., A. Fishes and E. Price. 1974. Ethylene oxide burns, a hospital
outbreak involving 19 womsn. Arch. Dermatol. 110: 924.
Blackford, J.L. 1974. Aery Ion itrile. Chemical Economics Handbook. Stanford
Research Institute, Nfenlo Park, California.
Blackford, J.L. 1976a. Propylene oxide. Chemical Economics Handbook.
Stanford Research Institute, Msnlo Park, California.
Blackford, J.L. 1976b. Ethylene oxide. Chemical Economics Handbook.
Stanford Research Institute, Msnlo Park, California.
Blackwood, J. and E. Erskine. 1938. Carboxide poisoning. U.S. Navy Med.
Bull. 36: 44-45.
Bogan, D.J. and C.W. Hand. 1978. Absolute rate constant, kinetic isotope
effect, and mechanism of the reaction of ethylene oxide with oxygen ( P)
atoms. J. Phys. Chem. 82: 2067-2073.
Bridie, A.L., C.J.M. Wblff and M. Winter. 1979. The acute toxicity of some
petrochemicals to goldfish. Water Res. 13: 623-626.
Bronsted, J.N., M. Kilpatrick and M. Kilpatrick. 1929. Kinetic studies on
ethylene oxides. J. Amer. Chem. Soc. 51: 428-461.
Brown, R.H. and C.J. Pumell. 1979. Collection and analysis of trace organic
vapour pollutants in ambient atmospheres: The performance of a Tenax GC
adsorbent tube. J. Chroma tog. 178: 79-90.
Brown, A.M., C. Bruch, E. Jackson, K. Krell, B. Page and M. Thomas. 1979.
Increased mutation frequency due to ethylene oxide adsorbed to plastics. Iri
vitro. 15: 220-221.
Brown, D. 1970. Determination of ethylene oxide and ethylene chlorohydrin in
plastic and rubber surgical equipment sterilized with ethylene oxide. J.
Assoc. Off. Anal. Chem. 53: 263-267.
Browning, J.B. 1982. Union Carbide Corporation. (Aug. 5) Letter to Office
of Toxic Substances, U.S. Environmental Protection Agency.
10-3
-------
Bruch, C. 1973. Industrial Sterilization. Phillips, G. and W. Miller, eds.
Duke University Press, Durham, North Carolina, pp. 119-123.
Bruch, C. and M. Koesterer. 1961. The microbicidal activity of gaseous
propylene oxide and its application to powdered or flaked foods. J. Food Sci.
26: 428.
Carlson, R.M. and R. Caple. 1977. Chemical/biological implications of using
chlorine and ozone for disinfection. EPA-600/3-77-066, Environmental Research
Laboratory, Duluth, Minnesota.
Carpenter, C., H. Smyth and U. Pozzani. 1949. The assay of acute vapor
toxicity, and the grading and interpretation of results on 96 chemical com-
pounds. J. Ind. Hyg. Toxicol. 31: 3^3-349.
Casteignau, G. and J .L. Halary. 1972. Identification et dosage par
chromatographic enphase gazause d'oxetanues disubstitues en 3 et de quelques
autres ethers-oxydes. Bull. Soc. Chem. Fr. 420-428.
Castro, C.E. and E.W. Bartnicki. 1968. Biodehalogenation. Epoxidation of
halohydrins. Epoxide opening and transhalogenation by a flavobacterium sp.
Biochemistry. 7: 3213-3218.
Cawse, J.N., J.P. Henry, M.W. Swartzlander and'P.H. Wadia. 1980. Ethylene
Oxide. In; Kirk-Othmer Encyclopedia of Chemical Technology, 3rd ed.
Grayson, M. and D. Eckroth, eds. John Wiley and Sons, Inc., New York.
vol. 9, pp. 432-471.
Chemical Week: 1982 Buyers' Guide Issue. 1981. McGraw-Hill, Inc., New York.
p. 417.
Clarke, C. , W. Davidson and J. Johnson. 1966. Hemolysis of blood following
exposure to an Australian manufactered plastic tubing sterilized by means of
ethylene oxide gas. Australia (South) New Zealand J. Surg. 36: 53-55.
Cobis, J. 1977. Correspondence from the Department of Medicine and Surgery
(Code 134D), Veterans Administration, Washington, D. C. , March 17, 1977.
(Cited in Glaser, 1977).
10-4
-------
Cogswell, S.A. 1980. Chemical Economics Handbook. Marketing Research Report
on Ethylene Oxide. SRI International, Menlo Park, CA. January, 1980. pp.
65*1.50323, 654.5032Q, 654.5032U.
Conan, L., B. Foucault, G. Siou, M. Chalgneau and G. LeMoan. 1979. Study of
the mutagenic action of ethylene oxide, ethylene glycol, and 2-chloroethanol
residues in plastic material sterilized by ethylene oxide. Ann. Falsif.
Expert. Chim. 72: 141-151.
Conway, R.A., G.T. Waggy, M.H. Spiegel and R.L. Berglund. 1983.
Environmental fate and effects of ethylene oxide. Environ. Sci. Technol. 17:
107-112.
Crump, K.S., H.A. Guess and L.L. Deal. 1977. Confidence intervals and test
of hypotheses concerning dose-response relations inferred from animal
carcinogenicity data. Biometrics. 33: 437-451.
Crutzen, P.J. and J. Fishman. 1977. Average concentrations of OH in the
troposphere and the budgets of CH., CO, H0, and CH-CC1 Geophys. Res. Lett.
4: 321-324. H d J 5
Gumming, R.B., G.A. Sega, C.Y. Horton and W.H. Olson. 1981. Degree of alkyl-
ation of DMA in various tissues of the mouse following inhalation exposure to
ethylene oxide. Environ. Mutagen. 3(3): 343.
Gumming, R.B., T.A. Michand, L.R. Lewis and W.H. Olson. 1984. Inhalation
mutagenesis in mammals. I. Patterns of unscheduled DNA synthesis induced in
germ cells of male mice by exposure to ethylene oxide in air. (In press).
(Mutagen. Res.).
Cupitt, L.T. 1980. Fate of toxic and hazardous materials in the air
environment. U.S. EPA. EPA-600/3-80-084. NTIS PB 80-221948, 35 pp.
Cupitt, L.T. 1983. Personal communication between Dr. L. Cupitt of ESRL and
Dr. D.A. Gray of Syracuse Research Corporation. March 1983.
Curme, G. and F. Johnston (eds.). 1952. Glycols. American Chemical Society
Monograph, Series 114. Reinhold, New York.
Cvetanovic, R.J. 1955. Mercury photosensitized decomposition of ethylene
oxide. Can. J. Chem. 33: 1684-1695.
10-5
-------
Darnall, K.R., A.C. Lloyd, A.M. Winer and J.N. Pitts, Jr. 1976. Reactivity
scale fbr atmospheric hydrocarbons based on reaction with hydroxyl radical.
Environ. Sci. Technol. 10: 692-696.
DeBont, J.A.M. and R.A.J.M. Albers. 1976.
An ton ie van Leeuwenhoek. 42(1): 78-80.
Microbial metabolism of ethylene.
Billing, W.L. 1977. Interphase transfer processes.
chloro me thanes, ethanes, ethylenes, propanes, and
aqueous solutions. Comparisons with theoretical predictions.
Technol. 11: 405-409.
II. Evaporation rates of
propylene from dilute
Environ. Sci.
Dimitriades, B. and S.B. Joshi. 1977. Application of reactivity criteria in
ox id ant-re la ted emission control in the U.S.A. In: Inter. Conf. on
Photochemical Oxidant Pollution and Its Control. Dimitriades, B., ed.
Research Triangle Park, North Carolina, pp. 705-711. EPA-600/3-77-011B.
Dobbs, A.J., B.C. Gilbert and R.O.C. Norman. 1971. Electron spin resonance
studies. Part XXVII. The geometry of oxygen-substituted alkyl radicals. J.
Chem. Soc. A: 124-135.
Dobinaon, B., W. Hoffman and B.D. Stack.
Groups. Pergamon Press, Oxford.
1969. The Determination of Epoxide
Dolovich, J. and B. Bell. 1978. Allergy to a produet(s)
gas: Demonstration of IgE and IgG antibodies and hapten
Allerg. Clin. Immunol. 62: 30-32.
of ethylene oxide
specificity. J.
Dow Chemical Co. 1977. Alkylene oxides from Dow.
110-551-77R, Midland, Michigan.
Dow Chemical Co. Form No.
Dunkelberg, H. 1979. On the oncogenic activity of ethylene oxide and
propylene oxide in mice. Br. J. Cancer. 39(4): 588-589.
Dunkelberg, H. 1981. Kanzerogene Aktivitat von Sthylenoxid und seinen
Reaktions-produkten 2-Chloroethanal, 2-Bromoethanol, Ethylenglykol und
Diethylenglykol. T. Kanzerogenitat von Ethylenoxid in Vergleich zu 1,2-
Propylenoxid bei Subkutaner Applikation an Mausen. Zbl. Bakt. Hyg. I. Abt.
Orig. B 174: 383-404.
10-6
-------
Dunkelberg, H. 1982. Carcinogenicity of ethylene oxide and 1 ,2-propylene
oxide upon intragastric administration to rats. Br. J. Cancer. 46: 924-933.
Eastham, A.M. and G.A. Latremouille. 1952. Ethylene oxide V reaction with
halide ions in neutral and acid solution. Can. J. Chem. 30: 169.
Ehrenberg, L. and A. Gustaffson. 1957. On the mutagenic action of ethylene
oxide and diepoxybutane in barley. Hereditas. 43: 595-602.
Ehrenberg, L. and T. Hallstrom. 196?. Haema to logic studies on persons occu-
pational ly exposed to ethylene oxide. In; International Atomic Energy Agency
Report SM 92/96, pp. 327-334.
Ehrenberg, L. and S. Hussain. 1981. Genetic toxicity of some important
epoxides. Mutat. Res. 86: 1-113.
Ehrenberg, L., A. Gustaffson and U. Lundquist. 1959. The mutagenic effect of
ionizing radiation and reactive ethylene derivatives in barley. Hereditas.
45: 351-368. (Cited in NIOSH, 1977).
Ehrenberg, L., K. Hiesche, S. Osterman-Golker and I. Wennberg. 1974. Evalua-
tion of genetic risks of alkylating agents: Tissue doses in the mouse from
air contaminated with ethylene oxide. Mutat. Res. 24: 83-103.
Ehrenberg, L. , A. Gustaffson and U. Lundqvist. 1956. Chemically induced
mutation and sterility in barley. Acta. Chem. Scand. 10: 492-494. (Cited in
NIOSH, 1977).
Embree, J. and C. Hine. 1975. Mutagenicity of ethylene oxide. Toxicol.
Appl. Pharmacol. 33: 172-173.
Embree, J., J. Lyon and C.H. Hine. 1977. The mutagenic potential of ethylene
oxide using the dominant lethal assay in rats. Toxicol. Appl. Pharmacol. 40:
261-267.
EOIC (Ethylene Oxide Industry Council of the Chemical Manufacturer's
Association). 1984. Research directed at the development of a sensitive
analytical method for the quantification of ethylene oxide and its metabolites
in biological samples. Final report. Prepared by Arthur D. Little, Inc.,
Cambridge, MA.
10-7
-------
Faberge, A.C. 1955. Types of chromosome aberrations induced by ethylene
oxide in maize. Genetics. 40: 571. (Cited in NIOSH, 1977).
Fahmy, 0. and M. Fahmy. 1970. Gene elimination in carcinogenesis: Reinter-
pretation of the somatic nutation theory. Cancer Res. 30: 195-205.
Fishbein, L. 1976. Potential hazards of fumigant residues. Environ. Health
Perspect. 14: 39-45.
Fomenko, V.N. and E.E. Strekalova. 1973. Mutagenic action of some industrial
poisons as a function of concentration and exposure time. Toksikol. Nov.
Prom. Khim. Veshchestv. 13: 51-57.
Francis, M. 1985. Letter to Herman Gibb, Carcinogen Assessment Group, U.S.
Environmental Protection Agency. Subject: Search request on methyl formate,
107-31-3. Environmental Mutagen, Carcinogen, and Teratogen Information
Program, Oak Ridge National Laboratory, Oak Ridge, TN.
Freed, V.H., C.T. Chiou and R. Haque. 1977. ChenDdynamics: Transport and
behavior of chemicals in the environment - A problem in environmental health.
Environ. Health Per spec. 20: 55-70.
Fritz, B., K. Lorenz, W. Steinert and R. Zellner. 1982. Laboratory kinetic
investigations of the tropospheric oxidation of selected industrial emissions.
Comm. Eur. Communities (Rep.) Eur. Phys. -Chem. Behav. Atmos. Pollut. 7624:
192-202.
Gaines, Personal Communication. Union Carbide Corporation. 203-794-2000.
Garry, V.V. , J. Hozier, D. Jacobs, R.L. Wade and D.G. Gray. 1979. Ethylene
oxide: Evidence of human chromosomal effects. Environ. Muta. 1: 375-382.
Garry, V.F., C.W. Opp, J.K. Wiencke and D. Lakatua. 1982. Ethylene oxide
induced sister chromatid exchange in human lymphocytes using a membrane
dosimetry system. Pharmacol. 25: 214-221.
Garry, V .F., C.W. Opp and J.K. Wiencke. 1981. Ethylene-oxide induced sister
chromatid exchange in human-lymphocytes using a membrane dosimetry system.
Environ. Mutagen. 3(3): 340.
10-8
-------
Generoso, W.M. , K.T. Cain, M. Krishna, C.W. Sheu and R.M. Gryder. 1980.
Heritable trans location and dominant-lethal nutation induction with ethylene
oxide in mice. Mil tat. Res. 73(1): 133-142.
Gilmour, R.H. 1978. Ethylene oxide treatment of cosmetics. Soap Perfum.
Cosnet. 51(11): 498-499.
Corner, R. and W.A. Noyes. 1950. Photochemical studies. XLJI. Ethylene
oxide. J. Am. Chem. Soc. 72: 101-108.
Goncarlovs, V. 1983. Personal communication between Dr. D.A. Gray of
Syracuse Research Corporation and V. Goncarlovs of the Office of Pesticides.
March, 1983.
Graedel, T.E. 1978. Chemical Conpounds in the Atmosphere. Academic Press,
New York. pp. 267, 272, 276.
Greaves-Walker, W. and C. Greeson. 1932. The toxicity of ethylene oxide. J.
Hyg. 32: 409-416.
Gritter, R.J. and E.G. Sabatino. 1964. Free-radical chemistry of cyclic
ethers. VII. Ultraviolet photolysis of epoxides and propylene sulfide in the
liquid phase. J. Org. Chem. 29: 1965-1967.
Gross, J., M. Haas and T. Swift. 1979. Ethylene oxide neurotoxicity: Report
of four cases and review of the literature. Neurology. 29: 978-982.
Hackett, P.L., M. G. Brown, R .L. Buschbom et al. 1982. Teratogenic study of
ethylene and propylene oxide and n-butyl acetate. Battelle, Pacific Northwest
Laboratories. Prepared for NIOSH, contract number 210-80-0013.
Hanson, C. and A.C. Leo. 1979. Substituent Constants for Correlation
Analysis in Chemistry and Biology. John Wiley and Sons, Inc., New York.
p. 176.
Hatch, G. G. , P.D. Manay, C. C. Christensen, R. Lagenbach, C.R. Goodheart and S.
Nesnow. 1982. Enhanced viral transformation of primary Syrian hamster lung
(V-79) cells exposed to ethylene oxide in sealed treatment chambers. Proc.
Am. Assoc. Cancer Res. 23: 74.
10-9
-------
Hawley, G.G. (ed. ). 1981. The Condensed Chemical Dictionary, 10th ed. , Van
No strand Reinhold Company, New York. p. 435.
Hemminki, K. , P. Mutinen, I. Saloniemi, M.L. Niemi and H. Vainis. 1982.
Spontaneous abortions in hospital staff engaged in sterilizing instruments
with chemical agents. Br. Med. J. 285: 1461-1463.
Hendry, D.G. , T. Mill, L. Piszkiewicz, J .A. Howard and H.K. Eigenmann. 1974.
A critical review of H-atom transfer in the liquid phase. J. Phys. Chem.
Ref. Data. 3: 937-978.
Hine, C. and V. Rowe. 1973. Industrial Hygiene and Toxicology. 2nd ed.
vol. 2, Interscience, New York. pp. 1600-1601.
Hine, C. , V.K. Rowe, E.R. White, K.I. Darmer and G.T. Youngblood. 1981.
Expoxy Compounds. In: Toxicology. Clayton, G.D. and F.E. Clayton, Eds.
Volume 2A of Patty's Industrial Hygiene and Toxicology. Third revised
edition. John Wiley and Sons, Inc., NY. pp. 2141-2258.
Hirose, T., R. Goldstein and C. Bailey. 1963. Hemolysis of blood due to
exposure to different types of plastic tubing and the influence of ethylene
oxide sterilization. J. Thorac. Cardiov. Surg. 45: 245-251.
Hirose, C. 1974. The microwave spectra and r0, rs, and rm structures of
ethylene oxide. Bull. Chem. Soc. Japan. 47: 1311-1318.
Hogstedt, C., 0. Rohlen, B.S. Berndtson, 0. Axelson and L. Ehrenberg. 1979a.
A cohort study of mortality and cancer incidence in ethylene oxide production
workers. Br. J. Ind. Med. 36: 276-280.
Hogstedt, C. , N. MaMgrist and B. Wadman. 1979b. Leukemia in workers exposed
to ethylene oxide. J. Am. Med. Assoc. 241(11): 1132-1133.
Hogstedt, C. , L. Aringer and A. Gustavasson. 1984. Ethylene oxide and
cancer: A literature review and follow-up of two epidemiologic studies.
Accepted for publication in Arbete och Halse.
Hollingsworth, R. , V. Rowe, F. Oyen, D. McCollister and H. Spencer. 1956.
Toxicity of ethylene oxide determined on experimental animals. Am. Med.
Assoc. Arch. Ind. Health. 13: 217-227.
10-10
-------
Holmberg, P.C. 1979. Central-nervous-system defects in children born to
mothers exposed to organic solvents. Lancet. 2: 8135: 177-179.
HoMberg, P.C. and M. Nurminen. 1980. Congenital defects of the nervous
system and occupational factors during pregnancy. A case-re fere nee study.
Am. J. Ind. fed. 1: 167-176.
Holmgren, A., N. Diding and G. Samuelsson. 1969. Ethylene oxide treatment of
crude drugs. Part V. Formation of ethylene chlorohydrin. Acta Pharmacol.
Sciencia. 6: 33-36.
Howe, R.B. 1983. GLOBAL 83. An experimental program developed for the U.S.
Environmental Protection Agency as an unpdate to GLOBAL 82: A computer program
to extrapolate quantal animal toxicity data to low doses (May 1982). K.S.
Crump and Co. , Inc. , Ruston, LA. Unpublished.
Hughes, K.J., R.W. Hum and F.G. Edwards. 1959. Separation and identifica-
tion of oxygenated hydrocarbons in combustion products from automotive
engines. In: Proc. 2nd Int. Symp. on Gas Cnromatography, East Lansing,
Michigan, p. 171-182.
Hughes, T.W. , D.R. Tierney and Z.S. Khan. 1979. Measuring fugitive emissions
from petrochemical plants. Chem. Eng. Prog. August 1979, pp. 35-39.
Hussain, S. and L. Ehrenberg. 1975. Prophage inductive efficiency of alkyl-
ating agents and radiations. Intl. J. Radiat. Biol. Relat. Stud. Phys. Chem.
Med. 27: 355-362.
I ARC (International Agency for Research on Cancer) 1976. Ethylene oxide.
In; IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to
Man. Lyon, France.
IARC (International Agency for Research on Cancer) 1982. IARC Monographs:
Evaluation of the Carcinogenic Risk of Chemicals to Humans. Supplement 4.
Lyon, France.
Ishishi, E. and Y. Kobayashi. 1973. Treatment of a waste solution obtained
in the manufacture of epoxide by chlorohydrination. Japan Kokai 73, 10,012,
5 pp.
10-11
-------
Jacobson, K., E. Hackley and L. Feinsilver. 1956. The toxicity of inhaled
ethylene oxide and propylene oxide vapors. Arch. Ind. Health. 13: 237-2M.
Jana, M.K. and K. Roy. 1975. Effectiveness and efficiency of methyl methane-
sulfonate and ethylene oxide for the induction of mutations in mice. Mutagen.
Res. 28(2): 211-215. (Cited in NIOSH, 1977).
Jay, W.M., T.R. Swift and D.S. Hull.
oxide exposure to cataract formation.
1982. Possible relationship of ethylene
Am. J. Ophthalmol. 93(6): 727-732.
Jefferson Chemical Co.
Houston, Texas.
Undated a. Ethylene oxide technical brochure.
Jefferson Chemical Co.
Houston, Texas.
Undated b. Propylene oxide technical brochure.
Jensen, E. 1977. Union Carbide Corp. (unpublished).
Register, 43: 3812, 1978).
(Cited in Federal
Jenssen, D. and C. Ramel. 1980. The micronucleus test as part of a short-
term mutagenicity test program for the prediction of carcinogenicity evaluated
by 143 agents tested. Mutat. Res. 75(2): 191-202.
Johnson and Johnson. 1982. Preliminary report of pilot research chromosome
study of workers at sites where ethylene oxide gas is utilized as a sterilant.
Unpublished report available from Dr. J. Paul Jones, Director of Health
Sciences, Johnson and Johnson, New Brunswick, NJ, p. 18.
Jones, A.R. and G. Wells. 1981. The comparative metabolism of 2-bromoethanol
and ethylene oxide in the rat. Xenobiotica. 11(11): 763-770.
Jones, L.A. and D.M. Adams. 1981. Mutations in Bacillus subtilis var. niger
(Bacillus globigii) spores induced by ethylene oxide. Sporulation
Germination, Proc. Int. Spore Conf. VIII. Levinson, H.S., A.L. Sonenshein,
and D.J. Tipper, Eds. American Society of Microbiology, Washington, DC.
Jones-Price, C., C.A. Kimmel, T.A. Marks, T.A. LeDoux, J.R. Reel, P.W. Fisher,
L. Langhoff-Paschke and M. Marr. Teratologic evaluation of ethylene oxide
(CAS No. 75-71-8) in New Zealand White rabbits. Final report (RB80--EO).
NIEHS Contract N01-ES-2127.
10-12
-------
Joshi, S.B., M.C. Dodge and J.J. Bufalini. 1982. Reactivities of selected
organic compounds and contamination effects. Atmos. Environ. 16: 1301-1310.
Joyner, R. 196*1. Chronic toxicity of ethylene oxide. Arch. Environ.
Health. 8: 700-710.
Jukes, T.H. 1964. Choline. Kirk-Othmer Encyclopedia of Chemical Technology,
2nd ed. Interscience, New York. 5: 405.
Kadoi, M. and I. Kataoka. 1968. Purification of a waste liquor from produc-
tion of an oxirane compound. Japan Patent 74, 25,119, 5 pp.
Kauhanen, K. 1977. Stanford Research Institute (unpublished). (Cited in
Federal Register, 43: 3912, 1978).
Klapproth, E.M. 1976. Ethylene glycol. Chemical Economics Handbook.
Stanford Research Institute, Menlo Park, California.
Kligerman, A.D., G.L. Erenson, M.E. Phelps and J.L. Wilmer. 1983. Sister-
chroraatid exchange induction in peripheral blood lymphocytes of rats exposed
to ethylene oxide by inhalation. Mutat. Res. 120: 37-44.
Kolmark, H.G. and B.J. Kilbey. 1968. Kinetic studies of mutation induction
by epoxides in Neurospora crassa. Mol. Gen. Genet. 101: 89-98.
Koskikallio, J. and E. Whalley. 1959. Effect of pressure on the spontaneous
and the base-catalyzed hydrolysis of epoxides. Can. J. Chem. 37: 783-787.
Krell, K. , E.D. Jacobson and K. Selby. 1979. Mutagenic effect on L5178Y
mouse lymphoma cells by growth in ethylene oxide-sterilized polycarbonate
flasks. In Vitro. 15(5): 326-328.
Kurginski, E.R. 1979. Personal Communication to C. Glasgow, U.S.
Environmental Protection Agency, October 22. Dow Chemical Co., Midland,
Michigan.
LaBorde, J.B. and C.A. Kimmel. 1980. The teratogenicity of ethylene oxide
administered intravenously to mice. Toxicol. Appl. Pharmacol. 56(1): 16-22.
10-13
-------
LaBorde, J.B., C.A. Kimmel, C. Jones-Price, T.A. Marks, and T.A. Ledoux.
1982. Teratogenic evaluation of ethylene chlorohydrin (ECH) in mice and
rabbits. Society of Toxicology Meeting, Boston, MA.
Lamty, G., R. Maleq, C. Selve, A. Sivade and J. Wylde. 1975. Mechanisms of
ring opening of oxirans by acids in aqueous and nonaqueous solvents. J.
Chem. Soc. Perkin II: 1119-1124.
Lambert, B. and A. Lindblad. 1980. Sister chromatid exchange and chromosome
aberrations in lymphocytes of laboratory personnel. J. Toxicol. Environ.
Health. 6(5-6): 1237-1243.
Landels, S.P. 1976. Fumigants and nematicides. Chemical Economics Handbook.
Stanford Research Institute, Menlo Park, California.
Lapkin, M. 1965. Epoxides. Kirk-Othmer Encyclopedia of Chemical Technology,
2nd ed. Interscience, New York. 8: 288-289.
Laurent, C.H., J. Frederic and F. Marechal. 1982. Etude des effects cyto-
genetiques d'intoxication a 1'oxyde d'ethylene. C.R. Soc. Biol. 176:
733-735.
Lee, W.R. Unpublished. Final report on project "Relation of mutation
frequency to dose of ethylene oxide in germ cells." Submitted to South
Carolina Pesticides Epidemic logic Study Center, Preventive Medicine Division,
Department of Family Practice, Medical University of South Carolina. 12 pp.
Lichtenstein, H.J. and G.H. Twigg. 1948. Trans. Faraday Soc. 44: 905.
(Cited in Mabey and Mill, 1978).
Lichtenwalter, G.D. and G.H. Riesser. 1964. Chlorohydrins. Kirk-Othmer
Encyclopedia of Chemical Technology, 2nd ed. 5: 304-324.
Liepins, R., F. Mixon, C. Hudak and T.B. Parsons. 1977. Industrial process
profiles for environmental use: Chapter 6. The Industrial Organic Chemicals
Industry. Washington, D. C.: U.S. Environmental Protection Agency. Avail.
NTIS PB281-478. pp. 6-370 to 6-373.
Lindgren, D. and L. Vincent. 1966. In: Advances in Pest Control Research.
vol. 5, Metcalf, R., ed. Interscience, New York. 85 pp.
10-14
-------
Lindgren, D. and K. Sulovska. 1969. The mutagenic effect of low concentra-
tions of ethylene oxide in air (abstract only). Hereditas. 63: 460. (Cited
in NIOSH, 1977).
Lindgren, D. , L. Vincent and H. Krohne. 1954. Relative effectiveness of ten
fumigants to adults of eight species of stored-product insects. J. Econ.
Entomol. 47: 923-926.
Lindgren, D.L., W.B. Sinclair and L.E. Vincent. 1968. Residues in raw and
processed foods resulting from post-harvest insecticidal treatment. Residue
Reviews. 21: 93-95.
Logan, J.A., M.J. Prather, S.C. Wofsy and M.B. McElroy. 1981. Tropospheric
chemistry: A global perspective. J. Geophys. Res. 86: 7210-7254.
Long, F.A. and J.G. Pritchard. 1956. Hydrolysis of substituted ethylene
oxides in H?0^° solutions. J. Araer. Chem. Soc. 78: 2663-2667.
Long, F.A., J.G. Pritchard and F.E. Stafford. 1957. Entropies of activation,
and mechanism for the acid-catalyzed hydrolysis of ethylene oxide and its
derivatives. J. Amer. Chem. Soc. 79: 2362-2364.
Loveless, A. and P. Wheat ley. 1966. In; Genetic and Allied Effects of
Alkylating Agents. Pennsylvania State University Press, University Park,
Pennsylvania, vol. 131.
Lynch, D.W., T.R. Lewis, W.J. Moorman, P.S. Sabharwal and J.R. Burg. 1982a.
Chronic inhalation toxicity of ethylene oxide in rats and monkeys. A
Preliminary Report. Presented at the 21st Annual Society of Toxicology
Meeting, Boston, MA, February 22-26, 1982. National Institute for
Occupational Safety and Health, Division of Biomedical and Behavioral Science,
4676 Columbia Parkway, Cincinnati, OH.
Lynch, D. , T. Lewis, W. Moorman, et al. 1982b. Carcinogenic and toxicologic
effects of inhaled ethylene oxide and propylene oxide in F344 rats. National
Institute for Occupational Safety and Health, Washington, D. C. (Unpublished
draft.)
Lynch, D.W. , T.R. Lewis, W.J. Moorman, P.S. Sabharwal, J.B. Lai and J.R. Burg.
1983. Toxic and mutagenic effects of ethylene oxide and propylene oxide on
spermatogenic function in Cynomologus monkeys. Society of Toxicology Meeting.
Las Vegas, NV.
10-15
-------
Mabey, W. and T. Mill. 1978. Critical review of hydrolysis of organic com-
pounds in water under environmental conditions. J. Phys. Chem. Ref. Data.
7(2): 383-415.
Mackay, D. and P.J. Leinonen. 1975. Evaporation of low-solubility contam-
inants from water bodies to atmosphere. Environ. Sci. Technol. 9: 1178-1180.
Mackey, J. 1968. Mutagenesis in volgare wheat. Hereditas. 59: 505-517.
(Cited in NIOSH, 1977).
Mantel, M. and M.A. Schneiderman. 1975. Estimating "safe" levels, a
hazardous undertaking. Cancer Res. 35: 1379-1386.
March, J. 1977. Advanced Organic Chemistry: Reactions, Mechanisms, and
Structure. 2nd ed. McGraw-Hill Book Co., New York.
Margosches, E.H. May 13, 1983. Memorandum to the Record U.S. Environmental
Protection Agency. EtO and the study of spontaneous abortions published in
the British Medical J. 20 November 1982.
Martis, L., R. Kroes, T.D. Darby and E.F. Woods. 1982. Disposition kinetics
of ethylene oxide, ethylene glycol and 2-chlorethanol in the dog. J. Toxicol.
Environ. Health. 10: 847-856.
Matter, B.E. and I. Jaeger. 1975. The cytogenetic basis of dominant lethal
mutations in mice. Studies with TEM, EMS, and 6-mercaptopurine. Mutat. Res.
33: 251-260.
MCA (Manufacturing Chemists Association). 1971. Chemical Safety Data Sheet
SD-38, Ethylene Oxide. Manufacturing Chemists Association, Washington, D.C.
McDonald, T., K. Kasten, R. Hervey, S. Gregg, A. Borgmann and T. Hurchison.
1977. Acute ocular toxicity for normal and irritated rabbit eyes and subacute
ocular toxicity for ethylene oxide, ethylene chlorohydrin, and ethylene
glycol. Bull. Parenter. Drug Assoc. 31: 25-32.
McLaughlin, R. 1946. Chemical burns of the human cornea. Am. J. Ophthal.
29: 1355-1362.
10-16
-------
Mehra, P.N. and S.K. Mann. 1975. Cytological effects of chemical mutagens on
Pterotheca falconeri. Part 2. Monofunctional alkylating agents. Nucleus
(Calcutta). L7(3): 167-182. (Cited in NIOSH, 1977).
Migliore, L., A.M. Rossi and N. Loprieno. 1982. Mutagenic action of
structurally related alkene oxides on Schizosaccharomyces pombe; The
influence, 'in vitro', of mouse-liver metabolizing system. Mutat. Res. 102:
425-437.
Mishmash, H.E. and C.E. Meloan. 1972. Indirect spectrophotometric determina-
tion of nanomole quantities of oxiranes. Anal. Chem. 44: 835-836.
Morgan, R.W., K.W. Claxton, B.J. Divine, S.D. Kaplan and V.B. Harris. 1981.
Mortality among ethylene oxide workers. J. Occup. Med. 23(11): 767-770.
Morris, J.C. 1975. Formation of halogenated organics by chlorination of
water supplies. EPA 600/1-75-002, U.S. Environmental Protection Agency,
Washington, D.C.
Moutschen, J., M. Moutschen-Dahmen and L. Ehrenberg. 1968. Note on the
chromosome breaking activity of ethylene oxide and ethyleneimine. Hereditas.
60: 267-269. (Cited in NIOSH, 1977).
NAS (National Academy of Science). 1976.
National Academy of Sciences, Washington, DC.
Vapor-phase organic pollutants.
NFPA (National Fire Prevention Association). 1975. Hazardous chemicals data.
Fire Protection Guide on Hazardous Materials. Boston, Massachusetts.
NIOSH (National Institute for Occupational Safety and Health). 1977. Special
occupational hazard review with control recommendations for the use of
ethylene oxide as a sterilant in medical facilities. Glaser, Z., 67 pp.,
Department of Health, Education, and Welfare Publication No. 77-200, National
Institute for Occupational Safety and Health, Cincinnati, Ohio.
NIOSH (National Institute for Occupational Safety and Heath).
Current Intelligence Bulletin No. 35. Ethylene Oxide (EtO):
Carcinogenicity, May 22, 1981.
1981. NIOSH
Evidence of
NTP (National Toxicology Program). 1981. National Toxicology Program, 1980.
Department of Health and Human Services. U.S. Public Health Service.
10-17
-------
Ong, E. 1948. Chemistry and Uses of Insecticides. Reinhold Publishing Co.,
New York. p. 150.
Oxirane Corporation. Undated. Propylene Oxide Technical Bulletin, Houston,
Texas.
Patty, F.A. (ed.). 1973- Industrial Hygiene and Toxicology. 2nd revised
edition, vol. 2. Interscience, New York. p. 1626.
Pellizzari, E.D., J.E. Bunch, R.E. Berkley and J. McRae. 1976. Collection
and analysis of trace organic vapor pollutants in ambient atmospheres. The
performance of a Tenax GC cartridge sampler for hazardous vapors. Anal.
Letters. 9: 45-63.
Pero, R.W., T. Bryngelsson, B. Widegren, B. Hogstedt and H. Welinder. 1982.
A reduced capacity for unscheduled DNA synthesis in lymphocytes from
individuals exposed to propylene oxide and ethylene oxide. Mutat. Res. 104:
193-200.
Pero, R.W., B. Widegren, B. Hogstedt and F. Mitelraan. 1981. In vivo and in
vitro ethylene oxide exposure of human lymphocytes assessed by chemical
stimulation of unscheduled DNA synthesis. Mutat. Res. 83: 271-289.
Pessayre, D. and A. Trevoux. 1978. Collapse after endovascular exploration:
The role of ethylene oxide or endotoxin? LaNouvelle Presse Medicale.
39: 3558.
Pfeiffer, E. and H. Dunkelberg. 1980. Mutagenicity of ethylene oxide and
propylene oxide and of the glycols and halohydrins formed from them during the
fumigation of foodstuffs. Toxicol. 18: 115-118.
Pfeilsticker, K., G. Fabricius and G. Timme. 1975. Simultaneous gas-chroma-
tographic determination of ethylene oxide, ethylene chlorohydrin, and ethylene
glycol in grain. Z. Lebensm-Unters. Forsch. 158: 21-25.
Phillips, C. and S. Kaye. 1949. The sterilizing action of gaseous ethylene
oxide. Part I. Review. Am. J. Hyg. 50: 270-279.
Poirier, V. and D. Papadopoulo. 1982. Chromosomal abberations induced by
ethylene oxide in a human amniotic cell line in vitro. Mutat. Res. 104:
255-260.
10-18
-------
Poothullil, J., A. Shimizu, R. Day and J. Dolovich. 1975. Anaphylaxis from
the product(s) of ethylene oxide gas. Ann. Intern. Med. 82: 58-60.
Pritchard, J.G. and F.A. Long. 1956. Hydrolysis of ethylene oxide deriva-
tives in deuterium oxide-water mixtures. J. Amer. Chera. Soc. 78: 6008-6013.
Pritchard, J.G. and I.A. Siddiqui. 1973. Activation parameters and mechan-
isms of the acid-catalyzed hydrolysis of epoxides. J. Chem. Soc. Perkin
II: 452-457.
Qazi, A.H. and N.H. Ketcham. 1977. A new method for monitoring personal
exposure to ethylene oxide in the occupational environment. Am. Ind. Hyg.
Assoc. J. 38: 635-647.
Rannug, U., R. Goethe and C.A. Wachtmeister. 1976. The mutagenicity of
chloroacetaldehyde, chloroethylene oxide, 2-chloroethanol, and chloroacetic
acid, conceivable metabolites of vinyl chloride. Chem. Biol. Interact.
12: 251-263.
Rathburn, R.E. and D.Y. Tai. 1981. Technique for determining the volatili-
zation coefficients of priority pollutants in streams. Water Res. 15: 243-
250. (Cited in Conway et al., 1983).
Reyniers, V., M. Sacksteder and L. Ashburn. 1964. Multiple tumors in female
germ-free inbred albino mice exposed to bedding treated with ethylene oxide.
J. Natl. Cancer Inst. 32: 1045-1057.
Roberts, J., L. Allison, P. Prichett and K. Riddle. 1943. Preliminary
studies on soil sterilization with ethylene oxide. J. Bacteriol. 45: 40.
Romano, S.J. and J.A. Renner. 1975. Comparison of analytical methods for
residual ethylene oxide analysis. J. Pharmacol. Sci. 64: 1412-1417.
Romano, S.J., J.A. Renner and P.M. Leitner. 1973. Gas chromatographic deter-
mination of residual oxide by head space analysis. Anal. Chem.
45: 2327-2330.
Rowell, R.M. and D.I. Gutzrner. 1975. Chemical modification of wood: Reac-
tions of alkylene oxides with southern yellow pine. Wood Sci. 7: 240-246.
10-19
-------
Rowell, R.M., D.I. Gutzmer, I.E. Sadio and R.E. Kinney. 1976. Effects of
alkylene oxide treatments on dimensional stability of wood. Wood Sci.
9: 51-54.
Rowlands, D.G. 1971. The metabolism of contact insecticides in stored
grains. II. 1966-1969. Residue Reviews. 34: 91-161.
Roy, K. and M.K. Jana. 1973. Chemically induced chlorophyll mutation in
rice. Indian Agric. 17(4): 301-308. (Cited in NIOSH, 1977).
Royce, A. and W. Moore. 1955. Occupational dermatitis caused by ethylene
oxide. Br. J. Ind. Med. 12: 169.
Salinas, E., L. Sasich, D.H. Hall, R.M. Kennedy and H. Morriss. 1981. Acute
ethylene oxide intoxication. Drug Intel. Clinical Pharmacy. 15(5): 384-386.
Schnorr, T.M. 1982. The health of health care workers: A proportionate
mortality study. A thesis in epidemiology presented to the graduate faculties
of the University of Pennsylvania in partial fulfillment of the requirements
for the degree of Master of Science.
Schultze, H.C. 1965. Ethylene Oxide. Kirk-Othmer Encyclopedia of Chemical
Technology, 2nd ed. Interscience, New York. 8: 523-558.
Scudamore, K.A. and S.G. Heuser. 1971. Ethylene oxide and its persistent
reaction products in wheat flour and other commodities: Residues from fumiga-
tion or sterilization, and effects of processing. Pest. Sci. 2(2): 80-91.
Sega, G.A., R.B. Gumming, J.G. Owens, C.Y. Horton and L.R. Lewis. 1981.
Alkylation pattern in developing mouse sperm, sperm DNA, and protamine after
inhalation of ethylene-oxide. Environ. Mutagen. 3(3): 371.
Seizinger, D.E. and B. Dimitriades. 1972. Oxygenates in exhaust from simple
hydrocarbon fuels. J. Air Pollut. Contr. Assoc. 22(1): 47-51.
Sexton, R. and E. Henson. 1949. Dermatological injuries by ethylene oxide.
J. Ind. Hyg. Toxicol. 31: 297-300.
Sexton, R. and E. Henson. 1950. Experimental ethylene oxide human skin
injuries. Arch. Ind. Hyg. Occup. Health. 2: 549-564.
10-20
-------
Shenderova, R.I., G.K. Alsopova, I.E. Shaklinazarov and V.A. Linetskii. 1972.
Purification of waste waters from simultaneous production of propylene oxide
and styrene oxide. Khirn. Prom. (Moscow). 48: 746-7^7.
Shupack, J.L., S.R. Andersen and S.J. Romano. 1981. Human skin reactions to
ethylene oxide. J. Lab. Clin. Med. 98(5): 723-729.
Sickles, J.E., II, R.S. Wright, C.R. Sutcliff, A.L. Blackard and D.P. Dayton.
1980. Smog chamber studies of the reactivity of volatile organic compounds.
Proc. Ann. Meet. Air Pollut. Contr. Assoc. Paper 80-501, 16 pp.
Sittig, M. 1962. Organic Chemical Processes. The Noyes Press, Pearl River,
New Jersey.
Sittig, M. 1965. Acrylic Acid and Esters. Noyes Development Corporation,
Park Ridge, New Jersey.
Skeehan, A. 1959. A portable sterilizer for the application of ethylene
oxide gas in ophthalmology. Am. J. Ophthalraol. 47: 86.
Smith, H.H. and T.A. Lotfy. 1954. Comparative effects of certain chemicals
on Tradescantia chromosomes as observed at pollen tube mitosis. Am. J. Bot.
41: 589-593. (Cited in NIOSH, 1977).
Smyth, H., J. Seaton and L. Fischer. 1941. Single dose toxicity of some
glycols and their derivatives. J. Ind. Hyg. Toxicol. 23: 259-268.
Snellings, W., J. Pringle, et al. 1979. Teratology and reproduction studies
with rats exposed to 10, 33, or 100 ppm of ethylene oxide. Abstract, Society
of Toxicology, 18th meeting, New Orleans, March 1979.
Snellings, W.M., C.S. Weil and R.R. Maronpot. 1981. Final report, ethylene
oxide, two years inhalation study. Submitted to the U.S. Environmental
Protection Agency by Bushy Run Research Center, Pittsburg, PA.
Snellings, W.M., J.P. Zelenak and C.S. Weil. 1982a. Effects on reproduction
in Fischer 344 rats exposed to ethylene oxide by inhalation for one
generation. Toxicol. Appl. Pharmacol. 63: 382-388.
10-21
-------
Snellings, W.M., R.R. Maronpot, J.P. Zelenak and C.P. Laffoon. 1982b.
Teratology study in Fischer 344 rats exposed to ethylene oxide by inhalation.
Toxicol. Appl. Pharmacol. 64: 476-481.
Snellings, W.M., C.S. Weil and R.R. Maronpot. 1984a. A subchronic inhalation
study on the toxicologic potential of ethylene oxide in B6C3F1 mice. Toxicol.
Appl. Pharmacol. 76: 510-518.
Snellings, W.M., C.S. Weil and R.R. Marenpot. 1984b. A two-year inhalation
study of the carcinogenic potential of ethylene oxide in Fischer 344 rats.
Toxicol. Appl. Pharmacol. 75: 105-117.
Spasovski, M., V. Khristeva, K. Pernov, V. Kirkov, T. Dryanovska, Z. Panova,
G. Bobev, N. Gincheva and S. Ivanova. 1980. Health state of the workers in
the production of ethylene and ethylene oxide. Khig. Zdraveopaz.
23(1): 41-47.
Spencer, E.F., Jr. 1971. Pollution control in the chemical industry,
Chapter 14. In; Industrial Pollution Control Handbook, H.F. Lund, ed.
McGraw-Hill Book Co., New York.
SRI International (Stanford Research Institute International). 1977. A Study
of Industrial Data on Candidate Chemicals for Testing. Contract No.
68-01-4109, U.S. Environmental Protection Agency, Office of Toxic Substances,
Washington, D.C.
SRI International (Stanford Research Institute International). 198la. 1981
Directory of Chemical Producers: United States of America. Menlo Park,
California: SRI International, pp. 594-595.
SRI International (Stanford Research Institute International). 198lb. 1981
Directory of Chemical Producers: United States of America. Supplement II.
Menlo Park, California: SRI International, p. 51.
Star, E.G. 1980. Mutagenic and cytotoxic effect of ethylene oxide on human
cell cultures. Zentralbl. Bakteriol. 170(5-6): 548-556.
Star, E.G., J. Caselitz and T. Loening. 1980. The effect of ethylene oxide
and chemical disinfectant residues upon the larynx and tracheal mucosa of
rabbits. Zentralbl. Bakteriol. 170(5-6): 539-547.
10-22
-------
Steinberg, S. 1977. Analytical methods for determining ethylene oxide in
plastics. Saf. Health Plastics, Natl. Tech. Conf. Soc. Plastics Eng.,
pp. 141-116.
Stijve, T., R. Kalsbach and G. Eyring. 1976. Determination of occurrence of
ethylene chlorohydrin residues in foods fumigated with ethylene oxide. Mitt.
Geb. Lebensmittelunters. Hyg. 67(4): 403-428.
Strekalova, E. 1971. Mutagenic action of ethylene oxide on mammals.
Toksikol. Nov. Prom. Khim. Veshchestv. 12: 72-78.
Sulovska, K., D. Lindgren, G. Eriksson and L. Ehrenberg. 1969. The mutagenic
effect of low concentrations of ethylene oxide in air. Hereditas.
62(1/2): 264-266. (Cited in NIOSH, 1977).
Sykes, G. 1964. Disinfection and Sterilization. J.P. Lippincott Co.,
Philadelphia, Pennsylvania, pp. 202-227.
Systems Application, Inc. 1982. Human exposure to atmospheric concentrations
of selected chemicals. Appendix A-14 Ethylene Oxide. U.S. Environmental
Protection Agency, Research Triangle Park, NC.
Tan, E.L., R.B. Gumming and A.W. Hsie. 1981. Mutagenicity and cytotoxicity
of ethylene oxide in the CHO/HGPRT system. Environ. Mutagen. 3: 683-686.
Tanooka, H. 1979. Application of Bacillus subtilis spores in the detection
of gas mutagens: A case of ethylene oxide. Mutat. Res. 64(6): 433-435.
Taylor, D.G. 1977a. NIOSH Manual of Analytical Methods, 2nd ed., vol. 2.
U.S. Department of Health, Education, and Welfare, National Institute for
Occupational Safety and Health, Cincinnati, Ohio.
Taylor, D.G. 1977b. NIOSH Manual of Analytical Methods, 2nd ed., vol. 3.
U.S. Department of Health, Education, and Welfare, National Institute for
Occupational Safety and Health, Cincinnati, Ohio.
Taylor, J. 1977. Dermatologic hazards from ethylene oxide. Curtis.
19: 189-192.
10-23
-------
Taylor, G. 1979. Mutagenicity testing of ethylene oxide
Communication to division directors, NIOSH, January 8, 1979.
(unpublished).
Thiess, A. 1963. Considerations of health damage through the influence of
ethylene oxide. Arch. Toxikol. 20: 127-140.
Thiess, A.M., H. Schwegler, I. Fleig and W.G. Stocker. 1981. Mutagenicity
study of workers exposed to alkylene oxides (ethylene oxide - propylene oxide)
and derivatives. J. Occup. Med. 23(5): 343-347.
Thiess, A.M., R. Frentzel-Beyme, R. Link and W.G. Stocker. 1982. Mortality
study on employees exposed to alkylene oxides (ethylene oxide/propylene oxide)
and their derivatives. In; Proceedings of the International Symposium on
Prevention of Occupational Cancer, Helsinki, Finland, p. 249-259.
TSCA (Toxic Substances Control Act). 1980. TSCA Chemical Assessment Series.
Chemical Screening: Initial Evaluations of Substantial Risk Notices,
Section 8(e). U.S. Environmental Protection Agency, Office of Pesticides and
Toxic Substances, Washington, D.C.
Tyler, T.R. and J.A. McKelvey. 1980. Dose-dependent disposition of C
labeled ethylene oxide in rats. Carnegie-Mellon Institute of Research,
Pittsburgh, PA.
U.S. EPA. 1976. Frequency of Organic Compounds Identified in Water.
600/4-76-, U.S. Environmental Protection Agency, Washington, D.C.
EPA-
U.S. EPA. 1980.
contaminants: Epoxides.
EPA-560/11-80-005.
Investigation of selected potential environmental
U.S. Environmental Protection Agency, Washington, DC.
U.S. EPA. 1984. Performance evaluation of the Sciex TAGA 600 mobile
laboratory for ambient air monitoring. Summary report. Advanced Analysis
Techniques Branch, Environmental Monitoring Div., Environmental Monitoring
Systems Laboratory, RTP, NC.
U.S. EPA. 1984. Porposed guidelines for carcinogen risk assessment.
Reg. 49(227): 46294-46301.
Fed.
10-24
-------
USITC (U.S. International Trade Commission). 1976. Synthetic Organic
Chemicals: United States Production and Sales, 197^. USITC Publication 776,
p. 203.
USITC (U.S. International Trade Commission). 1977a. Synthetic Organic
Chemicals: United States Production and Sales, 1975. USITC Publication 804,
p. 198.
USITC (U.S. International Trade Commission). 1977b. Synthetic Organic
Chemicals: United States Production and Sales, 1976. USITC Publication 833,
p. 303.
USITC (U.S. International Trade Commission). 1978. Synthetic Organic
Chemicals: United States Production and Sales, 1977. USITC Publication 920,
p. 361.
USITC (U.S. International Trade Commission). 1979. Synthetic Organic
Chemicals: United States Production and Sales, 1978. USITC Publication 1001,
p. 313.
USITC (U.S. International Trade Commission). 1980. Synthetic Organic
Chemicals: United States Production and Sales, 1979. USITC Publication 1099,
p. 269.
USITC (U.S. International Trade Commission). 1981. Synthetic Organic
Chemicals: United States Production and Sales, 1980. USITC Publication 1183,
p. 265.
USTC (U.S. Tariff Commission). 1971*. Synthetic Organic Chemicals: United
States Production and Sales, 1972. TC Publication 681, p. 206.
USTC (U.S. Tariff Commission). 1975. Synthetic Organic Chemicals: United
States Production and Sales, 1973. TC Publication 728, p. 205.
Van Duuren, B., L. Orris and N. Nelson. 1965. Carcinogenicity of epoxides,
lactones, and peroxy compounds, II. J. Natl. Cancer Inst. 35: 707-717.
Veith, G.D., K.J. Macek, S.R. Petrocelli and J. Carroll. 1979. Federal
Register 15926 (March 15).
10-25
-------
Verrett, M.J. 1974. Memorandum from M.J. Verrett, Reproduction Physiology
Branch, FDA, to Tibor Balazs, Bureau of Drug, FDA. Investigation of the toxic
and teratogenic effects of 2-chloroethanol to the developing chicken embryo.
Verschueren, K. 1977. Handbook of Environmental Data on Organic Chemicals.
Van Nostrand Reinhold, NY. p. 322.
Virtanen, P.O.I, and T. Kuokkanen. 1973. Effect of pressure on the rates of
acid-catalyzed methanolysis of 3-substituted 1,2-epoxypropanes. Suom.
Kemistilehti B. 46(10): 267-270.
von Oettingen, W. 1939. Ethylene oxide, In: Supplement to Occupation and
Health: Encyclopedia of Hygiene, Pathology, and Social Welfare. Geneva,
International Labor Office.
Waite, C., F. Patty and W. Yant. 1930. Acute responses of guinea pigs to
vapors of some new commercial organic compounds. IV. Ethylene oxide. Publ.
Health Rep. 45: 1832-1843.
Wallace, T.J. and R.J. Gritter. 1963. Free radical chemistry of cyclic
ethers. IV. Free radical rearrangement of epoxides. Tetrahedron.
19: 657-665.
Walpole, A. 1958. Carcinogenic action of alkylating agents. Ann. N.Y. Acad.
Sci. 68: 750-761.
Watson, W.A.F. 1966. Further evidence of an essential difference between the
genetical effects of mono- and bifunctional alkylation agents. Mutat. Res.
3: 455-457.
Weast, R.C. 1972. Handbook of Chemistry and Physics. 53rd ed. The Chemical
Rubber Co., Cleveland, Ohio. p. C-287.
Wesley, F., B. Rourke and 0. Darbishire. 1965. The formation of persistent
toxic chlorohydrins in foodstuffs by fumigation with ethylene oxide and with
propylene oxide. J. Food Sci. 30: 1037.
Whelton, R., H. Phaff, E. Mruk and C. Fisher. 1946. Control of micro-
biological food spoilage by fumigation with epoxides. Food Ind. 18: 23-25.
10-26
-------
Wolman, S.R. 1979. Mutational consequences of exposure to ethylene oxide.
J. Environ. Pathol. Toxicol. 2: 1289-1303.
Woodward, G. and M. Woodward. 1971. Toxicity of residuals from ethylene
oxide gas sterilization. Proc. of the Health Ind. Assoc. Tech. Symp.,
Washington, DC. pp. HJO-161.
Yager, J.L.W. 1982. Sister chromatid exchange as a biological monitor for
workplace exposure: A comparative study of exposure-response to ethylene
oxide. Berkeley, CA: The University of California. Northern California
Occupational Health Center Report No. NCOHC 82-B-2. March 1982. p. 163.
Ph.D. Dissertation.
Yager, J.L.W. and D. Benz. 1982. Sister chromatid exchanges induced in
rabbit lymphocytes by ethylene oxide after inhalation exposure. Environ.
Mutagen. 4: 121-134.
Yager, J.W., C.J. Hines and R.C. Spear. 1983. Exposure to ethylene oxide at
work increases sister chromatid exchanges in human peripheral lymphocytes.
Science. 219: 1221-1223.
Yakabova, Z.N., N.A. Shamove, F.A. Miftakhove and L.F. Shilova. 1976. Gyn-
ecological disorders in workers engaged in ethylene oxide production. Kazan.
Med. Zh. 57(6): 558-560.
Young, H. and R. Busbey. 1935. References to the use of ethylene oxide for
pest control. U.S. Department of Agriculture, Bureau of Entomology. Plant
Quarantine, April 1935.
Zamlauski, M.J. and G.J. Cohen. 1976. The effects of aortic infusion of
ethylene oxide renal function in the rat. Toxicol. Appl. Pharmacol.
38(2): 283-295.
10-27
------- |